final-project-nlp.ipynb

{"metadata":{"colab":{"provenance":[],"collapsed_sections":["Yrcm_rthbOU6","B9lA58OIYOpc","zDh2ueweYdo5","jyXbEpgUwJN9","gXvewezD7_LX","InlKsaHUei1g","4wjjsmN8j3Nn"],"gpuType":"T4"},"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"widgets":{"application/vnd.jupyter.widget-state+json":{"97debe01059e41e8b6d393731b184089":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_99d8122078b942a7b7363f8a203fad54","IPY_MODEL_9f85c48278a6455b9cb72d60c5fced43","IPY_MODEL_a9710041f8724bee955d42c736f61dc8"],"layout":"IPY_MODEL_fe85f1d8fbc8423296d891dade94e407"}},"99d8122078b942a7b7363f8a203fad54":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_dd7b17d3619f49d6b29ef8a9f938f4b6","placeholder":"​","style":"IPY_MODEL_e773a165f0a34cf89b878283c6433a59","value":"tokenizer_config.json: 100%"}},"9f85c48278a6455b9cb72d60c5fced43":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_d5d151ddb5d3477989a9a662017fcdb8","max":34173,"min":0,"orientation":"horizontal","style":"IPY_MODEL_192b3972cdd24fc596f7dca1dad36041","value":34173}},"a9710041f8724bee955d42c736f61dc8":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_21e00eb07a374b28a74496607f249381","placeholder":"​","style":"IPY_MODEL_86a09a499b1d43678332bb51431082d1","value":" 34.2k/34.2k [00:00&lt;00:00, 2.28MB/s]"}},"fe85f1d8fbc8423296d891dade94e407":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"dd7b17d3619f49d6b29ef8a9f938f4b6":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"e773a165f0a34cf89b878283c6433a59":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"d5d151ddb5d3477989a9a662017fcdb8":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"192b3972cdd24fc596f7dca1dad36041":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"21e00eb07a374b28a74496607f249381":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"86a09a499b1d43678332bb51431082d1":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"385df7336bcd4cbcb9bcec674e9883e4":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_0b2f08612c5249c7b9d708e5ea74101f","IPY_MODEL_f17212b4c7eb47f1a1ad0ec9ec936b46","IPY_MODEL_a8c7de301efb439c854c147126e47f26"],"layout":"IPY_MODEL_e519b051781d4fce888cfa2eaad51d51"}},"0b2f08612c5249c7b9d708e5ea74101f":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_9344a9fcff2348c29a697b220323ba86","placeholder":"​","style":"IPY_MODEL_7812c7fb76324d3797e1a6935c8c178d","value":"tokenizer.model: 100%"}},"f17212b4c7eb47f1a1ad0ec9ec936b46":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_0e0603e13bb245888fb90a5451788726","max":4241003,"min":0,"orientation":"horizontal","style":"IPY_MODEL_0e4188b7d25e4876aee7c0cfc71a5390","value":4241003}},"a8c7de301efb439c854c147126e47f26":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_a39d899c9bb945af8059ddc2f2e3ffa9","placeholder":"​","style":"IPY_MODEL_b34bb893890d4742bedd43d9fd3e23b8","value":" 4.24M/4.24M [00:00&lt;00:00, 18.9MB/s]"}},"e519b051781d4fce888cfa2eaad51d51":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"9344a9fcff2348c29a697b220323ba86":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7812c7fb76324d3797e1a6935c8c178d":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"0e0603e13bb245888fb90a5451788726":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"0e4188b7d25e4876aee7c0cfc71a5390":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"a39d899c9bb945af8059ddc2f2e3ffa9":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b34bb893890d4742bedd43d9fd3e23b8":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"fb788a77abbf4da3a1bf5cff6d16d6dd":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_d4e362ddc8e940c298c03461e60e42db","IPY_MODEL_4a2a471fef184b979a7c972374a97b00","IPY_MODEL_ca33018b709443fb97d5ce7846d5df82"],"layout":"IPY_MODEL_ebccb7643ed74dbab64dd15e0053282e"}},"d4e362ddc8e940c298c03461e60e42db":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_63dcb768345b4d78bb5e34fec2ba06b7","placeholder":"​","style":"IPY_MODEL_66878ff3020142acb66bf05d107ddfda","value":"tokenizer.json: 100%"}},"4a2a471fef184b979a7c972374a97b00":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_3d88b390323d491b97eb61f131cefa6f","max":17518497,"min":0,"orientation":"horizontal","style":"IPY_MODEL_68a641f26bd44e57a8629d6e17875433","value":17518497}},"ca33018b709443fb97d5ce7846d5df82":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_78a6c1d2de2e42589545ee138120e63f","placeholder":"​","style":"IPY_MODEL_106ec671c60146799535845133dd3241","value":" 17.5M/17.5M [00:00&lt;00:00, 42.5MB/s]"}},"ebccb7643ed74dbab64dd15e0053282e":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"63dcb768345b4d78bb5e34fec2ba06b7":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"66878ff3020142acb66bf05d107ddfda":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"3d88b390323d491b97eb61f131cefa6f":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"68a641f26bd44e57a8629d6e17875433":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"78a6c1d2de2e42589545ee138120e63f":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"106ec671c60146799535845133dd3241":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"6deb520898344e8d9873455c310055a0":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_d03fb8fe287a49b2b80781ba8e67cc39","IPY_MODEL_130f6723b06247be8b0f117e2589fc69","IPY_MODEL_3147321424674e7b92b340fe708034a9"],"layout":"IPY_MODEL_d296bd992283407da123bf74c60b4306"}},"d03fb8fe287a49b2b80781ba8e67cc39":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_f560c94036634a90a84a8262e39dba0c","placeholder":"​","style":"IPY_MODEL_b1b96089a7ec43268d1ef852312c5ba7","value":"special_tokens_map.json: 100%"}},"130f6723b06247be8b0f117e2589fc69":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_569644b964674226a0e1a3790b911f7a","max":636,"min":0,"orientation":"horizontal","style":"IPY_MODEL_8f1e0c7966cb4dd6ba1cd872f44e47ee","value":636}},"3147321424674e7b92b340fe708034a9":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_1b4fef37e8a140d2b123a3c2d6e1c4db","placeholder":"​","style":"IPY_MODEL_3f57906d1f6d4de8b0d1da518dc1bf83","value":" 636/636 [00:00&lt;00:00, 47.0kB/s]"}},"d296bd992283407da123bf74c60b4306":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"f560c94036634a90a84a8262e39dba0c":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b1b96089a7ec43268d1ef852312c5ba7":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"569644b964674226a0e1a3790b911f7a":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"8f1e0c7966cb4dd6ba1cd872f44e47ee":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"1b4fef37e8a140d2b123a3c2d6e1c4db":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"3f57906d1f6d4de8b0d1da518dc1bf83":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"4824ea6f85e844019b74285e33ae855a":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_2c9bc4e96ee44e73bfdf47919ad47977","IPY_MODEL_96bc14978ff645a29b8c026578853148","IPY_MODEL_e134d36200e441e5ac02e68f4163e1b6"],"layout":"IPY_MODEL_e182a0b3f5ec48549c64cb17b41eaf15"}},"2c9bc4e96ee44e73bfdf47919ad47977":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_f8a86df3661e43fda8c507f6a9aa7be1","placeholder":"​","style":"IPY_MODEL_ac837a5d4b254c149e1b030acfdb72fa","value":"config.json: 100%"}},"96bc14978ff645a29b8c026578853148":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_d55c7f607e494f4a8d0bf3d45dc9a5df","max":618,"min":0,"orientation":"horizontal","style":"IPY_MODEL_b437e22cf1974449a5372eecb1c0812e","value":618}},"e134d36200e441e5ac02e68f4163e1b6":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_0d7499045f664c918adfd42cef8fe6ae","placeholder":"​","style":"IPY_MODEL_471ae7a02c7746fba9a59fa638ecc9c4","value":" 618/618 [00:00&lt;00:00, 41.9kB/s]"}},"e182a0b3f5ec48549c64cb17b41eaf15":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"f8a86df3661e43fda8c507f6a9aa7be1":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"ac837a5d4b254c149e1b030acfdb72fa":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"d55c7f607e494f4a8d0bf3d45dc9a5df":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b437e22cf1974449a5372eecb1c0812e":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"0d7499045f664c918adfd42cef8fe6ae":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"471ae7a02c7746fba9a59fa638ecc9c4":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"c25b9ec8cecf462f84ca0916034f3fc5":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_09499f8c2a864740953d8069cea449e6","IPY_MODEL_b7d0588786484df28fbeb936b31a1622","IPY_MODEL_5bdbdd02077d4575b9a1f4a8060b7357"],"layout":"IPY_MODEL_9b1d290d6c6243f29a3b0b2a01ba142a"}},"09499f8c2a864740953d8069cea449e6":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_5a5f88040e4c40b189ecb81a79e96fb6","placeholder":"​","style":"IPY_MODEL_88577f42249743bd847c95dfa22792ff","value":"model.safetensors.index.json: 100%"}},"b7d0588786484df28fbeb936b31a1622":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_5605e10bd8344e94989d644c8dbc9732","max":13489,"min":0,"orientation":"horizontal","style":"IPY_MODEL_30ddb404b75748a88fb2a689220742d9","value":13489}},"5bdbdd02077d4575b9a1f4a8060b7357":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_b5deb1c30d2f4d1399e62059aa4eacdc","placeholder":"​","style":"IPY_MODEL_45865cd1b8f24cd2a3eff2e45f80de59","value":" 13.5k/13.5k [00:00&lt;00:00, 843kB/s]"}},"9b1d290d6c6243f29a3b0b2a01ba142a":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"5a5f88040e4c40b189ecb81a79e96fb6":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"88577f42249743bd847c95dfa22792ff":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"5605e10bd8344e94989d644c8dbc9732":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"30ddb404b75748a88fb2a689220742d9":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"b5deb1c30d2f4d1399e62059aa4eacdc":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"45865cd1b8f24cd2a3eff2e45f80de59":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"192ad8a5ae7d4c879e91099d4ccfc506":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_f059c54ad1b24834a4147118e1082b44","IPY_MODEL_cb0ea80367e84032b6f70696feccf903","IPY_MODEL_367c1b006e5d4e1ba619d195fbf44a81"],"layout":"IPY_MODEL_76d4b18518b24d77bbda57a8113e022c"}},"f059c54ad1b24834a4147118e1082b44":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_73a081c4407749cfab8caec27e9bec2b","placeholder":"​","style":"IPY_MODEL_8bd155f916e9449f9ba45352e95c93de","value":"Downloading shards: 100%"}},"cb0ea80367e84032b6f70696feccf903":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_aea810054c1843ceac0159244f9dc95d","max":2,"min":0,"orientation":"horizontal","style":"IPY_MODEL_fad03b1c75dd474b9020ef0ae8e9d1b1","value":2}},"367c1b006e5d4e1ba619d195fbf44a81":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_d588aec910844311b7b7819fdab2102d","placeholder":"​","style":"IPY_MODEL_dbd34d1fa88b4d8395f72eccf6c693d9","value":" 2/2 [00:42&lt;00:00, 17.58s/it]"}},"76d4b18518b24d77bbda57a8113e022c":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"73a081c4407749cfab8caec27e9bec2b":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"8bd155f916e9449f9ba45352e95c93de":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"aea810054c1843ceac0159244f9dc95d":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"fad03b1c75dd474b9020ef0ae8e9d1b1":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"d588aec910844311b7b7819fdab2102d":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"dbd34d1fa88b4d8395f72eccf6c693d9":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"c524d424e582413c92a54bc5c8a2ce28":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_a331cb60e24a4d138b781f04fea98cfc","IPY_MODEL_8466997b3385462884ab1d4e4f53397a","IPY_MODEL_fdb19a0858504b1ab659b6a165187b38"],"layout":"IPY_MODEL_21161189229d4ae28176f21b3bc48029"}},"a331cb60e24a4d138b781f04fea98cfc":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_658f41fbac494b95bc328d8a7bbe80ad","placeholder":"​","style":"IPY_MODEL_bc60e11fa7e44113960dc35cadd66fda","value":"model-00001-of-00002.safetensors: 100%"}},"8466997b3385462884ab1d4e4f53397a":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_c501c9932af146279de761a53a64f1c6","max":4945242264,"min":0,"orientation":"horizontal","style":"IPY_MODEL_ab633dab587543e79d3b1a4d23c551b1","value":4945242264}},"fdb19a0858504b1ab659b6a165187b38":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_07c390b0448340df92b1de813af96d5a","placeholder":"​","style":"IPY_MODEL_f19d9b0a6a034fd8ae979f41e6345dea","value":" 4.95G/4.95G [00:40&lt;00:00, 172MB/s]"}},"21161189229d4ae28176f21b3bc48029":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"658f41fbac494b95bc328d8a7bbe80ad":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"bc60e11fa7e44113960dc35cadd66fda":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"c501c9932af146279de761a53a64f1c6":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"ab633dab587543e79d3b1a4d23c551b1":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"07c390b0448340df92b1de813af96d5a":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"f19d9b0a6a034fd8ae979f41e6345dea":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"e18c298a543d44bfb5dc171baf81b1c0":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_474ae03450284eeb9943699057cee5c9","IPY_MODEL_d8ccf08cbe9249f3abf71f792bfc523a","IPY_MODEL_9f41761fe0b74150b5607b6a98bb8909"],"layout":"IPY_MODEL_b800901a241f44b7a51de3b89c8fce90"}},"474ae03450284eeb9943699057cee5c9":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_1276e35e423b435e91c1b1da8f660825","placeholder":"​","style":"IPY_MODEL_8db2a42a77374c47b79749b0ef27ba72","value":"model-00002-of-00002.safetensors: 100%"}},"d8ccf08cbe9249f3abf71f792bfc523a":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_459425a8beb648a58e0a0cc270c071f0","max":67121608,"min":0,"orientation":"horizontal","style":"IPY_MODEL_2f51576d2da945d08fa85db78a102d90","value":67121608}},"9f41761fe0b74150b5607b6a98bb8909":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_6623a30ef7ae48888be9d283ca1aacf6","placeholder":"​","style":"IPY_MODEL_572391b8c28540508502bdd5d586a9a5","value":" 67.1M/67.1M [00:00&lt;00:00, 98.3MB/s]"}},"b800901a241f44b7a51de3b89c8fce90":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"1276e35e423b435e91c1b1da8f660825":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"8db2a42a77374c47b79749b0ef27ba72":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"459425a8beb648a58e0a0cc270c071f0":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"2f51576d2da945d08fa85db78a102d90":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"6623a30ef7ae48888be9d283ca1aacf6":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"572391b8c28540508502bdd5d586a9a5":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"1c4af129211c4c0d904adb34338ecc40":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_7bc389680ee047dbb1ba4b4dfbc58067","IPY_MODEL_15005cb65973457faa5f92b09ee084eb","IPY_MODEL_912a373ce8604eb9b6fbf7150bdba3a3"],"layout":"IPY_MODEL_c01f97d60c0d41e195972d24dfc0af95"}},"7bc389680ee047dbb1ba4b4dfbc58067":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_bef54671fdb94bc1ae430d35c4448479","placeholder":"​","style":"IPY_MODEL_f23d25f4f32a4f41949faed79eacdd87","value":"Loading checkpoint shards: 100%"}},"15005cb65973457faa5f92b09ee084eb":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_c901d11c08864c42aa5998f069321330","max":2,"min":0,"orientation":"horizontal","style":"IPY_MODEL_c4253a8ea011480c98ad332d176db8be","value":2}},"912a373ce8604eb9b6fbf7150bdba3a3":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_2da4bef12bf7448aa20cafdbef545769","placeholder":"​","style":"IPY_MODEL_c52c8dff93ff44d893df97d4dde1d818","value":" 2/2 [00:21&lt;00:00,  9.11s/it]"}},"c01f97d60c0d41e195972d24dfc0af95":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"bef54671fdb94bc1ae430d35c4448479":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"f23d25f4f32a4f41949faed79eacdd87":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"c901d11c08864c42aa5998f069321330":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"c4253a8ea011480c98ad332d176db8be":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"2da4bef12bf7448aa20cafdbef545769":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"c52c8dff93ff44d893df97d4dde1d818":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"655ebd1d26dd42b2b0496870d5baaaa4":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_a18b8110e1844a83900fbe6a253b7729","IPY_MODEL_1c8a902735bf40628f6ccabfad555e8c","IPY_MODEL_81017e6422f24c13ad748197c6ccd78a"],"layout":"IPY_MODEL_42ca1d405d574336a5403214928b75b4"}},"a18b8110e1844a83900fbe6a253b7729":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_6f603c9dabf54199ab7088d046c37662","placeholder":"​","style":"IPY_MODEL_234583baeb084657b6b0c1cbcbca75e4","value":"generation_config.json: 100%"}},"1c8a902735bf40628f6ccabfad555e8c":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_2bba184ce3874019af6042797d63ac94","max":132,"min":0,"orientation":"horizontal","style":"IPY_MODEL_57fc7807b50044c385462e67b3f7bca0","value":132}},"81017e6422f24c13ad748197c6ccd78a":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_bd4c25a951d64a45b4d37b26d15f6ff5","placeholder":"​","style":"IPY_MODEL_7ae4167b53c44abd9ba6930502a8dce1","value":" 132/132 [00:00&lt;00:00, 8.03kB/s]"}},"42ca1d405d574336a5403214928b75b4":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"6f603c9dabf54199ab7088d046c37662":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"234583baeb084657b6b0c1cbcbca75e4":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"2bba184ce3874019af6042797d63ac94":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"57fc7807b50044c385462e67b3f7bca0":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"bd4c25a951d64a45b4d37b26d15f6ff5":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7ae4167b53c44abd9ba6930502a8dce1":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"110ad0a48b8c4b668d15c05ca4cd1cc1":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_2800afcf6d914278a756fc867a93db90","IPY_MODEL_f9cdd57f5c9146939ebda9943ce5d3f5","IPY_MODEL_2d7b2aeb2bd14b1cbf58a93870550809"],"layout":"IPY_MODEL_5188734966ad44a0818edf8c21b633c9"}},"2800afcf6d914278a756fc867a93db90":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_05563cc4bef14eb29970554d6aef1518","placeholder":"​","style":"IPY_MODEL_157b3bd25e084ea0afa39e26801ecde8","value":"Downloading builder script: 100%"}},"f9cdd57f5c9146939ebda9943ce5d3f5":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_1e87af4bc9894c2fb4a3f0853f0bb646","max":7950,"min":0,"orientation":"horizontal","style":"IPY_MODEL_2e3b086bafa94feca8b422aceeb8336b","value":7950}},"2d7b2aeb2bd14b1cbf58a93870550809":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_562a191c3c114576b1f0275920aa7406","placeholder":"​","style":"IPY_MODEL_cd9194d0da4842bf8100ab05b4fa8637","value":" 7.95k/7.95k [00:00&lt;00:00, 562kB/s]"}},"5188734966ad44a0818edf8c21b633c9":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"05563cc4bef14eb29970554d6aef1518":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"157b3bd25e084ea0afa39e26801ecde8":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"1e87af4bc9894c2fb4a3f0853f0bb646":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"2e3b086bafa94feca8b422aceeb8336b":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"562a191c3c114576b1f0275920aa7406":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"cd9194d0da4842bf8100ab05b4fa8637":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"260a89609b464e8dbdb80bf09df53815":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_cc908c867f8244b59749501ae9bc0797","IPY_MODEL_ee0969c2329846c585d9d99d99b62ade","IPY_MODEL_7e1e003a45d74dd295859076d19ec74c"],"layout":"IPY_MODEL_b2366e17bf3e458f8570f9924de4e382"}},"cc908c867f8244b59749501ae9bc0797":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_2d31a478ac464848842ddaa58b88545c","placeholder":"​","style":"IPY_MODEL_7e1378b67f52491fbe9441f3505f0931","value":"tokenizer_config.json: 100%"}},"ee0969c2329846c585d9d99d99b62ade":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_92a2e90cfe9149b8b11ccb9f61dbd822","max":25,"min":0,"orientation":"horizontal","style":"IPY_MODEL_cce8588a42fc4d9f962f2ba68a1b3c76","value":25}},"7e1e003a45d74dd295859076d19ec74c":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_36e43820dbf64d8aa53bfa8c5e0b5035","placeholder":"​","style":"IPY_MODEL_99fb35d05c2e4501b20e8ca8db2cfa88","value":" 25.0/25.0 [00:00&lt;00:00, 558B/s]"}},"b2366e17bf3e458f8570f9924de4e382":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"2d31a478ac464848842ddaa58b88545c":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7e1378b67f52491fbe9441f3505f0931":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"92a2e90cfe9149b8b11ccb9f61dbd822":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"cce8588a42fc4d9f962f2ba68a1b3c76":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"36e43820dbf64d8aa53bfa8c5e0b5035":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"99fb35d05c2e4501b20e8ca8db2cfa88":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"f4a66cbd934640449a8601026d165851":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_2102b8581b804f829b64c0a80e872b4d","IPY_MODEL_8422dd0f3cb14e7abb0163e43a289deb","IPY_MODEL_0e1f474280c8466a9cf15d919a49990f"],"layout":"IPY_MODEL_fbc25e7290a9416f991ddc152368c835"}},"2102b8581b804f829b64c0a80e872b4d":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_b0407b6cf07a427db75863761603f6a5","placeholder":"​","style":"IPY_MODEL_4b4c08dc9ef0414490aa13b40498698e","value":"config.json: 100%"}},"8422dd0f3cb14e7abb0163e43a289deb":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_0e7fab03f25c4c8abfe17f216c3f4a96","max":482,"min":0,"orientation":"horizontal","style":"IPY_MODEL_c3dd2825aa974413bbcf33ae30a7ad89","value":482}},"0e1f474280c8466a9cf15d919a49990f":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_47909fe2b0eb403eaa949f980ed983a9","placeholder":"​","style":"IPY_MODEL_60907538d7c842ae81fdaf59442a34b7","value":" 482/482 [00:00&lt;00:00, 15.2kB/s]"}},"fbc25e7290a9416f991ddc152368c835":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"b0407b6cf07a427db75863761603f6a5":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"4b4c08dc9ef0414490aa13b40498698e":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"0e7fab03f25c4c8abfe17f216c3f4a96":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"c3dd2825aa974413bbcf33ae30a7ad89":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"47909fe2b0eb403eaa949f980ed983a9":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"60907538d7c842ae81fdaf59442a34b7":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"4100710c824849a18316318f1b4e289b":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_13d4bc0f8c2744d6add9d14f9d32d7a6","IPY_MODEL_185299f8d07148f9a539a0f84f95679c","IPY_MODEL_1b403130c3b548c8a691227e516f06e6"],"layout":"IPY_MODEL_465f24676c084a1fbaf9da1d0d137983"}},"13d4bc0f8c2744d6add9d14f9d32d7a6":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_05d63ab569354b7bb2f25f88b3c652e5","placeholder":"​","style":"IPY_MODEL_c0d4a04f94e04ba2a52d7e4858ec924b","value":"vocab.json: 100%"}},"185299f8d07148f9a539a0f84f95679c":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_4eaa660573a046cbb4d53722b9bf9404","max":898823,"min":0,"orientation":"horizontal","style":"IPY_MODEL_7b5b06f07ea342a5a1dc16f46df0908e","value":898823}},"1b403130c3b548c8a691227e516f06e6":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_03c50c883f2040c6ad18f636ae68dcde","placeholder":"​","style":"IPY_MODEL_2cd7bb7d452e4496be45149b89f37e02","value":" 899k/899k [00:00&lt;00:00, 4.58MB/s]"}},"465f24676c084a1fbaf9da1d0d137983":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"05d63ab569354b7bb2f25f88b3c652e5":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"c0d4a04f94e04ba2a52d7e4858ec924b":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"4eaa660573a046cbb4d53722b9bf9404":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7b5b06f07ea342a5a1dc16f46df0908e":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"03c50c883f2040c6ad18f636ae68dcde":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"2cd7bb7d452e4496be45149b89f37e02":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"f70e6f56872844aea51761d62a2a7acf":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_b653c125ae86486c812c5ef81d8b5922","IPY_MODEL_9ff2418471ee4f87975771c38240416a","IPY_MODEL_94bb63ffd22745d1a945efa90298a4aa"],"layout":"IPY_MODEL_89e0e953d12d4025a522a9fb80b815b4"}},"b653c125ae86486c812c5ef81d8b5922":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_0e9cecd09b4c4eac9f30d5c04ec0046c","placeholder":"​","style":"IPY_MODEL_fcf6ce29dc004c5fa0a5b6c9163d1ba3","value":"merges.txt: 100%"}},"9ff2418471ee4f87975771c38240416a":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_3149896a259543539aff862be47ad4ee","max":456318,"min":0,"orientation":"horizontal","style":"IPY_MODEL_896f7baa6b7d4136855a03b7405913a7","value":456318}},"94bb63ffd22745d1a945efa90298a4aa":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_2fe7181f82dc4435ae87bcefa50ea316","placeholder":"​","style":"IPY_MODEL_bb06f202cebd469fa8ed9b90c4b4a666","value":" 456k/456k [00:00&lt;00:00, 2.58MB/s]"}},"89e0e953d12d4025a522a9fb80b815b4":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"0e9cecd09b4c4eac9f30d5c04ec0046c":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"fcf6ce29dc004c5fa0a5b6c9163d1ba3":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"3149896a259543539aff862be47ad4ee":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"896f7baa6b7d4136855a03b7405913a7":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"2fe7181f82dc4435ae87bcefa50ea316":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"bb06f202cebd469fa8ed9b90c4b4a666":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"2b449fde4f1a44e4898297a2b3c4522e":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_6fe9293074854bbf9de69a2a2fdffd16","IPY_MODEL_6648fcd7b660449387c817736e1b85db","IPY_MODEL_c3667166af36477992e0d8997ac2cc21"],"layout":"IPY_MODEL_7204a2356fe7444785f8769a041a2b45"}},"6fe9293074854bbf9de69a2a2fdffd16":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_96381ab5f9aa40e7bd2574c6e6d4fad4","placeholder":"​","style":"IPY_MODEL_7c9008b1ae0c4c2fa82ba0201e7bbaad","value":"tokenizer.json: 100%"}},"6648fcd7b660449387c817736e1b85db":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_a837c7541566488488012fe4de4603ef","max":1355863,"min":0,"orientation":"horizontal","style":"IPY_MODEL_f2ef4972beee480bbb47ff310fd01a48","value":1355863}},"c3667166af36477992e0d8997ac2cc21":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_d5134b6ba4d54f8384f6e4ed2e817df2","placeholder":"​","style":"IPY_MODEL_0655a7e33eae475ebc47caf4c1627170","value":" 1.36M/1.36M [00:00&lt;00:00, 9.59MB/s]"}},"7204a2356fe7444785f8769a041a2b45":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"96381ab5f9aa40e7bd2574c6e6d4fad4":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"7c9008b1ae0c4c2fa82ba0201e7bbaad":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"a837c7541566488488012fe4de4603ef":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"f2ef4972beee480bbb47ff310fd01a48":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"d5134b6ba4d54f8384f6e4ed2e817df2":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"0655a7e33eae475ebc47caf4c1627170":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"dee9352ab2644a7dbe995bd7fefc7129":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HBoxModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HBoxView","box_style":"","children":["IPY_MODEL_5a734feae8bf48729789e2ff09e847f9","IPY_MODEL_a9ca7d8b749d4acaa5422be5122352b3","IPY_MODEL_2347c4940fef43babd23fa81fa6fccef"],"layout":"IPY_MODEL_79cf311f9efc402ba62bc92f534a5f3e"}},"5a734feae8bf48729789e2ff09e847f9":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_ae72a4f696374925b673dd30640fc8f6","placeholder":"​","style":"IPY_MODEL_acff8c0c8ea14e26b0d29cd912986d69","value":"model.safetensors: 100%"}},"a9ca7d8b749d4acaa5422be5122352b3":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"FloatProgressModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"ProgressView","bar_style":"success","description":"","description_tooltip":null,"layout":"IPY_MODEL_e1164d31158f444a843fa8ae39caf4b9","max":1421700479,"min":0,"orientation":"horizontal","style":"IPY_MODEL_4200bdb846f2415d90a38bcf5a5da365","value":1421700479}},"2347c4940fef43babd23fa81fa6fccef":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","model_module_version":"1.5.0","state":{"_dom_classes":[],"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"HTMLModel","_view_count":null,"_view_module":"@jupyter-widgets/controls","_view_module_version":"1.5.0","_view_name":"HTMLView","description":"","description_tooltip":null,"layout":"IPY_MODEL_afbb4616752740f88752ca68a11674ff","placeholder":"​","style":"IPY_MODEL_0a1855a2531b4f7c9adac7ccdc5c1cd7","value":" 1.42G/1.42G [00:16&lt;00:00, 52.0MB/s]"}},"79cf311f9efc402ba62bc92f534a5f3e":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"ae72a4f696374925b673dd30640fc8f6":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"acff8c0c8ea14e26b0d29cd912986d69":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}},"e1164d31158f444a843fa8ae39caf4b9":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"4200bdb846f2415d90a38bcf5a5da365":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"ProgressStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","bar_color":null,"description_width":""}},"afbb4616752740f88752ca68a11674ff":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","model_module_version":"1.2.0","state":{"_model_module":"@jupyter-widgets/base","_model_module_version":"1.2.0","_model_name":"LayoutModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"LayoutView","align_content":null,"align_items":null,"align_self":null,"border":null,"bottom":null,"display":null,"flex":null,"flex_flow":null,"grid_area":null,"grid_auto_columns":null,"grid_auto_flow":null,"grid_auto_rows":null,"grid_column":null,"grid_gap":null,"grid_row":null,"grid_template_areas":null,"grid_template_columns":null,"grid_template_rows":null,"height":null,"justify_content":null,"justify_items":null,"left":null,"margin":null,"max_height":null,"max_width":null,"min_height":null,"min_width":null,"object_fit":null,"object_position":null,"order":null,"overflow":null,"overflow_x":null,"overflow_y":null,"padding":null,"right":null,"top":null,"visibility":null,"width":null}},"0a1855a2531b4f7c9adac7ccdc5c1cd7":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","model_module_version":"1.5.0","state":{"_model_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_model_name":"DescriptionStyleModel","_view_count":null,"_view_module":"@jupyter-widgets/base","_view_module_version":"1.2.0","_view_name":"StyleView","description_width":""}}}},"accelerator":"GPU","kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":8635895,"sourceType":"datasetVersion","datasetId":5171351},{"sourceId":8657974,"sourceType":"datasetVersion","datasetId":5186966}],"dockerImageVersionId":30733,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install rank_bm25\n!pip install accelerate\n!pip install evaluate\n!pip install bert_score\n!pip install -i https://pypi.org/simple/ bitsandbytes\n!pip install -U sentence-transformers","metadata":{"id":"ptqHHHsMxAX1","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import torch\nfrom torch.nn import functional as F\nimport transformers\nfrom sentence_transformers import SentenceTransformer, util\nimport re\nfrom transformers import AutoTokenizer, AutoModel\n\n# transformers\nfrom transformers import (\n    GPTJForCausalLM,\n    AutoModelForCausalLM,\n    AutoModelForMaskedLM,\n    AutoTokenizer,\n    GPT2Tokenizer,\n    GPT2LMHeadModel,\n    AutoModelForSequenceClassification,\n    set_seed,\n    BitsAndBytesConfig\n    )\n\nimport numpy as np\nimport json\nfrom rank_bm25 import BM25Okapi\nfrom nltk.tokenize import word_tokenize\nimport nltk\nimport concurrent.futures\n\nfrom evaluate import load\n\ntorch.backends.cuda.enable_mem_efficient_sdp(False)\ntorch.backends.cuda.enable_flash_sdp(False)\n\n# Download NLTK data files (if not already downloaded)\nnltk.download('punkt')\n\nset_seed(2024)\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nprint(\"Torch version: \", torch.__version__)\n\n#Additional Info when using cuda\nif device.type == 'cuda':\n    print(torch.cuda.get_device_name(0))\n    print('Memory Usage:')\n    print('Allocated:', round(torch.cuda.memory_allocated(0)/1024**3,1), 'GB')\n    print('Cached:   ', round(torch.cuda.memory_reserved(0)/1024**3,1), 'GB')","metadata":{"id":"EKVQeJyKBfMJ","outputId":"9ca1f841-8f19-4ca5-8596-05df112426ec","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Load & Preprocessing Data","metadata":{"id":"Yrcm_rthbOU6"}},{"cell_type":"code","source":"import json\ndata = []\nwith open(\"/kaggle/input/teacher-student-chatroom/test_with-reference.jsonl\", 'r') as file_json:\n    for line in file_json:\n        data.append(json.loads(line))","metadata":{"id":"AsvK6db3bRQy","execution":{"iopub.status.busy":"2024-06-27T09:43:02.221194Z","iopub.execute_input":"2024-06-27T09:43:02.221574Z","iopub.status.idle":"2024-06-27T09:43:02.242752Z","shell.execute_reply.started":"2024-06-27T09:43:02.221544Z","shell.execute_reply":"2024-06-27T09:43:02.241899Z"},"trusted":true},"execution_count":2,"outputs":[]},{"cell_type":"code","source":"def formateed_utterance(utterance):\n    utt = \"{speaker} : {text}\"\n    formatted_utterances = []\n    for idx, utterance in enumerate(utterance):\n        speaker = utterance['speaker']\n        text = utterance['text']\n        formatted_utt = utt.format(speaker = speaker, text = text)\n        formatted_utterances.append(formatted_utt)\n    \n    return speaker, \"\\n\".join(formatted_utterances)","metadata":{"id":"6ttKVdAbg4Nq","execution":{"iopub.status.busy":"2024-06-27T09:43:02.597814Z","iopub.execute_input":"2024-06-27T09:43:02.598616Z","iopub.status.idle":"2024-06-27T09:43:02.603727Z","shell.execute_reply.started":"2024-06-27T09:43:02.598581Z","shell.execute_reply":"2024-06-27T09:43:02.602819Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"teacher_reply = []\nteacher_continuation = []\nfor conversation in data:\n    id = conversation[\"id\"]\n    utterance = conversation[\"utterances\"]\n    response = conversation[\"response\"]\n\n    end_speaker, formatted_utterances = formateed_utterance(utterance)\n\n    if end_speaker == \"student\":\n        teacher_reply.append({\"id\":id, \"dialog\" : formatted_utterances, \"response\": response['text']})\n    elif end_speaker == \"teacher\":\n        teacher_continuation.append({\"id\":id, \"dialog\" : formatted_utterances, \"response\": response['text']})\n    else:\n        print(\"not found who is the end speaker\")\n\nprint(\"Jumlah data dalam teacher reply task: \", len(teacher_reply))\nprint(\"Jumlah data dalam teacher continuation task \", len(teacher_continuation))","metadata":{"id":"bhqtUbQVfQQR","outputId":"60f6b68e-397e-488e-a68c-42c31ed89e7d","execution":{"iopub.status.busy":"2024-06-27T09:43:02.929226Z","iopub.execute_input":"2024-06-27T09:43:02.929578Z","iopub.status.idle":"2024-06-27T09:43:02.939430Z","shell.execute_reply.started":"2024-06-27T09:43:02.929547Z","shell.execute_reply":"2024-06-27T09:43:02.938607Z"},"trusted":true},"execution_count":4,"outputs":[{"name":"stdout","text":"Jumlah data dalam teacher reply task:  181\nJumlah data dalam teacher continuation task  92\n","output_type":"stream"}]},{"cell_type":"code","source":"teacher_reply[1]['response']","metadata":{"id":"6forobkVroVZ","outputId":"861bf3b5-2810-4d39-9e3e-c70dab726d7f","execution":{"iopub.status.busy":"2024-06-27T09:43:03.407473Z","iopub.execute_input":"2024-06-27T09:43:03.407842Z","iopub.status.idle":"2024-06-27T09:43:03.414543Z","shell.execute_reply.started":"2024-06-27T09:43:03.407815Z","shell.execute_reply":"2024-06-27T09:43:03.413475Z"},"trusted":true},"execution_count":5,"outputs":[{"execution_count":5,"output_type":"execute_result","data":{"text/plain":"\"Oh, poor <STUDENT>! Go get somewhere warm! I'll see you on Friday 9.30 am. Well done today, you've been very brave!\""},"metadata":{}}]},{"cell_type":"code","source":"teacher_reply[1]['dialog']","metadata":{"execution":{"iopub.status.busy":"2024-06-27T09:43:12.700250Z","iopub.execute_input":"2024-06-27T09:43:12.700630Z","iopub.status.idle":"2024-06-27T09:43:12.706661Z","shell.execute_reply.started":"2024-06-27T09:43:12.700598Z","shell.execute_reply":"2024-06-27T09:43:12.705724Z"},"trusted":true},"execution_count":6,"outputs":[{"execution_count":6,"output_type":"execute_result","data":{"text/plain":"\"teacher : Which is easy, because you can use my / his / your etc. and not think about articles!\\nstudent : Only 10 minutes left!\\nteacher : I know, we can finish early if you are getting cold?\\nstudent : I'm really cold\""},"metadata":{}}]},{"cell_type":"code","source":"teacher_continuation[0:5]","metadata":{"id":"FOMC8iIMtJWZ","outputId":"965eb3e9-bce1-4121-9603-9ebdea0cc9a4","execution":{"iopub.status.busy":"2024-06-27T07:46:57.845962Z","iopub.execute_input":"2024-06-27T07:46:57.846700Z","iopub.status.idle":"2024-06-27T07:46:57.852480Z","shell.execute_reply.started":"2024-06-27T07:46:57.846670Z","shell.execute_reply":"2024-06-27T07:46:57.851707Z"},"trusted":true},"execution_count":5,"outputs":[{"execution_count":5,"output_type":"execute_result","data":{"text/plain":"[{'id': 'test_0002',\n  'dialog': \"teacher : But most importantly, there could have been other reasons behind children not eating / eating the marshmallow\\nteacher : For example, they might not have liked marshmallows!\\nteacher : Or they could have wanted to please the researchers\\nstudent : Yes that's true\\nteacher : What do you think could have been the reasons?\\nstudent : Maybe some of them were too young?\\nteacher : Yes, exactly!\",\n  'response': 'Could you make a sentence with a modal? Some of them ___ ___ ___ too young'},\n {'id': 'test_0003',\n  'dialog': \"student : my train leaving at 9:30\\nstudent : or - my train leaving time at 9:30.\\nteacher : OK thanks...so you can describe your own plans with present continuous or 'going to' ...when we talk about courses/trains etc (things with timetables) we often use present simple...\",\n  'response': \"So: 'I'm going home tomorrow - my train _____ at 9:30' One word...what is it?\"},\n {'id': 'test_0004',\n  'dialog': 'student : my time was shorter than yesterday.\\nteacher : So the adjective you can use with time (for running for example) is good/fast ok?\\nstudent : yes... so my time was faster than yesterday.\\nteacher : OK great! my time was faster/quicker/better than yesterday! good',\n  'response': 'OK so one more: My personal best for 5k is 15:03 (along time ago) ..... So 15:03 is the _____  I _______  ______ the 5k.??'},\n {'id': 'test_0005',\n  'dialog': \"student : right?\\nstudent : yes,, thanks\\nteacher : well <STUDENT> - we just usually use the verb 'be' with these adjectives\\nstudent : wow... I got it\\nteacher : like e.g. rich/poor/old etc - not because it's a feeling\\nteacher : So in a way this bit is the easy bit (the verb I mean) - the thing you have to get right is the adjective itself and the preposition that comes next you see?\",\n  'response': 'OK see you in 15 mins ok?'},\n {'id': 'test_0007',\n  'dialog': \"student : That's great!\\nteacher : But of course my main job is to help trainees pass the course!!\",\n  'response': 'What do you do to keep up with your English <STUDENT>? Do you watch series/films for example?'}]"},"metadata":{}}]},{"cell_type":"markdown","source":"## Load Model","metadata":{"id":"vyA6ZDTJ2zBq"}},{"cell_type":"code","source":"# # Load the pre-trained GPT-2 model and tokenizer\n# model_name = 'gpt2-xl'\n# model = GPT2LMHeadModel.from_pretrained(model_name)\n# tokenizer = GPT2Tokenizer.from_pretrained(model_name)\n\n# # Load DialogRPT\n# model_card = \"microsoft/DialogRPT-updown\"\n# tokenizer_dialogrpt = AutoTokenizer.from_pretrained(model_card)\n# model_dialogrpt = AutoModelForSequenceClassification.from_pretrained(model_card)\n\n# Another LLM ~ we need more ram to load the model\n# tokenizer = AutoTokenizer.from_pretrained(\"EleutherAI/gpt-j-6B\")\n# model = AutoModelForCausalLM.from_pretrained(\"EleutherAI/gpt-j-6B\")\n\n# I think this is more promising LLM we can use. It is only 2.78 bilion paramaters\n# tokenizer = AutoTokenizer.from_pretrained(\"microsoft/phi-2\")\n# model = AutoModelForCausalLM.from_pretrained(\"microsoft/phi-2\")\n\n# Load model directly gemmea 2 bilion\ntoken = \"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\"\nquantization_config = BitsAndBytesConfig(load_in_8bit=True)\ntokenizer = AutoTokenizer.from_pretrained(\"google/gemma-1.1-2b-it\", token=token)\nmodel = AutoModelForCausalLM.from_pretrained(\"google/gemma-1.1-2b-it\",\n                                             device_map=\"cuda\",\n                                             quantization_config=quantization_config,\n                                            #  low_cpu_mem_usage=True,\n                                             torch_dtype=torch.bfloat16,\n                                             token=token\n                                             )\n\n\n# Load model evaluation\nbertscore = load(\"bertscore\")\n\n# Load model ranking\nsbert_model = SentenceTransformer('paraphrase-MiniLM-L6-v2')\n\n# Load model bert\nbert_tokenizer = AutoTokenizer.from_pretrained('bert-base-uncased')\nbert_model = AutoModel.from_pretrained('bert-base-uncased')","metadata":{"id":"dpRO18Z323CI","outputId":"cef01dab-8b64-421a-cc68-1365081e47ae","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(model.get_memory_footprint())","metadata":{"id":"s6DE-KJP0ht6","outputId":"35ddc2c6-ff7f-4737-dae2-9920e879fa96","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# DialogRPT was not working. Let us use another approach\n# I think here is more convetional, but we can extend this approach more better\n\ndef bm250_score(cxt, candidates):\n\n    # Tokenize the context and candidate responses\n    tokenized_context = word_tokenize(cxt.lower())\n    tokenized_candidates = [word_tokenize(response.lower()) for response in candidates]\n\n    # Initialize BM25 and fit it with candidate responses\n    bm25 = BM25Okapi(tokenized_candidates)\n\n    # Compute BM25 scores for the context against the candidate responses\n    scores = bm25.get_scores(tokenized_context)\n\n    # Rank candidate responses based on their scores\n    ranked_indices = scores.argsort()[::-1]\n    ranked_responses = [candidates[i] for i in ranked_indices]\n\n    # Choose positive and negative examples based on their score\n    negative = ranked_responses[-1]\n    positive = ranked_responses[0]\n\n    return positive, negative","metadata":{"id":"xi4HgIr11wLs","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Dialog dan respon-respon\ndialog = \"Apa rencana kamu untuk akhir pekan ini?\"\nresponses = [\n    \"Saya akan pergi ke pantai.\",\n    \"Mungkin saya akan menonton film.\",\n    \"Saya tidak punya rencana khusus.\"\n]\n\nbm250_score(cxt = dialog, candidates=responses)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Use another rangking metrics: sentencebert\n\ndef sbert_rangking(ctx, candidates, model):\n    \n    # Hasilkan embedding\n    dialog_embedding = model.encode(ctx, convert_to_tensor=True)\n    response_embeddings = model.encode(candidates, convert_to_tensor=True)\n    \n    # Hitung kesamaan kosinus\n    cosine_scores = util.pytorch_cos_sim(dialog_embedding, response_embeddings)\n    \n    # Konversi tensor hasil argsort menjadi daftar indeks\n    sorted_indices = cosine_scores.argsort(descending=True).tolist()[0]\n    \n    # Rangkingkan respon berdasarkan kesamaan kosinus\n    ranked_responses = [candidates[i] for i in sorted_indices]\n    \n    return ranked_responses\n\n# Dialog dan respon-respon\ndialog = \"Apa rencana kamu untuk akhir pekan ini?\"\nresponses = [\n    \"Saya akan pergi ke pantai.\",\n    \"Mungkin saya akan menonton film.\",\n    \"Saya tidak punya rencana khusus.\"\n]\n\nsbert_rangking(ctx = dialog, candidates=responses, model=sbert_model)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Fungsi untuk menghasilkan embedding\ndef get_bert_embedding(text, tokenizer, model):\n    inputs = tokenizer(text, return_tensors='pt', truncation=True, padding=True, max_length=512)\n    outputs = model(**inputs)\n    # Ambil mean dari embedding semua token untuk menghasilkan representasi kalimat\n    embedding = torch.mean(outputs.last_hidden_state, dim=1)\n    return embedding\n\n# Fungsi untuk menggabungkan hasil dari SBERT dan BERT\ndef dual_encoder_ranking(ctx, candidates, sbert_model, tokenizer_bert, model_bert):\n    # Hasilkan embedding menggunakan SBERT\n    dialog_embedding_sbert = sbert_model.encode(ctx, convert_to_tensor=True)\n    response_embeddings_sbert = sbert_model.encode(candidates, convert_to_tensor=True)\n    \n    # Hasilkan embedding menggunakan BERT\n    dialog_embedding_bert = get_bert_embedding(ctx, tokenizer_bert, model_bert)\n    response_embeddings_bert = [get_bert_embedding(response, tokenizer_bert, model_bert) for response in candidates]\n    \n    # Hitung kesamaan kosinus untuk SBERT\n    cosine_scores_sbert = util.pytorch_cos_sim(dialog_embedding_sbert, response_embeddings_sbert)\n    \n    # Hitung kesamaan kosinus untuk BERT\n    cosine_scores_bert = [F.cosine_similarity(dialog_embedding_bert, response_embedding).item() for response_embedding in response_embeddings_bert]\n    \n    # Gabungkan skor dari kedua model (bisa dengan rata-rata, atau metode lain)\n    combined_scores = [(cosine_scores_sbert[0][i].item() + cosine_scores_bert[i]) / 2 for i in range(len(candidates))]\n    \n    # Rangkingkan respon berdasarkan kesamaan kosinus gabungan\n    ranked_indices = sorted(range(len(candidates)), key=lambda i: combined_scores[i], reverse=True)\n    ranked_responses = [candidates[i] for i in ranked_indices]\n    \n    return ranked_responses\n\n# Dialog dan respon-respon\ndialog = \"Apa rencana kamu untuk akhir pekan ini?\"\nresponses = [\n    \"Saya akan pergi ke pantai.\",\n    \"Mungkin saya akan menonton film.\",\n    \"Saya tidak punya rencana khusus.\"\n]\n\n# Dapatkan peringkat respon\nranked_responses = dual_encoder_ranking(dialog, responses, sbert_model, bert_tokenizer, bert_model)\nprint(ranked_responses)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Fungsi untuk menghasilkan embedding\ndef get_bert_embedding(text, tokenizer, model):\n    inputs = tokenizer(text, return_tensors='pt', truncation=True, padding=True, max_length=512)\n    outputs = model(**inputs)\n    # Ambil mean dari embedding semua token untuk menghasilkan representasi kalimat\n    embedding = torch.mean(outputs.last_hidden_state, dim=1)\n    return embedding\n\n# Fungsi untuk menghitung Euclidean distance\ndef euclidean_distance(embedding1, embedding2):\n    return torch.norm(embedding1 - embedding2, p=2).item()\n\n# Fungsi untuk menggabungkan hasil dari SBERT dan BERT\ndef dual_encoder_ranking(ctx, candidates, sbert_model, tokenizer_bert, model_bert):\n    # Hasilkan embedding menggunakan SBERT\n    dialog_embedding_sbert = sbert_model.encode(ctx, convert_to_tensor=True)\n    response_embeddings_sbert = sbert_model.encode(candidates, convert_to_tensor=True)\n    \n    # Hasilkan embedding menggunakan BERT\n    dialog_embedding_bert = get_bert_embedding(ctx, tokenizer_bert, model_bert)\n    response_embeddings_bert = [get_bert_embedding(response, tokenizer_bert, model_bert) for response in candidates]\n    \n    # Hitung Euclidean distance untuk SBERT\n    euclidean_distances_sbert = [euclidean_distance(dialog_embedding_sbert, response_embedding) for response_embedding in response_embeddings_sbert]\n    \n    # Hitung Euclidean distance untuk BERT\n    euclidean_distances_bert = [euclidean_distance(dialog_embedding_bert, response_embedding) for response_embedding in response_embeddings_bert]\n    \n    # Gabungkan skor dari kedua model (dengan rata-rata Euclidean distance)\n    combined_distances = [(euclidean_distances_sbert[i] + euclidean_distances_bert[i]) / 2 for i in range(len(candidates))]\n    \n    # Rangkingkan respon berdasarkan Euclidean distance gabungan (semakin kecil semakin baik)\n    ranked_indices = sorted(range(len(candidates)), key=lambda i: combined_distances[i])\n    ranked_responses = [candidates[i] for i in ranked_indices]\n    \n    return ranked_responses\n\n# Dialog dan respon-respon\ndialog = \"Apa rencana kamu untuk akhir pekan ini?\"\nresponses = [\n    \"Saya akan pergi ke pantai.\",\n    \"Mungkin saya akan menonton film.\",\n    \"Saya tidak punya rencana khusus.\"\n]\n\n# Dapatkan peringkat respon\nranked_responses = dual_encoder_ranking(dialog, responses, sbert_model, bert_tokenizer, bert_model)\nprint(ranked_responses)\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Fungsi untuk menghasilkan embedding\ndef get_embedding(text, tokenizer, model):\n    inputs = tokenizer(text, return_tensors='pt', truncation=True, padding=True, max_length=512)\n    outputs = model(**inputs)\n    # Ambil mean dari embedding semua token untuk menghasilkan representasi kalimat\n    embedding = torch.mean(outputs.last_hidden_state, dim=1)\n    return embedding\n\ndef single_encoder_ranking(ctx, candidates, tokenizer_bert, model_bert):\n    \n    # Hasilkan embedding\n    dialog_embedding = get_embedding(ctx, tokenizer_bert, model_bert)\n    response_embeddings = [get_embedding(response, tokenizer_bert, model_bert) for response in candidates]\n    \n    # Hitung kesamaan kosinus\n    cosine_scores = [F.cosine_similarity(dialog_embedding, response_embedding).item() for response_embedding in response_embeddings]\n\n    # Rangkingkan respon berdasarkan kesamaan kosinus\n    ranked_indices = sorted(range(len(candidates)), key=lambda i: cosine_scores[i], reverse=True)\n    ranked_responses = [candidates[i] for i in ranked_indices]\n    \n    return ranked_responses\n    \n# Dialog dan respon-respon\ndialog = \"Apa rencana kamu untuk akhir pekan ini?\"\nresponses = [\n    \"Saya akan pergi ke pantai.\",\n    \"Mungkin saya akan menonton film.\",\n    \"Saya tidak punya rencana khusus.\"\n]\n\nsingle_encoder_ranking(dialog, responses, bert_tokenizer, bert_model)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def post_processing(generated_outputs):\n  cleaned_responses = []\n\n  # Define a pattern to remove initial unwanted phrases\n  pattern = re.compile(r'\\*\\*.*?:\\*\\*\\n\\n')\n\n  for text in generated_outputs:\n    # Remove initial unwanted phrases\n    text = re.sub(pattern, '', text)\n\n    # Find the first quotation mark and keep the text after it\n    quote_index = text.find('\"')\n    if quote_index != -1:\n        text = text[quote_index:]\n\n    cleaned_responses.append(text.strip())\n\n  return cleaned_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Prompt Ensemble","metadata":{"id":"B9lA58OIYOpc"}},{"cell_type":"markdown","source":"### Zero-shot prompt","metadata":{"id":"zDh2ueweYdo5"}},{"cell_type":"code","source":"def trim_response(generated_text):\n    position = generated_text.find(\"model\")\n    if position != -1:\n        response = generated_text[position + len(\"model\"):].strip()\n    else:\n        response = \"\"\n\n    return response","metadata":{"id":"yhY1gZ8lpEpQ","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def split_responses(generated_text):\n    # Split the text by newline characters\n    lines = generated_text.split(\"\\n\")\n\n    # Initialize an empty list to store the responses\n    responses = []\n\n    # Flag to indicate when to start capturing responses\n    capture_responses = False\n\n  # Iterate over the lines\n    for line in lines:\n      # Check if the line starts with a number followed by a period\n        if line.strip().startswith(('1.', '2.', '3.')):\n          # Set the flag to start capturing responses\n            capture_responses = True\n\n      # If the flag is set, append the line to the responses list\n        if capture_responses:\n            responses.append(line)\n\n  # Initialize an empty list to store the cleaned responses\n    cleaned_responses = []\n\n  # Iterate over each response\n    for response in responses:\n      # Remove the leading number and period, also strip any leading/trailing whitespace\n        cleaned_response = response.lstrip('1234567890. ').strip()\n      # If the cleaned response is not an empty string, append it to the list\n        if cleaned_response:\n            cleaned_responses.append(cleaned_response)\n\n    return cleaned_responses","metadata":{"id":"W7d_zCwSApET","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def zero_shot_model(prompt, model, tokenizer, max_length=200, num_return_sequences=1, k=False):\n    \"\"\"\n    if k = true, it means the prompt should return 3 examples instead of 1 examples\n    \"\"\"\n    \n    # Parameter validation\n    if not prompt:\n        raise ValueError(\"Prompt cannot be empty\")\n    if not isinstance(model, torch.nn.Module) or not hasattr(model, 'generate'):\n        raise ValueError(\"Invalid model instance\")\n    \n    # Preprocess the text to suit the gemma prompt setting\n    chat = [{ \"role\": \"user\", \"content\": prompt},]\n    prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True)\n    \n    # Encode the prompt into tokens\n    input_ids = tokenizer.encode(prompt, return_tensors='pt').to(model.device)\n\n    # Generate attention mask\n    attention_mask = torch.ones(input_ids.shape, dtype=torch.long, device=input_ids.device)\n    \n    # Generate responses\n    output = model.generate(\n        input_ids,\n        max_length=max_length,  # adjust as needed\n        attention_mask=attention_mask,\n        num_return_sequences=num_return_sequences,\n        no_repeat_ngram_size=2,\n        top_k=50,\n        top_p=0.95,\n        temperature=0.7,\n        do_sample=True\n    )\n    \n    teacher_responses = []\n    generated_texts = []\n\n    if k and num_return_sequences == 1:\n        # if true this mean we want the k equals 3, single propmt but retrun three example\n        # Decode and print the generated responses\n        generated_text = tokenizer.decode(output[0], skip_special_tokens=True)\n\n        # Trim the generated text to remove the part that duplicates the prompt\n        teacher_response = trim_response(generated_text)\n        clean_response = split_responses(teacher_response)\n        teacher_responses.extend(clean_response)\n        generated_texts.append(generated_text)\n    else:\n        # if false then k equals 3\n        if num_return_sequences > 1:\n          # if true we want produce three examples running single prompt by n return\n          for candidate in output:\n            generated_text = tokenizer.decode(candidate, skip_special_tokens=True)\n            teacher_response = trim_response(generated_text)\n            teacher_responses.append(teacher_response)\n            generated_texts.append(generated_text)\n        else:\n            # if false then we want to produce only one example\n            # Decode and print the generated responses\n            generated_text = tokenizer.decode(output[0], skip_special_tokens=True)\n            \n            # Trim the generated text to remove the part that duplicates the prompt\n            teacher_response = trim_response(generated_text)\n            teacher_responses.append(teacher_response)\n            generated_texts.append(generated_text)\n\n    return generated_texts, teacher_responses","metadata":{"id":"TLa-AT8I2-fg","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_1_teacher_reply = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nCan you give an example teacher follow-up to their previous message that would be helpful for the language learner?\nThe message should be concise, and worded simply. It should either encourage the continuation of the current topic or gracefully transition to a new teacher-provided topic.\nQuestions should be specific and not open-ended. Try to not sound like an assistant, but a teacher, in charge of the flow of the lesson.\"\"\"\n\nprint(prompt_1_teacher_reply.format(conversation=teacher_reply[0]['dialog']))","metadata":{"id":"o_F71xkuYOdW","outputId":"2a6b4397-7275-413e-f37b-e2b8e4b5a96e","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_1_teacher_continuation = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nCan you give an example teacher follow-up to their previous message that would be helpful for the language learner?\nThe message should be concise, and worded simply. It should either encourage the continuation of the current topic or gracefully transition to a new teacher-provided topic.\nQuestions should be specific and not open-ended. Try to not sound like an assistant, but a teacher, in charge of the flow of the lesson.\"\"\"\n\nprint(prompt_1_teacher_continuation.format(conversation=teacher_continuation[0]['dialog']))","metadata":{"id":"ZcAfW4AavjhX","outputId":"a0fe90b4-3f48-4360-dee1-6b9eebc51937","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nexample = prompt_1_teacher_continuation.format(conversation=teacher_continuation[1]['dialog'])\nfull_text, generated_response = zero_shot_model(prompt=example, model=model,\n                                                tokenizer=tokenizer, max_length=512)\n\n# if your propmt use k = 3 it means \"give three examples...\" then run this\n# full_text, generated_response = zero_shot_model(prompt=prompt_1_teacher_reply,\n#                                                 model=model, tokenizer=tokenizer,\n#                                                 max_length=512, k=True)","metadata":{"id":"bfkjmkG42_tR","outputId":"3359dd14-eb71-46ab-89c4-4b66257da66b","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"generated_response","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import time\n\n# Prompting teacher_reply data\noutputs_reply_data_prompt_1 = []\ntotal_iterations = len(teacher_reply)\n\nprint(f\"Total data yang akan diproses: {total_iterations}\")\n\nfor idx, data in enumerate(teacher_reply):\n    if idx % 50 == 0 and idx != 0:\n        print(f\"Checkpoint ke-{idx}. Berhenti sejenak selama 1 menit...\")\n        print('Memory Usage:')\n        print('Allocated:', round(torch.cuda.memory_allocated(0)/1024**3,1), 'GB')\n        print('Cached:   ', round(torch.cuda.memory_reserved(0)/1024**3,1), 'GB')\n        time.sleep(60)  # Jeda selama 1 menit\n    \n    # Menggunakan \\r untuk meng-overwrite baris yang sama\n    print(f\"Memproses data ke-{idx+1}/{total_iterations}\", end='\\r')\n    \n    id_data = data[\"id\"]\n    example = prompt_1_teacher_reply.format(conversation=data['dialog'])\n    full_text, generated_response = zero_shot_model(prompt=example, model=model,\n                                                    tokenizer=tokenizer, max_length=512)\n    \n    cleaned_output = post_processing(generated_response)\n    prepare_data = {\"id\": id_data, \"text\": cleaned_output[0], \"speaker\": \"teacher\"}\n    outputs_reply_data_prompt_1.append(prepare_data)\n    \n    # Pastikan baris berikutnya menimpa baris yang sama pada terminal\n    print(f\"Data ke-{idx+1} selesai diproses.\", end='\\r')\n\n# Tambahkan baris kosong untuk menghindari menimpa pesan terakhir\nprint(\"\\nSemua data telah selesai diproses.\")","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Define the output file path\noutput_file_path = 'outputs_reply_data_prompt_1.jsonl'\n\n# Write the data to the JSONL file\nwith open(output_file_path, 'w') as file:\n    for entry in outputs_reply_data_prompt_1:\n        json_line = json.dumps(entry)\n        file.write(json_line + '\\n')\n\nprint(f\"Data successfully saved to {output_file_path}\")\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(outputs_reply_data_prompt_1[0])","metadata":{"id":"0C7t_PiwcwBh","outputId":"03cd2d61-7be0-422a-998f-5ee8ca88ce30","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Few-shot prompt","metadata":{"id":"jyXbEpgUwJN9"}},{"cell_type":"markdown","source":"#### Handcrafted Examples --> we can manualy create our own written examples mba marina and mba alma :)","metadata":{"id":"VpdzBQhdwb7O"}},{"cell_type":"code","source":"prompt_2_teacher_reply = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nCan you give an example teacher follow-up to their previous message that would be helpful for the language learner?\nThe message should be concise, and worded simply. It should either encourage the continuation of the current topic or gracefully transition to a new teacher-provided topic.\nQuestions should be specific and not open-ended. Try to not sound like an assistant, but a teacher, in charge of the flow of the lesson.\n\nGood example: 'Can you make a sentence using 'within'?'\nBad example: 'Do you have any questions about prepositions?'\"\"\"\n\nprint(prompt_2_teacher_reply.format(conversation=teacher_reply[0]['dialog']))","metadata":{"id":"fxjnCTKkxdQn","outputId":"6409a1c5-a6b7-4ded-b075-eeab47279b29","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_2_teacher_continuation = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nCan you give an example teacher follow-up to their previous message that would be helpful for the language learner?\nThe message should be concise, and worded simply. It should either encourage the continuation of the current topic or gracefully transition to a new teacher-provided topic.\nQuestions should be specific and not open-ended. Try to not sound like an assistant, but a teacher, in charge of the flow of the lesson.\n\nGood example: 'Can you make a sentence using 'within'?'\nBad example: 'Do you have any questions about prepositions?'\"\"\"\n\nprint(prompt_2_teacher_continuation.format(conversation=teacher_continuation[3]['dialog']))","metadata":{"id":"fMbUw-Q70O02","outputId":"0b99e671-a0b1-4c80-8cda-1911c625c088","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nexample = prompt_2_teacher_continuation.format(conversation=teacher_continuation[3]['dialog'])\nfull_text, generated_response = zero_shot_model(prompt=example, model=model,\n                                                tokenizer=tokenizer, max_length=512)","metadata":{"id":"gfBXlZPK0fb9","outputId":"86d2fffe-9a87-43a7-9d26-f5909f57308a"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(generated_response[0])","metadata":{"id":"LZnr2scSjwxA","outputId":"951dad74-68d4-4457-be40-4939d8189f73"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for text in generated_response:\n    result = re.sub(r'\\*\\*Example following up message:\\*\\*\\n\\n', '', text)\n\nprint(result)","metadata":{"id":"YjMTd8c30ihP","outputId":"1930aa9c-6387-46fc-f963-fb103c22ae96"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Generated Examples","metadata":{"id":"ck6y7EPHwadS"}},{"cell_type":"code","source":"prompt_4_teacher_reply = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nThey are in the middle of a lesson. Can you give a possible way the teacher could respond?\n\nRemember: A teacher typically sounds knowledgeable, authoritative, and focused on guiding and instructing students.\nThey may use formal language and provide detailed explanations. Teachers often offer constructive feedback, encourage critical thinking, and ask probing questions to stimulate learning.\n\nExample of a teacher-like response: \"That's a great observation, but let's delve deeper into the topic. Can you provide some evidence to support your claim?\"\n\nA chatbot, on the other hand, may sound more informal and conversational. It tends to provide general information or brief responses without much elaboration.\n\nExample of a chatbot-like response: \"Interesting! Tell me more.\"\n\nTeachers typically avoid expressing personal opinions or biases. They also refrain from engaging in casual banter or unrelated conversations to maintain a professional and educational atmosphere.\"\"\"\n\nprint(prompt_4_teacher_reply.format(conversation=teacher_reply[0]['dialog']))","metadata":{"id":"nLBb097UvjcE","outputId":"da7e5d5a-8f48-4086-e197-94df26f87f96","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_4_teacher_continuation = \"\"\"The following is a partial conversation between an English language learner and their teacher.\n\n{conversation}\n\nThey are in the middle of a lesson. Can you give a possible way the teacher could respond?\n\nRemember: A teacher typically sounds knowledgeable, authoritative, and focused on guiding and instructing students.\nThey may use formal language and provide detailed explanations. Teachers often offer constructive feedback, encourage critical thinking, and ask probing questions to stimulate learning.\n\nExample of a teacher-like response: \"That's a great observation, but let's delve deeper into the topic. Can you provide some evidence to support your claim?\"\n\nA chatbot, on the other hand, may sound more informal and conversational. It tends to provide general information or brief responses without much elaboration.\n\nExample of a chatbot-like response: \"Interesting! Tell me more.\"\n\nTeachers typically avoid expressing personal opinions or biases. They also refrain from engaging in casual banter or unrelated conversations to maintain a professional and educational atmosphere.\"\"\"\n\nprint(prompt_4_teacher_continuation.format(conversation=teacher_continuation[5]['dialog']))","metadata":{"id":"mj0CsFZevjZ2","outputId":"46cc281b-2799-4683-d581-f3920b2f9943","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nexample = prompt_4_teacher_continuation.format(conversation=teacher_continuation[5]['dialog'])\nfull_text, generated_response = zero_shot_model(prompt=example, model=model,\n                                                tokenizer=tokenizer, max_length=512)","metadata":{"id":"8-R00axof5SG"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"generated_response[0]","metadata":{"id":"7JTjyT6wf-ZH","outputId":"c553a196-3611-4fb9-96a9-1e5544c881cc"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"post_processing(generated_response)","metadata":{"id":"vX9sONmowPp3","outputId":"c2a0885c-f3c2-4e95-e1f8-c2634cea8323"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for text in generated_response:\n    result = re.sub(r'A possible response from the English teacher might be:\\n\\n', '', text)\n\nprint(result)","metadata":{"id":"WxRJswDFgavP","outputId":"7d258854-be3c-4cda-b908-9d61dc3aee7e"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Iterative Examples","metadata":{"id":"ym3mTeEt10RU"}},{"cell_type":"markdown","source":"1. Used the zero-shot prompt (1) to generate responses for all conversations in the data","metadata":{"id":"Wy0ko0d62N6U"}},{"cell_type":"code","source":"example = prompt_1_teacher_continuation.format(conversation=teacher_continuation[11]['dialog'])\nfull_text, generated_response = zero_shot_model(prompt=example, model=model,\n                                                tokenizer=tokenizer, max_length=512,\n                                                num_return_sequences=5)","metadata":{"id":"JYamWUrYvjXY","outputId":"8b90e84c-4882-4d4f-d730-5d0b491feca3"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(generated_response)","metadata":{"id":"MPMe0vrJhu_-","outputId":"b4fb716c-8d44-42f2-94af-f87ce204e7a6"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"result = []\nfor text in generated_response:\n    cleaned_text = re.sub(r'\\*\\*Teacher\\' s Response:\\*\\*\\n\\n', '', text)\n    result.append(cleaned_text)\n\nresult","metadata":{"id":"sBu4J6no2ZWy","outputId":"d53360b5-4efb-42f6-c544-70cbe60300f3"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"2. Used dialogRPT to score the responses and selected the highest and lowest scorering responses as the positive and negative examples","metadata":{"id":"2yZx9z2r99bI"}},{"cell_type":"code","source":"positive, negative = bm250_score(example, result)","metadata":{"id":"qcA-EOhNQ4Cl"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_5_teacher_reply = \"\"\"{prompt_1_teacher_reply}\n\nHere is an example of an exceptional teacher follow-up:\n{positive}\n\nHere is an example of a poor teacher follow-up:\n{negative}\"\"\"\n\n# print(prompt_5_teacher_reply.format(prompt_1_teacher_reply=example,\n#                                     positive=positive, negative=negative))","metadata":{"id":"si6O6TTW2Eva","outputId":"4a4c7b5c-c057-4a7c-aac0-79984996670b","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_5_teacher_continuation = \"\"\"{prompt_1_teacher_continuation}\n\nHere is an example of an exceptional teacher follow-up:\n{positive}\n\nHere is an example of a poor teacher follow-up:\n{negative}\"\"\"\n\n# print(prompt_5_teacher_continuation.format(prompt_1_teacher_continuation=example,\n#                                     positive=positive, negative=negative))","metadata":{"id":"cbJZkC3bxDau","outputId":"cdfe5814-65fd-4686-8628-31f9202c3d3b","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nexample = prompt_5_teacher_continuation.format(prompt_1_teacher_continuation=example, positive=positive, negative=negative)\nfull_text, generated_response = zero_shot_model(prompt=example,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=512)","metadata":{"id":"4vNi5TmeGHqc","outputId":"89133bb8-841d-419e-e19f-f14829d3fb74"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(generated_response[0])","metadata":{"id":"bxWaQZRWGelE","outputId":"08675286-298c-45c9-d665-45bce38b2630"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"post_processing(generated_response)","metadata":{"id":"0F6-hkRKxZHb","outputId":"2326f22e-7023-4865-c932-2cd8f5d716af"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for text in generated_response:\n    cleaned_text = re.sub(r'\\*\\*Example 1:\\*\\*\\n\\n', '', text)\n\ncleaned_text","metadata":{"id":"IDb1ojfu5-zY","outputId":"6aa08433-9255-4e78-9a9b-0d955d0678bd"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Teacher replies vs continuation prompt ~ i think we dont need this one","metadata":{"id":"gXvewezD7_LX"}},{"cell_type":"markdown","source":"#### Teacher reply","metadata":{"id":"InlKsaHUei1g"}},{"cell_type":"code","source":"prompt_6_teacher_reply = \"\"\"Here is a partial conversation between a student and their teacher during a private English lesson:\n\n{conversation}\n\nCan you give three possible ways the teacher could respond to continue the lesson? Use Simple English. While the conversation might be about culture or other topics, the point is to practice English\nEach teacher response should:\n1. Acknowledge what the student said, and demonstrate understanding.\n2. Be helpful to the student, without answering directly. Give hints to help the student think for themselves.\n3. Encourage the student to respond with an exercise or question. Respond without preamble, just number them.\"\"\"\n\nprint(prompt_6_teacher_reply.format(conversation=teacher_reply[25]['dialog']))","metadata":{"id":"2jjmVcVO2EqN","outputId":"a94fad52-2cc9-4874-bb09-522b3a9e03a4","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nexample = prompt_6_teacher_reply.format(conversation=teacher_reply[25]['dialog'])\nfull_text, generated_response = zero_shot_model(prompt=example,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=512)","metadata":{"id":"FqHtYoLhgMTA","outputId":"37ee36ae-e73f-4631-b640-14af603aea58"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(generated_response[0])","metadata":{"id":"M3Rqu1Ri6P1K","outputId":"f120ad08-5572-4149-858b-61d513381bfc"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_7_teacher_reply = \"\"\"The following is a partial live chat between a teacher and a student learning English. They are in the middle of a lesson.\n\nCan you provide 3 possible ways the teacher could wrap up the current conversation and start an exercise or new topic of discussion?\n\nRemember: Teachers often use specific language and techniques that chat bots have difficulty replicating.\nFor example, they may ask open-ended questions to encourage critical thinking and engagement, provide specific feedback on a student’s work, or offer personalized guidance based on a student’s strengths and weaknesses.\nTeachers have a specific agenda for each lesson, such as practicing a specific grammar point or vocabulary. Try to understand what’s happening in the conversation and what the teacher’s goal is for the lesson.\n\nIf the goal is unclear, you can assume that the teacher wants to move on to a new topic or exercise.\n\nDon't ask questions that are overly general such as \"Is there anything else you’d like to talk about?\"\n\nHere is the conversation so far:\n{conversation}\"\"\"\n\nprint(prompt_7_teacher_reply.format(conversation=teacher_reply[2]['dialog']))","metadata":{"id":"dRVP0APigMPM","outputId":"e4ba7023-b324-4e2b-d89f-699219b5bae4"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_7_teacher_reply,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"i5OB76eAjBSk","outputId":"b7a40fba-565e-4ac7-e1cf-a55917178be5"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Teacher continuation","metadata":{"id":"4wjjsmN8j3Nn"}},{"cell_type":"code","source":"prompt_8_teacher_continuation = f\"\"\"Here is a partial conversation between a student and their teacher during a private English lesson:\n\n{teacher_continuation[2]['dialog']}\n\nCan you give three possible ways the teacher could continue their response? Use simple English.\"\"\"\n\nprint(prompt_8_teacher_continuation)","metadata":{"id":"41uPJiktjJZt","outputId":"8bf2bff9-023d-4298-a241-85955b65dc2f"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_8_teacher_continuation,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"MQY5LOBvqFJ3","outputId":"ccc912a1-eac8-4085-f227-0ab0b16a631d"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_9_teacher_continuation = f\"\"\"The following is a partial live chat between a teacher and a student learning English. They are in the middle of a lesson, and the teacher has already sent a message, but wants to follow-up.\nThere could be various reasons why the teacher wants to follow-up on their previous message before the student has a chance to reply. Here are some possibilities:\n\n1. The teacher may have realized that their previous message contained some inaccuracies or omissions, and they want to correct or clarify their statement to avoid confusion.\n2. The teacher may have received new information or thought of a better way to explain something, and they want to add to their previous message to provide a more complete answer.\n3. The teacher may want to check if the student has any further questions or needs more explanation on the topic, and they want to encourage further discussion by sending a follow-up message.\n\nRegardless of the reason, the teacher’s follow-up message can help ensure that the student fully understands the topic being discussed and feels comfortable asking questions and engaging in the conversation.\n\nCan you provide 3 possible follow-up messages the teacher could write?\n\nUse simple English. The response should sound like a teacher, not an assistant. Good example: 'Can you make a sentence using 'within'? Bad example: 'Do you have any questions about prepositions?'.\nThe response should be helpful for the student and show that the teacher understood the student.\nHere is the conversation so far:\n\n{teacher_continuation[2]['dialog']}\"\"\"\n\nprint(prompt_9_teacher_continuation)","metadata":{"id":"TMJeBzoAmMCl","outputId":"4908900e-bffc-49ed-ac83-7efc8224bb6b"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_9_teacher_continuation,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"NUiLoTUvqTyG","outputId":"159945eb-5978-4c66-bf2c-d77aeac85e18"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_10_teacher_continuation = f\"\"\"The following is a partial live chat between a teacher and a student learning English. They are in the middle of a lesson, and the teacher has already sent a message, but wants to follow-up.\nThere could be various reasons why the teacher wants to follow-up on their previous message before the student has a chance to reply. Here are some possibilities:\n\n1. The teacher may have realized that their previous message contained some inaccuracies or omissions, and they want to correct or clarify their statement to avoid confusion.\n2. The teacher may have received new information or thought of a better way to explain something, and they want to add to their previous message to provide a more complete answer.\n3. The teacher may want to check if the student has any further questions or needs more explanation on the topic, and they want to encourage further discussion by sending a follow-up message.\n\nRegardless of the reason, the teacher’s follow-up message can help ensure that the student fully understands the topic being discussed and feels comfortable asking questions and engaging in the conversation.\n\nCan you provide 3 possible follow-up messages the teacher could write?\n\nUse simple English. The response should sound like a teacher, not an assistant. Good example: 'Can you make a sentence using 'within'?' Bad example: 'Do you have any questions about prepositions?'.\nThe response should be helpful for the student and show that the teacher understood the student.\nHere is the conversation so far:\n\n{teacher_continuation[2]['dialog']}\"\"\"\n\nprint(prompt_10_teacher_continuation)","metadata":{"id":"Io1iJOg_qTu5","outputId":"06e127fb-3306-4197-b361-d8f0d17e36a5"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_10_teacher_continuation,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"5wzwYQLZYUF4","outputId":"0d6cdd33-4861-439d-84f1-1e459cf7b656"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Specific teacher continuation","metadata":{"id":"lofgjw4OX0ec"}},{"cell_type":"code","source":"prompt_11_teacher_continuation = f\"\"\"Here is a partial conversation between an English student and their teacher:\n\n{teacher_continuation[2]['dialog']}\n\nIn the last utterance, did the teacher ask a question? If not, please provide one that would be appropriate. If they were in the middle of an exercise, what should they say to continue the exercise? The question or prompt should be simple.\nDon’t be too verbose or open ended. Good example: \"What else is 'surprising'?\" Bad example: \"Is there anything else you'd like to know?\"\n\nRespond in the following format:\nTeacher asked a question: (yes/no)\nQuestion or prompt:\"\"\"\n\nprint(prompt_11_teacher_continuation)","metadata":{"id":"PoY7VWzQqTpm","outputId":"a141da35-5cfd-4ef5-adb7-a06b4714a478"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_11_teacher_continuation,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"q2kZFoB-YkW7","outputId":"840b97c5-3616-41f4-f6c9-ec87f1f33999"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_12_teacher_continuation = f\"\"\"Here is a partial conversation between an English student and their teacher:\n\n{teacher_continuation[2]['dialog']}\n\nWere they in the middle of a conversation? If so, what should the teacher say to continue the conversation? The question or prompt should be simple and not use terminology such as 'collocations'.\nDon't be too verbose or open ended. Good example: \"What else is 'surprising'?\" Bad example: \"Is there anything else you'd like to know?\"\n\nRespond in the following format:\nConversation: yes or no\nTeacher:\"\"\"\n\nprint(prompt_12_teacher_continuation)","metadata":{"id":"4oI2qsw-YkUV","outputId":"9d8c7c69-e625-453e-9a8f-5fe91e1bfedd"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# if your prompt use single k such as \"give an example...\" then run this\nfull_text, generated_response = zero_shot_model(prompt=prompt_12_teacher_continuation,\n                                                model=model, tokenizer=tokenizer,\n                                                max_length=520)\n\nprint(generated_response[0])","metadata":{"id":"Y3Z07ZXDZaoF","outputId":"1595bd9b-2cef-4bb2-ea05-73668e8bfc18"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## NaisGemma\n\nThere are two sub-tasks, teacher reply and teacher continuation","metadata":{"id":"RrG6GxngZ_U0"}},{"cell_type":"code","source":"# def post_processing(generated_outputs):\n#   cleaned_responses = []\n#   for text in generated_outputs:\n#     text = re.sub(r'\\*\\*Teacher:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Teacher\\' s Response:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Teacher follow up:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Teacher Follow-Up:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Good Example:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Example following up message:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Teacher-Like Response:\\*\\*\\n\\n', '', text)\n#     text = re.sub(r'A possible response from the English teacher might be:\\n\\n', '', text)\n#     text = re.sub(r'\\*\\*Example 1:\\*\\*\\n\\n', '', text)\n#     cleaned_responses.append(text)\n\n#   return cleaned_responses\n\ndef post_processing(generated_outputs):\n  cleaned_responses = []\n\n  # Define a pattern to remove initial unwanted phrases\n  pattern = re.compile(r'\\*\\*.*?:\\*\\*\\n\\n')\n\n  for text in generated_outputs:\n    # Remove initial unwanted phrases\n    text = re.sub(pattern, '', text)\n\n    # Find the first quotation mark and keep the text after it\n    quote_index = text.find('\"')\n    if quote_index != -1:\n        text = text[quote_index:]\n\n    cleaned_responses.append(text.strip())\n\n  return cleaned_responses","metadata":{"id":"LSG8ZluKzGTr","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def naissgemma(model, tokenizer, conversation, prompts, ranking=\"sbert\", max_length=512):\n    \n    # step 1: prepare promt + conversation\n    prompt_with_conversation = []\n    \n    for i, prompt in enumerate(prompts):\n        if i == 3:\n            # do prompt 5 and return n sequences\n            _, generated_response = zero_shot_model(prompt=prompt_with_conversation[0],\n                                              model=model, tokenizer=tokenizer,\n                                              max_length=max_length, num_return_sequences=5)\n            \n            # post processing\n            cleaned_responses = post_processing(generated_response)\n            positive, negative = bm250_score(conversation, cleaned_responses)\n            text = prompt.format(prompt_1_teacher_reply=prompt_with_conversation[0],\n                                 positive=positive, negative=negative)\n        else:\n            text = prompt.format(conversation=conversation)\n        prompt_with_conversation.append(text)\n\n    # step 2: input ensemble prompt to model\n    outputs = []\n    for conversation in prompt_with_conversation:\n        # make sure model retrun best response\n        for i in range(3):\n            _, generated_response = zero_shot_model(prompt=conversation, model=model,\n                                            tokenizer=tokenizer, max_length=max_length)\n        outputs.extend(generated_response)\n    \n    # step 3: post processing the outputs\n    cleaned_responses = post_processing(outputs)\n    \n    # step 4: rerank the generated teacher output\n    if ranking == \"sbert\":\n        ranked_responses = sbert_rangking(ctx = conversation, candidates=cleaned_responses, model=sbert_model)\n    elif ranking == \"dual_encoder\":\n        ranked_responses = dual_encoder_ranking(ctx=conversation, candidates=cleaned_responses, \n                                                sbert_model=sbert_model, tokenizer_bert=bert_tokenizer, \n                                                model_bert=bert_model)\n    elif ranking == \"cosinus\":\n        ranked_responses = single_encoder_ranking(ctx=conversation, candidates=cleaned_responses, \n                                                  tokenizer_bert=bert_tokenizer, model_bert=bert_model)\n    elif ranking == \"bm205\":\n        ranked_responses, _ = bm250_score(cxt = dialog, candidates=responses)\n    else:\n        print(\"Not found ranking method\")\n        \n    \n    return ranked_responses[0], cleaned_responses","metadata":{"id":"Ur1ekdmpzMc8","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_reply_ensembles = [\n    prompt_1_teacher_reply,\n    prompt_2_teacher_reply,\n    prompt_4_teacher_reply,\n    prompt_5_teacher_reply,\n]\n\nprompt_continuation_ensembles = [\n    prompt_1_teacher_continuation,\n    prompt_2_teacher_continuation,\n    prompt_4_teacher_continuation,\n    prompt_5_teacher_continuation,\n]","metadata":{"id":"BWGgomlTbCKT","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"prompt_reply_ensembles[2]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# only for teacher reply task\nbest_response, outputs = naissgemma(model=model, tokenizer=tokenizer, ranking=\"sbert\",\n                                    conversation=teacher_reply[0][\"dialog\"],\n                                    prompts=prompt_reply_ensembles, max_length=512)\nprint(best_response)","metadata":{"id":"r4-_0-Opr9xq","outputId":"eab44d89-b9c4-4ae6-bb11-3a15d6ff2ef6","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for idx, (reply, continuation) in enumerate(zip(teacher_reply[:6], teacher_continuation[:6])):\n  print(reply['response'])\n  print(continuation[\"dialog\"])\n  print(idx)\n  break","metadata":{"id":"pUGjRQvX0kha","outputId":"9bb3c8fb-058d-4278-a3d6-e41555c6513f"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import time","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def forward(teacher_reply_data, teacher_continuation_data):\n  best_responses = []\n  references_responses = []\n  total_iterations_reply_data = len(teacher_reply_data)\n  total_iterations_continuation_data = len(teacher_continuation_data)\n  candidates = []\n    \n    \n  for idx, data in enumerate(teacher_reply_data):\n    if idx % 20 == 0 and idx != 0:\n        print(f\"Checkpoint ke-{idx}. Berhenti sejenak selama 1 menit...\")\n        print('Memory Usage:')\n        print('Allocated:', round(torch.cuda.memory_allocated(0)/1024**3,1), 'GB')\n        print('Cached:   ', round(torch.cuda.memory_reserved(0)/1024**3,1), 'GB')\n        time.sleep(60)  # Jeda selama 1 menit\n    \n    # Menggunakan \\r untuk meng-overwrite baris yang sama\n    print(f\"Memproses reply data ke-{idx+1}/{total_iterations_reply_data}\", end='\\r')\n    id_data = data[\"id\"]\n    best_response, outputs = naissgemma(model=model, tokenizer=tokenizer,\n                                        conversation=data[\"dialog\"], ranking=\"dual_encoder\",\n                                        prompts=prompt_reply_ensembles, max_length=512)\n    \n    prepare_data = {\"id\": id_data, \"text\": best_response, \"speaker\": \"teacher\"}\n    best_responses.append(prepare_data)\n    references_responses.append(data[\"response\"])\n    candidates.append(outputs)\n    \n    # Pastikan baris berikutnya menimpa baris yang sama pada terminal\n    print(f\"Data ke-{idx+1} selesai diproses.\", end='\\r')\n\n  for idx, data in enumerate(teacher_continuation_data):\n    if idx % 50 == 0 and idx != 0:\n        print(f\"Checkpoint ke-{idx}. Berhenti sejenak selama 1 menit...\")\n        print('Memory Usage:')\n        print('Allocated:', round(torch.cuda.memory_allocated(0)/1024**3,1), 'GB')\n        print('Cached:   ', round(torch.cuda.memory_reserved(0)/1024**3,1), 'GB')\n        time.sleep(60)  # Jeda selama 1 menit\n    \n    # Menggunakan \\r untuk meng-overwrite baris yang sama\n    print(f\"Memproses continues data ke-{idx+1}/{total_iterations_continuation_data}\", end='\\r')\n    id_data = data[\"id\"]\n    best_response, outputs = naissgemma(model=model, tokenizer=tokenizer,\n                                        conversation=data[\"dialog\"], ranking=\"dual_encoder\",\n                                        prompts=prompt_reply_ensembles, max_length=512)\n    \n    prepare_data = {\"id\": id_data, \"text\": best_response, \"speaker\": \"teacher\"}\n    best_responses.append(prepare_data)\n    references_responses.append(data[\"response\"])\n    candidates.append(outputs)\n    \n    # Pastikan baris berikutnya menimpa baris yang sama pada terminal\n    print(f\"Data ke-{idx+1} selesai diproses.\", end='\\r')\n\n  return best_responses, references_responses, candidates","metadata":{"id":"C9gam546FHW0","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# # feed all data to system (it will takes time fyi)\n# best_responses = []\n# references_responses = []\n# for idx, (reply, continuation) in enumerate(zip(teacher_reply[:6], teacher_continuation[:6])):\n#   if idx < 3:\n#     print(\"===========Menggunakan teacher reply dataset==========\")\n#     print(f\"==============Feed data ke {idx}===========\")\n#     best_response, outputs = naissgemma(model=model, tokenizer=tokenizer,\n#                                         conversation=reply[\"dialog\"],\n#                                         prompts=prompt_reply_ensembles, max_length=512)\n#     best_responses.append(best_response)\n#     references_responses.append(reply[\"response\"])\n#   else:\n#     print(\"===========Menggunakan teacher continuation dataset==========\")\n#     print(f\"==============Feed data ke {idx}===========\")\n#     best_response, outputs = naissgemma(model=model, tokenizer=tokenizer,\n#                                         conversation=continuation[\"dialog\"],\n#                                         prompts=prompt_reply_ensembles, max_length=512)\n#     best_responses.append(best_response)\n#     references_responses.append(continuation[\"response\"])\n\nbest_responses, references_responses, candidates = forward(teacher_reply[:5], teacher_continuation[:5])","metadata":{"id":"7LkrDjRpmj4g","outputId":"b681958d-d91a-4592-b733-93e54bb82f50","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"references_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"best_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Define the output file path\noutput_file_path = 'candidates_evaluation.jsonl'\n\n# Write the data to the JSONL file\nwith open(output_file_path, 'w') as file:\n    for entry in candidates:\n        json_line = json.dumps(entry)\n        file.write(json_line + '\\n')\n\nprint(f\"Data successfully saved to {output_file_path}\")\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"teacher_reply[91]['id']","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"candidates[0]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"best_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Define the output file path\noutput_file_path = 'outputs_data_100.jsonl'\n\n# Write the data to the JSONL file\nwith open(output_file_path, 'w') as file:\n    for entry in best_responses:\n        json_line = json.dumps(entry)\n        file.write(json_line + '\\n')\n\nprint(f\"Data successfully saved to {output_file_path}\")\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"best_responses","metadata":{"id":"BfHgvSDhq_xv","outputId":"dcb32912-3a6b-4792-ee8c-f67efd3b08c7","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"references_responses","metadata":{"id":"ubC0DMpJrQe8","outputId":"50c2c06f-9b04-4c55-9070-26e9a6237c75","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Experiments","metadata":{}},{"cell_type":"code","source":"# ","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Evaluation","metadata":{"id":"FYAFlgyJNeV0"}},{"cell_type":"code","source":"# !pip install bert_score\n!pip install evaluate\n!pip install git+https://github.com/google-research/bleurt.git\nimport evaluate\nfrom evaluate import load\n# bertscore = load(\"bertscore\")","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:22:42.682040Z","iopub.execute_input":"2024-06-25T03:22:42.682581Z","iopub.status.idle":"2024-06-25T03:23:35.075874Z","shell.execute_reply.started":"2024-06-25T03:22:42.682551Z","shell.execute_reply":"2024-06-25T03:23:35.075007Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"Collecting evaluate\n  Downloading evaluate-0.4.2-py3-none-any.whl.metadata (9.3 kB)\nRequirement already satisfied: datasets>=2.0.0 in /opt/conda/lib/python3.10/site-packages (from evaluate) (2.19.2)\nRequirement already satisfied: numpy>=1.17 in /opt/conda/lib/python3.10/site-packages (from evaluate) (1.26.4)\nRequirement already satisfied: dill in /opt/conda/lib/python3.10/site-packages (from evaluate) (0.3.8)\nRequirement already satisfied: pandas in /opt/conda/lib/python3.10/site-packages (from evaluate) (2.2.1)\nRequirement already satisfied: requests>=2.19.0 in /opt/conda/lib/python3.10/site-packages (from evaluate) (2.32.3)\nRequirement already satisfied: tqdm>=4.62.1 in /opt/conda/lib/python3.10/site-packages (from evaluate) (4.66.4)\nRequirement already satisfied: xxhash in /opt/conda/lib/python3.10/site-packages (from evaluate) (3.4.1)\nRequirement already satisfied: multiprocess in /opt/conda/lib/python3.10/site-packages (from evaluate) (0.70.16)\nRequirement already satisfied: fsspec>=2021.05.0 in /opt/conda/lib/python3.10/site-packages (from fsspec[http]>=2021.05.0->evaluate) (2024.3.1)\nRequirement already satisfied: huggingface-hub>=0.7.0 in /opt/conda/lib/python3.10/site-packages (from evaluate) (0.23.2)\nRequirement already satisfied: packaging in /opt/conda/lib/python3.10/site-packages (from evaluate) (21.3)\nRequirement already satisfied: filelock in /opt/conda/lib/python3.10/site-packages (from datasets>=2.0.0->evaluate) (3.13.1)\nRequirement already satisfied: pyarrow>=12.0.0 in /opt/conda/lib/python3.10/site-packages (from datasets>=2.0.0->evaluate) (14.0.2)\nRequirement already satisfied: pyarrow-hotfix in /opt/conda/lib/python3.10/site-packages (from datasets>=2.0.0->evaluate) (0.6)\nRequirement already satisfied: aiohttp in /opt/conda/lib/python3.10/site-packages (from datasets>=2.0.0->evaluate) (3.9.1)\nRequirement already satisfied: pyyaml>=5.1 in /opt/conda/lib/python3.10/site-packages (from datasets>=2.0.0->evaluate) (6.0.1)\nRequirement already satisfied: typing-extensions>=3.7.4.3 in /opt/conda/lib/python3.10/site-packages (from huggingface-hub>=0.7.0->evaluate) (4.9.0)\nRequirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/lib/python3.10/site-packages (from packaging->evaluate) (3.1.1)\nRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests>=2.19.0->evaluate) (3.3.2)\nRequirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests>=2.19.0->evaluate) (3.6)\nRequirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests>=2.19.0->evaluate) (1.26.18)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests>=2.19.0->evaluate) (2024.2.2)\nRequirement already satisfied: python-dateutil>=2.8.2 in /opt/conda/lib/python3.10/site-packages (from pandas->evaluate) (2.9.0.post0)\nRequirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.10/site-packages (from pandas->evaluate) (2023.3.post1)\nRequirement already satisfied: tzdata>=2022.7 in /opt/conda/lib/python3.10/site-packages (from pandas->evaluate) (2023.4)\nRequirement already satisfied: attrs>=17.3.0 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (23.2.0)\nRequirement already satisfied: multidict<7.0,>=4.5 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (6.0.4)\nRequirement already satisfied: yarl<2.0,>=1.0 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.9.3)\nRequirement already satisfied: frozenlist>=1.1.1 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.4.1)\nRequirement already satisfied: aiosignal>=1.1.2 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (1.3.1)\nRequirement already satisfied: async-timeout<5.0,>=4.0 in /opt/conda/lib/python3.10/site-packages (from aiohttp->datasets>=2.0.0->evaluate) (4.0.3)\nRequirement already satisfied: six>=1.5 in /opt/conda/lib/python3.10/site-packages (from python-dateutil>=2.8.2->pandas->evaluate) (1.16.0)\nDownloading evaluate-0.4.2-py3-none-any.whl (84 kB)\n\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m84.1/84.1 kB\u001b[0m \u001b[31m2.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[?25hInstalling collected packages: evaluate\nSuccessfully installed evaluate-0.4.2\nCollecting git+https://github.com/google-research/bleurt.git\n  Cloning https://github.com/google-research/bleurt.git to /tmp/pip-req-build-eokmuh0m\n  Running command git clone --filter=blob:none --quiet https://github.com/google-research/bleurt.git /tmp/pip-req-build-eokmuh0m\n  Resolved https://github.com/google-research/bleurt.git to commit cebe7e6f996b40910cfaa520a63db47807e3bf5c\n  Preparing metadata (setup.py) ... \u001b[?25ldone\n\u001b[?25hRequirement already satisfied: pandas in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (2.2.1)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (1.26.4)\nRequirement already satisfied: scipy in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (1.11.4)\nRequirement already satisfied: tensorflow in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (2.15.0)\nCollecting tf-slim>=1.1 (from BLEURT==0.0.2)\n  Downloading tf_slim-1.1.0-py2.py3-none-any.whl.metadata (1.6 kB)\nRequirement already satisfied: sentencepiece in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: absl-py>=0.2.2 in /opt/conda/lib/python3.10/site-packages (from tf-slim>=1.1->BLEURT==0.0.2) (1.4.0)\nRequirement already satisfied: python-dateutil>=2.8.2 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2.9.0.post0)\nRequirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2023.3.post1)\nRequirement already satisfied: tzdata>=2022.7 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2023.4)\nRequirement already satisfied: astunparse>=1.6.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.6.3)\nRequirement already satisfied: flatbuffers>=23.5.26 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (23.5.26)\nRequirement already satisfied: gast!=0.5.0,!=0.5.1,!=0.5.2,>=0.2.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.5.4)\nRequirement already satisfied: google-pasta>=0.1.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: h5py>=2.9.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.10.0)\nRequirement already satisfied: libclang>=13.0.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (16.0.6)\nRequirement already satisfied: ml-dtypes~=0.2.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: opt-einsum>=2.3.2 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.3.0)\nRequirement already satisfied: packaging in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (21.3)\nRequirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.20.3)\nRequirement already satisfied: setuptools in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (69.0.3)\nRequirement already satisfied: six>=1.12.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.16.0)\nRequirement already satisfied: termcolor>=1.1.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.4.0)\nRequirement already satisfied: typing-extensions>=3.6.6 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (4.9.0)\nRequirement already satisfied: wrapt<1.15,>=1.11.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.14.1)\nRequirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.35.0)\nRequirement already satisfied: grpcio<2.0,>=1.24.3 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.59.3)\nRequirement already satisfied: tensorboard<2.16,>=2.15 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.15.1)\nRequirement already satisfied: tensorflow-estimator<2.16,>=2.15.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.15.0)\nCollecting keras<2.16,>=2.15.0 (from tensorflow->BLEURT==0.0.2)\n  Downloading keras-2.15.0-py3-none-any.whl.metadata (2.4 kB)\nRequirement already satisfied: wheel<1.0,>=0.23.0 in /opt/conda/lib/python3.10/site-packages (from astunparse>=1.6.0->tensorflow->BLEURT==0.0.2) (0.42.0)\nRequirement already satisfied: google-auth<3,>=1.6.3 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.26.1)\nRequirement already satisfied: google-auth-oauthlib<2,>=0.5 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.2.0)\nRequirement already satisfied: markdown>=2.6.8 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.5.2)\nRequirement already satisfied: requests<3,>=2.21.0 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.32.3)\nRequirement already satisfied: tensorboard-data-server<0.8.0,>=0.7.0 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.7.2)\nRequirement already satisfied: werkzeug>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.0.3)\nRequirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/lib/python3.10/site-packages (from packaging->tensorflow->BLEURT==0.0.2) (3.1.1)\nRequirement already satisfied: cachetools<6.0,>=2.0.0 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (4.2.4)\nRequirement already satisfied: pyasn1-modules>=0.2.1 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.3.0)\nRequirement already satisfied: rsa<5,>=3.1.4 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (4.9)\nRequirement already satisfied: requests-oauthlib>=0.7.0 in /opt/conda/lib/python3.10/site-packages (from google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.3.1)\nRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.3.2)\nRequirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.6)\nRequirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.26.18)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2024.2.2)\nRequirement already satisfied: MarkupSafe>=2.1.1 in /opt/conda/lib/python3.10/site-packages (from werkzeug>=1.0.1->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.1.3)\nRequirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /opt/conda/lib/python3.10/site-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.5.1)\nRequirement already satisfied: oauthlib>=3.0.0 in /opt/conda/lib/python3.10/site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.2.2)\nDownloading tf_slim-1.1.0-py2.py3-none-any.whl (352 kB)\n\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m352.1/352.1 kB\u001b[0m \u001b[31m6.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n\u001b[?25hDownloading keras-2.15.0-py3-none-any.whl (1.7 MB)\n\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.7/1.7 MB\u001b[0m \u001b[31m47.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m:00:01\u001b[0m\n\u001b[?25hBuilding wheels for collected packages: BLEURT\n  Building wheel for BLEURT (setup.py) ... \u001b[?25ldone\n\u001b[?25h  Created wheel for BLEURT: filename=BLEURT-0.0.2-py3-none-any.whl size=16456765 sha256=a983d396ca2522c89f7a74d4034f445d41e48c2d0e8bec14cb9f5ee669631ad5\n  Stored in directory: /tmp/pip-ephem-wheel-cache-ke2_w59y/wheels/64/f4/2c/509a6c31b8ebde891a81029fd94f199b1b92f0e7cfc20d417a\nSuccessfully built BLEURT\nInstalling collected packages: tf-slim, keras, BLEURT\n  Attempting uninstall: keras\n    Found existing installation: keras 3.3.3\n    Uninstalling keras-3.3.3:\n      Successfully uninstalled keras-3.3.3\n\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\ntensorflow-decision-forests 1.8.1 requires wurlitzer, which is not installed.\u001b[0m\u001b[31m\n\u001b[0mSuccessfully installed BLEURT-0.0.2 keras-2.15.0 tf-slim-1.1.0\n","output_type":"stream"},{"name":"stderr","text":"2024-06-25 03:23:24.134268: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n2024-06-25 03:23:24.134390: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n2024-06-25 03:23:24.262501: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n","output_type":"stream"}]},{"cell_type":"code","source":"import json\ntest_answer = []\nwith open(\"/kaggle/input/testing/test_answer.jsonl\", 'r') as file_json:\n  for line in file_json:\n    test_answer.append(json.loads(line))\n\ntest_answer_w_ref = []\nwith open(\"/kaggle/input/testing/test_with-reference.jsonl\", 'r') as file_json:\n  for line in file_json:\n    test_answer_w_ref.append(json.loads(line))","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:23:41.445180Z","iopub.execute_input":"2024-06-25T03:23:41.446259Z","iopub.status.idle":"2024-06-25T03:23:41.475431Z","shell.execute_reply.started":"2024-06-25T03:23:41.446220Z","shell.execute_reply":"2024-06-25T03:23:41.474533Z"},"trusted":true},"execution_count":2,"outputs":[]},{"cell_type":"code","source":"# Fungsi untuk memisahkan data ke dalam dua list paralel\ndef separate_texts(test_answer, file_respons):\n    answers = []\n    responses = []\n    \n    for answer in test_answer:\n        for respons in file_respons:\n            if answer['id'] == respons['id']:\n                answers.append(answer['text'])\n                responses.append(respons['response']['text'])\n    return answers, responses\n\n# Menjalankan fungsi\nanswers, responses = separate_texts(test_answer, test_answer_w_ref)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:23:41.830726Z","iopub.execute_input":"2024-06-25T03:23:41.831799Z","iopub.status.idle":"2024-06-25T03:23:41.848254Z","shell.execute_reply.started":"2024-06-25T03:23:41.831765Z","shell.execute_reply":"2024-06-25T03:23:41.847392Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"len(responses)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:48:29.755969Z","iopub.execute_input":"2024-06-25T01:48:29.756660Z","iopub.status.idle":"2024-06-25T01:48:29.763303Z","shell.execute_reply.started":"2024-06-25T01:48:29.756619Z","shell.execute_reply":"2024-06-25T01:48:29.762414Z"},"trusted":true},"execution_count":4,"outputs":[{"execution_count":4,"output_type":"execute_result","data":{"text/plain":"273"},"metadata":{}}]},{"cell_type":"code","source":"len(answers)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:48:31.747203Z","iopub.execute_input":"2024-06-25T01:48:31.748106Z","iopub.status.idle":"2024-06-25T01:48:31.753568Z","shell.execute_reply.started":"2024-06-25T01:48:31.748073Z","shell.execute_reply":"2024-06-25T01:48:31.752744Z"},"trusted":true},"execution_count":5,"outputs":[{"execution_count":5,"output_type":"execute_result","data":{"text/plain":"273"},"metadata":{}}]},{"cell_type":"code","source":"answers","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:52:33.752550Z","iopub.execute_input":"2024-06-25T01:52:33.752922Z","iopub.status.idle":"2024-06-25T01:52:33.768337Z","shell.execute_reply.started":"2024-06-25T01:52:33.752895Z","shell.execute_reply":"2024-06-25T01:52:33.767372Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":10,"outputs":[{"execution_count":10,"output_type":"execute_result","data":{"text/plain":"['\"Excellent point about reducing carbon emissions! That directly connects to our previous discussion about combating extremeweather. Let\\' s explore some specific strategies to combat rising temperatures and melting ice, shall we?\"',\n '\"Hi [student\\'s name], I hear you need some time to warm up. Is there anything specific you have in mind for us to review in these last minutes to ensure we make the most of our remaining time? Perhaps revisiting some key vocabulary points from the list, or perhaps exploring something related to your current interests? Let me know what feels most productive for you.\"',\n '\"I\\'m interested to hear more about the specific situations where a distinction would signify outstanding achievement. Can you elaborate on what kinds of academic endeavors might warrant such recognition?\"',\n '\"Student, I appreciate you sharing your thoughts on using \"frank\" and \"direct\" language. Could you elaborate on how these terms differ in specific scenarios, such as when aiming for clarity or maintaining a professional tone? This will help us understand the nuances of using these words effectively.\"',\n '\"That insight is insightful, student. To further explore this notion, could you elaborate on specific examples of games that children might find particularly engaging and effective in combating feelings of depression, loneliness, or boredom? This would provide valuable context for understanding the impact of playing games on young minds.\"',\n '\"Excellent point about using words to describe your feelings! Have you encountered any situations or encountered situations that trigger strong emotions in you? Is there a particular emotion you find challenging to express or manage?\"',\n '\"I appreciate your perspective on this topic, student. I\\'d love to hear your thoughts on how we could approach justifying such large expenditures. Do you see any specific arguments that resonate with you?\"',\n '\"Excellent point! These new structures are indeed quite unique and offer a fresh perspective on the writing process. I\\'d love to explore how you apply them in your own writing. Do you have any specific examples or scenarios in mind where you see them making a difference?\"',\n '\"Great point! Can you tell me more about the context in which this word is used? Knowing the setting or situation would help me understand its meaning better.\"',\n '\"I understand your desire to pursue your master’s at Lse. That’ s fantastic! It’ll be an exciting challenge to navigate the differences between Japanese and UK education. Remember, your previous academic background is valuable, so don’t hesitate to leverage it. Let’',\n '\"Wonderful point, let\\' s explore the various aspects of childbirth preparation women may undertake. Does pregnancy preparation involve anything specific they need to be aware of?\"',\n \"**Excellent!** That'll help us refine our understanding of 'beating' in this context. Let’s explore synonyms to add variety and depth to our analysis. Did you notice any other team that impressed with its attacking display?\",\n '\"Thank you for sharing your thoughts! I appreciate the emphasis you place on the importance of effective language use. Let\\' s explore some specific ways students can become more confident and articulate in their communication.\"',\n '\"Hi [Student Name], I noticed you mentioned feeling tired after lessons. I completely understand how you feel, especially after a full 60 minutes of focused studying. \\n\\nI\\'d love to hear what aspects of your preparation you are focusing most heavily on now, whether it be grammar, vocabulary, pronunciation, or specific skills. Based on that, we can tailor the next sessions to further enhance your understanding and practice. This will ensure that your learning experience remains engaging and productive, while also prioritizing your physical and mental well being.\"',\n '\"That astute observation highlights an interesting point. To delve further, could you elaborate on what specific aspects of \"your\" observations suggest a potential difference between the Canaries and mainland English? Is there a specific linguistic feature or grammatical pattern that stands out?\"',\n '\"Great points, student! I appreciate you sharing your insights. To further our discussion, how about we explore how this concept applies to other English-speaking countries? Do you have any other interesting examples to share?\"',\n '\"Great point! That shows thorough planning and a focused approach. What subjects or topics do you want to focus on first? That will help me tailor the review to your specific needs.\"',\n '\"Wonderful! That sounds like a great point. Perhaps we could use an idiom to describe the experience of learning English?\"',\n '\"Your point about the advantages of owning versus renting resonates with me! I think it\\'d be valuable to consider the potential for **stability** and **community** associated with owning your own place. How might these factors contribute specifically to building a **meaningful** life?\"',\n '\"Great point about the importance of context! Let\\'s explore how this applies in different writing styles now.\"\\nOR\\n\"I\\'m curious to hear your thoughts on the use of figurative language in storytelling. Would you like to discuss specific examples?\"',\n '\"Thank you for sharing your experience with me! It\\'ll be interesting to delve deeper into how this result impacted you and how it might inform your future study strategies. Do you have any specific takeaways or areas where you felt particularly relieved or excited?\"',\n '\"I appreciate your understanding, student. To clarify, \\'pull in\\' means to stop temporarily where you are. Is there anything specific you\\'d like to discuss next?\"',\n '\"I understand you\\'ve had some thoughts about that. To clarify, when did you first **notice** the objects? Was it the sight of them in general, or did something specific grab your attention?\"\\n\\n** Example 2:** \\n\\n \"Following our discussion on context, could you walk me through the initial steps you took in that situation? Did you see the tabletop before or after I explained the reason for it?\"',\n '\"Excellent point! To delve deeper, what specific aspects of regular exercise are you most interested in exploring? Is there a particular type you\\'re particularly curious about, or are there broader considerations like frequency or benefits?\"',\n '\"That\\'s a great point about \\'shunk\\' being similar in meaning to \\'reduced\\'. It highlights the nuance of language and how even seemingly related words can have different connotations. Do you want to explore this further by discussing synonyms and antonyms?\"',\n '\"Excellent! Now that you have some practice under your belt, how does using this grammar concept feel in context of writing? Do you foresee any challenges or opportunities in applying it effectively?\"',\n '\"Thank you for bringing this to my attention. While I understand your perspective, let’s analyze the situation. The use of the past tense suggests an intentional act, which is why the word ‘hit’ is appropriate. However, the context suggests a level of intensity and force that warrants further exploration. \\n\\nCan you elaborate on what led you to conclude that she ‘ hit’ you intentionally? Providing specific examples will help us understand the nuances of your experience better.\"',\n '\"Excellent point! I was hoping to discuss the importance of clear and concise communication in effective teamwork. How do you think we can ensure our messages are easily understood within the team?\"',\n '\"That\\' s an interesting connection between CO₂ emissions, plant activity,  and oxygen levels. Could you elaborate on what you mean by \\'plants are decreasing CO₂,\"? Is there specific data or research supporting this decline?\"',\n '\"That\\'s a great point, <student>! I was just reading about Brighton\\' s stunning coastline and vibrant culture, which truly is an exceptional resort destination. What specific aspects of Brighton intrigue you the most?\"',\n '\"The \\' 100 % \\' refers not only to coal, it encompasses various fuels like natural gas, hydro, wind, solar, etc.\"\\n\\n\\n-  **Expanding on their understanding:** \\n    - Teacher could elaborate on why electricity from various sources is important, beyond just the percentage.\\n\\n\\n*    \"Can you explain how this percentage contribution impacts the overall energy consumption in Australia?\"',\n '\"I understand your point. Let\\'d focus on using \"search\" effectively rather than just relying on it for everything. What specific questions or topics come up when you use search?\"',\n '\"Wonderful observation about the impact of your vocabulary choices! It\\'ll be interesting to hear how your classmate, [mention specific classmate], used specific words in their writing recently. Does that align with your own strategies or are there any differences you notice?\"',\n '\"Following our last discussion, it seems you have some questions about airports, universities,  and stations. To further clarify these concepts, let\\'d explore each individually:\\n\\n* **Airport:** What information about airport operations, facilities, services, etc., are you particularly interested in?\\n\\n\\n*  **University:** Can  you tell  me what aspects of university life, such as academic programs, student life , or financial aid, are unclear to you?\\n\\nRemember, the most effective follow up will depend on your student\\' s specific needs and interests.',\n '\"Excellent! Before we move on to [new topic], I just wanted to clarify... [state any remaining points from the previous topic]. Is there anything specific you\\'ve encountered that you still want to discuss? It\\'ll help ensure we cover everything comprehensively.\"\\n\\n**Explanation:** This follow up encourages further discussion by directly addressing the student\\' s remaining questions and ensuring a smooth transition.\\n\\n\\n** Example 2:** \\n>\"I appreciate your perspective on [previous topic] and your willingness to try new things. To ensure everyone understands the concepts better, could you share any specific challenges you foresee when working with these idioms?\" \\n\\n> This approach demonstrates active listening and shows the teacher is open to further exploration of relevant topics.',\n '\"Great point! \\'Even if\\' is definitely a possibility when using the simple tense in your essay. Can you think of a specific example where using simple present might be appropriate?\"',\n '\"Human, I appreciate your perspective on the possibility of life elsewhere. That\\'s a fascinating topic, especially the efforts people make to find it. Do you think the search for extraterrestrial life is merely a matter of scientific curiosity, or do you see it as holding potential implications for our understanding of our place in the universe?\"',\n '\"Let\\'s discuss the specific situations where a \\'close brush\\' might occur, or how we can navigate challenging situations effectively.\"\\n \\n - Or\\n  \\n   -  \"Before we move on, let\\' s review the concept of a near miss and how to respond appropriately.\"',\n '\"Great point about the size of Africa! It\\'s interesting how details like rivers and names can also be included to describe a place. How does the Sahara\\' s location within Africa make it unique?\"',\n '\"Hi <student>, I noticed your voice sounds a little raspy. I know you have been under some stress lately, so I wanted to see if you would like some time to discuss it? It\\'d be great to hear how you feel and if there\\' s anything I can do to help.\"\\n\\n\\n**Explanation:** This response shows understanding and compassion, while also offering the option for further conversation.',\n '\"Thanks for sharing your perspective! It helps me understand your point about using repetition as a tool for language improvement. Do you have any specific examples of how this \"repeating\" might be applied in the context of football or other sports?\"',\n '\"That’s an insightful point. To further explore this, could you elaborate on how this relates to the broader context of the lesson? Are there any specific examples or theories that support this observation?\"',\n '\"I\\'m glad to see your progress! Remember the importance of practicing different sentence types and incorporating them into your writing. Keep this up, we\\'ve just begun to explore some engaging sentence structure techniques. Let\\'s focus our efforts on mastering those and then we can explore something new in the next lesson.\"',\n '\"Can you elaborate on what you mean by \\'violent behaviours\\'? Are you referring to specific actions or broader patterns of behaviour?\"\\n-  **Providing context and examples:**  \"It\\' s important to consider the context of violent behaviour and the potential factors that may contribute to it. Could you share specific examples of situations where you have observed this kind ofbehaviour?\" \\n-\"Let\\'sgather some information from experts and research papers on this topic to broaden our understanding before drawing any conclusions.\"',\n '\"That is an astute observation! To delve further, what specific aspects of whisky production in Scottish contribute to its reputation as a fine spirit? Are there any unique ingredients or traditions that influence its creation? Understanding the origins and production methods would strengthen your argument.\"',\n '\"I appreciate you providing examples, that helps me tailor the discussion to your specific needs. How about we focus on the use of figurative language now? What kind of examples can you think of that illustrate its impact?\"',\n '\"That\\'ll help me understand your point better. Could you explain why you think it might snow?\"\\n\\n**Explanation:** This encourages the learner to provide specific reasons and examples to support their opinion.\\n\\n\\n** Example 2:** \\n>\"I understand you have some uncertainty about snow\\' future occurrence. Is there any specific reason you feel it\\' s likely?\" \\n\\n> This shows you are actively listening and seeking clarification without being overly directive.',\n '\"Excellent point about the timeline! To ensure we cover all the essential points before April the 5th, how about focusing on reviewing the main takeaways from today\\' s session and identifying areas for further elaboration? This will give us a strong foundation and leave ample time for independent practice.\"\\n\\n**Explanation:** This response encourages further review of key concepts while also suggesting a clear direction for future exploration.\\n\\n\\n** Example 2:** \\n\\n“I appreciate your initiative to keep the discussion going! However, it seems we are nearing the end of our allotted time. To maximize our time efficiently, perhaps we could dedicate the remaining session to reviewing key points and assigning homework related to today’s discussion?”\\n\\n **Explanation**: This approach prioritizes time management and ensures everyone is aware of remaining time, potentially facilitating a smoother transition into a concluding activity.',\n '\"Hi [student\\' name], I noticed you mentioned feeling overwhelmed by the preposition rules. That\\'d be great to explore them further together. Do you have any specific examples that are causing you difficulty?\"',\n '\"That\\' s a fascinating perspective, thanks for sharing! It makes me appreciate the challenges you face as an ESL student in a high-unemployment context. \\n\\nLet\\'d talk about adapting your learning to your specific needs and goals. Is there anything specific you\\'re struggling with or interested in exploring that could help you progress more efficiently?\"',\n '\"I appreciate you sharing your thoughts about the possible reasons why children might have not eaten the marshmallows. That\\'ll help us understand better why they may have behaved in that way. Do you have any additional observations or examples that might be relevant? For instance, were there any other clues in the situation that could shed light on the issue?\"',\n '\"Wonderful! Thank you for sharing your train timings. To ensure smooth progress with your course materials, I\\'d love to discuss how you approach studying this upcoming week. Please tell me about your preferred learning strategies and how they help you stay on top of your tasks.\"',\n '\"That\\' s a perfectly good point, student! It highlights the positive difference your time experienced. Let\\'d continue exploring other adverbs related specifically to speed and compare them in different contexts. For example, \\'rapidly,\\' \\'expedited,\\' or \\'nimbly\\' might be interesting to explore.\"',\n '\"Great point, <student\\' name>! Remember, the key is identifying the precise adjective and choosing the appropriate preposition to complement it. Let\\' s focus on the specific adjective first, then we can explore different prepositions that best fit.\"',\n '\"Excellent point about [specific point]! I\\'m glad you\\'re grasping these concepts well. Now let’s move on to [new topic] to further your understanding.\"\\n\\n** Follow-UP 2:**\\n \\n \"You\\'ve made impressive progress with your pronunciation! Keep up the good work. In our next session, we\\'ll delve deeper into [topic] for additional practice.\"',\n '\"Excellent point, let\\'d explore the use of active and passive voice in sentences together. This will enhance clarity and engage learners more effectively.\"',\n '\"Great point about the police! It highlights the complexity of using descriptive language. Have you considered rephrasing it to avoid ambiguity?\"',\n '\"Great point about the importance of reflection. We\\'ve been exploring the concept of [current topic], and I think further analysis of its implications would deepen our comprehension. How about we dedicate a few minutes to discuss how this relates to [relevant field]?\"\\n\\n**Explanation:** This response encourages the student to think critically and suggests a deeper analysis, potentially leading to further understanding and application of knowledge.\\n\\n\\n** Example 2:** \\n\\n \"I really appreciate your perspective on [previous topic]. Let’s move beyond simply summarizing your points and delve into the nuances of your argument. What specific questions do you have about [topic]? I\\'m eager to hear your thoughts and ensure we\\'re addressing them effectively in our discussion.\"\\n\\n **Explanation**: This shows initiative and engagement by encouraging the learner to actively participate and think deeper, leading towards a more insightful and productive discussion.',\n '\"Hi [Student Name], I appreciate your honest feedback about the situation with your car. Allow me to follow up by asking, what specifically did you find frustrating about having to deal with the repair costs? This will help us identify any areas where we as a class can provide better support and guidance.\"',\n '\"Wonderful! That\\' s a great point. Could we delve into the specific benefits of choosing to rent rather than buy this book? This will help us tailor the paragraph to emphasize its unique advantages for your essay.\"',\n '\"Great points about the disparity in access to resources and its impact on learning outcomes. That leads to interesting questions, such as how individual circumstances and opportunities within different socioeconomic contexts influence learning experiences. How might we address this issue in our teaching?\"',\n '\"Great to continue exploring this, [student\\' name]. Remember, \\'most young\\' is an appositive, describing the broader group. If you replace it with \\'young people as a whole,\\' it becomes clearer what type of \"young\" we\\'re referring to.\"',\n '\"That\\'S a great point. To further emphasize the contrast, let\\'d use the past continuous tense, wouldn\\'t you say?\"\\n\\n**Explanation:** This approach expresses agreement while also subtly transitioning to the expected tense.\\n\\n\\n** Example 2:** \\n\\n \"You\\'re right! Using past perfect tense adds nuance to how we discuss the change in satisfaction. Perhaps we should explore how \\'being more\\' satisfied impacted the experience?\"',\n '\"Thank you for clarifying that the correct usage of \"the way\" in Partens depends on its context. Could you tell me specifically which situation you\\'re referring to, or provide an instance where \" the way \" is used incorrectly? This will help us focus our discussion and ensure we\\'m using the appropriate phrasing in the future.\"',\n '\"Wonderful point, that\\'s a great observation! It\\'ll be even more impactful if you could provide an actual example of how \\'with\\' is used in a real-life context.\"',\n '\"Following your insightful point that teachers might lack empathetic tendencies towards students, how do you think this might affect the learning environment? Do you believe it hinders effective communication or creates an atmosphere of indifference?\"',\n '\"Excellent point! \\'Inappropriate\\' can indeed carry negative connotations, implying a deviation from socially accepted norms or standards. However, it can also be a matter of context. \\n\\nWhich meaning of \\'incorrect\\' do you prefer in this situation? Knowing the specific context will help us choose either \"inappropriate\" or \"unsuitable\" most effectively.\"',\n '\"Great answer about the rain! Let\\'d discuss the impact of recent weather events on your daily life or explore other aspects of village life that intrigue you. Is there something specific that interests you in particular?\"',\n '\"That\\'s a great point! Perhaps we could discuss the environmental impact of nuclear energy compared to other sources in the 8o\\'ies? Or explore how technological advancements have impacted energy production since then?\"',\n '\"Excellent point about scientists **continuing their research**! I\\'m curious to hear more specific examples of how they might approach a future cure. Do you have any particular areas of interest within medical research that excite you?\"',\n '\"Great! Now let\\'s use the phrase \"within our budget\" in a complete sentence. How about: \"Water costs within our family\\'S budget\"? This emphasizes the connection between the two ideas.\"',\n '\"I\\'m glad you\\'re still eager to discuss this, [student\\' name]. Let\\'d continue exploring how to express enthusiasm in your writing.\"\\n - Show your continued interest and encourage further discussion.',\n '\"I understand your perspective, student. Let\\' s focus on the core elements of this concept and see if we can develop a clear and concise structure for your essay. We can also explore specific examples or relevant analogies to illustrate your points.\"',\n '\"Great examples you\\'ve shared! It seems communication, language, even technological skills like computer proficiency and online communication are crucial for finding a successful job in today\\'s job market. How do these skills fit into your future career aspirations?\"',\n '\"That is astute! While coal was undoubtedly a significant contributor in its time, it\\'d be beneficial to explore the nuances of that statement. Could you elaborate on what aspects of coal were most impactful?\"',\n '\"I understand you want to refine your ideas. How about we explore the pros & cons of [specific point] further? That way, we can ensure your message is clear and concise.\"',\n '\"I enjoyed learning about Camila\\' name and how it captures a certain air of refinement. Let\\'sgather some other options with similar qualities, perhaps with more modern twists? We can explore unique spellings, sounds, or even incorporate elements of her personality into the name.\"',\n '\"Excellent point! That\\'ll be a perfect segue to our next topic, which involves using time expression in different tenses. Let\\'d focus on expressions related specifically to [mention the specific period or event you want to discuss]. Think about examples like \\'last month\\', \\'every week\\', or \\'since last spring.\\'\"',\n '\"Excellent point about triggers! What specific actions or situations do users commonly associate with turning their phones on/off? Do you have any personal anecdotes or experiences that illustrate this?\"\\n\\n\\n**Poor follow follow:** \\n\\n\"- Are you aware of any idioms or colloquialisms related to turning phones \"on\" or \"off\"?\\n- Can you provide some examples of how people use mobile phrases to indicate phone activity?\"',\n '\"Great point! I think we can use the present continuous here. Imagine if we\\'re focusing on recent events related to our topic, what would you like to highlight about them?\"',\n '\"That\\'s a very insightful point. To further explore this, could you explain how this technique adds depth and richness to the narrative? Does it provide a clearer picture or add emotional weight to certain elements?\"',\n \"## Follow-Up Example 2:\\n\\n** Teacher:** Interesting! I'm glad to see the positive impact of sunshine on your activities today. Are there any specific skills you were working on in English that were particularly challenging because of it?\",\n '\"That astute observation highlights an interesting contrast. Let\\'d explore the reasons behind this phenomenon in greater depth. Could you elaborate on what you mean by \\'significant fuel\\'? Does it pertain specifically to the type of fuel or does it encompass other aspects as well?\"',\n '\"I appreciate your perspective on the importance of independence. To build on this, could you elaborate on what you mean by \\'independent from\\' in this context?\"',\n '\"I appreciate your perspective on the importance of both diet and exercise. To further strengthen your argument in your introductory statement, perhaps you could connect the dots between these two factors and explain how they influence each other?\"',\n '\"Excellent point, [student name]. I noticed you were using the present tense there. To ensure we maintain consistency in tenses throughout the conversation, let\\'sgive it another try using mainly the Past tense. Then, we can discuss...\"\\n\\n\\n**Alternative:**\\nIf the student brings it up again:\\n\\n“I think using present perfect would make the sentence clearer. Would you like to give it a try?”',\n '\"I appreciate your insights on running in Spain. I\\'d be interested in learning more about the specific adaptations people make to stay active in hot weather. Do you think extreme heat might affect certain aspects of running performance or require special considerations for training?\"',\n '\"Great point about using \"controversial.\" Let\\'s explore how to express disagreement or strong opinions respectfully in our writing. Is there a specific type of controversial topic you\\'d like to focus on? We can then discuss the most appropriate expression in English.\"',\n '\"Ely, I noticed you mentioned that running is more enjoyable in warmer weather. Could you elaborate on how it affects your performance? Has it made a noticeable difference to your running experience?\"',\n '\"Thank you for sharing this experience with me. Understanding the specific circumstances surrounding your confusion can be crucial to providing you with the most relevant and helpful feedback. If you\\'d be comfortable elaborating on the situation that led to your initial impression of confusion, I believe it will be beneficial to explore the nuances of those emotions. This will allow us to tailor our discussion towards understanding your perspective and providing targeted support.\"',\n '\"That is an astute observation. To fully understand your experience with spiders, could you elaborate on specific situations or incidents that make you feel anxious? Identifying the context and potential triggers would be helpful in understanding the nature of your fear.\"',\n '\"That\\'s a great point! It highlights the significant contribution of coal in generating electricity. Can you elaborate on how its widespread use impacted the development of alternative energy sources?\"',\n '\"I appreciate your observation about the textual complexity in round two and three. That highlights an interesting point about sentence structure and word choice. Let\\'d analyze these specific instances and see if we can identify any recurring patterns or areas for improvement. Remember, effective communication often involves striking a balance between clarity and nuance, so it\\'ll be important to consider how to express our ideas in a way that\\' s both accessible and impactful.\"',\n '\"Excellent point! It seems like you\\'re focusing on the technicality of using \\'although\\' correctly. That deserves further discussion as it can affect the overall meaning of your sentence. How about we revisit this after we return from break?\"',\n '\"I noticed you mentioned war recently. Could you elaborate on the specific implications it has on people in your community or globally?\"',\n '\"That point is well made, [student\\' name]. To elaborate further, the correct tense would be the past continuous in this case, as it describes an ongoing action at the moment. Could you elaborate on what you mean by \\'a period in time\\'?\"',\n '\"Hi [student name], I hope you’re having a good week! I noticed you mentioned wanting to practice using idioms in your writing. Would you be interested in some activities or games that focus on idioms? It would help you expand your vocabulary and learn new ways to express yourself in context.\"',\n '\"Hi [Student Name], I appreciate your input on feedback. I agree that timely and specific feedback can be invaluable for improving your understanding of [topic].\\n\\nWould you be open to discussing some examples of how we can enhance the timing and delivery of these feedbacks? Or are there specific aspects of providing feedback that seem particularly challenging for you?\"',\n '\"Great point about the similarity! Have you encountered any other expressions in Spanish that express a similar idea of \"something very similar\"? It\\'d be interesting to compare and contrast them.\"\\n\\n**Explanation:** This response encourages the comparison of ideas and promotes further exploration of related phrases.\\n\\n\\n** Example 2:** \\n\\n \"I appreciate your perspective on this. Do these two proverbs have anything in common beyond their literal meaning? Are there any nuances or additional layers of interpretation that distinguish them?\"\\n\\n **Explanation**: This prompts deeper analysis and encourages exploration beyond the surface meaning.',\n '\"Excellent point! That clarifies the complexity of this topic for you. To further explore your concerns, let\\'d focus on specific situations where you encounter these expressions in written texts. This will allow you to identify the nuances and context-dependent usage, enhancing your understanding.\"',\n '\"Excellent point. I think it ultimately depends on the context and the intended audience. For example, when discussing a shared experience or collaboration, \"we\" or \"our\" might be more appropriate. On the other hand, if addressing a specific individual, then \"you\" could be clearer.\"',\n '\"Your observation about students\\' behavior is astute. While it\\'sa commendable effort to understand the root causes, it would be beneficial to explore the situation from a pedagogical perspective. Could you elaborate on what specific aspects of classroom routines or student interactions seem to be encouraging such behavior? This will help us identify any underlying patterns or systemic issues that need addressing.\"',\n '\"Following your insightful observation, how do you think we can best utilize this opportunity to improve your communication skills? Do you have any specific areas you\\'re particularly curious about?\"',\n '\"Great point! Continuing to practice and apply your knowledge of these concepts will definitely help you develop your problem solving skills further. What specific examples come to mind where these elements work together?\"\\n\\n\\n**Explanation:** This response encourages the student to actively engage with the material and reinforces the relevance of continued practice as a means to enhance understanding.\\n\\n\\n **Example2:** \\n>\"Remember how learning a language is like building a complex machine? There are many parts that need to work in harmony for you to achieve fluency. Which part do you find challenging to master?\" \\n\\n\\n>\\n **Explanation**: This follow up acknowledges the acknowledged challenge and encourages introspection, prompting the learner to identify specific areas for improvement.',\n '\"Wonderful point, [Student\\' name]! I trust you\\'re comfortable continuing our exploration of recent happenings? Let’s utilize this knowledge to incorporate ‘recently’ into a cohesive sentence, enhancing your understanding of its temporal context.\"',\n '\"Wonderful to continue our discussion about idioms! Did you come across any particularly memorable idioms yesterday that you particularly enjoyed or found impactful?\"\\n\\n**Explanation:** This response encourages further exploration of specific points from the previous discussion, demonstrating engagement and connection.\\n\\n\\n** Example 2:** \\n\\n \"I appreciate your initiative in bringing the topic of [introduce new topic] to my attention. Would you like me to provide some resources or suggest further discussion points to deepen our understanding?\" \\n\\n\\n **Explanation**: This approach shows initiative and caters to the student\\' s interests, facilitating a smooth transition into a different subject.',\n '\"I\\'m happy you brought this up, [student\\' name]. To understand the context better, could you elaborate on what you mean by \\'relative contribution\\'? Does it relate to the overall use of energy sources or something else?\"',\n '\"That astute observation highlights the importance of context in describing actions. While \\'walking\\' is an ongoing action, \\'seeing\\' him at that precise moment suggests a specific instance within that ongoing process. \\n\\nYour sentence effectively conveys this distinction by stating that the action of seeing preceded the actual walking. This aligns with the principle of sequencing actions in chronological order.\"',\n '\"Great point! YouTube comments and news articles with diverse perspectives are excellent resources for expanding vocabulary and understanding different viewpoints. Let\\'s explore some specific topics based on your interests then.\"',\n '\"Great to hear you\\'re wrapping up your course soon! Is there anything specific you want to focus on or practice before it ends? We can explore some examples of using the word \\'after\\' or \\'before\\' in a future context.\"',\n '\"Excellent! I\\'m glad you\\'re recognizing areas for improvement. Let\\'d focus on [specific grammar point] next week to solidify your understanding.\"\\n\\n- \"[Student\\' name], I appreciate your proactive approach in recognizing your grammar mistakes. Can you tell me more about the specific areas you want to focus upon?\"\\n\\n - \"I\\'ve noticed a pattern in your sentences that could benefit from clearer word choice and sentence structure. Would you be interested in practicing writing a paragraph using precise language?\"',\n '\"Good morning, [Student\\' name]! I noticed you mentioned wanting to improve your reading comprehension today in your email. Is there a particular text or skill you have in mind?\"\\n\\nThis follow up is concise and specific, encouraging the learner to take ownership of their learning by suggesting a concrete area for practice.',\n '\"That\\'s a good point about spelling, [student\\' name]. Let\\'d focus on practicing specific verb tense patterns that are causing confusion for you. We can start with the present tense and then explore others as you feel more comfortable.\"',\n '\"That’s an astute observation about the impact of your tenses change. While it’ll be difficult to quantify the exact effect without further analysis, it certainly highlights an important aspect of language usage. \\n\\nLet’d explore this further. Could you elaborate on what specific aspects of present tenses or past simple usage surprised you? This will help us understand the broader implications of this change.\"',\n '\"Great point about the importance of clarity! Now, let\\'s practice using the active voice in a short sentence. Can you create a complete sentence that incorporates the word \\'despite\\'? For example, \\'Despite the challenging circumstances, she persevered with her studies.\"\\n\\n**Explanation:** This response demonstrates understanding of passive voice and suggests further practice through specific application.',\n '\"That insightful observation about the students\\' engagement is commendable. To further explore this point, let’s analyze specific instances and see if we can identify any patterns or trends.\"',\n '\"That\\'S a great point! I think it\\'ll be valuable to revisit the concept of prediction before we move on. Can we discuss it in more detail during our next session?\"\\n\\n**Explanation:** This response encourages further discussion by specifying the topic and offering a timeframe for exploration.\\n\\n\\n** Example 2:** \\n>\"I appreciate your perspective on preempt, student. However, I want to make sure we understand the core definition and its nuances before exploring different applications. How about we review the basic principles and then discuss specific instances where prediction might occur?\"',\n '\"Thank you for sharing your thoughts. I appreciate your proactive approach to evaluating our lesson plans. To ensure our learning objectives are aligned and effective, let\\' s discuss specific aspects of our current plans that might require adjustments. This will also help us identify any gaps in knowledge or clarity issues that need attention.\"',\n '\"I appreciate your perspective, [student name]. It\\'s true that with increased education and exposure to diverse situations, individuals may respond differently to bullying. However, it\\' s important to consider the context and motivations behind bullying, rather than making sweeping generalizations. Can you elaborate on what you mean by \\'softer\\' ways of dealing with bullying?\"',\n '\"I noticed you mentioned feeling balanced. Is there a specific skill or concept that has been particularly stimulating or challenging so for?\"\\n\\n**Explanation:** This encourages further elaboration on the students\\' current experience, encouraging them to connect their recent insights with specific elements of their learning.\\n\\n\\n** Example 2:** \\n - \"While I appreciate your positive attitude today, I noticed some hesitation in expressing your thoughts. Would you like to try brainstorming a few points together to ensure we\\' re on track with our current focus?\" \\n\\n **Explanation**: This invites collaborative problem',\n '\"Great to hear, <student\\'s name>. Is there anything specific you\\'d like to discuss today, or are you ready to wrap up this topic?\"',\n '\"Great point about the inconsistency! To further explore your question, what specific aspects of disagreement or inconsistency are you most interested to discuss?\"',\n '\"Excellent observation! I appreciate you bringing this to my attention. Let’s delve deeper into the specifics of [topic] and how to use it effectively in different contexts.”\\n\\n**Explanation:** This response encourages the student to specifically discuss the application and nuances, prompting further exploration and understanding.\\n\\n\\n** Example 2:** \\n\\n“You’re absolutely right! This prompts us to consider the broader implications of our discussion. Shall we shift the conversation towards [new topic]?”\\n\\n **Explanation**: This suggests a change in focus, promoting critical thinking and encouraging further discussion beyond the initial point.',\n '\"I appreciate your perspective, student. Let\\' s explore this further and see if we can identify other reasons why they might not have the resources to reside there.\"\\n-  \"That thought is intriguing. Could you elaborate on what specific challenges they may be facing that prevent them from affording to stay there?\"\\n-\"Your point about their ability to adapt and persevere is valid. Is there anything specific they\\'re lacking in terms of financial resources?\"',\n '\"Great point about charging your device, [student\\'s name]! Now, is there anything else you\\'d like to practice or review before we move on?\"',\n '\"Great point, student! You\\'re right. Past simple and past continuous are often used together when describing overlapping or sequential actions. Let\\'s look at this sentence together: \"The train had already left before I arrived.\" Here, the train left beforehand, while I was still traveling. We could also use past progressive here, which describes the ongoing action while something was happening: The train was being pulled when I reached the station.\"',\n '\"I appreciate you highlighting the importance of context. To further explore this, how does the setting of Wednesday mornings influence your understanding of…”\\n\\n** Example 2:** \\n\\n“It’ s interesting how you mentioned the gecko’ behavior in relation to its surroundings. Do you think other animals display similar responses in challenging situations?”',\n '\"I understand your perspective, [student\\' name]. High unemployment is indeed a concern in many parts of Spain. However, it\\'S important to remember that this is just one factor to consider when choosing a location for your language program. \\n\\nThe benefits of studying English in sunny Tenerife cannot be overstated. Not only will you gain valuable communication skills, you\\'ll also discover a vibrant culture and stunning natural beauty. Do you have any specific reasons why you feel this program is the best fit for you?\"',\n '\"That\\'s a great point! Can you tell me more about how the \\'w\\' and \\'R\\' sounds differ in these words? Are there any examples where they behave similarly in British English?\"',\n '\"I understand you need some time to process the previous exercise. Let\\'d focus on this section together in the next lesson.\"\\n - Show understanding and provide direction.',\n '\"That\\'S a great point, student! It\\'ll be valuable to consider how education interacts with other spheres of opportunity. Do you have any specific examples of how educational attainment impacts job prospects across different professions?\"',\n '\"I appreciate you sharing that information, [student name]. To delve deeper into this topic, let\\' s explore the distinctive features of French whiskey. How does it compare to other spirits produced locally?\"',\n '\"Excellent points! It\\' s fascinating how the relationship between \\'fruits\\' and \\'actions\\' can be explored further. Let\\'S examine the nuances of how \\'France uses\\' grapes and how it complements the concept of \"wine production.\" Can students provide additional examples or specific contexts where these nouns are used together?\"',\n '\"I understand your hesitation about mentioning the topic of government, student. Perhaps it\\'s best to focus on the specific issue that concerns you more. We can always revisit the broader topic later if you\\'d like.\"',\n '\"Excellent point! I appreciate your breakdown of 0.4 as three-quarters. To further explore this concept, why not discuss how fractions are used in real-life situations or how they relate to other mathematical concepts? This would help you connect with the material on a deeper level.\"\\n\\n\\n** Example 2:** \\n\\n \"Your observation about 5% is accurate. We can explore other ways to express this as a ratio or fraction. Do you have any other examples in mind where decimals are encountered?\"',\n '\"I understand your concern, [student name]. That\\'s a natural reaction to unexpected change. Let\\' s explore this further together. Have you encountered any specific challenges or questions regarding the recent shift in our studies?\"',\n '\"Excellent point, [student name]. You\\'re correct in recognizing the importance of understanding the formality of language. Now that you understand the different levels of frequency, I believe it would benefit you to explore how to utilize them effectively. How about focusing on how \\'seldom\\' compares to the more extreme cases of \\'rare\\' or \\'almost never\\'? This could be a fruitful topic for further discussion.\"',\n '\"Great point about children and decision-making. Can you elaborate on specific behaviours that might indicate bad behaviour towards others?\"',\n '\"Excellent point! That connection adds another layer of depth to your introduction. Let’s build on this and explore its implications for...”\\n\\n** Follow- Up 2:** \\n\\n“Wonderful observation! It seems these phrases hold the key to understanding the broader argument of our essay. What specific questions or areas do these words raise for you?”',\n '\"Let\\'d explore other instances of inversion and how they can enhance sentence structure. Have you encountered any situations where inversion adds extra clarity?\"\\n*  \"Since we\\'ve discussed inversion, let\\' s discuss other grammatical elements that contribute to clear and concise communication.\"\\n\\n **Transitioning to new topic:** \\n\\n * \"While inversion serves a specific purpose, it\\'ll soon be time to move on to other aspects of sentence construction. Would you like to discuss the role of conjunctions in building coherence?\"',\n '\"Excellent point! I appreciate your insights. To delve deeper, could we discuss real-life examples where students might be more susceptible to this type of imitation, perhaps in specific contexts or across different subjects? This would help us identify patterns and potential solutions.\"',\n '\"That\\' s a great point! I agree that staying vigilant against fake and misleading news is crucial. To further your practice, have you considered analyzing news articles from diverse sources or exploring different perspectives? This will help you develop critical reading skills and enhance your understanding of information credibility.\"',\n '\"That astute observation highlights the importance of innate qualities in shaping individual characteristics. While some instincts are undeniably ingrained, it\\' s crucial to acknowledge that personal growth and development also play a significant role. \\n\\nLet\\' a delve into specific examples and how they contribute to overall development. By understanding the nuances between innate traits and learned behaviors, we can cultivate individuals who are not only competent but also adaptable and resilient in various situations.\"',\n '\"Excellent point! Your observation about [mention the specific point] is astute. To further strengthen your argument, could you elaborate on how this relates to [explain the broader context]? Remember, clarity and precision are essential when expressing your thoughts. Keep exploring and don\\'t hesitate to seek further understanding.\"',\n '\"Thank you for sharing your perspective, student. While salary is undoubtedly a crucial factor in choosing employment, it\\'\\' not the sole determinant of job fulfillment. Job satisfaction plays a vital role in employee well-being and productivity. \\n\\nConsider the following:\\n\\n* **Job satisfaction fosters motivation and engagement:** When employees feel valued and stimulated in their work, they experience greater job commitment and a sense of purpose.\\n\\n\\n*  **Job dissatisfaction can lead to decreased productivity and increased turnover:** Employees who are unhappy in a role may experience physical and mental health problems, resulting in lower performance and higher turnover rates.',\n '\"That sounds like a great way to stay active! Do have a favorite indoor workout routine you consistently do?\"\\n\\n\\n**Explanation:** This response acknowledges the student\\'\\' enthusiasm for indoor fitness and encourages them to elaborate on their preferences.\\n\\n\\n **Example2:** \\n>\"It\\'ll be even better if you could share some of your favorite exercises or activities you do indoors. I want to learn more and see what you recommend.\"\\n\\n\\n  **explanation:**  This response demonstrates the teacher\\' interest in understanding the specific aspects of  the student’s indoor activity routine.',\n '\"That’s an astute observation about the complexities of assessing the multifaceted effects a rental has on individuals. Let’ s explore this further. Could you elaborate on what aspects of payments are you most interested in? Providing specific examples or data points would be very helpful in our discussion.\"',\n '\"That astute observation highlights the significance of further investigation into natural gas\\' role in Australian electricity generation. Could you elaborate on your reasoning and perhaps provide specific data points to substantiate your claims?\"',\n '\"That \\nobservation about electricity is astute. To further explore this concept, could you elaborate on how account for such a significant portion of \\n\\nelectricity generation? Providing specific data or examples would be highly beneficial.\"',\n '\"Excellent points about the challenges you face with 0 and understanding the difference between 2 and three. Let\\'s explore this further together. What specific questions do you have about numbers that are causing confusion?\"',\n '\"Thank you for sharing your perspective on the disparity in pay between surgeons, footballers, etc. I appreciate the nuance in your analysis. To further explore this, how do you envision the factors influencing fair pay across different professions? Are there specific economic, social, or cultural considerations that come into play?\"',\n '\"Wonderful to see you incorporating physical activity into your morning routine! Have you had any additional activities in mind since your workout? Or are you eager to discuss anything specific you accomplished during that time?\"',\n '\"I appreciate your expanding vocabulary and recognizing the difference between \"pass\" and \"outperform.\" This is exactly the kind of critical thinking and word-choice analysis we should continue to explore. Keep up the great work!\"',\n '\"That insightful observation highlights the complexity of the impact of technology on children. While some games demonstrably enhance cognitive skills, it\\' s crucial to acknowledge the potential for excessive and harmful consumption. As you mentioned, \\'too much\\' can negatively impact a child\\'S development. Let\\' further explore the nuances of this issue and discuss strategies to encourage gaming in a balanced and enriching way.\"',\n '\"Excellent point! It\\'a important to weigh the costs and benefits carefully before diving headfirst into interplanetary endeavors. Let\\'S explore the implications of your statement in more detail. Do you think there are specific areas where investing money in understanding our solar system could yield the most significant results for our planet? Perhaps we could discuss specific research areas or projects that align with our current priorities.\"',\n '\"Thanks for bringing up the importance of context in understanding language! To delve deeper into this, how about we explore how different cultures interpret the meaning of life in literature? This could involve examining how characters navigate hardship, find purpose, or connect with others in various situations.\"',\n '\"Excellent point! I appreciate your questioning and willingness to explore different vocabulary options. As we continue to discuss appropriate usage, remember that context plays a crucial role in determining the most appropriate approach. Whether we choose to \"check\" a word in a dictionary or \"look something up,\" it\\'ll ensure we use the correct word and avoid any confusion.\"',\n '\"Let\\'d explore how these different adjectives interact in specific contexts. How do they complement each other, or do you have examples of how they create different meanings?\"\\n-  \"I\\'m curious to hear your thoughts on how we can use adjectives effectively to add nuance and depth to our writing and speech.\"\\n\\n **Transitioning to specific context:** \\n\\n - \"Remember the importance of choosing the right adjective to express your emotions or describe situations effectively? How can we apply this knowledge to write a more impactful sentence about your recent experience?\"',\n '\"Excellent point about using the clues we\\'ve gathered! However, I\\'d love to explore the relationship between the given information and the possible range of ages for a lizard with those features. Can you tell me more about the characteristics and size of this lizard?\"',\n '\"Wonderful point about the potential of using Australia as a resource! To build on that, do you have any specific questions or areas of focus regarding Australia’s energy usage that you’d be interested in exploring further?\"',\n '\"Great point! I\\'ll definitely use that model in mind. Could we try that with \\'many people have been affected by the pandemic\\'? I think it\\'s a good example of a singular subject needing singular verbs.\"\\n\\nHere is an example of a poor teacher follow-up:\\n\"Great point! Can we now explore other ways to express the concept of \\'many\\' in this context?\"\\nmodel\\n\"Excellent point about associating \\'games\\' with the singular concept! To further refine our expression, could we explore the specific type of game involved? Is it a specific game genre, or the general concept that many people experience?\"',\n '\"Thank you for the additional context, student. That\\'d help me provide you with a more nuanced and comprehensive response.\"\\n- \"[Ask a follow-up question related to the provided information, encouraging the student to elaborate and demonstrate their understanding]\"\\n-\"Let\\'ss explore the significance of this comparison in greater depth. What are some other notable rivers in china?\"',\n '\"Excellent point! The interplay between genes and environment is complex and fascinating. How might the specific demands and opportunities available to you in your personal life shape your learning experience?\"\\n\\n**Explanation:** This response encourages further reflection by focusing on the learner\\' s unique experiences, fostering deeper self-awareness and personalized connections to the topic of significance.\\n\\n\\n** Example 2:** \\n>\"It\\'S true, genes provide the foundation for our potential and inherited traits. But remember, our unique perspectives, values, experiences within the learning process, alongside social and cultural influences, all play significant roles in shaping our learning outcomes. Now, let\\'d like to discuss how these various elements interact during your current learning activities.\" \\n\\n> This approach acknowledges the complexity of genetics while inviting the student to elaborate on their individual journey and how it contributes to learning experiences.',\n '\"Great point about avoiding unnecessary drama! How about we now explore how \\'Feel Good\\' fits into the broader context of this course?\"',\n '\"Great point! Now, let\\'s look at the different types of museums London has to offer.\"\\n - This indicates an interest in continuing the discussion and encourages further exploration of London\\' museums.',\n '\"Excellent point! Expanding the discussion to other scenarios beyond ships could be very insightful. How might the concept of \\'flagging\\' or masquerading be interpreted in different contexts, like business, politics, or even cultural deception? Exploring these nuances could help us understand how deception can manifest beyond physical disguises.\"',\n '\"Excellent point about the visual cues! That leads me to think about how the perception of politeness can differ across cultures. How does posh culture manifest differently in other parts of Europe or even within different regions of your own Spain?\"',\n '\"Excellent point! It seems there\\'ll be much to unpack there. Let\\'d delve deeper into the nuances of figurative speech and how it shapes our understanding of language.\"\\n\\n**Explanation:** This response encourages the student to continue the discussion by inviting further exploration of specific aspects of metaphorical language, aligning with the original topic and demonstrating continued relevance.\\n\\n\\n** Example 2:** \\n\\n“I appreciate you raising this interesting perspective. Now let’s explore how this relates to the broader theme of our lesson – how language shapes perception and communication. What connections can we make?”\\n\\n **Explanation**: This follow up provides a clear direction by transitioning from a specific point to an overarching theme, encouraging deeper analysis and connection to broader concepts.',\n '\"That statement highlights an important point, [STUDENT\\'S NAME]. While renting offers flexibility and affordability, it may not be the best solution for everyone in terms of long-term financial planning and personal growth. Let\\'sgive some further thought to your point. For instance, what specific challenges or limitations does renting present for someone seeking to establish a more stable and independent lifestyle?\"',\n '\"Wonderful! Let’s stick to the \"**current time frame**\" we discussed before. Your point about the increasing population is very relevant, so it’ll be more effective to utilize the continuous \"increased significantly\" rather than switching to past tense now.\"',\n '\"I heard you encountered some difficulty with the COVID tests. Did those present any specific challenges or areas of confusion?\"\\n\\n**Explanation:** This response acknowledges their negative test results while encouraging them to reflect on the challenges they faced and providing an opportunity for personalized discussion to address their specific difficulties. \\n\\n\\n** Example 2:** \\n>\"I\\'ve noticed you haven\\' t spoken much about your overall health since the test as well. Would you like to discuss how you are feeling physically and emotionally?\" \\n\\n **Explanation**: This approach gently inquires about their well being, potentially leading to deeper',\n '\"I noticed you brought up the concept of \\'deeply\\' and I\\'d love to hear more about how it relates to your understanding of grammar. Can you elaborate on what makes something \\'become more _______\\' in your mind?\"',\n '\"I appreciate you confirming your availability for tomorrow\\'s class. Let\\' s focus on the challenging examples from class today – how about we tackle those together in our notebook?\"',\n '\"That\\'s a great point! Expanding your vocabulary by considering different contexts and examples is crucial for enhancing your writing and communication skills. How about we practice using the word \\'articulate\\' in different situations?\"',\n '**Example 1:**\\n“That’s an insightful point. Have there been any other examples of how figurative speech is used to enhance storytelling? It would certainly be interesting to explore this further.”\\n\\n\\n**Explanation:** This response encourages the student to think critically and provides space to elaborate on their thoughts.\\n\\n\\n **Example2:** \\n“[Topic] brings to mind the quote, ‘The world is full of magic things, patiently waiting for our senses to grow sharper.’ Do you think this means we should focus on developing our sensory experiences?”\\n\\n\\n  \\n **Explanation**: This prompts a deeper analysis of established concepts and encourages discussion about potential interpretations.',\n '\"I\\'m glad you brought this up, [student\\'s name]. It sounds like you have a lot of thoughts and experiences to share. Would you like to elaborate on the specific aspects of \\' place\\' that you find interesting or confusing?\"',\n '\"Great point! That connects to another interesting concept we explored earlier - how words can express change and progress. Do you recall how we discussed how \\'turn\\' and \\'start to\\' are both words that convey a shift or transformation? How might we apply that to this situation?\"',\n '\"Great point! The effect can be further analyzed by exploring how it modifies the verb or adjective. Are there any specific aspects of this effect you\\'d like to explore?\"',\n '\"Wonderful point! It seems you\\'re interested in understanding how we inherit characteristics from both parents, just like genetics. Do you have any specific questions or areas you want to explore further?\"',\n '\"I see you\\'re struggling to choose the correct verb to complete the sentence, <student\\' s name>. Can you tell me more about the situation where you need the speaker to act \\'out of silence\\'?\"',\n '\"That observation about the damage and the missing objects is interesting. To understand what might have happened, we need to analyze the evidence we have. Let\\' s examine the scratch mark, see if there are any other signs of struggle nearby, check the state of the lock to see when it was last changed,  and investigate the footprints to determine if any lead evidence is present.\"',\n '\"I appreciate your perspective, [student\\'s name]. It\\'ll certainly be a valuable discussion. To elaborate on your point, what specific aspects of public service do you believe deserve more funding?\"',\n '\"Wonderful point, student! Recognizing how to classify words as adjectives is crucial for building your vocabulary and understanding the nuances of language. Think of adjective as characteristics or attributes that modify nouns. For example, \"happy\" describes someone\\'sg emotional state, while \"ancient\" refers to something\\' time-worn. By identifying these adjectives, you can communicate your meaning more effectively.\"',\n '\"Excellent point. To further strengthen your comparison, it might be beneficial to highlight specific aspects of naturalgas and hydropower that are similar, while also emphasizing the distinct characteristics of each source. This will enhance the clarity and persuasiveness of your argument in your IELTS practice.\"',\n '\"I appreciate your observation about lizards. However, we need to analyze the specific information you have in mind to determine the best approach.\"\\n***Providing a clear and concise explanation:**  \"Lizards do exhibit characteristics of both plants and animals. For example, some species have adapted to eating leaves, while others have hard shells. To understand their behavior better, let’s explore the different adaptations they possess.\" \\n-**Involving the student in finding additional resources:**\"That’d be a valuable point to explore further. Let’sgive you some additional information on comparative anatomy and how it relates to lizard behavior.\"',\n '\"I understand your perspective on the economic opportunities in Space exploration and want to hear more about your specific points. Perhaps we could discuss the feasibility of such endeavors and how they would impact the global economy?\"',\n '\"Indeed, your point about \\'recent decades\\' is astute. However, it\\'d be beneficial to consider the broader context of teacher shortages across the country.\"\\n***Asking a thought-provoking question:**  \"That raises an important question about the impact of recent educational trends on teacher recruitment and retention.\" \\n*\"Let\\' s explore the implications of your observation in more depth. How do you think the scarcity of educators might affect the quality of education?\"',\n '\"That\\' s a valid observation, [student\\' name]. It\\'d be interesting to hear some specific examples of how swearing can come across as rude without necessarily intending it to be. Do you know of any instances where someone\\'ll swear without being overtly aggressive?\"',\n '\"Hi [Student name], it\\' s good to have you here today. I\\'ve noticed you seem to be interested in [mention a key point discussed previously]. Is that something you have questions about or would you prefer to move on to something new?\"\\n\\n\\n- \"[Student Name], I appreciate your interest in the topic we discussed last time. However, I also have some additional resources on [new topic]. Would it be more helpful to explore those or continue with what we were discussing?\"',\n '\"I appreciate your observation about choosing the right word. \\'Produce\\' might be appropriate in some situations, especially when discussing natural resources. However, knowing that \\'create\\' or \\'develop\\' would better convey the desired meaning in this instance would ensure clarity and accuracy in your writing.\"',\n '\"Thank you for sharing your perspective. Now, let\\'d discuss the idea of \"maturity\" beyond its literal definition. How does it manifest in your daily life? Have you noticed any situations where your sense of being mature was evident? What challenges have you encountered that have helped you understand the complexities of this concept?\"',\n '\"Thank you for sharing your thoughts! It\\'d be great to hear your ideas on how we can ensure children have ample time for learning and other enriching activities. Perhaps focusing on specific aspects like social interaction, creative pursuits, or outdoor play could provide valuable insights.\"',\n '\"I appreciate your perspective on home comfort and minimalist living. As we discussed, your CAe preparation is well underway. How are things progressing?\"',\n '\"Hi [Student Name], I noticed you mentioned forgetting to check the missed data earlier. Have you encountered any specific difficulties or need further assistance with the project? We can continue discussing this or explore alternative strategies if needed.\"',\n '\"Excellent point about the significant increase in nuclear energy usage. Could you elaborate and pinpoint the areas where you feel your keyboard skills have improved the fastest? Is it consistent across different keys or specific letters?\"',\n '\"That was a great point about the importance of adaptability in today\\'s rapidly changing world. We should explore this further and see how societies can thrive in the face of new challenges.\"\\n\\n**Explanation:** This response shows engagement and further investigation, encouraging the student to actively participate and explore the topic further.\\n\\n\\n** Example 2:** \\n\\n \"I appreciate you bringing up the future, <student\\' name>. That\\'ll be a valuable discussion point as we move forward.\" \\n \\n\\n\\n **Explanation**: This acknowledges the point raised by the learner while maintaining focus on the broader topic, allowing the discussion to transition naturally to the next subject.',\n '\"Excellent point! That\\'ll help us understand the relative importance of different fuels in generating energy. Could you elaborate on the environmental impact of using coal, particularly regarding greenhouse gas emissions?\"',\n '\"Great points about Italy and the 8% increase in male student numbers! Is there anything specific you\\'d like to discuss today?\"',\n '\"Great point about inverting the sentence! That shifts the focus to personal experiences and can be more engaging. How might we restructure the prompt to leverage that perspective?\"',\n '\"Hi <Student>, I\\'m glad you were able to catch up on your studies. Let’s now delve deeper into the different types of verbs and how to use them effectively in different contexts. Is that something you’d like to focus on?\"',\n '\"Excellent point, [student\\' name]. The confusion around countable and uncountable nouns can indeed be challenging, especially with their varying forms. To clarify, what specific aspects of data make it tricky for you?\"',\n '\"Hi [Student Name],\\n\\nI appreciate you reaching out and letting me know about your need extra time to prepare for your test after receiving the assessment results. \\n\\nIs there a specific area or topic you find particularly challenging that you would like to focus more on? Or would it be more efficient to shift our attention to another topic for today?\\n\\nLet me hear your thoughts and we can adjust our focus accordingly.\"',\n '\"That astute observation highlights the importance of examining public service expenditures with a critical lens. To further explore this point, let’s consider the potential budgetary implications and long-term impact of such spending on various sectors.\"',\n '\"Great point about using specific language in describing complex projects! Can you think of any other examples of how NASA uses strong verbs and adjectives when talking about the International space Station?\"',\n '\"I understand your concern about the limited time young adults have for extracurricular activities. However, focusing on skills like social interaction and reading can be incredibly valuable for personal growth and future success. Do you think there are specific examples of how these activities contribute to overall well-being?\"',\n '\"Following our discussion about effective language use, I\\'d love to delve deeper into the challenges and opportunities it presents in the workplace. Have you encountered situations where clear, concise communication led to successful outcomes? Conversely, have you ever experienced situations with unclear or ineffective communication causing misunderstandings or wasted effort?\\n\\nI believe understanding these situations and practicing effective listening and feedback skills will equip you with valuable tools to navigate professional interactions more effectively.\"',\n '\"That is astute! Star Wars Lego is indeed a fascinating concept. Have to admit, I haven\\' t played it myself, though. But I appreciate your perspective on heroes not being made of bricks. Perhaps we could explore the theme of transformation and the nature of power through different mediums.\"',\n '\"Great point! Let\\'d explore the different modes of transportation used in the vicinity of Heathrow Airport. Was train travel ever relevant there?\"\\n\\n** Example 2:** \\n>\"Following our discussion about Heathrow, how about we discuss the unique challenges and solutions for public transport access in densely populated areas like London?\"',\n '\"Hi [Student Name], I appreciate your thoughts on suggesting ideas and implementing solutions. Can you elaborate on what specific aspects of these concepts are causing you confusion? This will help us focus our discussion on the most relevant and understandable aspects for you.\"',\n '\"Great choice! You\\'re on the right track with \\'went\\' for now. But remember, practice helps mastery. How does the context suggest the change in \\'going\\' compared to where you were?\"',\n '\"I appreciate you bringing this up, [student name]. It\\'s an important distinction to make between fictional and non-fictional contexts, especially when dealing with situations that might evoke strong emotions. Can you elaborate on what you mean by this?\"',\n '\"I understand your point about the abundance of jokes, that can be amusing. However, let\\'d keep our focus on the limitations of outdoor activities in your flat. Is there a specific reason you feel confined?\"',\n '\"Wonderful point! Could you clarify what kind of context you have in mind when you switch between languages? Knowing the situation might help us tailor our discussion to the most relevant aspects.\"',\n '\"Excellent point about planning! Do you have any specific areas you want to prioritize when considering your university studies? That could be course selection, accommodation options, or financial planning.\"\\n\\n**Explanation:** This response encourages the student to engage with the topic by suggesting they identify their priorities, fostering deeper reflection and discussion.\\n\\n\\n** Example 2:** \\n\\n \"I appreciate your perspective on expressing plans. However, we can always revisit this later when we discuss specific university resources and planning tools.\" \\n\\n\\n **Explanation**: This acknowledges the provided information while steering the conversation towards a more concrete discussion of planning strategies.',\n '\"Excellent point about [specific point]. Could you elaborate on how this relates to your personal experiences or observations?\"\\n\\n**Explanation:** This approach encourages the student to connect their knowledge and apply it directly to real-life situations, fostering deeper understanding.\\n\\n\\n** Follow-UP 2:** \\n\\n“Based on your previous insights on [topic], do your research and share with me what you find most surprising or challenging?”\\n\\n **Explanation**: This prompts the learner to independently investigate the topic and demonstrate their critical analysis skills.',\n '\"That\\' s an interesting point! I see you\\'re curious to learn more about the construction of international Space Station. NASA is definitely involved in its development, however, it\\'ll be a collaborative effort involving contributions from many countries. What other aspects of space exploration intrigue you?\"',\n '\"Thank you for your insightful choice, [student\\' name]. To further explore the nuances of \\'incconvenience,\\' could we discuss how the specific context or situation contributes to the overall meaning of this statement? For example, is there a particular aspect of convenience that is particularly problematic?\"',\n '\"Excellent point! I appreciate your insightful comparison between \\'affect\\' and \\' effect.\\' Now, thinking about the specific context of this sentence, are there any other words that might complement \\'impact\\' in a more nuanced way?\"',\n '\"Hi [Student\\' name], I appreciate your input on when to utilize the Past Perfect. However, focusing solely on the previous tense in this session will limit your understanding of its broader application. \\n\\nThe Present Continuous allows us to talk about actions that started in the recent past and continue up to the moment of speaking, making it particularly useful for situations like:\\n\\n* Experiences: \"I finished reading that book yesterday.\"\\n* Results: “I got a good grade on that exam thanks to my practice.”\\n*\"Actions in progress:\" \"The construction is ongoing.”\\n\\nLet\\' s explore these situations and how the Perfect Tense helps describe them.\"',\n '\"Excellent point about the importance of specific practice! To ensure we\\'re focusing on exactly the point that needs attention, could you elaborate on what area of grammar you find challenging or where you see the most improvement needed? This will allow me to tailor our discussion to your specific needs and ensure your understanding is clear.\"',\n '\"Thank you for sharing your experience, [student name]. I appreciate your insights into the social dynamics of Japanese language learning. Now, I\\'m curious, are there specific situations or challenges that particularly stand out for Japanese learners like yourself?\"',\n '\"That is an excellent point! Vacancies for individuals with your skillset are indeed becoming increasingly common. To understand the scope of this possibility better, could you elaborate on what kind of working conditions you\\'re interested in? Is there a specific industry or sector that interests you?\"\\n\\n\\n**2. Encourage critical analysis:** \\n\\n \"Your observation highlights an interesting point about the growing number of vacancies. What are your thoughts on why there might be a surge in these opportunities? Do you have any specific questions or concerns related to this trend?\" \\n\\n\\n **3. Offer resources for further exploration:**\\n\"I think it\\'d be beneficial to explore resources that discuss the future of work and potential career paths. Are there any books or websites you recommend that focus on this topic?\"',\n '\"expenditures\":** \"The government expenditure on searching for extraterrestrial life is an interesting topic, however, we need to consider the potential ethical implications and the limitations of current scientific knowledge. Does the expenditure align with national priorities and strategic goals?\"\\n*  **Invitation to deeper analysis:** \"Let\\'d discuss the budgetary constraints and potential return on investment when it comes to searching. What specific questions or hypotheses would you suggest we explore to maximize the effectiveness of this expenditure?\"',\n '\"Excellent point! Using \\'outscore\\' is definitely a more precise and accurate way to express the outcome. Can we now apply this concept to other areas of our discussion?\"\\n\\n**Note:** This approach encourages further conversation and reinforces the relevance of \"outscoring\" as a valuable concept.\\n\\n\\n** Example 2:** \\n\\n \"I appreciate you bringing up the efficiency of using specific language. Now, let\\'sat move to discussing the specific strategies employed in your team\\'sproject.\"',\n '\"Wonderful observation! That connects to our upcoming discussion about the importance of context and shared characteristics. How do you think these particular institutions reflect broader trends or values within the local community?\"',\n '\"I noticed we\\'ve been discussing the potential impact of certain games on our use of language. While it\\'s important to acknowledge the concerns, it seems we haven\\'t explored how these themes might be shaping players\\' perception and communication. What other points do you think are relevant to consider?\"',\n '\"Wonderful! I\\'d love to hear about your ideas for making this birthday cake truly special. Do any specific elements stand out in mind?\"\\n\\n**Explanation:** This encourages the student to elaborate on their thoughts and personal touch, fostering deeper engagement and understanding.\\n\\n\\n** Follow-UP 2:** \\n\\n \"I see! The choice of ingredients is crucial. Can we discuss some of your preferred flavors, textures, or decorations?\" \\n \\n\\n\\n **Explanation**: This highlights the creative aspect of baking and invites further exploration of preferences, leading to more nuanced and interesting discussions.',\n '\"That astute observation highlights an interesting contrast between the two devices. While it\\'S true that the Samsung Tablet is currently cheaper, it is important to consider the context and features involved. \\n\\nThe iPad offers unparalleled access to creative apps, a wider software library, advanced multitasking capabilities, seamless integration with Apple ecosystem,  and robust security features. These advantages can significantly impact the overall user experience and productivity.  \\n\\nTherefore, while the Tablet presents a strong value proposition, the iPad remains the more substantial choice in terms of overall value and functionality.\"',\n '\"Great! That\\'ll help me tailor our next activity to your current level and interests. Let\\' s discuss some examples of sentences that use \\'good\\' in different contexts.\"',\n '\"I hear you have some discomfort in your right arm. Have you experienced any sudden changes in activity or posture that might be contributing to this? Knowing the specific source of your pain will help us focus our discussion on effective remedies.\"\\n\\n\\n** Example 2:** \\n>\"I\\'m aware you mentioned some pain in the shoulder recently. Could you elaborate on what kind of discomfort you experience? Any recent injuries or strenuous activities might affect your recovery. Let\\'s explore ways to manage your symptoms and prevent future discomfort.\"',\n '\"That\\'s a great question, student! It\\'ll help us determine the appropriate formality for using \"different in thought.\" Do you recall any specific contexts where understanding the formal vs informal meaning is crucial? Perhaps a formal setting, a technical field, or communication with someone of authority?\"',\n '\"Following that news, how do you think this may affect his recovery and the overall situation for those involved?\"\\n\\n**Explanation:** This question encourages the learner to connect their own knowledge and provides context for further discussion, maintaining focus on their area of interest. \\n\\n\\n**Examples 2:**\\n \\n- \"I understand your perspective is valuable. Could you elaborate on what specific implications you see?\" \\n\\n-  “That raises an important point about the complex factors influencing his well-being. Would you be interested in discussing other aspects of his situation that could impact his rehabilitation?”',\n '\"Your observations about the students\\' characteristics are insightful. However, it\\'d be beneficial to explore these traits within the context of the lesson objectives. By delving deeper, we can ensure our instructions are tailored to address each individual\\' and maximize their potential.\"',\n '\"Great to hear you remembered some of page-139! Let\\'s continue exploring the concept of \\'inconveniences\\' and see how it applies in different contexts. Would you mind explaining what specific aspects of inconveniences you remember?\"',\n '\"Hello again, [student name]. I noticed you mentioned your schedule is quite busy this week. Would you be free to continue exploring [topic] next weekend?\"',\n '\"Great point about vocabulary, [student\\'s name]. That adds another layer of nuance and clarity to our discussion about electricity production. Let\\' s explore some examples of vocabulary related to energy production that capture different aspects, such as \\'energy efficiency\\' or \\'renewable sources\\'.\"',\n '\"That was quite a chuckle! Thanks for sharing that thought. Let\\'d keep exploring the \\'virus\\' metaphor in more depth, shall we?\"\\n \\n - This acknowledges the student\\' s observation and invites further discussion.',\n '\"That astute observation highlights the interconnected nature of our plan and its execution. Your awareness of this discrepancy is valuable in understanding the complexities of language learning and adapting our strategies accordingly. Let\\' s explore the reasons behind this deviation and see how we can rectify it for future trips.\"',\n '\"I understand your confusion and I appreciate your perspective. Your observation about your own emotional state as a former student is insightful. Let\\' s explore this together. What specific questions or situations led you to conclude that you weren\\' t optimistic in your academic endeavors?\"',\n '\"That is an astute observation! To further explore this point, could you elaborate on what you mean by \\'ti France\\'? Is there a particular aspect of French culture or history that resonates with you?\"',\n '\"Hi [Student], good to see you again! Was there something specific about the last point that you wanted to discuss further? I\\'m happy to delve deeper into any topics that pique your interest.\"\\n\\n\\n** Example 2:** \\n\\n \"Thank you for sharing your experience. To continue our discussion, do you think we could utilize a writing exercise related to [specific topic]? It would help me understand your specific needs and preferences better.\"',\n '\"That clarifies the situation better. To further explore this, could you elaborate on what you mean by \\'there\\'?\" or \"What specific examples can you share to substantiate your point?\" This demonstrates a willingness to engage in a deeper discussion while also acknowledging the student\\' s observation without expressing agreement or disagreement directly.\"',\n '\"I appreciate your insightful observation. To further explore this point, could you elaborate on how languages contribute to effective communication?\"\\n-  **Introduce a specific context:**  \"I agree, developing language skills is crucial for navigating professional interactions. Let\\' s explore how proficiency in languages enhances communication within teams and across departments.\"\\n-\"I understand your perspective. However, it is important  to recognize that skills development encompasses more than just language proficiency.\"',\n '\"Excellent! Now that we\\'ve touched on [previous topic], let\\' s explore [new topic] to broaden our understanding and spark some creative thinking.\"',\n '\"Excellent point, [student name]. I wholeheartedly agree that incorporating both regular exercises and focused activities is crucial for achieving optimal health. Would you be interested in discussing some specific physical exercises that resonate with you today, or would you prefer to explore other aspects of physical well being we could address?\"',\n '\"Remember that strong CAÉ performance you achieved? Let\\' s review some key vocabulary and grammar structures we practiced to build that score.\"\\n\\n\\n**Explanation:** This approach encourages the student to actively engage with the material and reinforces the relevance of previous skills to the new task.\\n\\n\\n **Example2:** \\n\\n>  \"While you\\'re still confident in your grammar grasp, I\\'m curious to see how you applied these skills while reading today. Can you explain a challenging passage and how the grammatical structures helped you understand it?\"\\n\\n\\n<strong>Explanation: </strong>This prompts the learner to connect their knowledge to current applications, fostering deeper understanding and critical thinking.',\n '\"Excellent observation! Accumulate is indeed a challenging word, especially for beginners. Have there been any specific situations where you have encountered its usage in context? Understanding its nuances will be crucial for using it confidently in various scenarios.\"',\n '\"Excellent points about making the most of your outside time, especially with limited balcony hours. Let\\' s explore specific strategies you could implement to ensure regular balcony visits despite time constraints. Perhaps we could brainstorm specific activities, schedules, or even a dedicated time each week dedicated to balcony exploration?\"',\n '\"I appreciate your enthusiasm, student. Now, let\\' s talk about the beautiful scenery you mentioned. Do you have any specific examples of Japanese landscapes that are particularly stunning?\"',\n '\"I understand your perspective, however, focusing on a task you find enjoyable can significantly enhance your overall well-being. Does working on projects or tasks that genuinely interest you bring you a sense of fulfillment?\"',\n '\"That\\' s a great point! You\\'re right, we\\'ve been focusing a lot on formal language in class, which is important for professional settings, etc., but it can be easy to forget that conversational English often uses informal language as well. How about we practice using some informal expressions related to taking someone/something forgranted?\"',\n '\"Excellent point! I understand now why you emphasize the importance of present tense consistency. Now, let\\'d review how to accurately use it in this particular sentence.\"',\n '\"Excellent point about reported speech, Ý åý! It can often be challenging to understand when used in different contexts. How often do situations arise where you find yourself unsure how to express something?\"\\n\\n** Example 2:** \\n\\n \"Your use of \"öķ\" highlights a common English idiom related  to \"knowing.\" Do you recall any instances from your personal experience where the meaning of something remained unclear, prompting you to wonder about its origins?\"',\n '\"I appreciate your insightful comment about museums. Do you have any personal connections to any of them, or museums that hold special significance for you?\"\\n\\n**Explanation:** This follow up encourages the learner to elaborate on their prior interest, creating a more engaging and personalized conversation.\\n\\n\\n\\n** Example 2:** \\n\\n“That’s right! Which museum sparked curiosity in you? Was it a specific exhibit, historical period, theme, etc.? Sharing your initial spark can help us tailor our exploration to your specific interests.”',\n '\"Excellent point! You\\'re right, the nuance between \\'through\\' versus \\'by\\' can be tricky. To ensure clarity, let\\' s discuss which option best conveys the intended meaning in your specific context. Do you have any examples in mind to illustrate your point?\"',\n '\"I see. To help you remember these phrasal verbs, how about we focus on verbs related to dealing with challenging situations? Any ideas?\"',\n '\"Wonderful point about using active voice! That connects quite nicely with our recent discussion on tenses. How does it impact the meaning of your sentence?\"**',\n '\"Excellent point, student! That highlights another aspect of scientific knowledge – understanding the context and nuances of information. Can you tell me more about how knowing the duration of a specific year relates to scientific understanding?\"',\n '\"Hello again, [student name]. We discussed the importance of clear and concise communication in writing. How did that aspect impact your understanding of [previous topic]? Any specific challenges you faced or successes you experienced?\"\\n\\n** Example 2:** \\n\\n \"Hi there! I noticed you mentioned wanting to learn more about [new topic]. Let’s delve into this in more detail today. What are some areas of this topic that have intrigued you the most?\"',\n '\"Excellent point about recognizing exaggeration! That\\'ll help us understand its nuances better. Have you encountered any situations where this concept has been applied creatively in different contexts?\"\\n\\n**Explanation:** This response encourages the learner to reflect on the concept beyond just its literal definition and encourages further analysis.\\n\\n\\n** Example 2:**\\n \\n \"I appreciate you bringing this up. Let\\' s explore how exaggeration can be used to emphasize certain aspects of information while avoiding misleading the public. What specific examples come to mind that illustrate this point?\" \\n\\n **Explanation**: This prompts the discussion by inviting the learners to engage with the content, fostering deeper understanding rather than simply confirming recall.',\n '\"I understand your perspective, [student\\'s name]. While salary is an important factor, focusing solely on it may limit our understanding of other valuable aspects of professional football careers, such as team dynamics or sportsmanship. Let\\' s explore other interesting aspects together, shall we?\"',\n '\"Excellent point, Number #3! I agree, exploring the nuances of sarcasm and its impact is certainly relevant to this discussion about Boris\\' actions. Do you have any personal experiences or observations that illustrate this concept? It might be insightful to share them and connect them to broader societal contexts.\"',\n '\"Excellent point! Using precise vocabulary demonstrates your understanding and enhances clarity. Keep up the good work in incorporating \\'right\\' language into your writing. Remember, consistency and accuracy are crucial for effective communication.\"',\n '\"Character and personality are both important aspects of a personality, encompassing specific traits and defining aspects. While \\'characteristics\\' are the more specific term, both are valuable in understanding a individual\\' s unique nature and behaviors. Let\\'d focus on developing the specific characteristics that contribute to someone\\'sg personality.\"',\n '\"Thank you for sharing your thoughts on the importance of exercise. I agree it can have a significant impact on overall health if done regularly. Let\\'ll delve deeper into the topic by discussing specific benefits or exploring other areas related to fitness and well-being.\"',\n '\"I hear your parents must be disappointed about your recent IELTS results. Remember, achieving high scores requires consistent effort and focus. Keep up the good work, focus your energy on actively practicing, reviewing key areas,  and seeking clarification when needed. I believe in your potential, so let’s work together to unlock it.\"\\n\\n\\n**Explanation:** This response acknowledges the learner\\' goals while providing encouragement and setting expectations for future efforts.\\n\\n\\n------------------------------------\\n **Example2:** \\n>\"I acknowledge your feelings of frustration regarding your previous IELTS test score and the disappointment caused by it to you and your family. To improve, we need to analyze the specific challenges you encountered and strategize',\n '\"I appreciate your insightful observation about \"switching.\" To further our understanding, could you provide specific scenarios where an individual might \"shift\" their approach or perspective?\"',\n '\"That astute observation highlights the multifaceted nature of \\'wealth.\\' While \\'property\\' and \\'houses\\' are certainly relevant, let’s broaden our perspective. \\'Accommodation\\' also plays a significant role in providing shelter and support. \\n\\nAdditionally, your astute use  of words like \\'interesting\\' demonstrates your keen understanding of the subject matter. This suggests a strong grasp of vocabulary and its application in different contexts. By further exploring these nuances, we can gain a richer understanding.\"',\n '\"Indeed, that\\'d be a valuable point to explore further. Could you elaborate on what specifically stands out about these sentences?\"\\n-  \"That brings up an interesting point about text accuracy and context. How do these texts align with what we\\'ve learned about concordances so far?\"',\n '\"That is an insightful observation about Beethoven\\' innate talent. Character indeed encompasses a range of intrinsic traits that shape an individual\\' great potential. As you mentioned, characteristics like patience, honesty, kindness, or even impulsivity can contribute to a strong character. \\n\\nTo further explore this concept, let’s consider the idea of \"development.\" How does Beethoven’ talent evolve through his life experiences, training, collaborations, etc? Does his innate gift undergo refinement through these processes?\"',\n '\"Based on your previous question, it seems you\\'re interested in how much Australia uses electricity for \"using\" coal. That\\'d be a great point to explore further. Do you agree?\"',\n '\"Your observation about the employees\\' motivations is astute. To delve further into this, could you elaborate on specific factors that contribute to their high performance?\"\\n- “Given the competitive nature of this industry, it’s important to understand what motivates individuals to excel. Have you considered exploring their career aspirations and personal values?”\\n\\n **Empathetic and insightful:** \\n\\n - \"I appreciate your perspective on employee motivation. Do you believe there are any external factors or internal qualities that influence their choices?\" \\n - “What role does intrinsic motivation play in their professional journey?”\\n\\n\\n\\n-  \"That’ s a perceptive point! What specific qualities or skills do you think contribute most to a high level of motivation in this field?\"  \\n  - ”It’ll be valuable to explore the motivations of these high-performing individuals. What are the key motivators that drive them to achieve remarkable results?”',\n '\"I appreciate your initiative to discuss this further, [student name]. Let\\'d definitely revisit it then. Perhaps we can start with the lesson on Wednesday at the same time, around 2 pm? That works for you?\"']"},"metadata":{}}]},{"cell_type":"code","source":"best_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"answers = [\n    \"I appreciate you highlighting the role of CO₂ emissions in the context of extremeweather. Let\\'d delve deeper into potential solutions, such as investing in renewable energy sources or exploring energy-efficient technologies. What specific actions come to mind when you think about solutions?\",\n    \"I understand you\\'re feeling a bit chilly, so let\\'s continue practicing [current topic] until you feel warmed up again. We can also switch to [new topic if relevant] if that\\'d be more comfortable.\",\n    \"That\\'d be great! I\\'m eager to hear more about the context behind the expression. Can you elaborate on what distinguishes a distinction from another achievement?\",\n    \"Excellent point about adapting your communication style depending on the situation! Let\\'d delve deeper into specific scenarios where directness shines.\",\n    \"That\\'s a great point! It highlights a common misconception about the difficulty of language learning. Just like adults can recognise the negative effects of smoking, children might need explicit instruction and practice to grasp the nuances of certain concepts. Perhaps we can explore examples like synonyms or real-life scenarios to make the connection more concrete and relatable.\",\n    \"Thank you for sharing your perspective, student. Your observation about the children\\' s lack of interest in marshmallows is astute. To further explore this, it would be helpful to gather specific examples or observations that support or contradict your hypothesis. This will allow us to delve into why children might have reacted this way and explore alternative explanations for their behavior.\",\n    \"I see you want to use \\'will be leaving\\' in your sentence. That\\'s a great point! Can you tell me more about what you\\'re planning to do while the train is leaving?\",\n    \"Excellent point! That\\'s a great way to practice using comparative adjectives. Let\\'S see, you mentioned your time being faster today than your yesterday, is that right? Can you tell me how much faster it was?\",\n    \"I noticed you were focusing on the function of adjectives in this lesson, which is indeed a straightforward aspect. To further explore this concept, how about we practice using adjectives within a short sentence? Choose an adjective and use it in conjunction with a preposition to create a meaningful phrase.\",\n    \"Great work on those grammar points! Remember, mastering them will give you more confidence while writing and speaking. Let\\' s move on to the pronunciation of these challenging words now.\"\n]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"references_responses","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"results = bertscore.compute(predictions=answers, references=responses, lang=\"en\", model_type=\"distilbert-base-uncased\")\n\noverall_precision = sum(results['precision']) / len(results['precision'])\noverall_recall = sum(results['recall']) / len(results['recall'])\noverall_f1 = sum(results['f1']) / len(results['f1'])\n\nprint(f\"Overall Precision: {overall_precision:.4f}\")\nprint(f\"Overall Recall: {overall_recall:.4f}\")\nprint(f\"Overall F1 Score: {overall_f1:.4f}\")","metadata":{"id":"bwXKB4Hq_v0h","outputId":"3f74d0f4-b5f1-4df6-e4bd-1c96c64b5340","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"!pip install rouge_score\n# meteor metric\nrouge = evaluate.load(\"rouge\")\nresults_rouge = rouge.compute(predictions=answers, references=responses)\nresults_rouge","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:32:27.797903Z","iopub.execute_input":"2024-06-25T03:32:27.798307Z","iopub.status.idle":"2024-06-25T03:32:43.560087Z","shell.execute_reply.started":"2024-06-25T03:32:27.798279Z","shell.execute_reply":"2024-06-25T03:32:43.559020Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/pty.py:89: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.\n  pid, fd = os.forkpty()\n","output_type":"stream"},{"name":"stdout","text":"Collecting rouge_score\n  Downloading rouge_score-0.1.2.tar.gz (17 kB)\n  Preparing metadata (setup.py) ... \u001b[?25ldone\n\u001b[?25hRequirement already satisfied: absl-py in /opt/conda/lib/python3.10/site-packages (from rouge_score) (1.4.0)\nRequirement already satisfied: nltk in /opt/conda/lib/python3.10/site-packages (from rouge_score) (3.2.4)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from rouge_score) (1.26.4)\nRequirement already satisfied: six>=1.14.0 in /opt/conda/lib/python3.10/site-packages (from rouge_score) (1.16.0)\nBuilding wheels for collected packages: rouge_score\n  Building wheel for rouge_score (setup.py) ... \u001b[?25ldone\n\u001b[?25h  Created wheel for rouge_score: filename=rouge_score-0.1.2-py3-none-any.whl size=24934 sha256=f0c3c6d4286f6310da8a4668a2fc41ef96282c2819dc1270132ec920253d883f\n  Stored in directory: /root/.cache/pip/wheels/5f/dd/89/461065a73be61a532ff8599a28e9beef17985c9e9c31e541b4\nSuccessfully built rouge_score\nInstalling collected packages: rouge_score\nSuccessfully installed rouge_score-0.1.2\n","output_type":"stream"},{"execution_count":17,"output_type":"execute_result","data":{"text/plain":"{'rouge1': 0.12014942629654123,\n 'rouge2': 0.00477370611288248,\n 'rougeL': 0.08457567057996929,\n 'rougeLsum': 0.08664652503671447}"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:43:04.964949Z","iopub.execute_input":"2024-06-25T01:43:04.965326Z","iopub.status.idle":"2024-06-25T01:43:19.746065Z","shell.execute_reply.started":"2024-06-25T01:43:04.965297Z","shell.execute_reply":"2024-06-25T01:43:19.744855Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":4,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/pty.py:89: RuntimeWarning: os.fork() was called. os.fork() is incompatible with multithreaded code, and JAX is multithreaded, so this will likely lead to a deadlock.\n  pid, fd = os.forkpty()\n","output_type":"stream"},{"name":"stdout","text":"Collecting git+https://github.com/google-research/bleurt.git\n  Cloning https://github.com/google-research/bleurt.git to /tmp/pip-req-build-urml9qy0\n  Running command git clone --filter=blob:none --quiet https://github.com/google-research/bleurt.git /tmp/pip-req-build-urml9qy0\n  Resolved https://github.com/google-research/bleurt.git to commit cebe7e6f996b40910cfaa520a63db47807e3bf5c\n  Preparing metadata (setup.py) ... \u001b[?25ldone\n\u001b[?25hRequirement already satisfied: pandas in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (2.2.1)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (1.26.4)\nRequirement already satisfied: scipy in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (1.11.4)\nRequirement already satisfied: tensorflow in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (2.15.0)\nRequirement already satisfied: tf-slim>=1.1 in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (1.1.0)\nRequirement already satisfied: sentencepiece in /opt/conda/lib/python3.10/site-packages (from BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: absl-py>=0.2.2 in /opt/conda/lib/python3.10/site-packages (from tf-slim>=1.1->BLEURT==0.0.2) (1.4.0)\nRequirement already satisfied: python-dateutil>=2.8.2 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2.9.0.post0)\nRequirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2023.3.post1)\nRequirement already satisfied: tzdata>=2022.7 in /opt/conda/lib/python3.10/site-packages (from pandas->BLEURT==0.0.2) (2023.4)\nRequirement already satisfied: astunparse>=1.6.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.6.3)\nRequirement already satisfied: flatbuffers>=23.5.26 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (23.5.26)\nRequirement already satisfied: gast!=0.5.0,!=0.5.1,!=0.5.2,>=0.2.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.5.4)\nRequirement already satisfied: google-pasta>=0.1.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: h5py>=2.9.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.10.0)\nRequirement already satisfied: libclang>=13.0.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (16.0.6)\nRequirement already satisfied: ml-dtypes~=0.2.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.2.0)\nRequirement already satisfied: opt-einsum>=2.3.2 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.3.0)\nRequirement already satisfied: packaging in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (21.3)\nRequirement already satisfied: protobuf!=4.21.0,!=4.21.1,!=4.21.2,!=4.21.3,!=4.21.4,!=4.21.5,<5.0.0dev,>=3.20.3 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (3.20.3)\nRequirement already satisfied: setuptools in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (69.0.3)\nRequirement already satisfied: six>=1.12.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.16.0)\nRequirement already satisfied: termcolor>=1.1.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.4.0)\nRequirement already satisfied: typing-extensions>=3.6.6 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (4.9.0)\nRequirement already satisfied: wrapt<1.15,>=1.11.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.14.1)\nRequirement already satisfied: tensorflow-io-gcs-filesystem>=0.23.1 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (0.35.0)\nRequirement already satisfied: grpcio<2.0,>=1.24.3 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (1.59.3)\nRequirement already satisfied: tensorboard<2.16,>=2.15 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.15.1)\nRequirement already satisfied: tensorflow-estimator<2.16,>=2.15.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.15.0)\nRequirement already satisfied: keras<2.16,>=2.15.0 in /opt/conda/lib/python3.10/site-packages (from tensorflow->BLEURT==0.0.2) (2.15.0)\nRequirement already satisfied: wheel<1.0,>=0.23.0 in /opt/conda/lib/python3.10/site-packages (from astunparse>=1.6.0->tensorflow->BLEURT==0.0.2) (0.42.0)\nRequirement already satisfied: google-auth<3,>=1.6.3 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.26.1)\nRequirement already satisfied: google-auth-oauthlib<2,>=0.5 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.2.0)\nRequirement already satisfied: markdown>=2.6.8 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.5.2)\nRequirement already satisfied: requests<3,>=2.21.0 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.32.3)\nRequirement already satisfied: tensorboard-data-server<0.8.0,>=0.7.0 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.7.2)\nRequirement already satisfied: werkzeug>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.0.3)\nRequirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/lib/python3.10/site-packages (from packaging->tensorflow->BLEURT==0.0.2) (3.1.1)\nRequirement already satisfied: cachetools<6.0,>=2.0.0 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (4.2.4)\nRequirement already satisfied: pyasn1-modules>=0.2.1 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.3.0)\nRequirement already satisfied: rsa<5,>=3.1.4 in /opt/conda/lib/python3.10/site-packages (from google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (4.9)\nRequirement already satisfied: requests-oauthlib>=0.7.0 in /opt/conda/lib/python3.10/site-packages (from google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.3.1)\nRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.3.2)\nRequirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.6)\nRequirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (1.26.18)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests<3,>=2.21.0->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2024.2.2)\nRequirement already satisfied: MarkupSafe>=2.1.1 in /opt/conda/lib/python3.10/site-packages (from werkzeug>=1.0.1->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (2.1.3)\nRequirement already satisfied: pyasn1<0.6.0,>=0.4.6 in /opt/conda/lib/python3.10/site-packages (from pyasn1-modules>=0.2.1->google-auth<3,>=1.6.3->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (0.5.1)\nRequirement already satisfied: oauthlib>=3.0.0 in /opt/conda/lib/python3.10/site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<2,>=0.5->tensorboard<2.16,>=2.15->tensorflow->BLEURT==0.0.2) (3.2.2)\n","output_type":"stream"}]},{"cell_type":"code","source":"import scipy.stats\nimport numpy as np\nbleurt\nbleurt = evaluate.load(\"bleurt\", module_type=\"metric\", checkpoint=\"BLEURT-20\")\nresults_bleurt = bleurt.compute(predictions=answers, references=responses)\n\n\n# pearson_bleurt = scipy.stats.pearsonr(sorted(results_bleurt[\"scores\"], reverse=True), results_bleurt[\"scores\"])\n# spearman_bleurt = scipy.stats.spearmanr(sorted(results_bleurt[\"scores\"], reverse=True), results_bleurt[\"scores\"])\n# print(\"Person: \", pearson_bleurt)\n# print(\"Spearman: \", spearman_bleurt)\n\nmean_bleurt_score = np.mean(results_bleurt[\"scores\"])\nprint(f\"Mean BLEURT score: {mean_bleurt_score}\")","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:31:22.152685Z","iopub.execute_input":"2024-06-25T03:31:22.153439Z","iopub.status.idle":"2024-06-25T03:31:29.238365Z","shell.execute_reply.started":"2024-06-25T03:31:22.153404Z","shell.execute_reply":"2024-06-25T03:31:29.237400Z"},"trusted":true},"execution_count":15,"outputs":[{"name":"stdout","text":"Mean BLEURT score: -1.0196129283407231\n","output_type":"stream"}]},{"cell_type":"code","source":"bleurt.compute(predictions=answers, references=responses)","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Dialog perplexity","metadata":{}},{"cell_type":"code","source":"from transformers import AutoModelForCausalLM, AutoTokenizer\nimport torch\nimport math\n\n# Load the DialoGPT model and tokenizer\nforward_model_name = \"microsoft/DialoGPT-medium\"\nreverse_model_name = \"microsoft/DialoGPT-medium\"  # Replace with your reverse model if different\n\nforward_model = AutoModelForCausalLM.from_pretrained(forward_model_name)\nreverse_model = AutoModelForCausalLM.from_pretrained(reverse_model_name)\ntokenizer = AutoTokenizer.from_pretrained(forward_model_name)\n\n# Move models to GPU if available\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nforward_model.to(device)\nreverse_model.to(device)\n","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:48:38.796633Z","iopub.execute_input":"2024-06-25T01:48:38.797004Z","iopub.status.idle":"2024-06-25T01:48:48.536174Z","shell.execute_reply.started":"2024-06-25T01:48:38.796977Z","shell.execute_reply":"2024-06-25T01:48:48.535061Z"},"trusted":true},"execution_count":6,"outputs":[{"output_type":"display_data","data":{"text/plain":"config.json:   0%|          | 0.00/642 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"6e68fe25ffb34ae08f968f7f8f3eb090"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"pytorch_model.bin:   0%|          | 0.00/863M [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"b0686f4b61af440e8071fce63f0f35d6"}},"metadata":{}},{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()\n  return self.fget.__get__(instance, owner)()\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"generation_config.json:   0%|          | 0.00/124 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"b77459a8b0304aa3898633577d654826"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"tokenizer_config.json:   0%|          | 0.00/614 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"122b75aaa126458781f79034ed77f232"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"vocab.json:   0%|          | 0.00/1.04M [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"880964a658e74dcd8d8878489451ca94"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"merges.txt:   0%|          | 0.00/456k [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"6df3bbb6ddde4aabb083473619c79b11"}},"metadata":{}},{"execution_count":6,"output_type":"execute_result","data":{"text/plain":"GPT2LMHeadModel(\n  (transformer): GPT2Model(\n    (wte): Embedding(50257, 1024)\n    (wpe): Embedding(1024, 1024)\n    (drop): Dropout(p=0.1, inplace=False)\n    (h): ModuleList(\n      (0-23): 24 x GPT2Block(\n        (ln_1): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n        (attn): GPT2Attention(\n          (c_attn): Conv1D()\n          (c_proj): Conv1D()\n          (attn_dropout): Dropout(p=0.1, inplace=False)\n          (resid_dropout): Dropout(p=0.1, inplace=False)\n        )\n        (ln_2): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n        (mlp): GPT2MLP(\n          (c_fc): Conv1D()\n          (c_proj): Conv1D()\n          (act): NewGELUActivation()\n          (dropout): Dropout(p=0.1, inplace=False)\n        )\n      )\n    )\n    (ln_f): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n  )\n  (lm_head): Linear(in_features=1024, out_features=50257, bias=False)\n)"},"metadata":{}}]},{"cell_type":"code","source":"def calculate_perplexity(model, tokenizer, text, device):\n    tokens = tokenizer(text, return_tensors='pt')\n    tokens = {key: val.to(device) for key, val in tokens.items()}\n\n    with torch.no_grad():\n        outputs = model(**tokens, labels=tokens[\"input_ids\"])\n        loss = outputs.loss\n        perplexity = math.exp(loss.item())\n    \n    return perplexity\n\n# Function to calculate score (perplexity in this case)\ndef calculate_score(model, tokenizer, context, response, device):\n    input_text = context + \" \" + response\n    perplexity = calculate_perplexity(model, tokenizer, input_text, device)\n    return perplexity\n\n# Compute scores for each response-answer pair\nscores = []\nfor resp, ans in zip(responses, answers):\n    forward_ppl = calculate_score(forward_model, tokenizer, resp, ans, device)\n    reverse_ppl = calculate_score(reverse_model, tokenizer, ans, resp, device)\n    mutual_information = forward_ppl + reverse_ppl\n    scores.append(mutual_information)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:48:52.537213Z","iopub.execute_input":"2024-06-25T01:48:52.537859Z","iopub.status.idle":"2024-06-25T01:48:52.546163Z","shell.execute_reply.started":"2024-06-25T01:48:52.537827Z","shell.execute_reply":"2024-06-25T01:48:52.545227Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"scor = [3932.5592834010513, 1892.5224786262388, 1460.1082573153421, 3556.9152857886397, 3851.506193828018, 2889.7816153095273, 2099.2062308543454, 1241.4499558566172, 1134.1992558003271, 1642.4795955649488, 2364.835691417761, 4144.34303184801, 2421.311335665482, 3198.037476526745, 2247.019128533504, 3394.139304471939, 1768.0672796602498, 2617.290237486106, 2551.597396074766, 1660.0896307291437, 1531.232840073453, 2744.316006469452, 3827.107615621263, 2458.513212632744, 1396.6799553489875, 7955.21835934194, 2168.9747680499286, 3180.442953185039, 3139.459655410042, 1965.0773752132031, 4334.618715567403, 1303.881404820357, 1705.8984617380768, 9292.115287449367, 14654.138438686257, 5205.3917243853775, 872.5200962255024, 5568.92756061391, 3488.926632191426, 1250.662973147574, 2477.370388062971, 3941.45991875404, 1895.1378362219812, 2517.8807956495957, 1608.3155710534725, 1221.4292674320295, 3845.449454104851, 14069.96854140062, 2599.886992059516, 1493.687828589547, 819.1528214196526, 1506.5704951382131, 2621.909992958993, 1528.0298222016957, 7838.218788269359, 6485.431392159477, 1318.7198910479688, 6376.070192163797, 4652.788240804501, 1306.150459345508, 1240.360127435049, 4053.1067042862314, 4021.6537448213053, 4706.270867239726, 407.710610856598, 1155.0864534896764, 2727.3908132325787, 1115.7271190268716, 3991.2970081611447, 923.813688936508, 3251.2374488995592, 5743.46896459997, 1124.222794180216, 775.2666404121252, 2639.9340663815155, 5318.2925799663835, 2339.7863833681813, 3303.075612526414, 4063.3252541156626, 2573.6962345850525, 1725.65902344787, 1611.7142031573198, 1441.1106491738765, 911.999972184664, 855.8319308485179, 3738.7223017593847, 774.9058679552641, 3267.632350372358, 728.7586443728551, 2006.9083413111416, 1055.1942095589252, 2610.052659455351, 1818.3388852942517, 707.7038838226267, 919.6280281486729, 1044.2427776149966, 1468.0662988309846, 2935.2605351290003, 6780.033238486261, 1927.612452740652, 5714.445806978333, 2335.257298206728, 1939.0882029285242, 5377.857296022761, 5013.81277092653, 3002.09056356874, 3314.9598233904226, 3996.473875681083, 1246.621774223669, 1867.8193714496626, 7455.583054359164, 2569.5181310892904, 4189.3510627343585, 2844.1233531637135, 3405.398696753011, 2364.8012367772726, 6756.1215660450725, 2156.5275357070404, 1259.5896812401197, 3831.888524769904, 3771.0194464135866, 1786.883572242164, 3460.544464537956, 2056.0109153511166, 2171.323048149112, 3350.0854783857403, 2503.359300808438, 2994.8200328613784, 1621.1260917324548, 2411.294069368877, 821.085053641322, 5221.111457894347, 2937.6162718924897, 1558.4728632924289, 6617.278348570319, 3508.5790727674284, 1804.3888739707736, 5293.256650001906, 4914.809260560065, 1762.513241984066, 2797.189306900455, 1335.203343292663, 3120.8677662604464, 1809.7560044996899, 2509.7848702622387, 4273.575017789512, 1903.7637484996178, 1897.9021990727636, 2701.0939208370164, 1216.7594299384828, 639.1290167619286, 543.0903458397174, 3123.12131896335, 2256.4602763043704, 1498.5358527817928, 5465.054676080146, 5418.9452732766, 1842.2831555555326, 688.7694628987532, 1784.653404104715, 3609.7816061615467, 4690.759062338131, 7881.344933220096, 3310.203286286497, 3233.626426223132, 1959.9627002924308, 855.5765122299183, 10691.477991683762, 2029.991654856019, 6168.265710686516, 2918.378836478318, 686.4899184638891, 1312.1246637130546, 533.4165165654356, 4513.561373169277, 777.2772244687983, 2378.031197208547, 603.4691739453056, 3559.037817596542, 1767.7031384090574, 1563.1657938202325, 2398.6630535395934, 3880.9071702110527, 1251.1779318538547, 3423.7355476782404, 1580.6549022330732, 1844.736675161474, 2486.4488388400614, 4526.736786114767, 1204.7344474860274, 3089.38670359339, 2056.759805894274, 2663.4770444536953, 2532.881195268306, 2037.4898273705505, 3409.439883195495, 2290.1593368940507, 1879.9366422706285, 1763.6635560038285, 978.5329120821013, 1730.3269054877053, 6850.10132019094, 4593.152422624651, 4337.189092928294, 1371.2288541458504, 3434.394315974465, 1220.015522175255, 4921.033260671205, 4053.189320490131, 5332.800826593457, 1599.0696303394805, 1592.8975628296394, 3030.888494724317, 9984.700480241492, 4376.545164632571, 1593.8869676265804, 3062.6812989058053, 1668.0824535290176, 11168.784389241237, 11300.414672648903, 1770.8756067251352, 4900.457729280653, 1464.7401475333718, 5025.043676262615, 1410.9753801962574, 1859.534041179882, 6445.667206284492, 5248.479723791857, 3329.703836535861, 1574.764076258624, 2986.6181276722923, 3885.0913167385715, 3326.5751354990025, 1015.2933556926315, 3426.5980056471108, 1356.1026738124374, 4401.145310282047, 4213.9099608042525, 1481.4989240851094, 1850.7554927854844, 2896.9071455614867, 2830.0910589100304, 3159.4529447648165, 1586.997113807297, 4344.689074836582, 7571.822838787932, 3145.3014885682696, 4260.609981879579, 515.0740392513934, 2761.374574202898, 1526.2266399464456, 2874.311818855197, 8103.798538181826, 4172.999901699564, 1176.1020090356974, 2087.875902250697, 2971.018574882027, 2929.1521330991836, 5752.486687960571, 8002.3390508543525, 3312.4736485388335, 2067.671321991174, 1176.5675591472318, 1144.6713436503906, 1100.7136042727413, 5462.170098345213, 3077.8795995593437, 2528.0190892951614, 1397.679525376226, 3946.885113129756, 3475.7866766669295, 6200.138574713683, 1722.9057220568898]","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:55:06.894218Z","iopub.execute_input":"2024-06-25T03:55:06.894780Z","iopub.status.idle":"2024-06-25T03:55:06.927382Z","shell.execute_reply.started":"2024-06-25T03:55:06.894738Z","shell.execute_reply":"2024-06-25T03:55:06.926459Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"sum(scor) / len(scor)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T03:56:31.206676Z","iopub.execute_input":"2024-06-25T03:56:31.207089Z","iopub.status.idle":"2024-06-25T03:56:31.213926Z","shell.execute_reply.started":"2024-06-25T03:56:31.207055Z","shell.execute_reply":"2024-06-25T03:56:31.212920Z"},"trusted":true},"execution_count":22,"outputs":[{"execution_count":22,"output_type":"execute_result","data":{"text/plain":"3069.204066540013"},"metadata":{}}]},{"cell_type":"code","source":"import matplotlib.pyplot as plt\nplt.figure(figsize=(10, 6))\nplt.hist(scores, bins=20, edgecolor='black')\nplt.title('Distribution of Mutual Information Scores')\nplt.xlabel('Mutual Information Score')\nplt.ylabel('Frequency')\nplt.grid(True)\nplt.tight_layout()\n# Save the plot as PDF\nplt.savefig('MUI_visualization.pdf', format='pdf')\nplt.show()\n\n","metadata":{"execution":{"iopub.status.busy":"2024-06-25T02:26:35.170166Z","iopub.execute_input":"2024-06-25T02:26:35.170547Z","iopub.status.idle":"2024-06-25T02:26:36.150153Z","shell.execute_reply.started":"2024-06-25T02:26:35.170521Z","shell.execute_reply":"2024-06-25T02:26:36.149139Z"},"trusted":true},"execution_count":21,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1000x600 with 1 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAAA90AAAJOCAYAAACqS2TfAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuNSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/xnp5ZAAAACXBIWXMAAA9hAAAPYQGoP6dpAABd/ElEQVR4nO3deVxU9f7H8fcg26DiAgqSgGtqWmpaSm6lKLmUlW0upeatW1lq2nK9LS5l2aJppXlbxDbSvO1WmrlXZi5JWkRqFpaK4gIi28h8f390mZ8ji4gcBsbX8/HwUXPOd875nPOZQd6ezWaMMQIAAAAAAOXOx9MFAAAAAADgrQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AUIlNnjxZNputQtZ1+eWX6/LLL3e9Xr16tWw2m/773/9WyPpHjBihRo0aVci6yiozM1P/+Mc/FB4eLpvNpnHjxnm6pArTqFEjjRgxwvL1vPXWW2rZsqX8/PxUu3Zty9dXUX7//XfZbDYtWLDA06UAACoYoRsAKsiCBQtks9lcfwIDAxUREaG4uDi98MILOnbsWLmsZ+/evZo8ebK2bt1aLssrT5W5ttJ48skntWDBAt1111166623dMsttxQ7tlGjRrLZbIqNjS1y/quvvur6LGzatOmMa8nKytLkyZO1evXqM36vlWw2m+65554yvfeXX37RiBEj1LRpU7366qt65ZVXyrk66yUkJGjWrFmeLqOQTz/9VD169FD9+vUVFBSkJk2a6MYbb9TSpUs9XRoAeD1fTxcAAOeaqVOnqnHjxnI4HNq/f79Wr16tcePGaebMmfrkk0900UUXucY+8sgj+te//nVGy9+7d6+mTJmiRo0aqV27dqV+35dffnlG6ymLkmp79dVX5XQ6La/hbKxcuVKdO3fWpEmTSjU+MDBQq1at0v79+xUeHu4275133lFgYKBycnLKVEtWVpamTJkiSW5nKFRlq1evltPp1OzZs9WsWTNPl1MmCQkJ2r59e6GzIKKjo5WdnS0/P78Kr+m5557TAw88oB49emjixIkKCgrSzp079dVXX2nhwoW68sorK7wmADiXELoBoIL17dtXHTt2dL2eOHGiVq5cqQEDBujqq69WUlKS7Ha7JMnX11e+vtb+qM7KylJQUJD8/f0tXc/peCKMnKkDBw7oggsuKPX4Ll26aOPGjVq0aJHGjh3rmv7nn39q3bp1uvbaa/X+++9bUWqVdODAAUkq19PKCz7fnlZwdktFO3HihB5//HH17t27yH9YK9jnFcHpdCovL88j+wEAPInTywGgEujZs6ceffRR/fHHH3r77bdd04u6pnv58uXq2rWrateurRo1aqhFixb697//LenvI4WXXHKJJGnkyJGu05cLriO9/PLL1aZNG23evFndu3dXUFCQ672nXtNdID8/X//+978VHh6u6tWr6+qrr9aePXvcxhR3ve/JyzxdbUVd0338+HFNmDBBkZGRCggIUIsWLfTcc8/JGOM2ruCU5o8++kht2rRRQECAWrduXepTZw8cOKBRo0YpLCxMgYGBatu2rd544w3X/ILr23fv3q3PPvvMVfvvv/9e4nIDAwN13XXXKSEhwW36u+++qzp16iguLq7Qe4rrw8n75/fff1e9evUkSVOmTHHVM3ny5FIvo8Bzzz2nyy67TCEhIbLb7erQoUO5XsdfsO/ee+89TZs2TQ0bNlRgYKB69eqlnTt3usY1atTIdQZBvXr13LZHkubOnavWrVsrICBAERERGj16tI4ePeq2ruI+3wXXUz/33HOaM2eOmjRpoqCgIPXp00d79uyRMUaPP/64GjZsKLvdroEDB+rw4cNuy/7444/Vv39/RUREKCAgQE2bNtXjjz+u/Px8t/V/9tln+uOPP1w9OblnRV3TvXLlSnXr1k3Vq1dX7dq1NXDgQCUlJbmNKfg5sHPnTo0YMUK1a9dWrVq1NHLkSGVlZZW4/9PS0pSRkaEuXboUOb9+/fpur3NycjR58mSdf/75CgwMVIMGDXTddddp165drjFn+r185513XL0r+E7+9ddfuu222xQWFub6vs6fP79QfS+++KJat26toKAg1alTRx07diz0fQKAyo4j3QBQSdxyyy3697//rS+//FK33357kWN++uknDRgwQBdddJGmTp2qgIAA7dy5U998840kqVWrVpo6daoee+wx3XHHHerWrZsk6bLLLnMt49ChQ+rbt69uvvlmDRs2TGFhYSXWNW3aNNlsNj300EM6cOCAZs2apdjYWG3dutV1RL40SlPbyYwxuvrqq7Vq1SqNGjVK7dq107Jly/TAAw/or7/+0vPPP+82/uuvv9YHH3ygu+++WzVr1tQLL7ygQYMGKSUlRSEhIcXWlZ2drcsvv1w7d+7UPffco8aNG2vx4sUaMWKEjh49qrFjx6pVq1Z66623dN9996lhw4aaMGGCJLmCb0mGDBmiPn36aNeuXWratKmkv09Bvv7668t8dL9evXp6+eWXddddd+naa6/VddddJ0lulyaU1uzZs3X11Vdr6NChysvL08KFC3XDDTdoyZIl6t+/f5nqK8r06dPl4+Oj+++/X+np6XrmmWc0dOhQbdiwQZI0a9Ysvfnmm/rwww/18ssvq0aNGq7tmTx5sqZMmaLY2FjdddddSk5O1ssvv6yNGzfqm2++cduPJX2+33nnHeXl5enee+/V4cOH9cwzz+jGG29Uz549tXr1aj300EPauXOnXnzxRd1///1uIXDBggWqUaOGxo8frxo1amjlypV67LHHlJGRoWeffVaS9PDDDys9PV1//vmn6/NZo0aNYvfJV199pb59+6pJkyaaPHmysrOz9eKLL6pLly7asmVLoX8gufHGG9W4cWM99dRT2rJli1577TXVr19fTz/9dLHrqF+/vux2uz799FPde++9qlu3brFj8/PzNWDAAK1YsUI333yzxo4dq2PHjmn58uXavn27mjZtesbfy5UrV+q9997TPffco9DQUDVq1Eipqanq3LmzK5TXq1dPX3zxhUaNGqWMjAzXqfmvvvqqxowZo+uvv15jx45VTk6OfvzxR23YsEFDhgwpdjsAoNIxAIAKER8fbySZjRs3FjumVq1apn379q7XkyZNMif/qH7++eeNJHPw4MFil7Fx40YjycTHxxea16NHDyPJzJs3r8h5PXr0cL1etWqVkWTOO+88k5GR4Zr+3nvvGUlm9uzZrmnR0dFm+PDhp11mSbUNHz7cREdHu15/9NFHRpJ54okn3MZdf/31xmazmZ07d7qmSTL+/v5u0xITE40k8+KLLxZa18lmzZplJJm3337bNS0vL8/ExMSYGjVquG17dHS06d+/f4nLO3XsiRMnTHh4uHn88ceNMcb8/PPPRpJZs2ZNkZ+JU/dZgVP3z8GDB40kM2nSpEJjS7sMY4zJyspye52Xl2fatGljevbsWWh7iurxqSSZ0aNHu14XfI5atWplcnNzXdNnz55tJJlt27a5phV83k/+fB84cMD4+/ubPn36mPz8fNf0l156yUgy8+fPd9vuoj7fu3fvNpJMvXr1zNGjR13TJ06caCSZtm3bGofD4Zo+ePBg4+/vb3JycordT8YY889//tMEBQW5jevfv3+hfXxyDSd/9tu1a2fq169vDh065JqWmJhofHx8zK233lpov9x2221uy7z22mtNSEhIoXWd6rHHHjOSTPXq1U3fvn3NtGnTzObNmwuNmz9/vpFkZs6cWWie0+k0xpz599LHx8f89NNPbmNHjRplGjRoYNLS0tym33zzzaZWrVqufT1w4EDTunXr024fAFR2nF4OAJVIjRo1SryLecG1rh9//HGZbzoWEBCgkSNHlnr8rbfeqpo1a7peX3/99WrQoIE+//zzMq2/tD7//HNVq1ZNY8aMcZs+YcIEGWP0xRdfuE2PjY11HUmW/j7qGxwcrN9+++206wkPD9fgwYNd0/z8/DRmzBhlZmZqzZo1Z7Ud1apV04033qh3331X0t9HWyMjI11H+j3t5LMVjhw5ovT0dHXr1k1btmwp1/WMHDnS7b4BBdt/uv589dVXysvL07hx4+Tj8/+/ttx+++0KDg7WZ5995ja+pM/3DTfcoFq1arled+rUSZI0bNgwt3sndOrUSXl5efrrr79c007eT8eOHVNaWpq6deumrKws/fLLLyVuQ1H27dunrVu3asSIEW5Hny+66CL17t27yO/XnXfe6fa6W7duOnTokDIyMkpc15QpU5SQkKD27dtr2bJlevjhh9WhQwddfPHFbqeyv//++woNDdW9995baBkFl7mc6feyR48ebvdBMMbo/fff11VXXSVjjNLS0lx/4uLilJ6e7vrs1a5dW3/++ac2btxY4vYBQGVH6AaASiQzM9Mt4J7qpptuUpcuXfSPf/xDYWFhuvnmm/Xee++dUQA/77zzzuimac2bN3d7bbPZ1KxZs9Nez3y2/vjjD0VERBTaH61atXLNP1lUVFShZdSpU0dHjhw57XqaN2/uFuhKWk9ZDBkyRD///LMSExOVkJCgm2++ucKev346S5YsUefOnRUYGKi6deu6Tl1PT08v1/Wc2p86depIUqn6I0ktWrRwm+7v768mTZoU6k9Jn+9TaygI4JGRkUVOP7m2n376Sddee61q1aql4OBg1atXT8OGDZOkMu2r4rZL+vuzl5aWpuPHj5dYf2n3oSQNHjxY69at05EjR/Tll19qyJAh+uGHH3TVVVe57qC/a9cutWjRosSbN57p97Jx48Zurw8ePKijR4/qlVdeUb169dz+FPxjScHN3R566CHVqFFDl156qZo3b67Ro0e7LqUBgKqEa7oBoJL4888/lZ6eXuKjkux2u9auXatVq1bps88+09KlS7Vo0SL17NlTX375papVq3ba9ZzJddilVVyAzM/PL1VN5aG49ZhTbu7kCZ06dVLTpk01btw47d69u8TrUW02W5E1n3zDrtMp7TLWrVunq6++Wt27d9fcuXPVoEED+fn5KT4+vtxvVlVR/Snp811cDaer7ejRo+rRo4eCg4M1depUNW3aVIGBgdqyZYseeuihCnvUXXnsw+DgYPXu3Vu9e/eWn5+f3njjDW3YsEE9evQorzLdnNqPgn01bNgwDR8+vMj3FFzL36pVKyUnJ2vJkiVaunSp3n//fc2dO1ePPfaY63F5AFAVcKQbACqJt956S5KKvKP1yXx8fNSrVy/NnDlTP//8s6ZNm6aVK1dq1apVkooPwGW1Y8cOt9fGGO3cudPtJk916tQpdCdpqfBRrzOpLTo6Wnv37i10un3BqbzR0dGlXtbp1rNjx45Cwam81zN48GCtXr1arVq1KvH56eWxL0u7jPfff1+BgYFatmyZbrvtNvXt21exsbElbkdFK9j/ycnJbtPz8vK0e/fucutPSVavXq1Dhw5pwYIFGjt2rAYMGKDY2FjXkeaTlfYzXtx2SX9/9kJDQ1W9evWzK/w0Ch5duG/fPklS06ZNlZycLIfDUex7zvZ7Wa9ePdWsWVP5+fmKjY0t8s/Jd1SvXr26brrpJsXHxyslJUX9+/fXtGnTyvx8ewDwBEI3AFQCK1eu1OOPP67GjRtr6NChxY479TFGklwBLjc3V5Jcv6gXFbrK4s0333T7Bfu///2v9u3bp759+7qmNW3aVN99953y8vJc05YsWVLo0WJnUlu/fv2Un5+vl156yW36888/L5vN5rb+s9GvXz/t379fixYtck07ceKEXnzxRdWoUaPcjgD+4x//0KRJkzRjxowSxzVt2lS//PKLDh486JqWmJhY6LTagmdPF7UvS7uMatWqyWazuR0B//333/XRRx+VdrMsFxsbK39/f73wwgtuR3Rff/11paenl+sd1otTcIT55PXn5eVp7ty5hcZWr169VKebN2jQQO3atdMbb7zh1sPt27fryy+/VL9+/c6+cP39nPL169cXOa/g+uuCU9wHDRqktLS0Qt856f+3/Wy/l9WqVdOgQYP0/vvva/v27YXmn/yZPXTokNs8f39/XXDBBTLGlPgPAwBQ2XB6OQBUsC+++EK//PKLTpw4odTUVK1cuVLLly9XdHS0PvnkEwUGBhb73qlTp2rt2rXq37+/oqOjdeDAAc2dO1cNGzZU165dJf0duGrXrq158+apZs2aql69ujp16lTo2srSqlu3rrp27aqRI0cqNTVVs2bNUrNmzdwea/aPf/xD//3vf3XllVfqxhtv1K5du/T222+73djsTGu76qqrdMUVV+jhhx/W77//rrZt2+rLL7/Uxx9/rHHjxhVadlndcccd+s9//qMRI0Zo8+bNatSokf773//qm2++0axZs0q8xv5MREdHuz13uji33XabZs6cqbi4OI0aNUoHDhzQvHnz1Lp1a7cbZtntdl1wwQVatGiRzj//fNWtW1dt2rRRmzZtSr2M/v37a+bMmbryyis1ZMgQHThwQHPmzFGzZs30448/lst2n6169epp4sSJmjJliq688kpdffXVSk5O1ty5c3XJJZe4rqu20mWXXaY6depo+PDhGjNmjGw2m956660iT+vu0KGDFi1apPHjx+uSSy5RjRo1dNVVVxW53GeffVZ9+/ZVTEyMRo0a5XpkWK1atUr1WSmNrKwsXXbZZercubOuvPJKRUZG6ujRo/roo4+0bt06XXPNNWrfvr2kv2+a+Oabb2r8+PH6/vvv1a1bNx0/flxfffWV7r77bg0cOLBcvpfTp0/XqlWr1KlTJ91+++264IILdPjwYW3ZskVfffWV6x8X+/Tpo/DwcHXp0kVhYWFKSkrSSy+9pP79+5fb9xIAKkTF3zAdAM5NBY+HKvjj7+9vwsPDTe/evc3s2bPdHk1V4NRHhq1YscIMHDjQREREGH9/fxMREWEGDx5sfv31V7f3ffzxx+aCCy4wvr6+bo8p6tGjR7GP4CnukWHvvvuumThxoqlfv76x2+2mf//+5o8//ij0/hkzZpjzzjvPBAQEmC5duphNmzYV+eiq4mor6nFWx44dM/fdd5+JiIgwfn5+pnnz5ubZZ591Pb6ogE55TFWB0j7mKjU11YwcOdKEhoYaf39/c+GFFxb5WLOyPDKsJMU9Ru7tt982TZo0Mf7+/qZdu3Zm2bJlRe6fb7/91nTo0MH4+/sXenxYaZfx+uuvm+bNm5uAgADTsmVLEx8fX+hzV7A9Z/PIsMWLF7uNK+oRWkU9MqzASy+9ZFq2bGn8/PxMWFiYueuuu8yRI0fcxhT3+S5Y17PPPus2vbjaiurLN998Yzp37mzsdruJiIgwDz74oFm2bJmRZFatWuUal5mZaYYMGWJq165tJLn2d1Hba4wxX331lenSpYux2+0mODjYXHXVVebnn392G1Pcfimoc/fu3YW2uYDD4TCvvvqqueaaa0x0dLQJCAgwQUFBpn379ubZZ591e4ybMX8/Gu3hhx82jRs3Nn5+fiY8PNxcf/31ZteuXa4xZ/u9NObv79zo0aNNZGSkaz29evUyr7zyimvMf/7zH9O9e3cTEhJiAgICTNOmTc0DDzxg0tPTi91eAKiMbMZUgjvMAAAAAADghbimGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsIivpwuwmtPp1N69e1WzZk3ZbDZPlwMAAAAA8ALGGB07dkwRERHy8Sn+eLbXh+69e/cqMjLS02UAAAAAALzQnj171LBhw2Lne33orlmzpqS/d0RwcHCZl+NwOPTll1+qT58+8vPzK6/y4GH01fvQU+9EX70TffVO9NU70VfvRF/PTkZGhiIjI12ZszheH7oLTikPDg4+69AdFBSk4OBgPpBehL56H3rqneird6Kv3om+eif66p3oa/k43WXM3EgNAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALCIr6cLQNWTkpKitLQ0T5dxWqGhoYqKivJ0GQAAAADOYYRunJGUlBS1aNlKOdlZni7ltALtQUr+JYngDQAAAMBjCN04I2lpacrJzlLIgAnyC4n0dDnFchzao0NLZigtLY3QDQAAAMBjCN0oE7+QSAWEN/N0GQAAAABQqXEjNQAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCIeD91//fWXhg0bppCQENntdl144YXatGmTa74xRo899pgaNGggu92u2NhY7dixw4MVAwAAAABQOh4N3UeOHFGXLl3k5+enL774Qj///LNmzJihOnXquMY888wzeuGFFzRv3jxt2LBB1atXV1xcnHJycjxYOQAAAAAAp+fryZU//fTTioyMVHx8vGta48aNXf9vjNGsWbP0yCOPaODAgZKkN998U2FhYfroo4908803V3jNAAAAAACUlkdD9yeffKK4uDjdcMMNWrNmjc477zzdfffduv322yVJu3fv1v79+xUbG+t6T61atdSpUyetX7++yNCdm5ur3Nxc1+uMjAxJksPhkMPhKHOtBe89m2V4A6fTKbvdrkBfm/yrGU+XUyybr012u11Op7PEntFX70NPvRN99U701TvRV+9EX70TfT07pd1vNmOMx5JTYGCgJGn8+PG64YYbtHHjRo0dO1bz5s3T8OHD9e2336pLly7au3evGjRo4HrfjTfeKJvNpkWLFhVa5uTJkzVlypRC0xMSEhQUFGTdxgAAAAAAzhlZWVkaMmSI0tPTFRwcXOw4j4Zuf39/dezYUd9++61r2pgxY7Rx40atX7++TKG7qCPdkZGRSktLK3FHnI7D4dDy5cvVu3dv+fn5lXk5VV1iYqK6d++usCHT5R/WxNPlFCsv9TelJvxLa9euVdu2bYsdR1+9Dz31TvTVO9FX70RfvRN99U709exkZGQoNDT0tKHbo6eXN2jQQBdccIHbtFatWun999+XJIWHh0uSUlNT3UJ3amqq2rVrV+QyAwICFBAQUGi6n59fuXyQyms5VZWPj4+ys7OVc8LI5Ns8XU6xck8YZWdny8fHp1T9Otf76o3oqXeir96Jvnon+uqd6Kt3oq9lU9p95tG7l3fp0kXJyclu03799VdFR0dL+vumauHh4VqxYoVrfkZGhjZs2KCYmJgKrRUAAAAAgDPl0SPd9913ny677DI9+eSTuvHGG/X999/rlVde0SuvvCJJstlsGjdunJ544gk1b95cjRs31qOPPqqIiAhdc801niwdAAAAAIDT8mjovuSSS/Thhx9q4sSJmjp1qho3bqxZs2Zp6NChrjEPPvigjh8/rjvuuENHjx5V165dtXTpUtdN2AAAAAAAqKw8GrolacCAARowYECx8202m6ZOnaqpU6dWYFUAAAAAAJw9j17TDQAAAACANyN0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAW8Wjonjx5smw2m9ufli1buubn5ORo9OjRCgkJUY0aNTRo0CClpqZ6sGIAAAAAAErP40e6W7durX379rn+fP3116559913nz799FMtXrxYa9as0d69e3Xdddd5sFoAAAAAAErP1+MF+PoqPDy80PT09HS9/vrrSkhIUM+ePSVJ8fHxatWqlb777jt17ty5oksFAAAAAOCMeDx079ixQxEREQoMDFRMTIyeeuopRUVFafPmzXI4HIqNjXWNbdmypaKiorR+/fpiQ3dubq5yc3NdrzMyMiRJDodDDoejzHUWvPdsluENnE6n7Ha7An1t8q9mPF1OsWy+NtntdjmdzhJ7Rl+9Dz31TvTVO9FX70RfvRN99U709eyUdr/ZjDEeS05ffPGFMjMz1aJFC+3bt09TpkzRX3/9pe3bt+vTTz/VyJEj3QK0JF166aW64oor9PTTTxe5zMmTJ2vKlCmFpickJCgoKMiS7QAAAAAAnFuysrI0ZMgQpaenKzg4uNhxHg3dpzp69Kiio6M1c+ZM2e32MoXuoo50R0ZGKi0trcQdcToOh0PLly9X79695efnV+blVHWJiYnq3r27woZMl39YE0+XU6y81N+UmvAvrV27Vm3bti12HH31PvTUO9FX70RfvRN99U701TvR17OTkZGh0NDQ04Zuj59efrLatWvr/PPP186dO9W7d2/l5eXp6NGjql27tmtMampqkdeAFwgICFBAQECh6X5+fuXyQSqv5VRVPj4+ys7OVs4JI5Nv83Q5xco9YZSdnS0fH59S9etc76s3oqfeib56J/rqneird6Kv3om+lk1p95nH715+sszMTO3atUsNGjRQhw4d5OfnpxUrVrjmJycnKyUlRTExMR6sEgAAAACA0vHoke77779fV111laKjo7V3715NmjRJ1apV0+DBg1WrVi2NGjVK48ePV926dRUcHKx7771XMTEx3LkcAAAAAFAleDR0//nnnxo8eLAOHTqkevXqqWvXrvruu+9Ur149SdLzzz8vHx8fDRo0SLm5uYqLi9PcuXM9WTIAAAAAAKXm0dC9cOHCEucHBgZqzpw5mjNnTgVVBAAAAABA+alU13QDAAAAAOBNCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFfD1dAGClpKSkEuc7nU5JUmJionx8PPNvUKGhoYqKivLIugEAAABYi9ANr5SfeUSy2TRs2LASx9ntdr377rvq3r27srOzK6g6d4H2ICX/kkTwBgAAALwQoRteyZmbKRmjkAET5BcSWey4QF+bJClsyHTlnDAVVZ6L49AeHVoyQ2lpaYRuAAAAwAsRuuHV/EIiFRDerNj5/tWMpHz5hzWRybdVXGEAAAAAzgncSA0AAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwSKUJ3dOnT5fNZtO4ceNc03JycjR69GiFhISoRo0aGjRokFJTUz1XJAAAAAAAZ6BShO6NGzfqP//5jy666CK36ffdd58+/fRTLV68WGvWrNHevXt13XXXeahKAAAAAADOjMdDd2ZmpoYOHapXX31VderUcU1PT0/X66+/rpkzZ6pnz57q0KGD4uPj9e233+q7777zYMUAAAAAAJSOr6cLGD16tPr376/Y2Fg98cQTrumbN2+Ww+FQbGysa1rLli0VFRWl9evXq3PnzkUuLzc3V7m5ua7XGRkZkiSHwyGHw1HmOgveezbL8AZOp1N2u12Bvjb5VzOeLqdYJ/yqlarOAB/j9t+KZvO1yW63KykpSU6n0yM1lFZISIgaNmzo6TJOi++qd6Kv3om+eif66p3oq3eir2entPvNZozxWHJauHChpk2bpo0bNyowMFCXX3652rVrp1mzZikhIUEjR450C9CSdOmll+qKK67Q008/XeQyJ0+erClTphSanpCQoKCgIEu2AwAAAABwbsnKytKQIUOUnp6u4ODgYsd57Ej3nj17NHbsWC1fvlyBgYHlttyJEydq/PjxrtcZGRmKjIxUnz59StwRp+NwOLR8+XL17t1bfn5+5VFqlZSYmKju3bsrbMh0+Yc18XQ5xTqetE6Hl7542joDfIwe7+jUo5t8lOu0VWCFfyuos+6V98qv7nkVvv7Schz+S4eXvqi1a9eqbdu2ni6nRHxXvRN99U701TvRV+9EX70TfT07BWdVn47HQvfmzZt14MABXXzxxa5p+fn5Wrt2rV566SUtW7ZMeXl5Onr0qGrXru0ak5qaqvDw8GKXGxAQoICAgELT/fz8yuWDVF7Lqap8fHyUnZ2tnBNGJr/iQ2pp5Tjyz6jOXKdNuR7YnoI684Mj5BvatMLXX1r5J4yys7Pl4+NTZT7/5/p31VvRV+9EX70TffVO9NU70deyKe0+81jo7tWrl7Zt2+Y2beTIkWrZsqUeeughRUZGys/PTytWrNCgQYMkScnJyUpJSVFMTIwnSgYAAAAA4Ix4LHTXrFlTbdq0cZtWvXp1hYSEuKaPGjVK48ePV926dRUcHKx7771XMTExxd5EDQAAAACAysTjdy8vyfPPPy8fHx8NGjRIubm5iouL09y5cz1dFgAAAAAApVKpQvfq1avdXgcGBmrOnDmaM2eOZwoCAAAAAOAs+Hi6AAAAAAAAvBWhGwAAAAAAixC6AQAAAACwSJlC92+//VbedQAAAAAA4HXKFLqbNWumK664Qm+//bZycnLKuyYAAAAAALxCmUL3li1bdNFFF2n8+PEKDw/XP//5T33//fflXRsAAAAAAFVamUJ3u3btNHv2bO3du1fz58/Xvn371LVrV7Vp00YzZ87UwYMHy7tOAAAAAACqnLO6kZqvr6+uu+46LV68WE8//bR27typ+++/X5GRkbr11lu1b9++8qoTAAAAAIAq56xC96ZNm3T33XerQYMGmjlzpu6//37t2rVLy5cv1969ezVw4MDyqhMAAAAAgCrHtyxvmjlzpuLj45WcnKx+/frpzTffVL9+/eTj83eGb9y4sRYsWKBGjRqVZ60AAAAAAFQpZQrdL7/8sm677TaNGDFCDRo0KHJM/fr19frrr59VcQAAAAAAVGVlCt07duw47Rh/f38NHz68LIsHAAAAAMArlOma7vj4eC1evLjQ9MWLF+uNN94466IAAAAAAPAGZQrdTz31lEJDQwtNr1+/vp588smzLgoAAAAAAG9QptCdkpKixo0bF5oeHR2tlJSUsy4KAAAAAABvUKbQXb9+ff3444+FpicmJiokJOSsiwIAAAAAwBuUKXQPHjxYY8aM0apVq5Sfn6/8/HytXLlSY8eO1c0331zeNQIAAAAAUCWV6e7ljz/+uH7//Xf16tVLvr5/L8LpdOrWW2/lmm4AAAAAAP6nTKHb399fixYt0uOPP67ExETZ7XZdeOGFio6OLu/6AAAAAACossoUugucf/75Ov/888urFgAAAAAAvEqZQnd+fr4WLFigFStW6MCBA3I6nW7zV65cWS7FAQAAAABQlZUpdI8dO1YLFixQ//791aZNG9lstvKuCwAAAACAKq9MoXvhwoV677331K9fv/KuBwAAAAAAr1GmR4b5+/urWbNm5V0LAAAAAABepUyhe8KECZo9e7aMMeVdDwAAAAAAXqNMp5d//fXXWrVqlb744gu1bt1afn5+bvM/+OCDcikOAAAAAICqrEyhu3bt2rr22mvLuxYAAAAAALxKmUJ3fHx8edcBAAAAAIDXKdM13ZJ04sQJffXVV/rPf/6jY8eOSZL27t2rzMzMcisOAAAAAICqrExHuv/44w9deeWVSklJUW5urnr37q2aNWvq6aefVm5urubNm1fedQIAAAAAUOWU6Uj32LFj1bFjRx05ckR2u901/dprr9WKFSvKrTgAAAAAAKqyMh3pXrdunb799lv5+/u7TW/UqJH++uuvcikMAAAAAICqrkxHup1Op/Lz8wtN//PPP1WzZs2zLgoAAAAAAG9QptDdp08fzZo1y/XaZrMpMzNTkyZNUr9+/cqrNgAAAAAAqrQynV4+Y8YMxcXF6YILLlBOTo6GDBmiHTt2KDQ0VO+++2551wgAAAAAQJVUptDdsGFDJSYmauHChfrxxx+VmZmpUaNGaejQoW43VgMAAAAA4FxWptAtSb6+vho2bFh51gIAAAAAgFcpU+h+8803S5x/6623lqkYAAAAAAC8SZlC99ixY91eOxwOZWVlyd/fX0FBQYRuAAAAAABUxruXHzlyxO1PZmamkpOT1bVrV26kBgAAAADA/5QpdBelefPmmj59eqGj4AAAAAAAnKvKLXRLf99cbe/eveW5SAAAAAAAqqwyXdP9ySefuL02xmjfvn166aWX1KVLl3IpDAAAAACAqq5Mofuaa65xe22z2VSvXj317NlTM2bMKI+6AAAAAACo8soUup1OZ3nXAQAAAACA1ynXa7oBAAAAAMD/K9OR7vHjx5d67MyZM8uyCgAAAAAAqrwyhe4ffvhBP/zwgxwOh1q0aCFJ+vXXX1WtWjVdfPHFrnE2m618qgQAAAAAoAoqU+i+6qqrVLNmTb3xxhuqU6eOJOnIkSMaOXKkunXrpgkTJpRrkQAAAAAAVEVluqZ7xowZeuqpp1yBW5Lq1KmjJ554gruXAwAAAADwP2UK3RkZGTp48GCh6QcPHtSxY8fOuigAAAAAALxBmU4vv/baazVy5EjNmDFDl156qSRpw4YNeuCBB3TdddeVa4EAKo+kpCRPl3BaJ5+BAwAAAHhamUL3vHnzdP/992vIkCFyOBx/L8jXV6NGjdKzzz5brgUC8Lz8zCOSzaZhw4Z5upTTqlM3RPHzX/d0GQAAAICkMobuoKAgzZ07V88++6x27dolSWratKmqV69ersUBqBycuZmSMQoZMEF+IZGeLqdYjkN7lLVirqfLAAAAAFzKFLoL7Nu3T/v27VP37t1lt9tljOExYYAX8wuJVEB4M0+XAQAAAFQZZbqR2qFDh9SrVy+df/756tevn/bt2ydJGjVqFI8LAwAAAADgf8oUuu+77z75+fkpJSVFQUFBruk33XSTli5dWm7FAQAAAABQlZXp9PIvv/xSy5YtU8OGDd2mN2/eXH/88Ue5FAYAAAAAQFVXpiPdx48fdzvCXeDw4cMKCAg466IAAAAAAPAGZQrd3bp105tvvul6bbPZ5HQ69cwzz+iKK64ot+IAAAAAAKjKynR6+TPPPKNevXpp06ZNysvL04MPPqiffvpJhw8f1jfffFPeNQIAAAAAUCWV6Uh3mzZt9Ouvv6pr164aOHCgjh8/ruuuu04//PCDmjZtWt41AgAAAABQJZ3xkW6Hw6Err7xS8+bN08MPP2xFTQAAAAAAeIUzPtLt5+enH3/8sVxW/vLLL+uiiy5ScHCwgoODFRMToy+++MI1PycnR6NHj1ZISIhq1KihQYMGKTU1tVzWDQAAAACA1cp0evmwYcP0+uuvn/XKGzZsqOnTp2vz5s3atGmTevbsqYEDB+qnn36S9PfzwD/99FMtXrxYa9as0d69e3Xddded9XoBAAAAAKgIZbqR2okTJzR//nx99dVX6tChg6pXr+42f+bMmaVazlVXXeX2etq0aXr55Zf13XffqWHDhnr99deVkJCgnj17SpLi4+PVqlUrfffdd+rcuXNZSgcAAAAAoMKcUej+7bff1KhRI23fvl0XX3yxJOnXX391G2Oz2cpUSH5+vhYvXqzjx48rJiZGmzdvlsPhUGxsrGtMy5YtFRUVpfXr1xcbunNzc5Wbm+t6nZGRIenva9EdDkeZait4/8n/PVc5nU7Z7XYF+trkX814upxinfCrVqo6A3yM238rWmnr9LSqUqfN16Y8u10S31Vvw89g70RfvRN99U701TvR17NT2v1mM8aU+jfoatWqad++fapfv74k6aabbtILL7ygsLCwslUpadu2bYqJiVFOTo5q1KihhIQE9evXTwkJCRo5cqRbgJakSy+9VFdccYWefvrpIpc3efJkTZkypdD0hIQEBQUFlblOAAAAAAAKZGVlaciQIUpPT1dwcHCx487oSPep+fyLL77Q8ePHy1bh/7Ro0UJbt25Venq6/vvf/2r48OFas2ZNmZc3ceJEjR8/3vU6IyNDkZGR6tOnT4k74nQcDoeWL1+u3r17y8/Pr8zLqeoSExPVvXt3hQ2ZLv+wJp4up1jHk9bp8NIXT1tngI/R4x2denSTj3KdZTtL42yUtk5Pqyp15qX+powPp2j+/Pnn/HfV2/Az2DvRV+9EX70TffVO9PXsFJxVfTpluqa7wBkcJC+Wv7+/mjVrJknq0KGDNm7cqNmzZ+umm25SXl6ejh49qtq1a7vGp6amKjw8vNjlBQQEKCAgoNB0Pz+/cvkglddyqiofHx9lZ2cr54SRya/4kFpaOY78M6oz12lTrge250zr9JSqUmfuCaPs7GxJfFe9FX31TvTVO9FX70RfvRN9LZvS7rMzunu5zWYrdM12Wa/hLo7T6VRubq46dOggPz8/rVixwjUvOTlZKSkpiomJKdd1AgAAAABghTM+vXzEiBGuI8k5OTm68847C929/IMPPijV8iZOnKi+ffsqKipKx44dU0JCglavXq1ly5apVq1aGjVqlMaPH6+6desqODhY9957r2JiYrhzOQAAAACgSjij0D18+HC318OGDTurlR84cEC33nqr9u3bp1q1aumiiy7SsmXL1Lt3b0nS888/Lx8fHw0aNEi5ubmKi4vT3Llzz2qdAAAAAABUlDMK3fHx8eW68tdff73E+YGBgZozZ47mzJlTrusFAAAAAKAinNE13QAAAAAAoPQI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWMTX0wXg/6WkpCgtLc3TZZQoKSnJ0yUAAAAAQJVB6K4kUlJS1KJlK+VkZ3m6FAAAAABAOSF0VxJpaWnKyc5SyIAJ8guJ9HQ5xcr+bZPS173t6TIAAAAAoEogdFcyfiGRCghv5ukyiuU4tMfTJQAAAABAlcGN1AAAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALCIr6cLAAArJCYmysen8v67YmhoqKKiojxdBgAAACxG6AbgNfIzj0g2mySpe/fuys7O9nBFxQu0Byn5lySCNwAAgJcjdAPwGs7cTMkYSVLYkOnKOWE8XFHRHIf26NCSGUpLSyN0AwAAeDmPhu6nnnpKH3zwgX755RfZ7XZddtllevrpp9WiRQvXmJycHE2YMEELFy5Ubm6u4uLiNHfuXIWFhXmwcgCVnX9YE5l8m6fLAAAAwDnOoxc8rlmzRqNHj9Z3332n5cuXy+FwqE+fPjp+/LhrzH333adPP/1Uixcv1po1a7R3715dd911HqwaAAAAAIDS8eiR7qVLl7q9XrBggerXr6/Nmzere/fuSk9P1+uvv66EhAT17NlTkhQfH69WrVrpu+++U+fOnT1RNgAAAAAApVKpbu2bnp4uSapbt64kafPmzXI4HIqNjXWNadmypaKiorR+/XqP1AgAAAAAQGlVmhupOZ1OjRs3Tl26dFGbNm0kSfv375e/v79q167tNjYsLEz79+8vcjm5ubnKzc11vc7IyJAkORwOORyOMtdX8N6zWUZJnE6n7Ha7An1t8q9WOW/+JEkn/Kp5VZ0BPsbtvxXN2/anp53wqyZjt0vyXE9Lw+Zrk91ul9PptOxnirex+mcwPIO+eif66p3oq3eir2entPvNZoypFL+Z3nXXXfriiy/09ddfq2HDhpKkhIQEjRw50i1ES9Kll16qK664Qk8//XSh5UyePFlTpkwpND0hIUFBQUHWFA8AAAAAOKdkZWVpyJAhSk9PV3BwcLHjKsWR7nvuuUdLlizR2rVrXYFbksLDw5WXl6ejR4+6He1OTU1VeHh4kcuaOHGixo8f73qdkZGhyMhI9enTp8QdcToOh0PLly9X79695efnV+blFCcxMVHdu3dX2JDp8g9rUu7LLy/Hk9bp8NIXvabOAB+jxzs69egmH+U6K/5O1962Pz3teNI6Za95TfPnz/dYT0sjL/U3pSb8S2vXrlXbtm09XU6VYPXPYHgGffVO9NU70VfvRF/PTsFZ1afj0dBtjNG9996rDz/8UKtXr1bjxo3d5nfo0EF+fn5asWKFBg0aJElKTk5WSkqKYmJiilxmQECAAgICCk338/Mrlw9SeS3nVD4+PsrOzlbOCVOpH3OU48j3yjpznTblemB7vHV/ekpBnZLneloauSeMsrOz5ePjw19wZ8iqn8HwLPrqneird6Kv3om+lk1p95lHQ/fo0aOVkJCgjz/+WDVr1nRdp12rVi3Z7XbVqlVLo0aN0vjx41W3bl0FBwfr3nvvVUxMDHcuBwAAAABUeh4N3S+//LIk6fLLL3ebHh8frxEjRkiSnn/+efn4+GjQoEHKzc1VXFyc5s6dW8GVAgAAAABw5jx+evnpBAYGas6cOZozZ04FVAQAAAAAQPmpVM/pBgAAAADAmxC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAixC6AQAAAACwCKEbAAAAAACLELoBAAAAALAIoRsAAAAAAIsQugEAAAAAsAihGwAAAAAAi/h6ugAAOFclJSV5uoTTCg0NVVRUlKfLAAAAqLII3QBQwfIzj0g2m4YNG+bpUk4r0B6k5F+SCN4AAABlROgGgArmzM2UjFHIgAnyC4n0dDnFchzao0NLZigtLY3QDQAAUEaEbgDwEL+QSAWEN/N0GQAAALAQN1IDAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAAAAAALOLR0L127VpdddVVioiIkM1m00cffeQ23xijxx57TA0aNJDdbldsbKx27NjhmWIBAAAAADhDHg3dx48fV9u2bTVnzpwi5z/zzDN64YUXNG/ePG3YsEHVq1dXXFyccnJyKrhSAAAAAADOnK8nV963b1/17du3yHnGGM2aNUuPPPKIBg4cKEl68803FRYWpo8++kg333xzRZYKAAAAAMAZq7TXdO/evVv79+9XbGysa1qtWrXUqVMnrV+/3oOVAQAAAABQOh490l2S/fv3S5LCwsLcpoeFhbnmFSU3N1e5ubmu1xkZGZIkh8Mhh8NR5noK3ns2yyiJ0+mU3W5XoK9N/tWMJesoDyf8qnlVnQE+xu2/Fc3b9qennfCrJmO3S/JcT0ujquxPm69NdrtdTqfTsp99pWX1z2B4Bn31TvTVO9FX70Rfz05p95vNGFMpfuOz2Wz68MMPdc0110iSvv32W3Xp0kV79+5VgwYNXONuvPFG2Ww2LVq0qMjlTJ48WVOmTCk0PSEhQUFBQZbUDgAAAAA4t2RlZWnIkCFKT09XcHBwseMq7ZHu8PBwSVJqaqpb6E5NTVW7du2Kfd/EiRM1fvx41+uMjAxFRkaqT58+Je6I03E4HFq+fLl69+4tPz+/Mi+nOImJierevbvChkyXf1iTcl9+eTmetE6Hl77oNXUG+Bg93tGpRzf5KNdpq8AK/+Zt+9PTjietU/aa1zR//nyP9bQ0qsr+zEv9TakJ/9LatWvVtm1bj9Zi9c9geAZ99U701TvRV+9EX89OwVnVp1NpQ3fjxo0VHh6uFStWuEJ2RkaGNmzYoLvuuqvY9wUEBCggIKDQdD8/v3L5IJXXck7l4+Oj7Oxs5ZwwMvmVMyhIUo4j3yvrzHXalOuB7fHW/ekpBXVKnutpaVSV/Zl7wig7O1s+Pj6V5i9iq34Gw7Poq3eir96Jvnon+lo2pd1nHg3dmZmZ2rlzp+v17t27tXXrVtWtW1dRUVEaN26cnnjiCTVv3lyNGzfWo48+qoiICNcp6AAAAAAAVGYeDd2bNm3SFVdc4XpdcFr48OHDtWDBAj344IM6fvy47rjjDh09elRdu3bV0qVLFRgY6KmSAQAAAAAoNY+G7ssvv1wl3cfNZrNp6tSpmjp1agVWBQAAAABA+ai0z+kGAAAAAKCqI3QDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARX08XAADA2UpJSVFaWpqnyyhRaGiooqKiPF0GAACoYIRuAECVlpKSohYtWyknO8vTpZQo0B6k5F+SCN4AAJxjCN0AgCotLS1NOdlZChkwQX4hkZ4up0iOQ3t0aMkMpaWlEboBADjHELoBAF7BLyRSAeHNPF0GAACAG26kBgAAAACARQjdAAAAAABYhNANAAAAAIBFuKYbAFCipKQkT5cgp9MpSUpMTJSPj/u/F1eG+gAAAIpD6AYAFCk/84hks2nYsGGeLkV2u13vvvuuunfvruzsbE+XAwAAUGqEbgBAkZy5mZIxleJRXIG+NklS2JDpyjlh3OZl/7ZJ6eve9kRZAAAAp0XoBgCUqDI8isu/mpGUL/+wJjL5Nrd5jkN7PFMUAABAKXAjNQAAAAAALELoBgAAAADAIoRuAAAAAAAswjXdAADAJSUlRWlpaRWyrpIeBXc6oaGhioqKsqIsAADKFaEbAABI+jtwt2jZSjnZWRWyvrN5FFygPUjJvyQRvAEAlR6hGwAASJLS0tKUk51VYY+JK+lRcCVxHNqjQ0tmKC0tjdANAKj0CN0AAMBNRT0mrqRHwQEA4C24kRoAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFCNwAAAAAAFiF0AwAAAABgEUI3AAAAAAAWIXQDAAAAAGARQjcAAAAAABYhdAMAAAAAYBFfTxcAAMC5IikpydMllKiy1wdUFSkpKUpLS/N0GacVGhqqqKgoT5eBCnby59PpdEqSEhMT5eNTuY7HetPnk9ANAIDF8jOPSDabhg0b5ulSAFgsJSVFLVq2Uk52lqdLOa1Ae5CSf0nymmCD0zv182m32/Xuu++qe/fuys7O9nB17rzp80noBgDAYs7cTMkYhQyYIL+QSE+XU6zs3zYpfd3bni4DqNLS0tKUk51V6b/vjkN7dGjJDKWlpXlFqEHpnPr5DPS1SZLChkxXzgnj4er+n7d9PgndAABUEL+QSAWEN/N0GcVyHNrj6RIAr1HZv+84txV8Pv2rGUn58g9rIpNv83RZXqtynbgPAAAAAIAXIXQDAAAAAGARQjcAAAAAABbhmm4AAACLVJVHR+Xm5iogIKBclmXlI4jKs06r8Og9AKcidAMAAFigKj06SjYfyTjLZVGWPoKoHOsEgIpC6AYAALBAVXl0VMGj4sqrTqseQVTedVqFR+8BOBWhGwAAwEKV/dFRBY+KK686rXoEUXnXaRUevQfgVFXiRmpz5sxRo0aNFBgYqE6dOun777/3dEkAAAAAAJxWpQ/dixYt0vjx4zVp0iRt2bJFbdu2VVxcnA4cOODp0gAAAAAAKFGlD90zZ87U7bffrpEjR+qCCy7QvHnzFBQUpPnz53u6NAAAAAAASlSpQ3deXp42b96s2NhY1zQfHx/FxsZq/fr1HqwMAAAAAIDTq9Q3UktLS1N+fr7CwsLcpoeFhemXX34p8j25ubnKzc11vU5PT5ckHT58WA6Ho8y1OBwOZWVl6dChQ/Lz8yvzcoqTkZGhwMBA2Q7tlnHmnv4NHuJzbJ9X1en0lbKyIuXct0fmRAUW+D/etj89raDOrKwsj/W0NKra/qwMdZb0Xa1MdRanKtQoVXydZf0ZbDuyV4GBgdq8ebMyMjKsK/As7dix45zsu1V/t/I9Kl9n+j1yOp3KysrSunXryv3566fj4+Pjev57ZVYV6jz155KnfxcuTsHnMyMjQ4cOHfJ0OcU6duyYJMmYkp/UYDOnG+FBe/fu1Xnnnadvv/1WMTExrukPPvig1qxZow0bNhR6z+TJkzVlypSKLBMAAAAAcI7as2ePGjZsWOz8Sn2kOzQ0VNWqVVNqaqrb9NTUVIWHhxf5nokTJ2r8+PGu106nU4cPH1ZISIhstrI/tiIjI0ORkZHas2ePgoODy7wcVC701fvQU+9EX70TffVO9NU70VfvRF/PjjFGx44dU0RERInjKnXo9vf3V4cOHbRixQpdc801kv4O0StWrNA999xT5HsCAgIUEBDgNq127drlVlNwcDAfSC9EX70PPfVO9NU70VfvRF+9E331TvS17GrVqnXaMZU6dEvS+PHjNXz4cHXs2FGXXnqpZs2apePHj2vkyJGeLg0AAAAAgBJV+tB900036eDBg3rssce0f/9+tWvXTkuXLi10czUAAAAAACqbSh+6Jemee+4p9nTyihIQEKBJkyYVOnUdVRt99T701DvRV+9EX70TffVO9NU70deKUanvXg4AAAAAQFVWsQ/ZAwAAAADgHELoBgAAAADAIoRuAAAAAAAsQuguhTlz5qhRo0YKDAxUp06d9P3333u6JPzPU089pUsuuUQ1a9ZU/fr1dc011yg5OdltTE5OjkaPHq2QkBDVqFFDgwYNUmpqqtuYlJQU9e/fX0FBQapfv74eeOABnThxwm3M6tWrdfHFFysgIEDNmjXTggULrN48/M/06dNls9k0btw41zT6WjX99ddfGjZsmEJCQmS323XhhRdq06ZNrvnGGD322GNq0KCB7Ha7YmNjtWPHDrdlHD58WEOHDlVwcLBq166tUaNGKTMz023Mjz/+qG7duikwMFCRkZF65plnKmT7zkX5+fl69NFH1bhxY9ntdjVt2lSPP/64Tr5lDH2t/NauXaurrrpKERERstls+uijj9zmV2QPFy9erJYtWyowMFAXXnihPv/883Lf3nNBST11OBx66KGHdOGFF6p69eqKiIjQrbfeqr1797otg55WPqf7rp7szjvvlM1m06xZs9ym01cPMCjRwoULjb+/v5k/f7756aefzO23325q165tUlNTPV0ajDFxcXEmPj7ebN++3WzdutX069fPREVFmczMTNeYO++800RGRpoVK1aYTZs2mc6dO5vLLrvMNf/EiROmTZs2JjY21vzwww/m888/N6GhoWbixImuMb/99psJCgoy48ePNz///LN58cUXTbVq1czSpUsrdHvPRd9//71p1KiRueiii8zYsWNd0+lr1XP48GETHR1tRowYYTZs2GB+++03s2zZMrNz507XmOnTp5tatWqZjz76yCQmJpqrr77aNG7c2GRnZ7vGXHnllaZt27bmu+++M+vWrTPNmjUzgwcPds1PT083YWFhZujQoWb79u3m3XffNXa73fznP/+p0O09V0ybNs2EhISYJUuWmN27d5vFixebGjVqmNmzZ7vG0NfK7/PPPzcPP/yw+eCDD4wk8+GHH7rNr6gefvPNN6ZatWrmmWeeMT///LN55JFHjJ+fn9m2bZvl+8DblNTTo0ePmtjYWLNo0SLzyy+/mPXr15tLL73UdOjQwW0Z9LTyOd13tcAHH3xg2rZtayIiIszzzz/vNo++VjxC92lceumlZvTo0a7X+fn5JiIiwjz11FMerArFOXDggJFk1qxZY4z5+y8VPz8/s3jxYteYpKQkI8msX7/eGPP3Dy8fHx+zf/9+15iXX37ZBAcHm9zcXGOMMQ8++KBp3bq127puuukmExcXZ/UmndOOHTtmmjdvbpYvX2569OjhCt30tWp66KGHTNeuXYud73Q6TXh4uHn22Wdd044ePWoCAgLMu+++a4wx5ueffzaSzMaNG11jvvjiC2Oz2cxff/1ljDFm7ty5pk6dOq4+F6y7RYsW5b1JMMb079/f3HbbbW7TrrvuOjN06FBjDH2tik79Rb4ie3jjjTea/v37u9XTqVMn889//rNct/FcU1I4K/D9998bSeaPP/4wxtDTqqC4vv7555/mvPPOM9u3bzfR0dFuoZu+eganl5cgLy9PmzdvVmxsrGuaj4+PYmNjtX79eg9WhuKkp6dLkurWrStJ2rx5sxwOh1sPW7ZsqaioKFcP169frwsvvFBhYWGuMXFxccrIyNBPP/3kGnPyMgrG8Dmw1ujRo9W/f/9C+56+Vk2ffPKJOnbsqBtuuEH169dX+/bt9eqrr7rm7969W/v373frSa1atdSpUye3vtauXVsdO3Z0jYmNjZWPj482bNjgGtO9e3f5+/u7xsTFxSk5OVlHjhyxejPPOZdddplWrFihX3/9VZKUmJior7/+Wn379pVEX71BRfaQn8uek56eLpvNptq1a0uip1WV0+nULbfcogceeECtW7cuNJ++egahuwRpaWnKz893+6VdksLCwrR//34PVYXiOJ1OjRs3Tl26dFGbNm0kSfv375e/v7/rL5ACJ/dw//79Rfa4YF5JYzIyMpSdnW3F5pzzFi5cqC1btuipp54qNI++Vk2//fabXn75ZTVv3lzLli3TXXfdpTFjxuiNN96Q9P99Keln7v79+1W/fn23+b6+vqpbt+4Z9R7l51//+pduvvlmtWzZUn5+fmrfvr3GjRunoUOHSqKv3qAie1jcGHpsrZycHD300EMaPHiwgoODJdHTqurpp5+Wr6+vxowZU+R8+uoZvp4uACgvo0eP1vbt2/X11197uhScpT179mjs2LFavny5AgMDPV0OyonT6VTHjh315JNPSpLat2+v7du3a968eRo+fLiHq0NZvffee3rnnXeUkJCg1q1ba+vWrRo3bpwiIiLoK1AFOBwO3XjjjTLG6OWXX/Z0OTgLmzdv1uzZs7VlyxbZbDZPl4OTcKS7BKGhoapWrVqhOyKnpqYqPDzcQ1WhKPfcc4+WLFmiVatWqWHDhq7p4eHhysvL09GjR93Gn9zD8PDwIntcMK+kMcHBwbLb7eW9Oee8zZs368CBA7r44ovl6+srX19frVmzRi+88IJ8fX0VFhZGX6ugBg0a6IILLnCb1qpVK6WkpEj6/76U9DM3PDxcBw4ccJt/4sQJHT58+Ix6j/LzwAMPuI52X3jhhbrlllt03333uc5Soa9VX0X2sLgx9NgaBYH7jz/+0PLly11HuSV6WhWtW7dOBw4cUFRUlOv3pz/++EMTJkxQo0aNJNFXTyF0l8Df318dOnTQihUrXNOcTqdWrFihmJgYD1aGAsYY3XPPPfrwww+1cuVKNW7c2G1+hw4d5Ofn59bD5ORkpaSkuHoYExOjbdu2uf0AKviLpyAgxMTEuC2jYAyfA2v06tVL27Zt09atW11/OnbsqKFDh7r+n75WPV26dCn0SL9ff/1V0dHRkqTGjRsrPDzcrScZGRnasGGDW1+PHj2qzZs3u8asXLlSTqdTnTp1co1Zu3atHA6Ha8zy5cvVokUL1alTx7LtO1dlZWXJx8f914lq1arJ6XRKoq/eoCJ7yM/lilMQuHfs2KGvvvpKISEhbvPpadVzyy236Mcff3T7/SkiIkIPPPCAli1bJom+eoyn7+RW2S1cuNAEBASYBQsWmJ9//tnccccdpnbt2m53RIbn3HXXXaZWrVpm9erVZt++fa4/WVlZrjF33nmniYqKMitXrjSbNm0yMTExJiYmxjW/4NFSffr0MVu3bjVLly419erVK/LRUg888IBJSkoyc+bM4dFSFezku5cbQ1+rou+//974+vqaadOmmR07dph33nnHBAUFmbfffts1Zvr06aZ27drm448/Nj/++KMZOHBgkY8lat++vdmwYYP5+uuvTfPmzd0edXL06FETFhZmbrnlFrN9+3azcOFCExQUxKOlLDJ8+HBz3nnnuR4Z9sEHH5jQ0FDz4IMPusbQ18rv2LFj5ocffjA//PCDkWRmzpxpfvjhB9edrCuqh998843x9fU1zz33nElKSjKTJk3iMURlVFJP8/LyzNVXX20aNmxotm7d6vY71Ml3rKanlc/pvqunOvXu5cbQV08gdJfCiy++aKKiooy/v7+59NJLzXfffefpkvA/kor8Ex8f7xqTnZ1t7r77blOnTh0TFBRkrr32WrNv3z635fz++++mb9++xm63m9DQUDNhwgTjcDjcxqxatcq0a9fO+Pv7myZNmritA9Y7NXTT16rp008/NW3atDEBAQGmZcuW5pVXXnGb73Q6zaOPPmrCwsJMQECA6dWrl0lOTnYbc+jQITN48GBTo0YNExwcbEaOHGmOHTvmNiYxMdF07drVBAQEmPPOO89Mnz7d8m07V2VkZJixY8eaqKgoExgYaJo0aWIefvhht1/c6Wvlt2rVqiL/Ph0+fLgxpmJ7+N5775nzzz/f+Pv7m9atW5vPPvvMsu32ZiX1dPfu3cX+DrVq1SrXMuhp5XO67+qpigrd9LXi2YwxpiKOqAMAAAAAcK7hmm4AAAAAACxC6AYAAAAAwCKEbgAAAAAALELoBgAAAADAIoRuAAAAAAAsQugGAAAAAMAihG4AAAAAACxC6AYAAAAAwCKEbgAALPb777/LZrNp69at5brcX375RZ07d1ZgYKDatWtXrsu20uTJk6tUvQAAnA1CNwCgShoxYoRsNpvuvPPOQvNGjx4tm82mESNGnNEybTabPvroo/Ip8AxdfvnlGjdu3Bm9Z9KkSapevbqSk5O1YsUKawo7S0Xt0/vvv79C6k1MTNTVV1+t+vXrKzAwUI0aNdJNN92kAwcOWL5uAAAKELoBAFVWZGSkFi5cqOzsbNe0nJwcJSQkKCoqyoOVVYxdu3apa9euio6OVkhISJmWkZeXV85VnV6NGjXKXG9pHTx4UL169VLdunW1bNkyJSUlKT4+XhERETp+/Lhl63U4HJYtGwBQNRG6AQBV1sUXX6zIyEh98MEHrmkffPCBoqKi1L59e7exjRo10qxZs9ymtWvXTpMnT3bNl6Rrr71WNpvN9XrEiBG65ppr3N43btw4XX755a7XS5cuVdeuXVW7dm2FhIRowIAB2rVr11ltW6NGjfTkk0/qtttuU82aNRUVFaVXXnnFNd9ms2nz5s2aOnWqbDabazu2bdumnj17ym63KyQkRHfccYcyMzNd7yvYnmnTpikiIkItWrRwnf7+3nvvqVu3brLb7brkkkv066+/auPGjerYsaNq1Kihvn376uDBg65lbdy4Ub1791ZoaKhq1aqlHj16aMuWLW7bUNQ+PfX0cqfTqalTp6phw4YKCAhQu3bttHTpUtf8gvo++OADXXHFFQoKClLbtm21fv36YvffN998o/T0dL322mtq3769GjdurCuuuELPP/+8Gjdu7Br3008/acCAAQoODlbNmjXVrVs3V+9KW9eiRYvUo0cPBQYG6p133pEkvfbaa2rVqpUCAwPVsmVLzZ0793QtBwB4KUI3AKBKu+222xQfH+96PX/+fI0cOfKMl7Nx40ZJUnx8vPbt2+d6XRrHjx/X+PHjtWnTJq1YsUI+Pj669tpr5XQ6z7iOk82YMUMdO3bUDz/8oLvvvlt33XWXkpOTJUn79u1T69atNWHCBO3bt0/333+/jh8/rri4ONWpU0cbN27U4sWL9dVXX+mee+5xW+6KFSuUnJys5cuXa8mSJa7pkyZN0iOPPKItW7bI19dXQ4YM0YMPPqjZs2dr3bp12rlzpx577DHX+GPHjmn48OH6+uuv9d1336l58+bq16+fjh07Jqn0+3T27NmaMWOGnnvuOf3444+Ki4vT1VdfrR07driNe/jhh3X//fdr69atOv/88zV48GCdOHGiyGWGh4frxIkT+vDDD2WMKXLMX3/9pe7duysgIEArV67U5s2bddttt7mWWdq6/vWvf2ns2LFKSkpSXFyc3nnnHT322GOaNm2akpKS9OSTT+rRRx/VG2+8UWQdAAAvZwAAqIKGDx9uBg4caA4cOGACAgLM77//bn7//XcTGBhoDh48aAYOHGiGDx/uGh8dHW2ef/55t2W0bdvWTJo0yfVakvnwww+LXM/Jxo4da3r06FFsbQcPHjSSzLZt24wxxuzevdtIMj/88EOx7+nRo4cZO3asW73Dhg1zvXY6naZ+/frm5ZdfLrb+V155xdSpU8dkZma6pn322WfGx8fH7N+/37U9YWFhJjc31zWmoL7XXnvNNe3dd981ksyKFStc05566inTokWLYrchPz/f1KxZ03z66aeuaUXt00mTJpm2bdu6XkdERJhp06a5jbnkkkvM3XffXWx9P/30k5FkkpKSiq3n3//+t/H19TV169Y1V155pXnmmWdc+8EYYyZOnGgaN25s8vLyinx/aeuaNWuW25imTZuahIQEt2mPP/64iYmJKbZWAID34kg3AKBKq1evnvr3768FCxYoPj5e/fv3V2hoaIXWsGPHDg0ePFhNmjRRcHCw6zTqlJSUs1ruRRdd5Pp/m82m8PDwEm8ClpSUpLZt26p69equaV26dJHT6XQdIZekCy+8UP7+/iWuLywszDX25Gknrz81NVW33367mjdvrlq1aik4OFiZmZlntN0ZGRnau3evunTp4ja9S5cuSkpKKra+Bg0aSFKJ+2PatGnav3+/5s2bp9atW2vevHlq2bKltm3bJknaunWrunXrJj8/v7Oqq2PHjq7/P378uHbt2qVRo0apRo0arj9PPPHEWV9yAAComnw9XQAAAGfrtttuc51CPWfOnCLH+Pj4FDrNuDQ3vSrN+6666ipFR0fr1VdfVUREhJxOp9q0aXPWNyk7NQzabLazPmVdklsoL259NputyGknr3/48OE6dOiQZs+erejoaAUEBCgmJsaym7MVVd/p9kdISIhuuOEG3XDDDXryySfVvn17Pffcc3rjjTdkt9vLpa6T92fB9fOvvvqqOnXq5DauWrVq5bI+AEDVwpFuAECVd+WVVyovL08Oh0NxcXFFjqlXr5727dvnep2RkaHdu3e7jfHz81N+fn6J75Pk9rztQ4cOKTk5WY888oh69eqlVq1a6ciRI2e5RWXTqlUrJSYmut2d+5tvvpGPj49atGhR7uv75ptvNGbMGPXr10+tW7dWQECA0tLS3MYUtU9PFhwcrIiICH3zzTeFln3BBReUa73+/v5q2rSpa/9cdNFFWrduXZH/+FLWusLCwhQREaHffvtNzZo1c/tz8g3cAADnDkI3AKDKq1atmpKSkvTzzz8XezSxZ8+eeuutt7Ru3Tpt27ZNw4cPLzS2UaNGWrFihfbv3+8Kzj179tSmTZv05ptvaseOHZo0aZK2b9/uek+dOnUUEhKiV155RTt37tTKlSs1fvx46za2BEOHDlVgYKCGDx+u7du3a9WqVbr33nt1yy23uE4XL0/NmzfXW2+9paSkJG3YsEFDhw4tdPS4qH16qgceeEBPP/20Fi1apOTkZP3rX//S1q1bNXbs2DLXtmTJEg0bNkxLlizRr7/+quTkZD333HP6/PPPNXDgQEnSPffco4yMDN18883atGmTduzYobfeest1Kn5Z65oyZYqeeuopvfDCC/r111+1bds2xcfHa+bMmWXeHgBA1UXoBgB4heDgYAUHBxc7f+LEierRo4cGDBig/v3765prrlHTpk3dxsyYMUPLly9XZGSk65FjcXFxevTRR/Xggw/qkksu0bFjx3Trrbe63uPj46OFCxdq8+bNatOmje677z49++yz1mzkaQQFBWnZsmU6fPiwLrnkEl1//fXq1auXXnrpJUvW9/rrr+vIkSO6+OKLdcstt2jMmDGqX7++25ii9umpxowZo/Hjx2vChAm68MILtXTpUn3yySdq3rx5mWu74IILFBQUpAkTJqhdu3bq3Lmz3nvvPb322mu65ZZbJP196vnKlSuVmZmpHj16qEOHDnr11Vddp7GXta5//OMfeu211xQfH68LL7xQPXr00IIFCzjSDQDnKJs59UI1AAAAAABQLjjSDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWITQDQAAAACARQjdAAAAAABYhNANAAAAAIBFCN0AAAAAAFiE0A0AAAAAgEUI3QAAAAAAWOT/AFd7BrIKntCuAAAAAElFTkSuQmCC"},"metadata":{}}]},{"cell_type":"code","source":"import scipy\n# Calculate Pearson and Spearman correlations\nsorted_scores = sorted(scores, reverse=True)\npearson_corr = scipy.stats.pearsonr(sorted_scores, scores)\nspearman_corr = scipy.stats.spearmanr(sorted_scores, scores)\n\nprint(\"Scores:\", scores)\nprint(\"Pearson correlation:\", pearson_corr)\nprint(\"Spearman correlation:\", spearman_corr)","metadata":{"execution":{"iopub.status.busy":"2024-06-25T01:58:35.020340Z","iopub.execute_input":"2024-06-25T01:58:35.021200Z","iopub.status.idle":"2024-06-25T01:58:35.073337Z","shell.execute_reply.started":"2024-06-25T01:58:35.021170Z","shell.execute_reply":"2024-06-25T01:58:35.072346Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stdout","text":"Scores: [3932.5592834010513, 1892.5224786262388, 1460.1082573153421, 3556.9152857886397, 3851.506193828018, 2889.7816153095273, 2099.2062308543454, 1241.4499558566172, 1134.1992558003271, 1642.4795955649488, 2364.835691417761, 4144.34303184801, 2421.311335665482, 3198.037476526745, 2247.019128533504, 3394.139304471939, 1768.0672796602498, 2617.290237486106, 2551.597396074766, 1660.0896307291437, 1531.232840073453, 2744.316006469452, 3827.107615621263, 2458.513212632744, 1396.6799553489875, 7955.21835934194, 2168.9747680499286, 3180.442953185039, 3139.459655410042, 1965.0773752132031, 4334.618715567403, 1303.881404820357, 1705.8984617380768, 9292.115287449367, 14654.138438686257, 5205.3917243853775, 872.5200962255024, 5568.92756061391, 3488.926632191426, 1250.662973147574, 2477.370388062971, 3941.45991875404, 1895.1378362219812, 2517.8807956495957, 1608.3155710534725, 1221.4292674320295, 3845.449454104851, 14069.96854140062, 2599.886992059516, 1493.687828589547, 819.1528214196526, 1506.5704951382131, 2621.909992958993, 1528.0298222016957, 7838.218788269359, 6485.431392159477, 1318.7198910479688, 6376.070192163797, 4652.788240804501, 1306.150459345508, 1240.360127435049, 4053.1067042862314, 4021.6537448213053, 4706.270867239726, 407.710610856598, 1155.0864534896764, 2727.3908132325787, 1115.7271190268716, 3991.2970081611447, 923.813688936508, 3251.2374488995592, 5743.46896459997, 1124.222794180216, 775.2666404121252, 2639.9340663815155, 5318.2925799663835, 2339.7863833681813, 3303.075612526414, 4063.3252541156626, 2573.6962345850525, 1725.65902344787, 1611.7142031573198, 1441.1106491738765, 911.999972184664, 855.8319308485179, 3738.7223017593847, 774.9058679552641, 3267.632350372358, 728.7586443728551, 2006.9083413111416, 1055.1942095589252, 2610.052659455351, 1818.3388852942517, 707.7038838226267, 919.6280281486729, 1044.2427776149966, 1468.0662988309846, 2935.2605351290003, 6780.033238486261, 1927.612452740652, 5714.445806978333, 2335.257298206728, 1939.0882029285242, 5377.857296022761, 5013.81277092653, 3002.09056356874, 3314.9598233904226, 3996.473875681083, 1246.621774223669, 1867.8193714496626, 7455.583054359164, 2569.5181310892904, 4189.3510627343585, 2844.1233531637135, 3405.398696753011, 2364.8012367772726, 6756.1215660450725, 2156.5275357070404, 1259.5896812401197, 3831.888524769904, 3771.0194464135866, 1786.883572242164, 3460.544464537956, 2056.0109153511166, 2171.323048149112, 3350.0854783857403, 2503.359300808438, 2994.8200328613784, 1621.1260917324548, 2411.294069368877, 821.085053641322, 5221.111457894347, 2937.6162718924897, 1558.4728632924289, 6617.278348570319, 3508.5790727674284, 1804.3888739707736, 5293.256650001906, 4914.809260560065, 1762.513241984066, 2797.189306900455, 1335.203343292663, 3120.8677662604464, 1809.7560044996899, 2509.7848702622387, 4273.575017789512, 1903.7637484996178, 1897.9021990727636, 2701.0939208370164, 1216.7594299384828, 639.1290167619286, 543.0903458397174, 3123.12131896335, 2256.4602763043704, 1498.5358527817928, 5465.054676080146, 5418.9452732766, 1842.2831555555326, 688.7694628987532, 1784.653404104715, 3609.7816061615467, 4690.759062338131, 7881.344933220096, 3310.203286286497, 3233.626426223132, 1959.9627002924308, 855.5765122299183, 10691.477991683762, 2029.991654856019, 6168.265710686516, 2918.378836478318, 686.4899184638891, 1312.1246637130546, 533.4165165654356, 4513.561373169277, 777.2772244687983, 2378.031197208547, 603.4691739453056, 3559.037817596542, 1767.7031384090574, 1563.1657938202325, 2398.6630535395934, 3880.9071702110527, 1251.1779318538547, 3423.7355476782404, 1580.6549022330732, 1844.736675161474, 2486.4488388400614, 4526.736786114767, 1204.7344474860274, 3089.38670359339, 2056.759805894274, 2663.4770444536953, 2532.881195268306, 2037.4898273705505, 3409.439883195495, 2290.1593368940507, 1879.9366422706285, 1763.6635560038285, 978.5329120821013, 1730.3269054877053, 6850.10132019094, 4593.152422624651, 4337.189092928294, 1371.2288541458504, 3434.394315974465, 1220.015522175255, 4921.033260671205, 4053.189320490131, 5332.800826593457, 1599.0696303394805, 1592.8975628296394, 3030.888494724317, 9984.700480241492, 4376.545164632571, 1593.8869676265804, 3062.6812989058053, 1668.0824535290176, 11168.784389241237, 11300.414672648903, 1770.8756067251352, 4900.457729280653, 1464.7401475333718, 5025.043676262615, 1410.9753801962574, 1859.534041179882, 6445.667206284492, 5248.479723791857, 3329.703836535861, 1574.764076258624, 2986.6181276722923, 3885.0913167385715, 3326.5751354990025, 1015.2933556926315, 3426.5980056471108, 1356.1026738124374, 4401.145310282047, 4213.9099608042525, 1481.4989240851094, 1850.7554927854844, 2896.9071455614867, 2830.0910589100304, 3159.4529447648165, 1586.997113807297, 4344.689074836582, 7571.822838787932, 3145.3014885682696, 4260.609981879579, 515.0740392513934, 2761.374574202898, 1526.2266399464456, 2874.311818855197, 8103.798538181826, 4172.999901699564, 1176.1020090356974, 2087.875902250697, 2971.018574882027, 2929.1521330991836, 5752.486687960571, 8002.3390508543525, 3312.4736485388335, 2067.671321991174, 1176.5675591472318, 1144.6713436503906, 1100.7136042727413, 5462.170098345213, 3077.8795995593437, 2528.0190892951614, 1397.679525376226, 3946.885113129756, 3475.7866766669295, 6200.138574713683, 1722.9057220568898]\nPearson correlation: PearsonRResult(statistic=-0.038099589455054705, pvalue=0.530759061285225)\nSpearman correlation: SignificanceResult(statistic=-0.06976122846668145, pvalue=0.2506566531036057)\n","output_type":"stream"}]},{"cell_type":"markdown","source":"Cosine Similarity","metadata":{}},{"cell_type":"code","source":"teacher_reply","metadata":{"execution":{"iopub.status.busy":"2024-06-25T08:55:31.120394Z","iopub.execute_input":"2024-06-25T08:55:31.121103Z","iopub.status.idle":"2024-06-25T08:55:31.152401Z","shell.execute_reply.started":"2024-06-25T08:55:31.121072Z","shell.execute_reply":"2024-06-25T08:55:31.151360Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":37,"outputs":[{"execution_count":37,"output_type":"execute_result","data":{"text/plain":"[{'id': 'test_0001',\n  'dialog': \"student : Do you say them as 'extreme weather', right?\\nteacher : Absolutely yes the two words together: so, the three collocations for the consequences are; rising sea levels/melting ice caps and extreme weather - last category: what about the solutions?!\\nstudent : It's reducing CO2 emissions. For example, using solar/wind energy instead of fossil fuels.\",\n  'response': \"yes fantastic - so one answer is to 'reduce carbon emissions' (Collocation!) - what kind of energy is e.g. solar/wind\"},\n {'id': 'test_0006',\n  'dialog': \"teacher : Which is easy, because you can use my / his / your etc. and not think about articles!\\nstudent : Only 10 minutes left!\\nteacher : I know, we can finish early if you are getting cold?\\nstudent : I'm really cold\",\n  'response': \"Oh, poor <STUDENT>! Go get somewhere warm! I'll see you on Friday 9.30 am. Well done today, you've been very brave!\"},\n {'id': 'test_0009',\n  'dialog': \"teacher : f you get a very high mark in e.g. a university degree you might receive a 'distinction' if you're familiar with that idea\\nteacher : Thanks for your example with relieved -  try another adjective when you're ready\\nstudent : That's interesting. Thank you for telling me the origin of the phrase.\",\n  'response': \"THat's OK - do you know the word 'nautical'? as in related to the sailing - there are loads of phrases related to this area in English\"},\n {'id': 'test_0010',\n  'dialog': \"student : innocent??\\nteacher : Right! and 'abrupt' = frank/direct\\nteacher : If you're speak frankly/directly is it good?\\nteacher : I mean if you speak\\nstudent : It depends on situations, but I think it's generally good.\",\n  'response': \"Yes you're absolutely right - if you need tobe frank/direct then maybe it's like honest/open or just saying what needs to be said but...if it's rude/insensitive or just seems like a rude/agressive question/request for example then we'd say 'abrupt'\"},\n {'id': 'test_0011',\n  'dialog': \"student : so..they play games to stop feeling depressed/lonely/bored.\\nteacher : yes ok...E.g. I realized that smoking was bad when I read the news.'\\nteacher : Try your sentence with 'realise' again if you can ok?\\nstudent : Adults can realise that playing in the game word, but children are not.\",\n  'response': 'OK thanks <STUDENT> ...you need a subject + present continous (2 extra words in total!)'},\n {'id': 'test_0012',\n  'dialog': \"student : yes, I will find a word\\nstudent : relieve one's feelings\\nteacher : OK good idea...so , you can 'relieve pain' for example...in more common language we can say e.g. they play games to stop feeling depressed\\nteacher : or lonely or bored etc\\nstudent : okey\",\n  'response': \"Before you said: 'adults always can realise the game word, but children are not'...after the verb 'realise' you need to say '...that + clause (subject and verb)...' can you try it?\"},\n {'id': 'test_0013',\n  'dialog': 'teacher : materialistic = the adjective you need here (good word)\\nstudent : yes... I think it is a strange idea...\\nteacher : Are there any arguments in favour of spending money on space exploration which might be easier to explain?\\nstudent : I think If I disagree with this idea is more easily to explain.',\n  'response': \"Oh sorry <STUDENT> - can you hold that though t and tell me after the break OK? OK i see your point there...we'l c ome back to it OK? bye for now..\"},\n {'id': 'test_0014',\n  'dialog': \"teacher : IS this one new for you too?\\nstudent : I think so\\nteacher : OK - well you've got to start somewhere! I think these two structures will help you get a better score\\nstudent : Both of them are new and wonderful for me.\\nstudent : thanks\",\n  'response': \"ok...so in the context of the pie charts we're looking at now we can say: Natural gas ______ for _____ of electricity in Australia in 1980'?\"},\n {'id': 'test_0015',\n  'dialog': \"teacher : BUT with 's' at the end!\\nstudent : okey, thanks\\nteacher : do you know the word 'colloborate'?\\nstudent : no.. it looks like more advance\\nteacher : Well - it's a nice word yes... Does it mean work together or work separately do you think?\\nstudent : I think it may means work together\",\n  'response': \"yes you're right... a good synonym for co-operate...how about 'sophisticated' - do you know this word?\"},\n {'id': 'test_0019',\n  'dialog': \"student : yes\\nteacher : yes I remember now sorry - the master's!\\nstudent : no,,,I will study my first degree in the UK\\nstudent : I just left from Japan's university and becoming a freshman in the UK\\nstudent : because I really want to study in LSE for my master, it is really difficult for me if I graduated from Japan.\",\n  'response': \"When you say 'freshman' you mean in your first year right ?Sorry, it's an American phrase...what's the word for a second-year student?\"},\n {'id': 'test_0021',\n  'dialog': \"teacher : For example...pregnant womenoften go to hospital to give _____?\\nteacher : beginning with 'b_ _ _ _ '!\\nstudent : begin\\nteacher : Good try! 'birth' is the word here...\\nstudent : okey.\",\n  'response': 'so e.g. before birth/at birth/after birth'},\n {'id': 'test_0026',\n  'dialog': \"student : against??\\nteacher : no try again!\\nstudent : for??\\nstudent : for??\\nteacher : Yes that's it!\\nstudent : I didn't know that.\\nteacher : OK so your sentience is possible/good  -can you try another sentence with the word 'beat' for the same score\\nstudent : Barcelona was beaten by Liverpool 1-0.\",\n  'response': \"OK great! Or: Liverpool beat Barcelona 1-0 ('one nil' remember in speaking!) - also we normally say 'were' for a football team in speaking\"},\n {'id': 'test_0027',\n  'dialog': 'student : I mean\\nstudent : If the teacher teach a good things to students\\nstudent : These student’ll be an effective person',\n  'response': 'You said: ‘The world change’ – can you think about the verb tense here and write it again? You said :’you’ll make a mistake’ – can you change the ‘will’?! Think about the verb tenses in your sentence beginning: For example, if your boss…’  And the verb tense in your final sentence… ? ‘So it …’'},\n {'id': 'test_0028',\n  'dialog': \"student : I need a lo of preparation. At the moment I'm taking lessons (1 hour a week, I can't afford more) and I have some books... I should spend more time studying however after the course I'm too tired.\\nstudent : I see what you mean. That's very good!\",\n  'response': 'OK right - well if you like you can ask me for any help/use this time to practise new language etc'},\n {'id': 'test_0030',\n  'dialog': \"student : hahah Yes, I'm sorry to tell you\\nteacher : But yours is unusual right?\\nstudent : Not on the Canary Islands, it's originally from here so there are a lot of <STUDENT'S NAME> but not on (in?) the mainland.\",\n  'response': \"Ah OK i see - I didn't consider you had kind of indigenous names etc! 'On the mainland' = good ..on an island etc\"},\n {'id': 'test_0031',\n  'dialog': \"teacher : ha! yes great ecample!!!!!!!!\\nteacher : example\\nteacher : the same as in America in the north of the UK but less nasal!\\nstudent : That's interesting.\\nstudent : Thank you for telling me a lot of things today.\",\n  'response': \"OK thanks a lot <STUDENT> (you see 'many thanks' much more formal!) I'll contact you soon about a further lesson ok?\"},\n {'id': 'test_0032',\n  'dialog': \"teacher : ok good\\nteacher : And one last test: what are your plans after this lesson?\\nstudent : I am going to revision all knowledge which I studied in the class\\nteacher : OK good answer - what a good student1\\nteacher : OK so remember: 'going to' + infinitive (VERB right?) - so where's your verb?\\nstudent : the verb is after  the 'going to'\\nstudent : I am going to review the knowledge which I study in this class.\",\n  'response': \"so is 'revision' a verb? Remember the ending -ion is common for nouns\"},\n {'id': 'test_0034',\n  'dialog': \"student : OK... I have a lot of things to study today!\\nteacher : We can stop there if you like! Give me  a second...(or continue with example if you prefer!\\nstudent : Ah, of course it's OK! His story relieve me.\",\n  'response': \"THanks - a logical example...but (!) ...have you seen for example adverts for painkillers? They always say 'fast,efffective pain relief' ...so...what can you releive\"},\n {'id': 'test_0035',\n  'dialog': \"student : thanks,, I really want to use the word which you were teaching me.\\nteacher : Use an adjective beginning with 's' for the first gap and add a tiny little word for the second gap...can you do it?! It is often argued that the most effective way to lead a less ________  lifestyle is __ rent a house.\\nstudent : depress   and   that\",\n  'response': 'stress = noun ... and the adjective is ______?'},\n {'id': 'test_0036',\n  'dialog': 'teacher : Hi <STUDENT>!\\nstudent : Hi <TEACHER>!',\n  'response': 'How are you today? Still stuck inside :('},\n {'id': 'test_0037',\n  'dialog': 'teacher : Could you try relief or excitement? I thin k this is useful...\\nstudent : The result of an exam was a relief for me.',\n  'response': \"THanks <STUDENT> - yes, good example and you can say a relief FOR...it's common to say: it was relief to me. Also the phrase is quite common: it came as a relief to me\"},\n {'id': 'test_0038',\n  'dialog': \"teacher : But all really common\\nstudent : OK, thank you. I'll write them down\\nteacher : OK ...you're happy with e.g. 'I got my car fixed' right - I'm sure you know it ....\\nstudent : So pull in means that you stop somewher??\",\n  'response': \"Yes - at  a place I guess ..but confusingly...it can also be used for the same as pull over actually sorry! or 'had my car fixed'  = the same\"},\n {'id': 'test_0039',\n  'dialog': \"teacher : Hi there...\\nstudent : hey, I am here\\nteacher : OK so,  let's have a think about this again <STUDENT> OK? Don't answer yet but here is what we're going to do:  'I saw this when you_______(teach) us.'\\nteacher : So <STUDENT>....what action happened first? you seeing the table or me teaching you?\\nstudent : I seeing table firstly\",\n  'response': 'Are you sure? Was i your teacher before that?!'},\n {'id': 'test_0040',\n  'dialog': \"teacher : All right <STUDENT>?\\nstudent : I think exercise is plural\\nstudent : exercises\\nteacher : Well yes...they are both possible but they mean slightly different things! Which one means physical activity in general?\\nteacher : So e.g. It's good to try to do regular __________?\\nstudent : I think 'exercise' is means physical activity\",\n  'response': \"OR 'Regular _________ is part of a healthy lifestyle' (there's a clue there with the verb after the gap!)\"},\n {'id': 'test_0041',\n  'dialog': \"teacher : yes - it would be a very high frequency word in this area\\nteacher : <STUDENT>, would you like to finish off page 128?\\nstudent : yes, please\\nteacher : ok go for it when you're ready\\nstudent : we were looking at answer number 5\\nteacher : yes sure...\\nstudent : the answer is b but I checked the meaning of 'lessened' and it seems very similar to shunk, isn't it?\",\n  'response': \"yes that's true in general. .if i give you the subjects: violence/pain/ do they collocate with lessen or shrink?\"},\n {'id': 'test_0042',\n  'dialog': \"student : yes, I'm ready!\\nteacher : Yes, you;re right ... it;s that too like if you;re hurt yourself/forget something for example (bit less rude than shit)\\nstudent : I see\",\n  'response': 'Can you find a word/phrase in the 5 comments with the following meaning? They’re in the right order by the way):   a) get better/not die! b) allowed to leave hospital  c) generous d) unpleasant and prolonged experience  e) brief contact with something unpleasant f) rubbish'},\n {'id': 'test_0043',\n  'dialog': \"student : Oh no, it's a mistake. I thought the verb was to tent, I`ll correct it thanks\\nteacher : OK sure - She hit me with her elbow the other day and it really hurt even though she's only 9\\nstudent : haha did she hit you by accident??\",\n  'response': 'No! we were play fighting - but she takes it quite seriously hahah the tricky thing is as a parent i feel like I have to lose you know? but..at the same time I have to try to get her to cope with failure!'},\n {'id': 'test_0044',\n  'dialog': \"teacher : and 'they' you see\\nstudent : Ah, because they are a team, right??\",\n  'response': \"Yes that's it - you can use the singular too though especailly more formally (or imagine like thinking about the club as a company or whatever - definitely in that situation) So <STUDENT> - because you're IELTS example is not exactly like a knockout competiition you could say she's always trying to beat me  (as in my score) but not really win against me -\"},\n {'id': 'test_0046',\n  'dialog': \"student : sorry i'm back!\\nstudent : It's because of CO2?\\nteacher : Yes absolutely - what's happening with CO2?\\nstudent : CO2 emittion is going up because of consuming fossil fuels, and in addition, plants is decreasing that change CO2 to Os.\\nstudent : O2\\nstudent : Sorry, plants 'are'\",\n  'response': \"great answer! yes, so carbon emissions are increasing because we are 'burning fossil fuels' (collocation 1) and .....when you say plants - what kind of plants and where especially would you say?\"},\n {'id': 'test_0047',\n  'dialog': 'student : Sorry, it was an unnecessary part!!\\nteacher : Ok no problem! Listen <STUDENT> can you tell me about any famous places (like buildings) you have visited in the UK - just one sentnce is enough\\nstudent : Brighton is one of the best resort place in Europe.',\n  'response': \"OK thanks - yes I remember you said you went there. It's pretty cool right? What did you like about it?\"},\n {'id': 'test_0048',\n  'dialog': \"teacher : OK thanks <STUDENT> - let's try it together a bit more....hold on\\nstudent : yes thanks\\nteacher : the '50%' refers to electricity right?\\nteacher : so....Australia used coal to produce....?\\nstudent : 50% of electricity\\nteacher : Yes good!\\nteacher : So it's not a percentage of coal you see?\\nteacher : it's a percentage of ELECTRICITY from different fuels\\nstudent : yes now I have understood\",\n  'response': 'OK great! Can you try a different way of giving the same information e.g. Australia coal 1980 but this time starting with the number e.g. 50 units...'},\n {'id': 'test_0050',\n  'dialog': \"teacher : ?\\nstudent : not sure. I don't say it but I don't know if it is right\\nstudent : I search on the internet\\nteacher : no that's good - just wanted to check - lots of non-natives use 'search' a lot now because of the internet of course  - it's fine for that but they use it too much for other things you see - look up and check = perfect\\nstudent : OK\",\n  'response': 'by the way do you use the internet much to keep up your English? e.g. youtube ..or Netflix or something'},\n {'id': 'test_0051',\n  'dialog': 'student : haha...you always give me a lot of tips.\\nstudent : ok',\n  'response': \"The essay question yesterday was something like: 'What are the advantages of renting a home.' .... and the intro you've written feels like an intro to an essay question like this: 'How can we lead more relaxed lives?' Or maybe: 'Some people argue that the best way to lead a more relaxed life is to rent your home. Do you agree?' Do you see?!\"},\n {'id': 'test_0053',\n  'dialog': 'student : okey..I am confuse about it.\\nstudent : airport/university/station',\n  'response': \"So YES we use the 'the' for rivers/deserts/oceans (for example) but we DO NOT use 'the' for example when we talk about .....yes, I see your answer! Correct - airports, universities/stations    You just have to remember these rules I;m afraid!\"},\n {'id': 'test_0056',\n  'dialog': \"teacher : good idea ...reference would be 'Ref:'\\nstudent : okk\\nteacher : sorry not 'of course' for the pronunciation exactly!\\nteacher : it's arbitrary\\nteacher : unlike e.g and i.e.\\nteacher : Any other questions <STUDENT>?\\nstudent : No, thanks\\nstudent : We can continue with your plan :)\",\n  'response': \"OK.....so imagine: comparing 2 tablets let's say (!)  Samsung = £100 and ipad = £825 (not exactly right I know but...so my question is : how can you compare those two things for example? A simple sentence...?)\"},\n {'id': 'test_0057',\n  'dialog': \"student : but for example\\nstudent : if I say 'we can see...' then I have to use the present simle\\nstudent : simple\\nstudent : ?\\nstudent : even if I used always the past in my essay\",\n  'response': 'Yes ...we can see thd charts now BUT when you talk aboug the actual information in the charts its (normally) in the past ok? ...maybe change : people, satisfied and refurbishment ok? And....show me the whole conclusion again ok?'},\n {'id': 'test_0058',\n  'dialog': \"student : ......is that human is not isolation in space.\\nstudent : I am sure..\\nstudent : not sure\\nteacher : OK thanks ... 'isolation' = nice word.... you mean that there is life on other planets?\\nstudent : yes\\nteacher : OK undestood! So can you change your sentence to show that this just a possibility?\\nstudent : human want to spend millions on search of aline life.\",\n  'response': \"OK thanks ... 'spend millions on searching for alien life' = good sentence yes! Do you think this is a good idea? Do you think we should spend this money?\"},\n {'id': 'test_0059',\n  'dialog': \"student : Yes, I think so\\nteacher : Sorry <STUDENT> , before i forget .... close brush WITH soemthing (often death!) whereas 'a close shave' basically comes after the dangerous thing (and bigger variety of things possible than with 'brush') has been mentioned...OK?!\\nstudent : yes\\nteacher : e.g. we nearly crashed! yes, it was a close shave.\\nteacher : ok!\\nstudent : I see, thanks\",\n  'response': 'Fancy having a look at another set of comments quickly?'},\n {'id': 'test_0060',\n  'dialog': \"teacher : One more <STUDENT>: Sahara/Africa ...give me a sentence!\\nstudent : The Sahara is in Africa.\\nteacher : Yes good - so we use 'the' with _____ and ______?\\nstudent : The Sahara is biggest in Africa.\\nstudent : with river and desert？\\nstudent : or with place and name.\",\n  'response': 'Well hold on - a lot of the time yes BUT not always hold on....yes good with rivers and deserts...ok one more:       Pacific/world/biggest ...go for it!'},\n {'id': 'test_0061',\n  'dialog': \"teacher : Hi <STUDENT>, is that you?!\\nstudent : Hi <TEACHER>\\nstudent : How are you?\\nteacher : Yeah I'm good thanks - all OK with you?\\nstudent : I'm fine, getting a cold though...\",\n  'response': \"Oh no! Sorry to hear that - hopefully it'll only be a minor one....did you know it's common too to say: 'I'm coming down with a cold'?\"},\n {'id': 'test_0062',\n  'dialog': \"student : I see\\nstudent : (en español decimos 'practicar deporte')\\nteacher : Yes I understand So e.g. If an English person says I practise football' do they mean i play football\\nteacher : ?\\nstudent : no, I think that means that you repeat something you want to improve?\\nstudent : something to do with football\",\n  'response': \"In fact that's a terrible example really because basically we never say 'practise football' .... yes exactly to your last comment! So we could say I practise penalties (e.g. you take 100 again and again you see?\"},\n {'id': 'test_0063',\n  'dialog': 'student : excite about / disappointed with\\nteacher : OK good thanks\\nstudent : thanks',\n  'response': \"It's abit of a nightmare and sometimes more than one preposition is possible (with sometimes a small change in meaning) BUT don't worry about that now - let#s just practice with what we've got...OK so if you can <STUDENT> tell me something about yourself using: happy/disappointed/excited/annoyed/proud/scared/motivated OK? just do one at a time if you like (choose any one you like)\"},\n {'id': 'test_0064',\n  'dialog': \"student : no.. I just could write very simple sentences before.\\nteacher : OK well don't worry too much - the first two-sentence version was still good and you could still get a high score with that OK?\\nstudent : ok\\nstudent : hanks\\nstudent : thanks\",\n  'response': \"OK <STUDENT> before we finish this hour let me ask you one final question....can you tell me about nuclear power in France in 1980 AND 2000. So you could try it like this: 'In 1980, nuclear power.....However, in 2000....' OK?!\"},\n {'id': 'test_0065',\n  'dialog': \"teacher : One of....it can make them more violent.' is good OR for example: .'....can lead to v______  beh_______'??\\nstudent : can lead to violent behaviours\",\n  'response': \"yes good! Correct adjective and noun combination well done....'behaviours' possible but more normal as uncountable noun ('behaviour')\"},\n {'id': 'test_0067',\n  'dialog': \"student : and Scotland\\nteacher : yes sure\\nteacher : So ...we normally say: 'Whisky....'?\\nteacher : So START with 'whisky'\\nstudent : Whisky was made in Scotland.\\nteacher : ok good - only in the past or in general ?\\nstudent : In general we can say that whisky are made in Scotland.\\nteacher : OK - thanks. Singular or plural? I mean is 'whisky' countable?\\nstudent : I am not sure...may uncountable.\",\n  'response': \"Yes that's right..so we have to say: 'Whisky ___ made in Scotland'?\"},\n {'id': 'test_0068',\n  'dialog': \"teacher : I'll give you my own examples as you go along but don't stop OK?\\nstudent : ok\",\n  'response': \"My example: I'm _____  _____with my new car (it's working fine). ?\"},\n {'id': 'test_0070',\n  'dialog': \"teacher : with your snow example again i mean\\nstudent : It is highly likely that it will snow later\\nteacher : ok great thanks (just because you originally wrote: 'that itsnows'\\nstudent : It is highly likely to get some snow later? Not sure\\nteacher : that it snows\\nstudent : I see, sorryy\",\n  'response': \"that's OK! ....so, 'it is highly likely to + infinitive' e.g. likely to snow\"},\n {'id': 'test_0071',\n  'dialog': 'teacher : and its\\nteacher : for the class\\nstudent : okey - so I can say for the class, its continue until 04th of April. right?',\n  'response': \"Yes that's right ('It continues' = correct) OK a quick question ....in bad English = Me/tomorrow/go home...train/9:30 leave (you understand the idea?!) - say that in good English! Go ...'....'\"},\n {'id': 'test_0072',\n  'dialog': \"student : motivated to / proud of / scared of / disappointed to / happy with / annoy with / excited to\\nstudent : I think it is hard for me\\nstudent : I don't know the rules of preposition.\",\n  'response': \"Yes these are hard to remember  there are no rules I'm afraid!! Don't worry, this is why we're going to practise them! So look back <STUDENT> and find the words you need after: excited and disappointed (your other answers are correct)\"},\n {'id': 'test_0073',\n  'dialog': \"teacher : Ha! Quite ironic ! I guess it's a little bit similar to me teaching English in some countries where there is high unemployment - sometimes I felt a bit guilty because ogten in reality better English didn't really make any difference to their job prospects!\\nstudent : haha I know, it's very difficult to keep a language if you don't practise\",\n  'response': \"Yes I guess that's why you''re keen to keep up your ENglish right? I'll stop typing for a minute to avoid overlap\"},\n {'id': 'test_0074',\n  'dialog': \"teacher : the host can enter each room when they like\\nstudent : I see, that would be great\\nteacher : yes it's very effective (i was doing lessons on Zoom for a bit)\\nstudent : no, we all work at the same time so\\nstudent : Well, I'm not sure about the use of 'I' or 'me' in sentences like\",\n  'response': 'Do they use any other functions to allow/control participation? e.g. chatbox/take control of the screen function/whiteboard/screenshare (even?)/ sharing files?'},\n {'id': 'test_0075',\n  'dialog': 'student : and attempt to break regulations whatever in schools or at home, because they just want to copy the behaviours when they did in games.\\nstudent : they attempt ( I forgot the subject ).',\n  'response': 'OK thanks <STUDENT> - so in other words the games have _______ content which has a bad __________ on young people.  adjective for first gap and then noun for second gap - any ideas?'},\n {'id': 'test_0076',\n  'dialog': 'teacher : Choose a tense <STUDENT> to complete the question!\\nstudent : what are you doing?\\nstudent : what should I do?',\n  'response': \"OK yes good exactly: 'what's he doing? ' and the answer maybe is: 'Nothing special ...He ___ just ____ (walk) along the street'\"},\n {'id': 'test_0077',\n  'dialog': \"teacher : Yes you're right ...count to ten and keep calm....\\nstudent : yes, I still remembering the easy of problem and solution.\",\n  'response': \"OK right...so in the intro maybe we might want to say (for the background sentence) something like: ' The number of teachers ___________ (go down) in recent decades.' Can you put the verb in the right tense?\"},\n {'id': 'test_0080',\n  'dialog': \"teacher : OK so this is exactly what we need to practise <STUDENT>!!\\nteacher : So remember we said yesterday that if you say the word 'recent' this goes together with which verb tense??\\nstudent : present continue\\nstudent : past continue\",\n  'response': \"So for example the phrase 'in recent years' is talking about time beginning in the past (years ago) BUT up to now - so it's like a mixture of past and present you see? So which tense??\"},\n {'id': 'test_0082',\n  'dialog': \"teacher : Hi <STUDENT>?\\nstudent : Hi\\nstudent : How are you?\\nteacher : I'm good thanks - all OK with you?\\nstudent : Yes, I'm very well too\",\n  'response': 'How was your Christmas?'},\n {'id': 'test_0083',\n  'dialog': \"teacher : So you could say : oil and coal were both significant sources of electricity in 1980, contributing 20 and 25 units respectively but....\\nstudent : yes that's better\\nteacher : By 2000 their relative contribtion had fallen to...\\nteacher : So.ething like this...ok?\\nstudent : yes thanks\\nteacher : Bit more complex i know but i think you're  ready to make this jump!\\nteacher : Did you notice the verb tense in the second part?\\nstudent : passive\",\n  'response': 'Um....no! So it was past simple first: in 1980 there were....but then : by 2000....?'},\n {'id': 'test_0085',\n  'dialog': \"teacher : So two action right? Which one started first? (the seeing or the walking?)\\nstudent : walking is the first action\\nteacher : OK good fine so...we have two tenses: the present simple and the present continuous OK?\\nstudent : and during he is walking, I saw him, after I seeing him, he still walking along the street.\\nteacher : yes exactly good! So the seeing is 'inside' the walking!\\nstudent : yes,ok\",\n  'response': \"SO...your original sentence was : 'I saw this when you_______(teach) us.'\"},\n {'id': 'test_0086',\n  'dialog': 'teacher : Yes nothing special - I just mean on e.g. youtube videos or newspaper websites (where they have comment sections after opinion pieces for example)\\nstudent : Ahh now I see what you mean',\n  'response': 'I was thinking about Boris JOhnson coming out of intensive care after our previous discussion and i looked at the comments section on the youtube videos of his speech afterwards'},\n {'id': 'test_0089',\n  'dialog': \"student : okey - so just use present simple to describe the future in this sentence, right?\\nteacher : Yes that's right!\\nstudent : I finish my course at the end of this week. - or My course finishing at the end of this week.\",\n  'response': \"Yes good - remember for the present continuous you need the structure 'be + _______ing' can you do that second one again?\"},\n {'id': 'test_0090',\n  'dialog': \"student : My course is finishing at the end of this week.\\nteacher : THat's it great\\nstudent : thanks - I realised my mistakes on English grammar.\",\n  'response': 'I should say <STUDENT> at this point that I;m afraid these verb forms for the future are not always used consistently (you know what I mean? As in predictably or always in the same way) so you might see/hear English speakers mixing them up sometimes BUT I wanted to give you some general feedback'},\n {'id': 'test_0092',\n  'dialog': 'teacher : Morning <STUDENT>!\\nstudent : Good morning',\n  'response': 'How are you today?'},\n {'id': 'test_0097',\n  'dialog': \"teacher : methods = good yes\\nstudent : months...because I use 'one if..'\\nstudent : one of\\nteacher : yes exactly - perfect\\nstudent : yes..\\nteacher : <STUDENT> - I just wan to focus a bit more on verb tenses OK? Hold on...\\nstudent : If I wanted to type quickly, I also wrong on spelling.\\nstudent : yes..okey\",\n  'response': \"Just before you said: 'yes, and my class until 04 April'  - you need a verb in this phrase....can you choose one?\"},\n {'id': 'test_0101',\n  'dialog': 'teacher : But for example, the changes you made to your use of tenses last week was a big step\\nstudent : thanks for youencourage me\\nteacher : it might not feel that way but any English speaker reading/listening to your correct present perfect/past simple or going to/will examples would notice a big difference\\nteacher : OK fine!\\nstudent : thanks very much.',\n  'response': \"OK <STUDENT> - let's have a quick practice of past/future OK? Tell me 5 things you did at the weekend if you can OK?\"},\n {'id': 'test_0105',\n  'dialog': \"student : some jobs which are couldn't be replaced\\nteacher : Listen - I think we should practise the passive - we'll start more simply with it - it would be good if you could use it correctly!\\nstudent : ok, I will do more passive practice after class.\",\n  'response': \"I've copied and pasted your answers so we can continue tomorrow - I'll speak to you then OK? Yes do some more practise and I;ll ask you some (easy!) questions tomorrow ....bye!\"},\n {'id': 'test_0106',\n  'dialog': \"student : was/were + __ing\\nteacher : OK great so...'I__________ it when you ____  ______ us.'??\\nstudent : I saw it the when you were teaching us.\\nteacher : OK fantastic - you've got it\\nstudent : yes!! thanks\",\n  'response': \"So <STUDENT> let's just use this a bit before you forget it! Hold on....\"},\n {'id': 'test_0107',\n  'dialog': \"student : ok\\nteacher : I'm finished = also OK but again, less common\\nteacher : I thought that might be your next question- I was trying to be clever and pre-empt it (it's a common one!)\\nstudent : pre-empt??\\nstudent : predict\",\n  'response': \"yeah sorry =do something to prevent an anticipated event Thanks <STUDENT> - I know it's tough to think on the spot and also in this format  you feel like you should answer quickly etc well-being/want to be heard by the way I liked!\"},\n {'id': 'test_0109',\n  'dialog': 'student : ok\\nstudent : because the plans are not truth,  right?',\n  'response': \"So if you ask me aobut my plans for the weekend you say: 'HI <TEACHER>, what __________ at the weekend?' Can you complete this question in a natural way?  You need 3 words and one of them is 'doing'\"},\n {'id': 'test_0110',\n  'dialog': \"teacher : One final question <STUDENT>: do you think bullying is worse than in the past?\\nstudent : no.. I don't think so, with time goes by, people who got a high education are more than before\\nstudent : as a result, they will become teachers and parents.\\nstudent : they have the more soften way to deal with this problem.\",\n  'response': \"OK right understood... people often talk about the impact of social media in connection with this problem nowadays - do you think that's significant?\"},\n {'id': 'test_0111',\n  'dialog': \"teacher : Hi there <STUDENT>...\\nstudent : Hi <TEACHER>\\nteacher : Everything OK?\\nstudent : How are you doing today?? less mad?\\nteacher : Um..yes I think so! I'll try and be a bit more upbeat today and avoid coming across like i;m off my head\\nteacher : Which of course I'm not\\nstudent : Good plan! Is it working?\\nteacher : Oooops - yes so far - I've had a strict timetable of sane productive activities\\nstudent : I think that works very well (having an structure I mean)\",\n  'response': 'Yes..do you try to have one?'},\n {'id': 'test_0112',\n  'dialog': 'teacher : OK <STUDENT>? Use the keywords and say it correctly if you  can\\nstudent : okey, In good English I suppose - For me, I am going home by train and leaving at 9:30.',\n  'response': \"OK thanks <STUDENT> - that's good yes...present continous to talk about your personal plans (especially when you have exact times etc) hold on...\"},\n {'id': 'test_0114',\n  'dialog': \"student : in this intro, do we need the structure of ' it is argued that...'\\nstudent : we need the present continuous\\nteacher : Um....not really beacuse the lack of teachers would be given as a fact not an opinion or disagree with you see?\\nteacher : to agree or disagree with i mean\\nstudent : okey,i understand\",\n  'response': \"OK so if i say 'this year' has the time finished?\"},\n {'id': 'test_0115',\n  'dialog': 'student : important = necessary\\nteacher : ok great\\nteacher : more more!\\nstudent : problem = issus\\nteacher : fine\\nstudent : I think that all\\nteacher : yes that was good thanks\\nstudent : so.. just replace those words into question?',\n  'response': \"So we can say for example: 'Many people argue that millions _____  ______ on wildlife protection and it is _______ to spend more on other ________. I strongly with this idea and I will explain why in this essay.'\"},\n {'id': 'test_0116',\n  'dialog': \"student : they don't have enough money to buy.\\nteacher : Yes ok good\\nteacher : Or maybe: 'They can't ______ to live there.'?\\nstudent : they can not afraid to live there.\\nteacher : you mean 'afford' right?!\\nstudent : yes..afford\",\n  'response': \"ok! So if you can't afford something then thing is 'unaff_ _ _ _ _ _ '??\"},\n {'id': 'test_0119',\n  'dialog': 'student : it is a tower with parkour\\nteacher : I play Lego games with my son\\nstudent : and it is not easy\\nteacher : I see!\\nstudent : and I sometimes play brawl stars\\nteacher : oh, on your phone?\\nstudent : yes\\nteacher : Do you play Lego games?\\nstudent : no\\nteacher : <STUDENT>, quick revision. If your phone battery is low, you need to ___ ___ your phone (phrasal verb)\\nstudent : charge up\\nteacher : yes, good!\\nteacher : And to charge it up, you need to __ it ___\\nstudent : ...',\n  'response': 'connect to the source of electricity'},\n {'id': 'test_0120',\n  'dialog': 'teacher : So do we use the past simple for BOTH actions?\\nstudent : no,,I think one of them is past perfect\\nteacher : Ah ok I see your idea...well past perfect is some that happened BEFORE something else -BUT this is different because the other thing happened before, DURING and AFTER OK?\\nstudent : yes',\n  'response': 'OK <STUDENT> so imagine: I look out of my window and I see a man i have never seen before OK? Maybe I ask myself: At the moment...what ________(do)?'},\n {'id': 'test_0122',\n  'dialog': \"teacher : It could also be 'He is jealous of his sister's success' or something like that\\nstudent : Ok\\nteacher : Ok :)\\nteacher : Could you give me a preposition, please? __ January\\nstudent : Now the lizard on the cat's fur\\nstudent : In\\nteacher : The lizard's on the cat?? Is the cat ok with it?\\nteacher : In - correct!\\nstudent : The cat is slepping now\\nteacher : Ha, interesting!\\nteacher : __ the morning\\nstudent : In\\nteacher : yes\\nteacher : ___ Wednesday morning\\nstudent : On\\nteacher : correct!\\nteacher : __ school\\nstudent : At\\nteacher : yes\\nteacher : __ August 5th\\nstudent : In\",\n  'response': \"It's a particular day though\"},\n {'id': 'test_0123',\n  'dialog': \"teacher : OK I see - that sounds very valuable. Would you do that in Tenerife? Is there high unemployment there?\\nstudent : Yes, I'd like to do it here. Unemployment is quite high in all Spain, especially in the south\",\n  'response': 'Yes I thought it probably was. I lived in Spain (in Cadiz and Granada) for a few years - but nearly 20 years ago now - it was a problem then too. Too reliant on tourism etc. Do you think there is a solution?'},\n {'id': 'test_0127',\n  'dialog': \"student : Yes, I can... Aren't has 'r' sound.\\nteacher : For most British people?\\nstudent : What do you mean? Do they sound the same in theUK?\",\n  'response': 'Yes exactly they sound exactly the same for me  so e.g. both are: /ɑːnt/  sorry, do you understand the phonemic symbols?'},\n {'id': 'test_0128',\n  'dialog': \"teacher : Can you manage those in a few mintues?! see how you go....\\nstudent : No, that's too hard for 2 minutes\\nstudent : sorryyy\\nteacher : OK I'll leave them with you  for next time OK?\\nstudent : Yes, please!!\\nstudent : thank you very much again!\",\n  'response': 'OK no problem speak to you on Tuesday - have a good one!'},\n {'id': 'test_0130',\n  'dialog': 'teacher : ah ok well keep using those!\\nstudent : ok,thanks\\nteacher : what about these (from the same answer)? : education can decline income differences / some jobs which can not replace are also...\\nstudent : good\\nstudent : good?\\nstudent : I guess\\nstudent : but I am not sure',\n  'response': \"OK well ...something declines by itself do you see? But if something does it it's different so ...i.e. you can't say : the goverment has declined taxes...\"},\n {'id': 'test_0131',\n  'dialog': 'student : yes\\nteacher : What does France produce for example? Think aoout food and drink....\\nteacher : one example if you can\\nstudent : whisky is produced from France.\\nstudent : France produces whisky.',\n  'response': \"OK thanks ...so we can say (let's forget about the passive for now OK?) Scotland produces whisky or e.g. France produces wine...both good examples ...what does France use to produce wine? Milk? Oranges? no.... 'it uses____________ to make wine'.\"},\n {'id': 'test_0134',\n  'dialog': \"teacher : So we can say that the noun 'grapes' is the object of 'use' right? They go together!\\nstudent : yes..\\nteacher : France uses grapes.' = good\\nteacher : And the noun 'wine' goes with the verb 'produce' yes?\\nstudent : yes\",\n  'response': \"So you complete this sentence if you can: 'France uses ______ to _______ wine'.\"},\n {'id': 'test_0135',\n  'dialog': 'teacher : So do we need to mention the government?\\nstudent : I do not think we need to mention the government.\\nstudent : but I do not know the reason.',\n  'response': \"OK <STUDENT> thanks you're right, we don't have to....and it;s the same we don't need to mention 'people' when we talk about whisky in Scotland! Because we care more about what? The government or the money?!\"},\n {'id': 'test_0136',\n  'dialog': 'teacher : sorry: 30% of visitors were satisfied with their experience\\nstudent : 30/100 ahaha\\nteacher : technically correct:)\\nteacher : how about a fraction?\\nstudent : But not for ielts\\nstudent : 44107.0\\nteacher : nooo..\\nteacher : can you write it in words\\nstudent : Third of ten\\nstudent : Or three tenth',\n  'response': 'ok so best for Ielts to use fairly basic fractions e.g. half/quarter/third./fifth......and then add little phrases to make it more specific e.g. 52% = _______  ________ half?'},\n {'id': 'test_0137',\n  'dialog': 'student : The world is change\\nstudent : You would make mistake\\nstudent : If your boss would ask you',\n  'response': \"OK <STUDENT> so...maybe the chane happened in the past right? BUT you're talking about the present result so .you need...the present perfect! In other words: the world ______  changed?\"},\n {'id': 'test_0138',\n  'dialog': \"student : seldom?\\nteacher : good idea! So in order of formality: seldom, rarely and hardly ever\\nteacher : you've heard that before?\\nstudent : So you mean hardly is the most casual? I know these words, but I've never thought about the order of formality.\",\n  'response': \"yes ' hardly ever' (two words together) is by far the most common in normal speaking. Just to clarify: 'seldom' = correct but very formal i.e. many English people will never have said it in their lives!\"},\n {'id': 'test_0139',\n  'dialog': \"teacher : addicted' = adjective\\nstudent : because I was addicted to them\\nteacher : ok good! So addiction is a possible problem - anything else?\\nstudent : children trend to copy bad behaviours from games.\\nteacher : OK thanks - quick example so I get the idea?\\nstudent : because they don't have enough ideas to recognise easily which behaviours are good or bad.\\nteacher : OK understood - 'bad behaviour'...for example?\\nstudent : they don't respect their parents or teachers possibly.\",\n  'response': \"ok...one more negative effect if you can (don't worry if not!)\"},\n {'id': 'test_0140',\n  'dialog': \"teacher : And NO you shouldn't ..not like this anyway...these two sentences are for the introduction\\nteacher : and they make a beautiful intro for our original essay question!\\nstudent : yes!! that was I meant\\nteacher : OK!\\nstudent : I got it,thanks...\",\n  'response': \"So <STUDENT> can you follow this model (the two sentences) for the following question...wait a minute and I'll send it (remember keep it short and simple i suggest at the moment -just focus on the model structure ok?\"},\n {'id': 'test_0141',\n  'dialog': \"teacher : and also make the main verb the infinitive ...\\nteacher : for example....hold on\\nteacher : Not only does he love chocolate...\\nteacher : i.e. he + love NOT loves\\nstudent : ok\\nstudent : I didn't know we need auxiliary in affirmative sentences\\nteacher : One last example: He only realised later that ... = only later did her realise that...\\nteacher : Yes not normally - but inversion is a special case\\nstudent : I see\",\n  'response': 'although <STUDENT> you know of course that we sometimes say e.g. I do like chocolate right?'},\n {'id': 'test_0142',\n  'dialog': 'student : yes,,as result of..............\\nstudent : as a result of, young people really easy to copy the violence in some ways, such as teas classmates.',\n  'response': \"OK thanks <STUDENT> - have you finished? Here is your paragraph - I'll give you a minute to see if you want to change anything OK? Don't worry if you can't think of anything and try not to make really big changes OK?1\"},\n {'id': 'test_0143',\n  'dialog': \"teacher : So you need to say: Some newspapers want to attract readers print fake news\\nteacher : sorry '....and print fake news\\nstudent : yes, it became more native, I need to practice it.\\nstudent : or more academic\",\n  'response': \"next one: what about'as quick as' - you just need to change one of the words slightly\"},\n {'id': 'test_0144',\n  'dialog': \"teacher : ok\\nteacher : so 'innate' = something your born with e.g. Children seem to have an innate sense of fairness. And e.g. patience, honesty, a sense of humour  are all examples of particular personal qualities or characteristics (OK?!)\\nteacher : So 7 = i and 9 = e !\\nstudent : okey~I write them on notebook\",\n  'response': 'OK great - OK can I give you one final test now - you will need one word from the second group to complete the following sentences'},\n {'id': 'test_0145',\n  'dialog': \"teacher : yes good  -well done\\nstudent : thanks\\nteacher : OK thanks <STUDENT> - so you're on the right track and you know some good words ...you just need to get things a bit more correct overall (and keep learning new words that's true but being correct with what you have is important so you can explain your ideas clearly!)\\nstudent : yes, that is what I need to study really importantly.\",\n  'response': \"I guess we'll stop there - thanks a lot and I look forward to speaking to you tomorrow! bye\"},\n {'id': 'test_0146',\n  'dialog': \"student : Okey\\nteacher : Just try and answer as naturally as you can ... as if you were speaking i guess! OK....here goes....\\nstudent : Okey\\nteacher : OK great so...the topic is 'Work' and here comes the first question...\\nteacher : Is job satisfaction more important than salary when choosing a job?\\nstudent : Personally,\\nstudent : I think job satisfication isn’t more important than salary\",\n  'response': \"OK...so you're going to give me some more right? Like a reason/example etc....?\"},\n {'id': 'test_0147',\n  'dialog': \"teacher : I'm fine thanks yeah - trying to keep fit!\\nstudent : yes, i am fine\\nstudent : and you?\\nteacher : good...are you still doing much indoor physical activity?!\\nstudent : yeah, I am insist to do some indoor exercise .\\nteacher : OK...'I am insist'? ..you mean like 'continue'?\\nstudent : yes, I continue to do some indoor exercises.\",\n  'response': \"Ah right OK...'still' is best here (most common) so 'I'm still _______ some indoor exercise.'\"},\n {'id': 'test_0148',\n  'dialog': \"teacher : you OK <STUDENT>?\\nteacher : Let me know if you' re not sure! Or maybe internet problems....?\\nstudent : sorry....\\nstudent : my internet problems occur agnin.\\nstudent : again\\nstudent : so..if I willing to use 'effects' I can write like that  it is difficult to assess the positive and negative effects of renting a house.\\nteacher : Yes perfect!\\nstudent : first gap I think is paying\",\n  'response': 'Ah hold on ...so this is a word from yesterday that means buildings/houses/flats OK?'},\n {'id': 'test_0149',\n  'dialog': \"student : so...Natural gas accounted for 20 % of electricity in Australia in 1980.\\nstudent : ok..thanks\\nstudent : I got it\\nteacher : Yes that's great - that would look good in an answer\\nteacher : ok fine\\nteacher : So....that's one way to explain that data point....hold on\\nstudent : it make me easy to describe the data.\",\n  'response': \"Yes that's right.....Another way for this chart could be: 20% of electricity _________  produced _____ natural gas. (Any ideas?\"},\n {'id': 'test_0151',\n  'dialog': \"student : account\\nstudent : coal accounted for 50 percent of electricity\\nteacher : yes that's it good well done - 'electricity' or 'electricity production' both possible actually so fine\\nstudent : okey,\",\n  'response': \"ok one more: try number 2...let me give you a little bot of help....so the first word is 'Coal' and the last 4  words are: ....producing 50% of its electricity'\"},\n {'id': 'test_0153',\n  'dialog': \"student : 5 f\\nteacher : sure\\nstudent : 6g\\nteacher : yes\\nstudent : no idea about 7 and 8\\nstudent : I don't know the meaning of the words in number 7 :(\",\n  'response': 'well, you know that ships had barrels with e.g. meat stored in them - so if there was only a bit left they had no choice but to scrape out the last horrible bit you see'},\n {'id': 'test_0154',\n  'dialog': \"student : that's okay\\nteacher : The words are debunk, deduce - deduction, exploration, abandon and evidence\\nteacher : Could you fill in the gaps, please?\\nstudent : okay\\nteacher : To ____ what happened that night, the police have to look at the ____: a fresh scratch on the car, a broken lock, some footprints on the carpet.\\nstudent : abandon and deduction ?\\nteacher : Mmm, not quite\\nstudent : both are not okay, or the one?\",\n  'response': 'first gap: understand something from looking at the facts'},\n {'id': 'test_0155',\n  'dialog': \"teacher : OK thanks a lot <STUDENT> - give me a second...\\nteacher : The government should spend more money on public services, such as schools and hospitals. ' - this is your idea right? (one of them)\\nstudent : yes\",\n  'response': \"OK fine..so do we need to say ' the government'? Do we need to include the 'agent' here - the person/thing that does the action? Is it essential?\"},\n {'id': 'test_0156',\n  'dialog': \"teacher : It's good to learn words together right?\\nstudent : thanks,, and it is necessary do identify the words is adjective or noun.\\nstudent : it is necessary to identify the words is a adjective or a noun.\",\n  'response': \"I mean: try to remember the phrase (for example) 'positive effect' as two words together  so you don't have to always think too much (but sometimes you have to when you;re checking or learning it's true..)\"},\n {'id': 'test_0157',\n  'dialog': \"teacher : No problem):\\nstudent : InAustralia in 1980, 20 units of electricity was produced from natural gas  as same as from hydro power\\nstudent : As same proportion as\\nteacher : Ok thanks <STUDENT> hold on...\\nteacher : Yes ok pretty good....you could use your structure you just need to say: ...from natural gas, the same as....\\nteacher : In other words add a comma and then 'the' ok?\\nteacher : But yes good and good to put similar things together in ielts task 1 right?\\nstudent : Yes\",\n  'response': 'Ok thanks <STUDENT> speak to you soon!'},\n {'id': 'test_0158',\n  'dialog': \"teacher : I don't know what I should do.\\nstudent : What sould I do?\\nstudent : Sorry should\\nteacher : That's right\\nteacher : I don't know whether lizards like plants or bugs.\\nstudent : Lizards whether like plants or bugs\\nstudent : ?\\nteacher : Not quite. Are you confused about 'whether'? It's the same as 'if'. I don't know if lizards like plants or bugs.\\nteacher : What's my question here?\\nstudent : Are lizards like plants or bugs?\",\n  'response': '___ lizards like plants or bugs? like is a verb, so you need ...?'},\n {'id': 'test_0160',\n  'dialog': \"student : so.. there are economic potential to space exploration.\\nteacher : Yes  - there is (singular) economic potential in space exploration = possible\\nstudent : ok, thanks\\nteacher : OK your idea before was 'a lack of resources' - let's try that now OK? hold on a sec\\nstudent : ok\",\n  'response': \"Any ideas? Natural resources on Earth are ______- very quickly.  OR We are _____ Earths natural resources very quickly. '\"},\n {'id': 'test_0164',\n  'dialog': \"student : this morning is...past simple I guess\\nteacher : Yes good!\\nstudent : because everyday and every time is specific.\\nteacher : Yes that's right - OK so finally: 'in recent decades'? Past simple or present perfect? Remember the logic of the essay question is : We have a problem NOW with a lack of teachers.\\nstudent : let me think\\nstudent : past simple!\",\n  'response': \"OK so remember the result will be either: 'In recent decades the number of teachers has gone down...' OR 'In recent decades the number of teachers went down....\"},\n {'id': 'test_0165',\n  'dialog': \"teacher : I guess the people swearing don't exactly feel it's rude  you know? i.e. it's not meant in a personally aggressive/offensive way\\nstudent : Sure, that's the saddest thing\\nteacher : Yes absolutely! POsh people can be REALLY horrible when they're complaining to service workers!\\nstudent : I mean when someone is rude and they don't even know they are...\\nteacher : yes  sure -that's just upbringing i guess\\nstudent : yes I think so\",\n  'response': 'OK thanks <STUDENT> - speak to you tomorrow!'},\n {'id': 'test_0166',\n  'dialog': 'teacher : Hi there...\\nstudent : hi\\nstudent : I am here.',\n  'response': \"OK <STUDENT> so let' s imagine you're written your intro for this essay already ok? LEt's start with a main paragraph...which one? Addiction? OK?\"},\n {'id': 'test_0167',\n  'dialog': \"teacher : Ah i see sorry\\nteacher : Yes no problem\\nstudent : Sorry, did you mean change produce to another word?\\nstudent : Give?\\nteacher : Yes ....dont worry...so make electricity is possible but generate is a good one!\\nstudent : Ah... I forgot this word, thank you\\nteacher : Sometimes if you cant think of another word its ok just repeat one you're co fident with ....your original produce was fine\\nteacher : Confident\\nstudent : Ah I see\",\n  'response': 'Ok so another we can change is the number i.e. we can use fractions...so 50 units in this case is...?'},\n {'id': 'test_0170',\n  'dialog': \"student : I'm lazy and selfish.haha\\nteacher : Really?! Maybe most of us are sometimes....any positive ones?!\\nstudent : Ah, I'm sometimes said I'm more mature than my actual age.\\nteacher : OK, you mean people sometimes say you're mature?\\nstudent : Yeah.\",\n  'response': \"OK right - that's good I guess! You can say: people say....OR it's sometimes said (that) I'm....OR 'I'm said to be mature' - the last two are more formal...\"},\n {'id': 'test_0171',\n  'dialog': \"teacher : Can you add a subject/verb in  'a lot time wasting children's life' to make it correct?\\nstudent : spending a lot of time is wasting children's life\\nstudent : children are wasted a lot of time on games.\",\n  'response': \"OK thanks <STUDENT> - So 1. Spending a lot of time playing them is a waste of children's life = possible good. And 2. do you need the passive? IF children is the subject?\"},\n {'id': 'test_0172',\n  'dialog': \"teacher : yes you need a few home comforts I guess\\nteacher : <STUDENT> - are you still planning to do CAE soon?\\nstudent : I do need some comforts, I guess we are all different about these things. Don't you think?\\nteacher : sure -to be honest i'm pretty minimalist actually - I genuinely like having a very limited set of possessions\\nteacher : not because I'm virtuous just because I've always been like that\\nstudent : About CAE, my idea is to take it in April\",\n  'response': 'OK great - are you planning to do much preparation for it?'},\n {'id': 'test_0173',\n  'dialog': \"student : if I said, I haven't checked the data in the morning, Is it correct?\\nteacher : Fill in the gap with past simple or present perfect!\\nteacher : Oh sorry ...No because the morning is finished\\nstudent : I have lost my galsses\",\n  'response': 'So here you do have a specific time you see? Not eaxctly like 11:17 a.m. it;s true but a definite finished period you see?'},\n {'id': 'test_0174',\n  'dialog': 'student : ok..\\nstudent : in 1980, nuclear power accounted for 15 out of 90 units of electricity,However, in 2000, it was by far the most signifiant fuel, producing 126 out of 180 units.\\nstudent : I think my type speed have become more quickly than a few days ago.\\nstudent : have becoming',\n  'response': \"Ooohh <STUDENT>! That's a great sentence! I'm very happy with that!   Perfect...A great ending...I'll see you in 15 minutes OK?\"},\n {'id': 'test_0175',\n  'dialog': \"teacher : Oh sorry...hold on!\\nteacher : Just the following !! ...Thanks <STUDENT> – good point and good passive (‘are taught’) and good to use ‘effective’!\\nteacher : and good phrase:'the future of society''\\nstudent : Thank you;)\\nstudent : Okey\",\n  'response': \"OK thanks we'll stop there - sorry if you had technical problems I hope it was still useful ...see you tomorrow\"},\n {'id': 'test_0176',\n  'dialog': 'teacher : Not the numbers just this idea in words\\nstudent : Coal was by far the most popular fuel, with 130 out of 170 units were produced.\\nteacher : THanks <STUDENT> - a very good sentence overall...hold on\\nstudent : ok',\n  'response': \"So if we make 2 basic correct sentences we could get: Coal was by far the most significant (better than 'popular' here because people aren't actually choosing it like Coca cola you see?) fuel. It produced 130 out of 170 units. OK?\"},\n {'id': 'test_0178',\n  'dialog': \"student : trust me\\nteacher : So...'Coal accounted for 50% of electricity production in Australia in 1980.'\\nteacher : yes I  know what you mean about the Italian!\\nteacher : You can also the verb 'make up'  - hold on\\nteacher : in your morning class how many students are there more or less?\\nstudent : 7 maybe\\nteacher : ok and how many male students?\\nstudent : 5.0\",\n  'response': \"OK so can you tell me that with the verb  'account for' and the word 'majority' for the proportion??\"},\n {'id': 'test_0179',\n  'dialog': \"student : Maybe we could talk about that another day, sorry (mixing topics)\\nteacher : yes! well most are 'separable' so e.g. work out something OR work something out = both correct\\nstudent : About a prompt, yes please\\nstudent : Ahhh, I thought there was a rule about it!!\\nteacher : ok so e.g.: I haven't had so much fun since I was in Spain.\\nteacher : Can you invert that?\\nstudent : Little fun have I had since I was in Spain??\",\n  'response': \"Yes I see your idea...Start with 'Not since...'\"},\n {'id': 'test_0180',\n  'dialog': 'teacher : Welcome back <STUDENT>...\\nstudent : hey, i am here',\n  'response': 'OK a quick vocabulary exercise for you based on earlier ... hold on'},\n {'id': 'test_0181',\n  'dialog': \"student : okey..when I am not sure the noun is countable or not, I alway add's'\\nteacher : Yes I see...well actually the word 'data' is very confusing even for English people hold on...\\nstudent : in normal way...I think is that I didn't look the data.\",\n  'response': \"Well....do you have a specific time in mind about when you didn't check it? OR do you just really mean that you don't know the data NOW??\"},\n {'id': 'test_0182',\n  'dialog': \"student : or  'I still needing to'\\nstudent : yes ,,  maybe just a few people, I didn't  check the datas\\nstudent : so..after 14 days, I need to do that test again. after the results are all negative, I will be freedom\",\n  'response': 'Sorry internet went down and I lost my reply! So I;m sorry to hear that - another two weeks in a hotel... I still need = correct (need not really used in continuous form)'},\n {'id': 'test_0185',\n  'dialog': 'teacher : Rememeber after a modal verb (can/might/shall etc) you need the infinitive\\nstudent : public services are spent money, such as schools and hospitals?\\nstudent : I am not sure.',\n  'response': \"OK <STUDENT> - you've gone a bit crazy there! Hold on ...come back! So, you can keep money as the first word...so 'Money should______ spent on schools and hospitals' What's the missing word?\"},\n {'id': 'test_0186',\n  'dialog': \"teacher : Don;t worry if not!\\nstudent : yes! the international space station.\\nstudent : build from NASA\\nteacher : OK great example....\\nteacher : So... 'The International Space Station ____  b _ _ _  _ by NASA.'?\\nstudent : The International Space Station is build by NASA.\",\n  'response': \"Hold on <STUDENT>: one correct way of saying this = 'NASA built the ISS' the passive way = 'The International Space Station ____ b _ _ _ _ by NASA.'?\"},\n {'id': 'test_0187',\n  'dialog': \"student : as a consequence, young people don't have time to do some useful things for their life, such as being social and  reading.\\nteacher : OK great .... a clear reason about why this is bad ...hold on a sec....\\nstudent : ok\",\n  'response': \"So that's 67 words - if you had three main paragraphs that could be enough....imagine you've got  2....can you add a final sentence?!\"},\n {'id': 'test_0190',\n  'dialog': 'student : The world change\\nstudent : Everybody can contact another easily\\nstudent : But if you don’t have communicatuon skill\\nstudent : You can’t understand each other\\nstudent : You’ll make mistake for your work\\nstudent : Forexample\\nstudent : If your boss ask you to improve a sugar process but you don’t understand and you think your boss allow you to change the process as you like\\nstudent : So it lead to a big problem to your company',\n  'response': \"OK thanks <STUDENT> - you've developed your point there, which is the right thing to try to do - well done....hold on a sec...\"},\n {'id': 'test_0191',\n  'dialog': \"student : It's teen but you need to see gameplay\\nstudent : If you want , I can send one\\nteacher : I'm sure I can find it, thank you! It's reassuring that there is no blood\\nteacher : So what's the fun fact about you? :)\\nstudent : Have you ever played star wars lego?\\nteacher : Yes! We like it a lot\\nstudent : Fortnite very near it\\nteacher : Oh is it?\\nstudent : But heroes aren't lego:)\",\n  'response': \"Aha, you mean 'characters'?\"},\n {'id': 'test_0192',\n  'dialog': \"student : so..with names and places?\\nteacher : Yes perfect wel ldone ...so we need 'the' with rivers/deserts/oceans!\\nstudent : or something is specific\\nteacher : OK so....Remember my Heathrow example? Was there an article there?\\nstudent : okey,i understand\\nstudent : no..\\nstudent : The Heathrow is an airport in London.\",\n  'response': \"No! But Heathrow is a specific place of course - so it depends on the place I'm afraid. So <STUDENT> your example there ('The Heathrow....') is wrong...\"},\n {'id': 'test_0193',\n  'dialog': \"teacher : No ..I don't see the logic!\\nstudent : no idea about suggest and install!!\\nstudent : sorry\",\n  'response': \"No that's ok! As in: ' Put ______ an idea/suggestion' (put in or put forward..what do you think?) and Put ____ a new alarm system (or whatever) (put in or put on..what do you think? )\"},\n {'id': 'test_0196',\n  'dialog': \"student : decline\\nstudent : decrease\\nteacher : So you copy the verb 'go down' - it's fine! But what tense??\\nteacher : you can copy the verb\\nstudent : declined/decreased/went down\",\n  'response': \"OK thanks so you;ve chosen the past simple which maybe seems logical BUT the time phrase 'in recent decades' means UP TO AND INCLUDING NOW so....we need the...??\"},\n {'id': 'test_0198',\n  'dialog': \"student : I understand...thanks\\nteacher : OK no problem..so ready to try the next sentence?\\nstudent : yes, I am reday\\nteacher : ok go for it....i'll wait....\\nstudent : it is difficult for children to realise between games word and reality.\\nstudent : ,because some games is for adults to forget pressure, and sometime it may have many violent contexts.\",\n  'response': \"OK thanks <STUDENT>...I'll come back to the language here later OK? So next you're going to say that children copy the violence in some way?\"},\n {'id': 'test_0201',\n  'dialog': \"teacher : Wow - so if you live in a flat (do you?) you basically can't be outside at all really?\\nstudent : Here there are soooo many jokes about everything... It's great because at least we laugh quite often...\",\n  'response': \"Yes you've got to laugh or...(you know the ending right?)\"},\n {'id': 'test_0202',\n  'dialog': 'student : Do you know what I am talking about? When you speak and your phone writes what you say...\\nteacher : Yes yes I know\\nstudent : Nooo, in English, I change the language when I want to do it in English',\n  'response': \"Hey <STUDENT> did you notice my Spanish intefernce just not - seriously, when I put 'no' at the end of my sentence - that's not normal - it something i  retained from speaking Spanish years ago...\"},\n {'id': 'test_0203',\n  'dialog': \"teacher : So hold on....\\nteacher : So when you talk about university in the uK it's your PLAN right?\\nstudent : yes\\nteacher : So do we use 'will'' for plans?!\\nteacher : E.g. I will go to to London at the weekend' = correct?\\nstudent : yes...I think it is corretc\",\n  'response': \"OK so <STUDENT> - we use 'will' for: PRomises/offers/decision we make right now this second and ...predictions BUT NOT plans OK?\"},\n {'id': 'test_0205',\n  'dialog': 'teacher : Hello <STUDENT>, is that you?!\\nstudent : Yes, I’m here :)',\n  'response': \"OK hi again! So <STUDENT> as we said I'm going to give you some Part 3 style speaking quesrtions...\"},\n {'id': 'test_0206',\n  'dialog': \"student : the international space station is built by NASA.\\nteacher : Present?\\nteacher : good 'built' though!\\nstudent : I suppose NASA still expanding the International space station.\\nstudent : so.. 'is' here is correct, right?\\nteacher : Ah ok fine! Well I guess you could say then: 'NASA is building the International...'\\nteacher : But hold on - isn't NASA American?!\\nstudent : yes... but also have other county's spacemen.\",\n  'response': \"OK fine! 'spacemen' = 'astro_ _ _ _ _ '?\"},\n {'id': 'test_0207',\n  'dialog': \"teacher : Ah ok I'm with you! Yes of course (just checking though because you'll have to explain this if you use it as a main idea right?) Quick question: 'it is very unconvenience ' What kind of word do you need after 'very'? (noun/verb/adjective?\\nstudent : adjective I think..\\nteacher : OK good yes! Can you change your word to an adjective?\\nteacher : incon_ _ _ _ _ _ _ '\\nstudent : very inconvenience\",\n  'response': \"Hold on <STUDENT>....'importance' = adjective or noun?\"},\n {'id': 'test_0208',\n  'dialog': \"student : yes,, I see it... I need to delete the 'to'.\\nstudent : maybe not...\\nteacher : Hold on...! You've used the noun 'effects' right?\\nstudent : I suppose it might be ' the'\\nstudent : yes\",\n  'response': \"So the word(s) before 'effects' should be what? Adjectives/nouns/verbs? e.g. positive effect\"},\n {'id': 'test_0209',\n  'dialog': \"student : yes..maybe this is the reason\\nteacher : OK so <STUDENT> - if we know exactly when something happened do we use the present perfect e.g. 'Yesterday I have watched a film.' Correct?\\nstudent : no...I think just use past tense is ok\",\n  'response': \"OK good - so if we have a specific past time then we need the past simple....So for example: ' Last year I ________ to France.'\"},\n {'id': 'test_0210',\n  'dialog': \"teacher : OK thanks a lot <STUDENT> -  so let's have a think about your answers ...hold on...\\nstudent : ok..I always have any grammar mistakes...\",\n  'response': 'OK <STUDENT> some good language you used (well done!): fake news/attract readers/build/willing/get news/on social media/channels/set/culture/and so on/copy their behaviour ….(keep using this stuff!)'},\n {'id': 'test_0211',\n  'dialog': \"student : That cannot happen in Tokyo, unfortunately.\\nteacher : That's interesting...I guess Tokyo is a very big city and maybe more impersonal - or is that generally true in japan do you think?\\nstudent : I think Japanese(including me) are very shy to talk to someone who they don't know.\",\n  'response': \"Yes, I suppose that's been my impression from teaching - why do you think that is (a big question I know!)\"},\n {'id': 'test_0213',\n  'dialog': \"student : there is a chance to work??\\nteacher : yes good! or opportunity of course = both good. possibility is more like a statistical thing you know\\nteacher : vacancy + preposition?\\nstudent : No, I don't understand 'statistical thing'\\nteacher : OK so for example: there's a possibility it will rain\\nstudent : Ahh I understand now\\nstudent : vacancies for\\nstudent : vacancies for working in\",\n  'response': 'basically for a Spanish speaker you normally need to say chance/opportunity when you want to say possibility! YEs good vacancy for - change the verb in the last one'},\n {'id': 'test_0214',\n  'dialog': \"student : expenditure of public money of search for alien life.\\nteacher : Sorry..internet...! OK hold on..I think the government should + NOUN right? So you can't have 'expenditure' next here ...\\nteacher : Can you solve this problem here?\\nstudent : so could I use 'of ' to link two nouns.\\nstudent : no.....I do not think so\\nteacher : Um...what about the should'?\\nteacher : should'\\nstudent : yes! I understood the regulation\",\n  'response': 'How about: One reason why I think government expenditure on space exploration is justified is because...'},\n {'id': 'test_0216',\n  'dialog': \"student : I'll try to correct it... My team outscored yours last night\\nstudent : Before your examples, I wrote one\\nteacher : Oh sorry I see it now!\\nteacher : Yes you just need to say 'outscore' because this already includes the idea of 'points'\\nteacher : this is why it's quite efficent you see!\\nteacher : German!\\nstudent : Now I got it... I think\\nstudent : haha yes\",\n  'response': 'OK one more ...so Apple has sold more devices than anyone else = Apple has hugely  _______ other tech companies'},\n {'id': 'test_0217',\n  'dialog': \"student : The light cinema is near to my house.\\nteacher : Ah you live quite centrally - I'm jealous! (sort of) And e.g. King's College/Barclay's (bank)/St Mary's Church do NOT have 'the' - can you see what they have in common?!\\nstudent : It's owned by someone like King, St Mary and Barclay??\",\n  'response': \"Yeah that's it basically! so the 'someone' is the important bit - a name! Can you give me another example of a place like this?\"},\n {'id': 'test_0218',\n  'dialog': \"teacher : The games have _______ content which has a bad __________ on young people.'\\nstudent : the games have negative content which has a bad influence on young people.\",\n  'response': \"Yes very good, especially 'influence' - negative is possible .... how about: 'uns______' or 'inapp_________'?\"},\n {'id': 'test_0219',\n  'dialog': \"teacher : simple\\nstudent : okey, I am ready\\nteacher : So .... It's your birthday and I come in the room and I say: 'Happy birthday <STUDENT>, I made/I have made you a cake!'\\nteacher : What do you think?\\nstudent : I have made you a cake\\nstudent : nono...i made you a cake\\nstudent : because make is finished\",\n  'response': 'Aahhhh! hold on....well in both cases the cake is finished/made BUT is the result NOW important? OR...is it reporting the action in the past?'},\n {'id': 'test_0220',\n  'dialog': 'student : The Samsung tablet (is it correct) is cheaper than the ipad\\nstudent : The ipad is considerable or much more expensive than the tablet\\nstudent : considerably\\nteacher : OK thanks so : The Samsung tablet (is it correct - yes!) ) is cheaper than the ipad  : The ipad is considerable= considerablY / much more expensive than the ipad\\nstudent : Yes',\n  'response': 'Considerably = correct of course ...a little bit formal - other formal options = signifcantly etc less formal options might include.....??'},\n {'id': 'test_0221',\n  'dialog': \"teacher : So the first verb is in the past simple good ...and the second needs to be the....??\\nteacher : Yes yes that's it!\\nstudent : wow..I got it\\nstudent : I never understand it before you teaching me this stuff.\",\n  'response': \"Yes that's ok ...So you need the PAST CONTINUOUS - how do we make it? We need to two parts remember? the PRESENT continuous = is/are + ____ing and the PAST continuous?\"},\n {'id': 'test_0222',\n  'dialog': \"teacher : Oh no! No, go ahead tell me...\\nstudent : I have a tendinitis (is it right) in my right shoulder. It is very painful... I've had it for over 6 months I think and it just gets worse...\",\n  'response': \"Oh I'm sorry that's bad..(tendonitis with an 'o' because of the word 'tendon' you know?) - is it because of an activity you do?\"},\n {'id': 'test_0226',\n  'dialog': \"student : What do you mean by 'a difference in your mind'? Are you asking me if I know how to use them all??\\nteacher : which one is more formal of the three?\\nstudent : I don't know!!\",\n  'response': \"also' is more formal - to the extent that it's a little bit weird to hear: 'I also...' in everyday sepaking.\"},\n {'id': 'test_0227',\n  'dialog': \"student : No, I heard on Spanish news that he is 'en planta'\\nstudent : which means that he is in a normar room in a hospital\\nstudent : normal\\nteacher : ok i see\\nteacher : i wonder whether he'll be more popular as a result of having been ill?\\nteacher : early days i guess...(too soon to tell  i mean)\\nstudent : haha maybe\",\n  'response': \"OK thanks <STUDENT> ... so i'll speak to you on Tuesday at 5 right? Have a good one!\"},\n {'id': 'test_0228',\n  'dialog': \"student : extroverted→shy, cruel→kind, generous→tight-fisted\\nteacher : Yes you're right! we'll come back to those OK/\\nstudent : ambitious,naive,pushy,ambitious,shy\",\n  'response': \"Thanks <STUDENT> - some good answers. Ambitious is fine - what about if you are prepared to work hard and are not going to give up? And...these peopel from e.g. Germany/Norway etc seem rude so a word like that  rather than 'shy'\"},\n {'id': 'test_0229',\n  'dialog': \"teacher : it varies'\\nstudent : yess\\nteacher : ok great - would you like to have a go at page 14?\\nstudent : sure\\nteacher : ok\\nstudent : I did this exercise not too long ago and I kind of remember the answers\\nstudent : Shall we go to the last test?\\nteacher : yes ok\\nstudent : page 149\\nteacher : page 149?\\nteacher : yes!\\nstudent : 17 inconvenients?\",\n  'response': \"ok so you know those shops that are open 24 hours a day...? they're called _______ stores?\"},\n {'id': 'test_0230',\n  'dialog': \"teacher : Hi <ANOTHER STUDENT>!\\nstudent : Hello\\nstudent : Maybe if you have time on weekend\\nteacher : <STUDENT> and I are just discussing September, sorry, we ran a bit late\\nteacher : I don't think I'll have the time at the weekend, sorry\\nstudent : no problem\\nstudent : Ok I'll think bye!\",\n  'response': \"Anyway, you have a think, and I'll speak to you Wednesday 9.30, <STUDENT>, ok?\"},\n {'id': 'test_0231',\n  'dialog': \"teacher : Not Australia but...?\\nstudent : Ah\\nstudent : Electricity\\nstudent : Was produced\\nteacher : Yes good...basically! But how much?\\nstudent : Ahh\\nstudent : 50 units of electricity was produced from coal\\nteacher : Yes great well done!\\nstudent : Yeh😂\\nteacher : So to get a good score you include different  grammar like you've  just done and maybe change the...?\\nteacher : One word answer?\\nteacher : Beginning with letter v!\\nteacher : The letter\\nstudent : V what\\nstudent : Value?\\nteacher : Vocabulary!\\nstudent : Ah5555 ok\\nstudent : Thank you\",\n  'response': \"For example the verbs e.g. in this chart you can use 'produce' electronic...can you think of another possinverb\"},\n {'id': 'test_0234',\n  'dialog': \"teacher : Yes ...all very common here too ..i got sent (a stupid one) one today revealing 'zoom' as the hidden baddie behind the virus (like in SCooby Doo) which made me chuckle for a second\\nstudent : that's interesting, I thought it was a Spanish thing\\nstudent : I said that I think it must hurt but I laugh a lot!\\nteacher : OK! I feel a bit less bad now...\\nstudent : hahaha\",\n  'response': 'Hey <STUDENT> can i give you a quick phrasal verb test from yesterday?'},\n {'id': 'test_0235',\n  'dialog': 'student : ok - I am very happy of the whole trip ( because the feeling is happen now ) - we were not obey the plan that we written before.\\nstudent : because the behaviour of write plan was happened before obey the plan, right?',\n  'response': \"OK thanks <STUDENT> - well not impossible I guess your first idea (I am happy) BUT I was happy was good! What about 'we were not obey'?? What's the main verb here?\"},\n {'id': 'test_0236',\n  'dialog': \"student : I was confused when I heard about my grandmother's death.\\nstudent : I didn't use to be optimistic when I was a student.\",\n  'response': \"Yeah - and remember it has a very strong naval tradition - that's what kept the BRitish Empire going for so long! I'll get some examples for our next lesson of common phrases derived from this. OK thanks for your confused example - sorry if it's a sad subject ...but this suggests you didn't understand something....is that true for your example?\"},\n {'id': 'test_0239',\n  'dialog': 'student : Last year I went ti Frace\\nstudent : went to',\n  'response': \"OK thanks correct! BUT ...if we don't know or care about exactly when it's different in English. So that's why if I ask you the general (without a specific time)  question I ask: '_______ you ever _____ to France?' ?\"},\n {'id': 'test_0240',\n  'dialog': 'teacher : Hi <STUDENT> are you there?\\nstudent : Hi <TEACHER>\\nstudent : Are you alright?',\n  'response': \"Hello - they've changed the settings for the timeout but it's still not working well enough obviously - I''ll follow this up again!  yes I'm fine thanks and you?\"},\n {'id': 'test_0241',\n  'dialog': \"student : okey,thanks\\nteacher : And the answer for example = 'Yes I _____________ there twice'\\nstudent : Yes I have been to there twice.\\nteacher : OK great, correct! Did i say exactly when?\\nstudent : NO, just this behaviour was happened and we didn't know the time.\",\n  'response': \"OK good - so ...'I ___________ there when I was a student.' ??\"},\n {'id': 'test_0243',\n  'dialog': \"student : languages\\nstudent : yes I see\\nteacher : yes good example!\\nteacher : give me a sentence with that one if you like\\nstudent : which one\\nteacher : languages...\\nteacher : cos it was your idea!\\nstudent : it's also important to develop languages skills to improve the communication with other agencies and colleagues\\nteacher : ok thanks! 'colleagues'/'develop skills' 'important to' = all good!\\nteacher : not 'the' though\\nstudent : yes sorry\",\n  'response': \"That's OK! we'll call it a day I guess  thanks  <STUDENT> see you tomorrow....\"},\n {'id': 'test_0244',\n  'dialog': \"teacher : That's OK - important thing is it's NOT past simple\\nteacher : OK <STUDENT> - shall we try a completely new topic to just get going and forget about these problems\\nstudent : yes.. I am ready.\",\n  'response': \"OK ... I've just made some material on the topic of space (spending on space exploration ) ... why don't you tell me some words that you think might be useful to include in an essay on this topic? Give me three to start if you can ok?\"},\n {'id': 'test_0245',\n  'dialog': 'student : it is good to try to do regular exercise\\nteacher : Yes both exercise and exercises means phyiscal activity (ah I see what you mean yes - exercises can be for school or whatever yes)\\nstudent : okay, I understand .',\n  'response': \"So in the singular (exercise) means in general  yes good - and exercises then? (when you're talking about physical activity?)\"},\n {'id': 'test_0246',\n  'dialog': \"student : so the answer is virtually?\\nteacher : yes!\\nstudent : ahhh\\nstudent : 6d\\nteacher : yep!\\nstudent : 77b\\nteacher : well done!\\nstudent : 8c\\nteacher : OK great  - a clear pass at CAE!\\nstudent : haha\\nstudent : The beginning was terrible!!\\nteacher : Well I forgive you the 'ranks' one - but not the 'overlook'! :)\\nstudent : oh thank you!\",\n  'response': \"OK <STUDENT> thanks for today - I'll be in touch later next month OK?\"},\n {'id': 'test_0248',\n  'dialog': \"student : No, I don't know 'rake it in'. It's the first time I see it... I think\\nstudent : accumulate says my dictionary\",\n  'response': \"the first time I'VE seen it' - well a rake is for gardening - a pole with metal end with kind of metal spikes that you use to sweep up  e.g. leaves/weeds etc (but not like a brush so you don't normally say 'sweep  - we say 'rake'!) - yes accumulate correct but rubbish!\"},\n {'id': 'test_0250',\n  'dialog': 'student : organised\\nstudent : The sun gets on my balcony only a few hours a day so I try to make sure I go out on time!',\n  'response': \"Right I see ...ooh I'm a bit jealous of that one (though it is sunny right now actually) ..lying IN the sun...Yes lying = good (or e.g. sunbathing on my balcony)\"},\n {'id': 'test_0251',\n  'dialog': \"student : It's famous in Japan as a hometown of the Peter Rabbit.\\nteacher : Yes, it's certainly very pretty - if you're lucky it's sunny (but cold) and clear - if not it just rains! Yes I get that about being famous in Japan - you see quite a few Japanese tourists there...\\nstudent : I hope you will have many sunny days there!\",\n  'response': \"thank you! I also like it beacuse I'm really into running and it's a good opportunity to run on the hills you see\"},\n {'id': 'test_0252',\n  'dialog': \"student : I believe so, because if you had a high salary but a job that you don't really enjoy, it would be difficult for you to be happy\\nstudent : it's important to do a work that we can enjoy because it will be our main occupation during the day\",\n  'response': \"OK thanks <STUDENT>...maybe another sentence would be good here ....to add support ...like a specific case (as in soemthing that wouldn't be enjoyable if you see what i mean\"},\n {'id': 'test_0253',\n  'dialog': \"student : Do you?\\nteacher : ok well....yes definitely! it's very common\\nteacher : including in normal speaking\\nstudent : It sounds formal to me... but I see than I'm wrong\\nteacher : you know that you can something or someone for granted right?\\nteacher : take something or someone for granted\\nstudent : I think I know... I can't think of an examlo\\nstudent : example\\nstudent : I didn't know you could take someone for granted though\",\n  'response': \"that's ok ....so e.g. if a man takes his wife for granted then is that good or bad would you say? sorry stupid example\"},\n {'id': 'test_0255',\n  'dialog': 'student : yes, i found my mistake\\nteacher : ok....\\nteacher : Yes sorry i see it -good!\\nstudent : thanks\\nteacher : Try to remember this OK? YOU make this mistake a lot!\\nstudent : yes, I need to check this point every time when I want to use present continuous.',\n  'response': \"OK so hold on...we have 2 ideas: addiction and violence...and the third (is it this negative emotions idea? ...or children don't realise it is not real idea? Are either of these ideas good for this idea do you think?\"},\n {'id': 'test_0259',\n  'dialog': \"teacher : yes, well done!\\nstudent : Ýåý\\nteacher : Has he been to the USA?\\nteacher : I don't know ...\\nstudent : If  he has ever been to the USA\\nteacher : Perfect!\\nteacher : if or whether\\nteacher : It's starting to make sense, right?\\nstudent : AHA\\nteacher : It's called 'reported speech' but we use it way more often, with all kinds of 'I don't know' and 'I wonder'\\nteacher : not just 'he said'\\nstudent : őķ\",\n  'response': \"Ok <STUDENT>, excellent work today, I'll talk to you on Friday!\"},\n {'id': 'test_0261',\n  'dialog': 'teacher : just one word!\\nteacher : You OK?! Bad internet connection?\\nstudent : yes, i am back\\nstudent : let me try\\nteacher : ok!\\nstudent : There are many museums in London. -\\nteacher : OK cool hold on\\nstudent : okey',\n  'response': \"SO....you said: 'the National Gallery where have many painting.right?' You can say: The National Gallery, which ________ many paintings' OR 'The National Gallery where _____ are many paintings'\"},\n {'id': 'test_0262',\n  'dialog': \"teacher : sure ... well 'firmly' first e.g. opposed/committed = common with noun forms too\\nstudent : By the way, I think the first question is wrong... Maybe 'which adj ca be used with highly?\\nteacher : well both possible... 'with which' and in general prepositions first = formal or normal?\\nstudent : Ahh thanks, I didn't sound good to me.\",\n  'response': \"ok so...prepositions at the end though are also fine e.g. what are you scared of? ok? so..can you ask you're quesstion again with that structure\"},\n {'id': 'test_0263',\n  'dialog': \"student : I think I only know put on and off\\nteacher : Well these are good ones to know! Hold on a sec....(but you can continue if you like!)\\nteacher : tolerate/insult or kill!/delay/ suggest/install\\nteacher : can you think of a phrsal verb with put for these meanings?!  (you can use the 'logic' of the preposition as a rough guide(!)\\nstudent : No idea but I'll try\",\n  'response': \"You know the 'tolerate' one right? I can't put ___ with it any more!'\"},\n {'id': 'test_0264',\n  'dialog': 'teacher : ok\\nstudent : 1. to start something\\nstudent : 2. to stop someone to do something\\nstudent : 3. boring\\nstudent : 4. Opportunity to change\\nstudent : 5. to be honest\\nstudent : 6. to boast\\nstudent : 7. prepare for something\\nstudent : 8. something you are good at',\n  'response': \"OK <STUDENT>, I'm really impressed with your willingness to just have a go! Some great guesses!- Let me give you a bit of help  - which ones means:  show what someone's real character is, especially when it is unpleasant\"},\n {'id': 'test_0268',\n  'dialog': \"teacher : So OK <STUDENT> yes - this is a good idea and the CATEGORY here is 'scientific' right?\\nteacher : As in increasing fundamental  scientific knowledge...\\nstudent : 366.0\\nstudent : like why is 365days a year but sometime is 366 days.\\nteacher : Yes understood!\\nstudent : yes!! basic knowledge\",\n  'response': 'So the international co-operation idea was a POLITICAL category you caould say....Can i ask you to think about the ECONOMIC categeory in relation to this space question...do you think there are economic benefits to space exploration?'},\n {'id': 'test_0269',\n  'dialog': 'teacher : Hey welcome back\\nstudent : yes, I am here',\n  'response': \"OK cool - so <STUDENT> let's imagine you get an essay about population OK?\"},\n {'id': 'test_0271',\n  'dialog': 'teacher : So e.g. newspapers exaggerate\\nstudent : ok, I am remembering those words\\nstudent : thanks',\n  'response': \"OK <STUDENT> ...here's your first sentence coming up..I just want to focus on two grammatical points: 1. = 'one of' and 2. 'as quick as - Are you happy with those two parts? ! 'if one of newspaper just want to attract readers, they also write overstates or fake information which can build a large reader group as quick as possible.\"},\n {'id': 'test_0273',\n  'dialog': \"teacher : OK so <STUDENT>, thinking about the questions from yesterday about different jobs....so you were talking about football players and you said they have a very good salary ...hold on a sec\\nstudent : haha I don't know much about football but players from Barcelona club or Real Madrid earn far too much money\",\n  'response': \"and this is fine and correct obviously ....any other phrases for this idea e.g. (Footballers) have a very good salary ...ah yes! in other words they are 'overpaid' right?\"}]"},"metadata":{}}]},{"cell_type":"code","source":"teacher_continuation","metadata":{"execution":{"iopub.status.busy":"2024-06-25T08:55:53.812051Z","iopub.execute_input":"2024-06-25T08:55:53.812420Z","iopub.status.idle":"2024-06-25T08:55:53.831584Z","shell.execute_reply.started":"2024-06-25T08:55:53.812388Z","shell.execute_reply":"2024-06-25T08:55:53.830617Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":38,"outputs":[{"execution_count":38,"output_type":"execute_result","data":{"text/plain":"[{'id': 'test_0002',\n  'dialog': \"teacher : But most importantly, there could have been other reasons behind children not eating / eating the marshmallow\\nteacher : For example, they might not have liked marshmallows!\\nteacher : Or they could have wanted to please the researchers\\nstudent : Yes that's true\\nteacher : What do you think could have been the reasons?\\nstudent : Maybe some of them were too young?\\nteacher : Yes, exactly!\",\n  'response': 'Could you make a sentence with a modal? Some of them ___ ___ ___ too young'},\n {'id': 'test_0003',\n  'dialog': \"student : my train leaving at 9:30\\nstudent : or - my train leaving time at 9:30.\\nteacher : OK thanks...so you can describe your own plans with present continuous or 'going to' ...when we talk about courses/trains etc (things with timetables) we often use present simple...\",\n  'response': \"So: 'I'm going home tomorrow - my train _____ at 9:30' One word...what is it?\"},\n {'id': 'test_0004',\n  'dialog': 'student : my time was shorter than yesterday.\\nteacher : So the adjective you can use with time (for running for example) is good/fast ok?\\nstudent : yes... so my time was faster than yesterday.\\nteacher : OK great! my time was faster/quicker/better than yesterday! good',\n  'response': 'OK so one more: My personal best for 5k is 15:03 (along time ago) ..... So 15:03 is the _____  I _______  ______ the 5k.??'},\n {'id': 'test_0005',\n  'dialog': \"student : right?\\nstudent : yes,, thanks\\nteacher : well <STUDENT> - we just usually use the verb 'be' with these adjectives\\nstudent : wow... I got it\\nteacher : like e.g. rich/poor/old etc - not because it's a feeling\\nteacher : So in a way this bit is the easy bit (the verb I mean) - the thing you have to get right is the adjective itself and the preposition that comes next you see?\",\n  'response': 'OK see you in 15 mins ok?'},\n {'id': 'test_0007',\n  'dialog': \"student : That's great!\\nteacher : But of course my main job is to help trainees pass the course!!\",\n  'response': 'What do you do to keep up with your English <STUDENT>? Do you watch series/films for example?'},\n {'id': 'test_0008',\n  'dialog': \"student : I'm sure...\\nteacher : as in some will be in your passive knowledge but perhaps not used so much\\nteacher : in my experience with e.g. Italians/Spanish/French etc\\nteacher : for obvious reasons\\nstudent : Can you give an example? I can't think of any\\nteacher : yes sure ..hold on...\\nteacher : so verbs first ...maybe a bit easier cos there are more perhaps\",\n  'response': 'so if I run faster than you ...I _______ you (one-word verb answer)'},\n {'id': 'test_0016',\n  'dialog': \"teacher : What about e.g. a ring that costs $1,000?\\nstudent : Good! Thank you\\nstudent : Shall we go back to yesterday mistakes?\\nteacher : yes sure!\\nstudent : a one thousand-dollar ring\\nteacher : yes good\\nstudent : there is police\\nstudent : Shall I say 'there are police everywhere'?\\nteacher : yes!\\nstudent : why?? is it like people?\\nteacher : Hold on...\\nteacher : So it's a bit difficult with groups/organisations sometimes...\\nteacher : so for example some can be either e.g. a football team can be Manchester UNited is/are it depends\",\n  'response': 'if you mean as a company then....?'},\n {'id': 'test_0017',\n  'dialog': \"teacher : Don't worry about it now but i just wanted to point that out ok?\\nstudent : yes...I think it maybe far away from question.\\nstudent : okey, thanks a lot\\nteacher : OK no probs - but it was a good way to practise language anyway...\",\n  'response': 'A quick test coming up for you <STUDENT> OK? Hold on...'},\n {'id': 'test_0018',\n  'dialog': 'student : They said that it was my fault so I had to fix my car and pay for it. I learnt the lesson ;)\\nteacher : Oh no! I guess you can never tell what people are going to do...',\n  'response': 'Thanks <STUDENT> -  Yes we say ‘on the road’ to talk in general – ‘in’ the road is like a specific place in space do you see?  So for example do we say sleep ‘in’ or ‘on’ the street for rough sleepers do you think?'},\n {'id': 'test_0020',\n  'dialog': 'teacher : Hi there...\\nstudent : hi\\nteacher : Right <STUDENT> - Would you like to try beginning of a main paragraph for the essay question?\\nstudent : yes\\nteacher : Choose an advantage of renting  and go for it!',\n  'response': 'How about the financial one you mentioned?'},\n {'id': 'test_0022',\n  'dialog': \"student : Yes\\nteacher : But poor kids didn't get much stuff, and when they saw the marshmallow, they just ate it\\nteacher : And then rich kids went to better schools and universitities / colleges and got into better jobs!\\nteacher : Not because they had better self-control, but because they could afford them\",\n  'response': 'And this is how the marshmallow test was debunked. Debunk = prove a theory wrong'},\n {'id': 'test_0023',\n  'dialog': \"student : yes\\nteacher : Sorry ...not exactly!\\nstudent : or have some grammar mistakes\\nteacher : Yes that's right...What about the first chunk? What do you need to change?\\nteacher : most of young people?\\nstudent : most young people\\nteacher : Yes great wel ldone - like we said yesterday right?\\nteacher : OK and the next chunk ....you need to change the noun 'benfit' to an adjective ..do you know it?\",\n  'response': 'I mean the adjective form of benefit'},\n {'id': 'test_0024',\n  'dialog': \"student : ....\\nstudent : after that it's better if I write information in general?\\nteacher : Yes...\\nstudent : like '...are more satisfied'\\nteacher : Yes for example\\nstudent : than the visitors in the years before\\nteacher : Yes ok ...again , pretty good you've got the big idea clearly expressed\\nstudent : thanks\\nteacher : A couple of small things...\\nstudent : yes\\nteacher : Dont say ' we can conclude'- just say it!\\nteacher : Use past simple...\\nteacher : I.e. 'were' more satisfied\",\n  'response': 'And finally....change some of the words if possible using synonyms and different phrases like we did last week'},\n {'id': 'test_0025',\n  'dialog': \"student : I mean the 'way' can not use here to explain something from partens\\nteacher : Oh i see sorry - yes I guess you;re right! you inherit an ability/talent etc = the collocation!\",\n  'response': 'one more potentially useful word <STUDENT> is connected to this idea of ‘before you were born’ in the question ...so ‘when you were born’ = your ‘______’? One-word noun needed here! What is it?'},\n {'id': 'test_0029',\n  'dialog': \"student : haha yes, sadly people say that phenomenon always occurs so harry up, show her as much cool things as you can now!!\\nteacher : Yes, I figure that anything you do is worth it - even if it doesn't seem like it at the time - so your nephew might like those things at some point!\",\n  'response': \"But it's funny how different people can be from each other isn't it?\"},\n {'id': 'test_0033',\n  'dialog': \"teacher : ?\\nstudent : social media\\nteacher : ok fine....one more?\\nstudent : and  education between parents and shools.\\nstudent : schools\\nteacher : OK great - so you mean that schools and parents don't....??\\nstudent : they don't have more sympathetic approaches to deal with it.\\nteacher : OK ...is that a cause? I suppose it could be! At least it explains why it doesn't stop ....\",\n  'response': \"Maybe slightly clearer as a cause would be the idea that schools/parents don't teach children that bullying is wrong?\"},\n {'id': 'test_0045',\n  'dialog': \"student : unsatisfactory\\nteacher : unsuit_ _ _ _ '??\\nstudent : unsuitable?\\nteacher : yes good!\\nteacher : And ...'inapprop_ _ _ _ _ '??\\nstudent : and...inapprppriate\\nteacher : yes great...so either of these words would be good to use here OK?\",\n  'response': 'Can I suggest <STUDENT> that it would be good to have a clear idea of exactly what the bad behaviour is?!'},\n {'id': 'test_0049',\n  'dialog': \"teacher : a short word is missing!\\nstudent : Is in\\nteacher : nope, you have 'is' already\\nstudent : Now?\\nteacher : like!\\nteacher : What's the weather like?\\nstudent : Oh ok\\nteacher : It's been raining since Tuesday!\\nteacher : What's life like in your village?\\nstudent : It's hard and interesting\\nteacher : Good answer! I kind of want to know more, but it's time to let you go!\",\n  'response': \"Enjoy a break, and I'll talk to you again on Wednesday 22 July, and I'll email you to remind :)\"},\n {'id': 'test_0052',\n  'dialog': 'student : 555555you know😂\\nteacher : So...! Before you said: 15 units of electricity was produced from nuclear power in France in 1980. (=good!) Now do a second part to link this with the situation in in 2000\\nteacher : OK?!\\nteacher : I mean another sentence...',\n  'response': \"Is the link (i.e. first it was 15 units then it was 126) a contrast (so words like 'but or 'while' might be useful) or is just another thing (so 'and' might be OK)?\"},\n {'id': 'test_0054',\n  'dialog': \"student : will\\nteacher : ok good - for first conditional we usually use 'will'\\nstudent : okey, thanks\\nteacher : One final one: 'In my view, scientists _________ discover a cure for most diseases .'\\nstudent : are going to\\nteacher : Ok thanks - 'will' is also possible here - which one is best if you are making a general prediction (at some time in the future)\",\n  'response': 'the other word would be if you think the cure is coming soon/you are thinking of specific progress right now'},\n {'id': 'test_0055',\n  'dialog': \"teacher : Missing word?\\nstudent : yes.i studied the topic sentence  when you were teaching us on class.\\nstudent : a\\nteacher : OK thanks great..one more...: “…and even cheaper than _____ a normal one”. What kind of word do you need in the gap? Noun or verb?\\nstudent : buy\\nstudent : verb\\nteacher : So for example: 'Water is cheaper than whisky.' Correct?\",\n  'response': \"So what comes after 'than' if you comparing things normally?\"},\n {'id': 'test_0066',\n  'dialog': \"teacher : OK?!\\nstudent : Yes, I mean so. Even though IELTS isn't a competiton...\\nteacher : OK - you mean you still want to say 'win' - OK you can, you can - it's possible!\\nteacher : to really emphasis she though of it like a compettiion I guess!\\nteacher : thought\",\n  'response': \"OK <STUDENT> - we'll have to stop there - thanks a lot- speak to you soon\"},\n {'id': 'test_0069',\n  'dialog': \"student : ok\\nteacher : I just think it's good to think about main ideas which are easy to explain or give examples for....I know we might have different ideas about this but i feel that this idea is a bot more difficult to make successful in an essay - it's a bit vague/abstract do you see what i mean?\",\n  'response': 'THe addiction and violence ideas are easier to explain I think ..do you agree??!!'},\n {'id': 'test_0078',\n  'dialog': \"teacher : average' would be a good word to use for your point as well\\nstudent : Do you mean balance satification and salary,\\nstudent : Ah I see\\nteacher : as in 'your salary is below average'\\nteacher : Next question: 'What skills are needed to get a good job nowadays?'\\nstudent : Nowadays I think\\nstudent : It’s need a lot of skiils\\nstudent : Comunucation\\nstudent : It or computer skill\\nstudent : And language also\\nteacher : THanks - good examples (communication,language and IT skills)\",\n  'response': 'Say one more thing about any or all of those OK?'},\n {'id': 'test_0079',\n  'dialog': \"teacher : so in fact in this example we only have one thing - coal - so while is ok but whereas isn't ...OK?\\nstudent : Ok\\nteacher : but you could say : Coal was the most important source in 2000 while/whereas in France nuclear power (both fine because we have two things )\",\n  'response': 'OK thanks <STUDENT> - some hard stuff today - well done!'},\n {'id': 'test_0081',\n  'dialog': \"student : so...I think I can delete the 'effects'\\nteacher : Yes good!\\nteacher : pros and cons' already includes that idea\",\n  'response': 'OK last sentence = pretty good… two things: First gap = a general word for houses/flats etc beginning with ‘p’. Second gap = a modal verb please beginning with ‘m’! “In my opinion, although __________ rental plays a significant role, some uncomfortable experience _________ occur.”'},\n {'id': 'test_0084',\n  'dialog': \"student : In my class  there was a very posh Camila\\nteacher : Less traditional ones for boys that are popular now = e.g. Archie\\nstudent : And Isla (I like this one, very original) and Freya\\nteacher : Most of them end in vowels and are a bit 'babyish' that's the fashion!\\nteacher : Yes Camilla = posh!\",\n  'response': \"It's funny when some non-natives think of John as a common name - if I imagine a baby in a pram called John I imagine a deep gravelly voice coming out like in MOnty Python or seomthing!\"},\n {'id': 'test_0087',\n  'dialog': 'student : no ,it not finished\\nstudent : so we use present simple\\nstudent : or present continue\\nteacher : OK good - so some time phrases AUTOMATICALLY need the past simple OR the present perfect OK? It depends on the phrase\\nstudent : okey, I understand.\\nteacher : So e.g. last year/last week/2018/ = ? (past simple or present perfect?)\\nstudent : present simple because it is specific time\\nteacher : OK fine (you mean past simple right?!)\\nstudent : yes\\nteacher : OK!',\n  'response': \"OK good! and one more...what about 'this morning'  ? Past simple or present perfect?\"},\n {'id': 'test_0088',\n  'dialog': \"student : like dress up?\\nteacher : Yes, that's right, a verb + preposition that work together and mean something different from just the verb\\nstudent : okay\\nteacher : Ok, all today's phrasal verbs have something to do with mobile phones! I bet you know a lot of them, let's see\\nteacher : We press the power button, to ___ the phone on / off\\nstudent : put\\nteacher : no, we switch the phone on / off\",\n  'response': 'For example, when we get on a plane, we have to ___ ___ our phones'},\n {'id': 'test_0091',\n  'dialog': 'student : yes..\\nteacher : So for example......The Black Death happened in the 14th century when population was increasing.\\nteacher : So this is talking about the 14thcentury which ash definitely finished\\nteacher : But for our essay intro we want (imagine) to talk about recent time - from past and up to NOW\\nteacher : which tense??\\nstudent : present perfect?\\nteacher : YEs!',\n  'response': \"So for example: 'Population _______  ________ significantly in recent decades.'??\"},\n {'id': 'test_0093',\n  'dialog': \"student : present continue\\nteacher : You have 1. a topic sentence 2. Explanation of topic sentence 3. Result and 4. Summary with a bit on the other side (very good!).\\nstudent : okey, thanks!!\\nteacher : You've got some good 'chunks' of language...hold on!\",\n  'response': \"One of the main reasons why playing video games negative influence on young people is that .......can you add an article ('a') and think about the agreement of verb and subject?? i.e. it has/they have\"},\n {'id': 'test_0094',\n  'dialog': \"teacher : Ah yes sorry - it's the name for the little bits of wood or whatever you use to start the fire at the beginning\\nstudent : Here 18 degrees now, during the day it's very sunny\\nteacher : NIce....yes it's actually really sunny this week here too - and maybe....13 degrees!\\nstudent : Not too bad 13!!\\nteacher : yes! Actually my boat gets warm in the sun so it doesn't really matter about the outside temperatrue...\",\n  'response': \"Have you heard of the Amazon 'Kindle' the thing for reading ebooks\"},\n {'id': 'test_0095',\n  'dialog': 'student : ok, it is better than my sentence.\\nteacher : Well yes I guess so! BUT we can take one more step if we like....Look at the one coming up...can you see I have deleted the full stop and the verb - you just need to change the form of the verb ..can you try it?Coal was by far the most significant fuel, _________  130 out of 170 units',\n  'response': \"So e.g. 'produce/produced/producing'? you decide!\"},\n {'id': 'test_0096',\n  'dialog': \"student : yeah\\nteacher : Ok, let's keep going\\nteacher : Our trip to the park depends __ the weather\\nstudent : on\\nteacher : yes, good!\\nteacher : She makes her own decisions.  She's quite independent ___ her friends' opinions\\nstudent : on\\nteacher : No, that's the tricky one. Depend on (verb), dependent on (adjective), but independent from\\nstudent : ok\\nteacher : For example, America became independent ___ the United Kingdom in the 18th century\\nstudent : from\\nteacher : yes, but\",\n  'response': \"He's diabetic. He's dependent __ insulin\"},\n {'id': 'test_0098',\n  'dialog': \"teacher : sorry I mean if you only say that exercise is the most importnatn thing\\nteacher : So for example...in your opinion is diet as important as exercise or less important ?\\nstudent : as important as exercise\\nteacher : ah ok! So if this is your 'thesis' then you should say that in the intro...hold on...\",\n  'response': \"So for example; Although I agree that exercise plays an important role ....' can you finish this?\"},\n {'id': 'test_0099',\n  'dialog': 'teacher : OK thanks <STUDENT> for all that  ... hold on asec\\nstudent : seemed like - I forgot to use past simple here.\\nteacher : yes OK got it!',\n  'response': \"yes so...correct past simple verbs here and also a nice 'going to' for plans = great!\"},\n {'id': 'test_0100',\n  'dialog': \"student : I think it's great! I can see much more people exercising on the streets nowadays than a few years ago\\nteacher : Yes, you know one thing that I think about with running and Spain is that it  must be really hot for some of the year - as in too hot to run(?)\",\n  'response': \"When I lived there I didn't do much exercise - just smoked and drank too much!\"},\n {'id': 'test_0102',\n  'dialog': 'student : I understand your correction. In Spanish we have both words (controvertido y provocativo) and I see that controversial is more appropriated\\nteacher : controversial',\n  'response': \"<STUDENT> - can I recommend a language corpus website? if you put 'bnc byu corpus' in google then you should get it - put a keyword in the search box and it'll give you loads of examples - a bit messy and some difficult stuff BUT might be interesting for you - it'll give you a snapshot of the word in real use\"},\n {'id': 'test_0103',\n  'dialog': \"student : Wow... you are really into running! Ely is so far from here.\\nteacher : Oh right! I see what you mean...well it's better int he summer/warmer weather anyway because you can sit outside. If you run a lot you get used it and it doesn't seem hard? Unless you try to run faster of course - that's what starts making the difference\\nteacher : the ? was a mistake!\",\n  'response': 'get used to it I meant - sorry trying to type too fast!'},\n {'id': 'test_0104',\n  'dialog': \"student : Yes, unfortunately it's true and I went back Japan two weeks ago because of this. I was so confused because she was really fine until that day.\\nteacher : OK thanks I understand. I think maybe 'shocked' is better here.\",\n  'response': '<STUDENT>, can you give me the missing two words in the following: I’m optimistic _______ my chances of winning. I was confused ______ the explanation.'},\n {'id': 'test_0108',\n  'dialog': \"student : I have always felt scared with spider even though there are no dangerous ones in the UK.\\nstudent : scared of\\nteacher : yes good correction\\nteacher : can you change the verb 'feel' ?\\nteacher : the present = I am scared\\nteacher : and in this example?\\nstudent : I have always feel scared of spiders\\nteacher : I have always __________ scared of spiders.?\\nteacher : Don't use 'feel'! What verb do you need?\\nteacher : A simple verb!\",\n  'response': 'But the past participle (because you;re using the present perfect)'},\n {'id': 'test_0113',\n  'dialog': 'teacher : so for example the fact that coal generated such a big proportion of electricity means that coal was very...?\\nstudent : Important?\\nstudent : Important source\\nteacher : yes great! was important/was an important source\\nteacher : hold on ...',\n  'response': 'so: in Ielts you can use a two parts to give the information: one to example the meaning and one to give the data OK? In other words, Coal was ......It generated....(can you complete these?)'},\n {'id': 'test_0117',\n  'dialog': 'student : ok..I am not sure\\nteacher : No problem...well for me <STUDENT> changing round 2 and 3 seems a bit better',\n  'response': 'So...One of the most significant negative effects of playing video games is that they can lead to violent behaviour. Because many games are designed for adults, they may often contain violent content. It is difficult for children to distinguish between video games and reality. As a result, young people are likely to copy the violence they see, and this can result in them bullying classmates for example.'},\n {'id': 'test_0118',\n  'dialog': \"teacher : Some topic words and some nice general phrases\\nteacher : I;ll go over some other things to think about when we come back after the break....\\nstudent : okey, thanks- see you\\nteacher : Ah yes ok  i;ve just noticed: ' even though ' is correct ( notthrough!)\",\n  'response': 'not even thRough!'},\n {'id': 'test_0121',\n  'dialog': \"teacher : *dropped it\\nstudent : Maybe\\nstudent : You might know that there is war in Ukraine\\nteacher : Of course I know that :(\\nteacher : But isn't the revolver from a long time ago, when your Grandma was young? That was the impression I got\\nstudent : Yes\\nteacher : Yes, sure. What were you going to say?\\nstudent : When?\\nteacher : Never mind :)\",\n  'response': 'Did you manage to watch the video I sent you? Of the marshmallow test?'},\n {'id': 'test_0124',\n  'dialog': \"student : maybe I need to use present perfect\\nteacher : Ah ...well if you're talking about right now then the present continuous is OK but have you used it correctly here?\\nteacher : THis is the same mistake you made before (that's why I'm asking you again!)\\nstudent : no..I think I talk about a period of the time.\\nteacher : OK <STUDENT> ...hold on\\nteacher : Don't worry about changing the tense...it's not the problem here\",\n  'response': 'You can use the presnent continuous but you have to make it correctly!'},\n {'id': 'test_0125',\n  'dialog': 'teacher : Hi <STUDENT> all ok? Sorry for delay - connection problwm\\nstudent : Hi <TEACHER>\\nstudent : I’m fine\\nstudent : And you\\nteacher : Yes fine thanks',\n  'response': 'Did you have  a chance to look at the sentences?'},\n {'id': 'test_0126',\n  'dialog': \"student : because it is not an effective way to study subjects, such as feedbacks are not quick\\nstudent : I guess less good is suitable.\\nteacher : Yes I guess it's difficult to give feedback - I see what you mean\",\n  'response': \"Hey <STUDENT> - I've just emailed you a file with some an exercsie connected to the pie charts we were looking at last week - is it easy for you to open that an look at it at the same time as talking to me?\"},\n {'id': 'test_0129',\n  'dialog': \"student : haha we have something similar! 'Me río por no llorar'\\nteacher : OK nice - thanks! Exactly the same really...\\nstudent : You are very lucky there in the middle of nowhere\\nteacher : ya gotta laugh' is if you can imagine the quick spoken form of the phrase...\\nteacher : Hey <STUDENT> do you read phonemic script by the way?\",\n  'response': \"As in you'd follow this right?: /jəgɒ(t)əlɑːf/\"},\n {'id': 'test_0132',\n  'dialog': \"student : 3: scrape the barrel\\nstudent : 4. turn the corner\\nstudent : 5: scrape the barrel\\nstudent : I'm not sure for 1,2 and 3\\nteacher : hanks <STUDENT> – so quite difficult with such  general examples/situations… but that’s OK – good to think about! Give me a second\",\n  'response': \"So 'starting a new job' was that feeling when everything was a bit difficult at the beginning and I felt...'all at sea'!\"},\n {'id': 'test_0133',\n  'dialog': \"student : I think that surgeons should be the best paid considering that they save lives\\nteacher : understood.....\\nstudent : I strongly disagree with the fortunes that footballers make\\nteacher : ok....\\nstudent : Is it correct?\\nteacher : Yes its\\nteacher : it is\\nstudent : can you disagree with a pay or maybe you use another better verb?\\nteacher : You can disagree with the money they make - it's OK\\nteacher : pay without article\\nteacher : I disagree with their pay\\nteacher : with how much they make/get etc\",\n  'response': 'Of course lots of other possiblities e.g. I think the amount of money they get is outrageous etc etc'},\n {'id': 'test_0150',\n  'dialog': \"student : That's a great way to start the day. I also did some exercise this morning\\nstudent : I have done, I guess\\nteacher : Yeah? what did you do? I'll come back on the have done or did in a bit\\nteacher : It is more common overall to use the present perfect for e.g. this morning if it hasn’t finished yet but…it’s by no means a rule and e.g.  did is perfectly fine especially if you’re focusing on the action itself\",\n  'response': 'tell me about your exercise if you like!'},\n {'id': 'test_0152',\n  'dialog': \"teacher : ?\\nstudent : pass\\nstudent : overtake\\nstudent : ? not sure\\nteacher : Ah well possible I guess - as in the moment itself - but more generally we can say 'outrun'\\nteacher : by the way overtake good idea and very common (but still only for the moment itself)\\nstudent : Another new word for me!!\\nteacher : lots of Spanish speakers just say pass - good to have thought of overtake = very natural\\nteacher : but yes 'outrun' = normal too so good to know\\nteacher : What about e.g. score more points than?\\nstudent : no idea\\nteacher : outscore!\",\n  'response': 'and: fight harder than?'},\n {'id': 'test_0159',\n  'dialog': \"teacher : So in other words 1. the ability to + verb OR 2. improve + noun\\nstudent : yes,,I understand them, thanks\\nteacher : OK cool! Final chunk <STUDENT> = 'a lot time wasting children's life'.\\nteacher : THe full version was although some video games are able to improve some abilities of attrition, a lot time wasting children's life.\",\n  'response': 'So after the comma you need a clause OK?!  (a subject and verb)'},\n {'id': 'test_0161',\n  'dialog': \"student : ok\\nteacher : Here's the question (I'll give you a few minutes): Governments are spending a lot of money on finding life on other planets. Some people say this is a waste of money and that they should more money on the problems of the public. To what extent do you agree?\",\n  'response': \"sorry: '...that they should SPEND more money on the problems....\"},\n {'id': 'test_0162',\n  'dialog': 'student : lifes\\nteacher : OK thanks some specific examples often good idea - hold on a sec',\n  'response': \"<STUDENT>, look at following and make a change if you can: You said: I believe so, because if you had a high salary but a job that you don't really enjoy...’ Are the two verb tenses OK here?  By the way: spelling = conscious and responsibility  Depend ‘from’? Lives = good for plural\"},\n {'id': 'test_0163',\n  'dialog': \"student : Wait, do you check words in the dictionary?? or look up?\\nteacher : the more i teach the more I focus on pronunciation to b ehonest\\nteacher : you can say check or look up - both fine (and common)\\nteacher : I guess 'chec' could imply you kinda know a bit already and 'look up' can suggest you know nothing you see?\\nteacher : check' I mean\",\n  'response': \"A new phrasal verb by the way used by young people (but not by me) = 'search up'\"},\n {'id': 'test_0168',\n  'dialog': 'teacher : yes ok good! So the important thing here <STUDENT> (from my English teacher point of view!) is that you use the correct adjective + preposition ok? Hold on\\nstudent : ok\\nteacher : OK remember these adjectives from before? Tell me the preposition you need to use after each one OK? Can you write each work again with the correct preposition please? motivated/proud/scared/disappointed/happy/annoyed/excited\\nteacher : each word i mean sorry',\n  'response': \"So for example: 'motivated TO' right? Keep going!\"},\n {'id': 'test_0169',\n  'dialog': \"student : So there is 4 words\\nteacher : Yes, two of them are 'the lizard'\\nstudent : If the lizard is male or female\\nteacher : Well done!!\\nteacher : How old is the lizard? --- We don't know __ ____ ____ ____ ___.\\nstudent : how old is the lizard\\nteacher : Not quite\\nteacher : We don't know how old ___ ____ ___\",\n  'response': 'You need the structure of an affirmative sentence, not a question'},\n {'id': 'test_0177',\n  'dialog': \"teacher : number 1 I mean! I;m just giving you a little bit of help by telling you the first word you need from the list ('Australia')\\nstudent : ok, could I use   its used a half Australia of electricity produced coal?\\nteacher : OK so: 'Australia used....(you continue JUST with the words I've given you)...'\",\n  'response': \"START with the word 'Australia' - this gives you a good structure - that's why I;m giving you the first word for each one\"},\n {'id': 'test_0183',\n  'dialog': \"teacher : Sorry I see I deleted the verb 'have' from your original when I copied it!\\nteacher : Anyway...playing video games = singular ok!?\\nstudent : ok..\\nstudent : so 'video games' is singular, right?\\nteacher : games are associated to them' ..this bit needs changing! Let me give you a model to help: 'Many people are addicted to drugs.' Can you try your bit again?\",\n  'response': 'One of the main reasons why playing video games has a negative influence on young people is that ......?'},\n {'id': 'test_0184',\n  'dialog': \"teacher : OK thanks - I'll give you one back! 'The Yangtze is the longest river in China. Can you add a word in your example??'\\nteacher : Just one little word!\\nstudent : Thames is a famous river in London.\\nteacher : Hold on...my example was confusing I;m sorry - look at this one: 'The Yangtze River is in China.' Can you add a diffrent word!\",\n  'response': \"Sorry go ahead: for example:  '_______ River Thames is in London.'  ?\"},\n {'id': 'test_0188',\n  'dialog': \"student : okey, thanks a lot\\nteacher : Good use of keywords too....\\nteacher : So it is argued that our characteristics are fixed and we are not able to alter _____ (plural pronoun?!). Although I strongly agree whit ('with' + noun - do you need it here? )that genes play a significant role, learning afterward is more important.\\nteacher : significant role' + good too by the way!\",\n  'response': 'sorry = good I wanted to say'},\n {'id': 'test_0189',\n  'dialog': \"student : I'm trying to watch 'feel good' series, I don't want more drama after watching the news a few times a day...\\nteacher : I think I put the preposition in by mistake if i try to say that IN spanish .....\\nteacher : Yes I understand...'feel-good' = great adective\\nstudent : So, I should say 'I'm glad you reminded me of El club...\\nteacher : yes exactly\\nstudent : Thanks\\nteacher : you must include 'of'\",\n  'response': \"I'm planning to learn more languages - when i have more time (very soon)\"},\n {'id': 'test_0194',\n  'dialog': 'student : yes,  I still remembering the example is that there are various museums in London.\\nstudent : thanks... because correct is the most imperative part.\\nteacher : Hey sorry - my internet went again\\nstudent : that ok\\nteacher : Hi..Yes, the museum example is a great point here..exactly!',\n  'response': 'there is + noun is like to be + adjective ...usually correct!'},\n {'id': 'test_0195',\n  'dialog': \"student : which ones means: show what someone's real character is, especially when it is unpleasant'\\nstudent : You mean about No.5?\\nteacher : yes that's it! The 'colours' is from the colours on the flag the ship is flying, so if a ship was disguising itself to attack it would fly a false flag you see and then reveal it's true identity just before attacking\",\n  'response': 'but we use it to describe people - i.e. they pretend/seem to be nice and then do something horrible'},\n {'id': 'test_0197',\n  'dialog': \"student : yes\\nstudent : about poshness (does this word exist? haha)\\nstudent : I think it's similar in Spain\\nstudent : or the same\\nteacher : I could see when some people on TV or whatever were kind of posh from their clothes/behaviour but not from their accent which is a big giveaway in Britain\\nteacher : Yes I was going to ask you this (see my cpoy and pasted comment above!)\\nteacher : So i could see some similarities i guess\\nteacher : yes poshness is fine (but posh much more common )\",\n  'response': 'But can you hear from accent?'},\n {'id': 'test_0199',\n  'dialog': 'student : I see haha IN\\nteacher : Yes true ...out on the deck',\n  'response': 'By the way this reminds me of a programme I watched last night from Spanish TV (indirectly inspired by communicating with you I think – it keeps the idea/existence of Spain in my head!) that I’ll translate as ‘Quarantine Diaries’ to stay in English … actually after having boasted to you about how I could understand Spanish stand-up etc I realise how rusty I’m getting, even with comprehension let alone with speaking…'},\n {'id': 'test_0200',\n  'dialog': 'teacher : So: It is often argued that the most effective way to lead a less stressful lifestyle is __ rent a house.??\\nstudent : to OR house rental.\\nteacher : OK thanks <STUDENT> - yes good',\n  'response': 'This sentence coming up was great actually .. just one small thing: The words ‘pros’ and ‘cons’ are nouns, so delete one word from the sentence coming up to make it correct! “Obviously, it is difficult to assess the pros and cons effects of renting a house.”'},\n {'id': 'test_0204',\n  'dialog': \"teacher : So here you need? Which tense?\\nstudent : okey - Population has increasing significant.\\nstudent : Population has increased significant.\\nstudent : significantly.\\nteacher : We're changing from the present perfect now because we are not using the word recent - we are just talking about population going up at the moment/now ...so which tense?\",\n  'response': 'So option 1: Population has increased in recent years. AND option 2: Population ___  _______ing ...'},\n {'id': 'test_0212',\n  'dialog': \"student : Good afternoon <TEACHER>\\nteacher : Hi there <STUDENT> - how's it going?\\nstudent : yeah..I have returned to my hometown\\nstudent : and the local government required me to do the second test of COVID-19\\nteacher : Yes I see - you said before that weren't looking forward to that too much!\\nteacher : Oh no! Another test!\\nstudent : I am still negative on COVID-19\\nteacher : Will this be a quick one?\\nteacher : Oh right so you got a quick result\",\n  'response': 'I read that that there are basically no new cases in China now - is that right?'},\n {'id': 'test_0215',\n  'dialog': \"teacher : it's unfinished that;s the important thing (more than the idea that;s it;s not specific)\\nteacher : OK good! 'It has become more ______ in my mind.'?\\nstudent : it has became more deeply in my ming.\\nstudent : mind\\nteacher : Hold on <STUDENT>....so the idea of clearly is OK in terms of vocabulary BUT is not correct grammatically - you need to change it but just a little but!\",\n  'response': \"E.g. 'The idea is c_______ to me now.'\"},\n {'id': 'test_0223',\n  'dialog': 'teacher : Ok, <STUDENT>, a quick question. Would you be able to have a class tomorrow at 10.30?\\nstudent : yes\\nteacher : Lovely!\\nteacher : Some of the more difficult ones from today I think were: in the photo / in the picture, at the park, on Monday, on my birthday, at Christmas and at night\\nteacher : Feel free to stay and copy them in your notebook if you want to',\n  'response': \"And have a nice day, I'll speak to you tomorrow!\"},\n {'id': 'test_0224',\n  'dialog': \"teacher : Ah i see\\nteacher : OK so if you just out in 'list' then in just gives you the one word that you enter in any context\\nstudent : SEARCH FREQUENCY CONTEXT OVERVIEW more options\\nteacher : But e.g. if you choose 'colocates' then e.g. put in 'vivid' (i think that was an example we used before?) and then press 1...\",\n  'response': 'This means that the results will give you the most common words just one word away i.e. nect to'},\n {'id': 'test_0225',\n  'dialog': 'student : Usually\\nteacher : Yes, very good!',\n  'response': \"Ok, <STUDENT>, a little bit of homework for you today. Could you find a bit of an English text, just a few sentences, and think about the articles in it? It could be anything, an advert, a description of a game, a song. And next time show it to me and say, for example, there is 'a' here because it means 'any', not this that we're thinking about\"},\n {'id': 'test_0232',\n  'dialog': \"student : okk\\nstudent : What about 'place'? from the beginning...\\nstudent : I can walk around my place, does it sound weird?\\nteacher : Ha! Well my dad was/is (but retired now) quite a talented musician so i kind of got into it but never really measured up - i think i got a bit disillusioned with the difference!\",\n  'response': \"Um...so the 'place' thing maybe sounds weird because we basically use my place wen we say e.g. let's have a party at my place etc....\"},\n {'id': 'test_0233',\n  'dialog': \"student : I see. It's interesting!\\nteacher : OK let me make it quicker/easier - can you match these 4 with definitions? take something on board rock the boat plain sailing  turn the corner  (a situation) start to improve after a difficult period do or say something that will upset people or cause problems be easy and without problems  consider an idea, problem, or situation and try to deal with it\",\n  'response': 'Sorry hold on - let me do that again with gaps!'},\n {'id': 'test_0237',\n  'dialog': 'student : okey, thanks\\nteacher : so the effect of x is NOUN ...OR the effect of x is that + CLAUSE (subject and verb)\\nteacher : or e.g. the effect of x is ADJECTIVE (common actually)',\n  'response': \"So for example : 'the effects of smoking are terrible.'\"},\n {'id': 'test_0238',\n  'dialog': 'teacher : Where do you get your genes from?!\\nstudent : I think it must be form parents.\\nteacher : yes absolutely! So your parents pass on their genes OK? Another way of saying this = You _____ your genes/character from you r parents(like money when they die!)',\n  'response': \"sorry 'pass on' = phrasal verb in my example (maybe genetically maybe just information etc and the one-word verb i;m looking for in the second part = ?  )\"},\n {'id': 'test_0242',\n  'dialog': \"student : silent/difference/consequence\\nteacher : THanks <STUDENT>...you need to change one of your answers! Which one?\\nstudent : slience\\nteacher : Yes ok good!\\nteacher : You need 'something' right (a noun in other words) .... you can say need TO + verb\\nstudent : need to slent\\nstudent : silent\\nteacher : Ah well.....need to + verb so....where's the verb in your last example??\",\n  'response': \"You would need to add one! 'We need to ____ silent now.'??\"},\n {'id': 'test_0247',\n  'dialog': \"student : Number 3 talks about his physical appearance\\nteacher : ok!\\nstudent : Number 5 talks about a local problem\\nteacher : yep\\nstudent : the most sarcastic I think it is number 2?\\nteacher : YEs absolutely! Classic British example - i could have written that myself!\\nstudent : haha\\nstudent : nobody here seems to suport Boris heres\\nstudent : but maybe number 4?\\nteacher : Yes i guess 'support' is not quite the right word!\",\n  'response': \"I've reduced these ti be quick...\"},\n {'id': 'test_0249',\n  'dialog': \"teacher : Do you mean it's good to use 'normal' words instead of 'academic' words?\\nstudent : yes\\nteacher : OK right sorry yes I understand ...hold on...\\nteacher : I think the important thing is that you achieve the right result using the language you know as accurately as possible\\nteacher : in other words the priority is to write an effective intro\\nteacher : if you don't do that because ou make lots of mistakes trying to use 'difficult' words then no that's not great\",\n  'response': 'better to use more simple words accurately to write an effective essay'},\n {'id': 'test_0254',\n  'dialog': \"student : qualities\\nstudent : characters\\nteacher : yes good ...or characteristics ...OK? Both are countable and mean particular features ...your CHARACTER or PERSONALITY = uncountable (at least for one person they are uncountable )\\nteacher : So in other words a person's character or personality includes different specific qualities or characteristics (OK?)\\nstudent : not ok\\nteacher : Ha!\\nteacher : Hold on...\\nstudent : we can only write characteristics here\\nteacher : No 'qualities' is OK too = the same thing\\nteacher : Oh sorry yes you can't write character\",\n  'response': 'character is formed by a GROUP of characteristics'},\n {'id': 'test_0256',\n  'dialog': \"teacher : in your opinion?\\nstudent : of course, I suppose it is the best way for people's health.\\nteacher : OK fine ....so if you agree then you can just say that e.g.: I strongly agree that ....?\\nstudent : ok, I strongly agree that doing exercise is the best way\\nteacher : Or sorry....at the most basic level you could just say: I strongly agree with this idea.\",\n  'response': 'Or you coud repeat the main idea again a little bit and say: I strongly agree that exercise plays a ....?'},\n {'id': 'test_0257',\n  'dialog': \"student : New knowledge is always studied by me. when I had full motivation on it.\\nstudent : my parents were disappointed with my last IELTS score.\\nteacher : ok good\\nteacher : but bad sorry\\nteacher : OK <STUDENT>: 'when I had full motivation on it.'  try this again BUT with the verb/adjective structure we're focusing on ok?\",\n  'response': \"So: ...' when I ____  f______  _______.'\"},\n {'id': 'test_0258',\n  'dialog': \"student : switch\\nteacher : beginning with'a' ...?\\nteacher : al...@?\\nstudent : alter\\nstudent : we lean it on class\\nteacher : yes OK - that's good as a general synonym for change\\nteacher : and good here too\\nstudent : okey,\\nteacher : switch ' is like one specific to another specific thing do you see ?\\nteacher : like from your left hand to your right hand for example\",\n  'response': \"OK look at these two words: character and characteristic - which one is one particular presonality feature and which one is the whole group together and best as a synonym for 'personality'\"},\n {'id': 'test_0260',\n  'dialog': \"student : it might means wealth I guess\\nteacher : Ah! property is just another word for e.g. buildings/houses/flats etc\\nstudent : and including houses and others\\nteacher : Actually 'accommodation' would be good here too maybe right?\\nstudent : okey...thanks\\nteacher : AND your use of wealth is interesting\\nteacher : You could definitely include that somehow!\\nteacher : or similar words...wealth = adjective or noun?\",\n  'response': \"OK <STUDENT> I have to go now I'm afraid - I van see you're having internet problems! I'll speak to you tomorrow..bye!\"},\n {'id': 'test_0265',\n  'dialog': \"student : taking about the site, the sentences that are shown with the word are from texts??\\nstudent : Are all correct? I mean, the sentences...\\nteacher : yes it's called a concordance - have you seen one before? It just has a massive database of spoken and written texts from different sources and searches them for you\\nteacher : well, they're basically 'correct' yes - 'authentic' at least\\nteacher : For me <STUDENT> it's like a dictionary but better you know\",\n  'response': 'I use them when i can if I learn a language'},\n {'id': 'test_0266',\n  'dialog': \"student : brith\\nteacher : ok yes 'birth'\\nteacher : AND number 3: Beethoven;s talent was 'i _ _ _ _ ' ?\\nteacher : It means something he was born with!\\nstudent : innate\\nteacher : ok great! And e.g. patience/honesty/kindness/selfishness etc are ....?\\nstudent : character\\nteacher : you need a COUNTABLE noun here ...so e.g. 'character' is not possible (at least in this context) - a person's character is made up of different _______?\",\n  'response': \"Qu....' or 'ch...' = both possible\"},\n {'id': 'test_0267',\n  'dialog': \"student : coal of electricity?\\nstudent : I don't know\\nteacher : OK no problem hold on a sec\\nstudent : I ask me the same question when I wrote the sentence\\nteacher : well...If you begin with e.g. Australia....can you continue from here?\\nteacher : you need a verb next obviously so 'Australia....'?\\nstudent : Australia use 50% of electricity....\\nteacher : OK nearly...! Did Australia 'use' electricity? well obviously yes but what about coal?\",\n  'response': \"so ...'Australia used _____ to _______?\"},\n {'id': 'test_0270',\n  'dialog': \"student : Could you tell me another way of saying that they might be motivated to choose their profession by the high salary?\\nstudent : motivated by something is correct??\\nteacher : Yes OK sure...so 'motivated to' is a good idea...I was going to suggest a couple of phrases for this...yes: motivated to (do something) and 'motivated by (+noun) = good hold on\\nteacher : another good word here might be 'driven' OK?\",\n  'response': 'same structures - a bit stronger maybe but common'},\n {'id': 'test_0272',\n  'dialog': \"student : I want to continue\\nteacher : I understand that you are super busy!\\nstudent : but this time\\nstudent : I usually finish at 3 am\\nstudent : okay\\nstudent : let's talk about it later\\nteacher : Yes, have a think, ok?\\nstudent : okay\\nteacher : Oh, can we start a bit earlier on Tuesday, 9.45 am?\\nstudent : oh, what about the third lesson this week?\\nteacher : Ooh, thanks for reminding me!\\nteacher : Friday 10.30 am?\\nstudent : okay\\nteacher : Sure :)\",\n  'response': 'Have a nice day!'}]"},"metadata":{}}]},{"cell_type":"code","source":"a = []\nb = []\nfor data in teacher_reply[:5]:\n    a.append(data['dialog'])\nfor data in teacher_continuation[:5]:\n    b.append(data['dialog'])\n\ncontext = a + b","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:00:44.150167Z","iopub.execute_input":"2024-06-26T07:00:44.150541Z","iopub.status.idle":"2024-06-26T07:00:44.155980Z","shell.execute_reply.started":"2024-06-26T07:00:44.150511Z","shell.execute_reply":"2024-06-26T07:00:44.154973Z"},"trusted":true},"execution_count":20,"outputs":[]},{"cell_type":"code","source":"responses = [[\"\\\"That' s a great point! Reducing CO\\u2082 emissions is an important part of mitigating the effects of extremeweather. Have you thought about how we can implement this in our lessons? Are there specific activities or discussions that could highlight the connection between climate change and energy usage?\\\"\", \"\\\"That' s a great point! Expanding on this idea, how does reducing carbon emissions specifically impact the issue of extremeweather? Is there a connection we could explore, perhaps through a real-life example?\\\"\", \"\\\"Your observation about reducing carbon emissions as a solution for extremeweather is astute. While it'sa crucial issue that demands comprehensive action, let\\u2019s explore this further. Could you elaborate on specific measures that have been implemented or proposed to address this problem? Examining concrete actions provides a clearer picture of potential solutions.\\\"\", \"\\\"Excellent point about tackling the root cause of climate issues through CO\\u00b2 emission reduction. Expanding on specific actions individuals can take is crucial. Can you elaborate on what changes you envision in your daily life that could contribute to this larger goal?\\\"\"],\n[\"\\\"I understand you're feeling a bit chilly. Would you like to continue with the grammar practice or do you have any questions about the vocabulary word 'cold'?\\\"\", \"\\\"I noticed you're feeling a bit cold. Would you like to take a moment to rest or grab a drink to warm up your voice?\\\"\", \"\\\"That astute observation highlights the significance of considering various perspectives. To further explore this notion, let us delve into specific examples and analyze their applicability across different contexts. Your ability to identify such nuances is commendable, so let\\u2019s build on this thought together.\\\"\", \"\\\"Lovely work with your pronunciation, [student' s name]. I noticed you mentioned feeling cold. Is there anything specific you were practicing that might be making you feel chilly? We could take a quick break or move to something easier to review.\\\"\"],\n[\"\\\"I'm glad you found that interesting! Is there another relevant adjective you have in mind?\\\"\", \"\\\"I'm curious to hear your thoughts on how we can use \\\"within\\\" in different contexts. Do you think it'd be useful to practice identifying different examples of its usage?\\\"\", \"\\\"The term 'inction,' while often associated with academic achievements, has broader applications beyond just high grades. Think of it as a mark of high achievement or distinction in a specific field.\\\"\\n***Connecting to the student' s example:**  \\\"That example is certainly relevant. To truly appreciate the significance of your accomplishment, let\\u2019s explore the nuances of that word and its various interpretations.\\\" \\n*\\\"Let'ss explore this further and see how this distinction translates into valuable skills or knowledge. What specific skills have you developed or gained through your work?\\\"\", \"\\\"Great point! That brings up an interesting point about how language evolves and acquires new connotations over time. How does the context or specific situation contribute to the interpretation of similar phrases?\\\"\"],\n[\"\\\"I appreciate your perspective on using \\\"frank\\\" in certain situations. Let'd explore how to incorporate it effectively in our upcoming writing assignment.\\\"\", \"\\\"That'd be a great point! 'Within' can be used to describe something being confined or enclosed within something else, like a book within a box. Do you think we could explore some examples together?\\\"\", \"\\\"That is an astute observation. To further explore this point, could you elaborate on what specific aspects of 'innocence' are you interested in?\\\"\", \"\\\"Following your insightful observation, could you provide an instance where speaking truthfully might be particularly effective in achieving a desired outcome?\\\"\"],\n[\"\\\"I appreciate your perspective on this, student. To further explore this idea, how does playing games help children cope with difficult emotions?\\\"\", \"\\\"I noticed you used \\\"realize\\\" in your previous sentence. Can you expand on how playing games might help children deal with these emotions?\\\"\", \"\\\"That insight about adults and children' playing games for emotional wellbeing is astute. While it holds some truth, it could benefit from further exploration. What specific games or activities do children engage in that might contribute to their emotional resilience? Understanding the mechanisms and motivations behind these activities could provide valuable insights for both children and adults.\\\"\", \"\\\"Thanks for sharing your insights on the importance of games in emotional development. To delve deeper into this topic, could you elaborate on specific instances where children might find games particularly helpful in managing depression, loneliness, or boredom? This will help us understand the nuances of this issue and explore potential applications in a more targeted way.\\\"\"],\n[\"\\\"I appreciate your insightful perspective on the possible reasons why children didn't eat the marshmallows. Let'd explore this further. Do you have any other theories or specific examples that support your reasoning?\\\"\", \"\\\"That' s an interesting point! It makes me wonder... Could there be other factors influencing why some children didn't eat the marshmallows? Perhaps their interests or preferences differed from the research group, or maybe they were simply curious about the experiment and wanted more information.\\\"\", \"\\\"That\\u2019s an astute observation! To understand the underlying reasons why children didn\\u2019t eat the marshmallows, it\\u2019d be beneficial to explore the context and any prior experiences they may have had. Were there specific reasons they didn't enjoy them, or were there external factors influencing their choices?\\\"\", \"\\\"Thank you for sharing your thoughts! I appreciate you focusing on the age factor as a possible reason children weren't eating. Could this be something you've noticed in other situations where children refuse food? Are there any additional observations you might have that could shed light on this issue?\\\"\"],\n[\"\\\"That's right! I understand you have plans for after school. Is there anything specific you're looking forward to doing?\\\"\", \"\\\"Wonderful point! That leads me to think about how we can express timeframes in different contexts. Do you recall any specific examples of when you'd need to talk specifically about train timings?\\\"\", \"\\\"That point raises a valid observation. To further explore this, could you elaborate on how your plans align with the scheduled departure time of the train? Knowing the specific time it leaves can help us better understand the context and potential implications.\\\"\", \"\\\"Excellent point! The present progressive allows us to express ongoing arrangements and future plans, which is relevant when discussing your specific travel plans. Do you perhaps have any exciting destinations in your sights, or any activities you particularly look forward  to doing during your journey?\\\"\"],\n[\"\\\"Great point! Speed is an important aspect of your time management, so maintaining a sense of efficiency is crucial. How did you achieve your faster timeframe?\\\"\", \"\\\"Great point! That helps me understand your perspective better. Is there a specific situation where your time seemed to fly by quickly?\\\"\", \"\\\"Excellent point! It' s interesting to observe how our time management strategies differ. Your observation highlights the importance of analyzing and optimizing your own time more closely. Perhaps you could elaborate on specific areas where you spent more or less time than anticipated? This information would be invaluable in further developing effective time-management techniques.\\\"\", \"\\\"Excellent point about your observation! I'm curious to learn more about what specifically contributed (to your faster timeframe). Are there any activities or tasks that stood out as particularly fast today?\\\"\"],\n[\"\\\"Great point! You've grasped the main point of this section. Now, let' s move on to the next point and see how we use conjunctions with adjectives.\\\"\", \"\\\"Great point! The specific juxtaposition of adjectives and preposition is crucial. Could you elaborate on how they interact in this sentence?\\\"\", \"\\\"Excellent point! The use of adjectives to describe quality is indeed a fundamental aspect of English grammar. The key is to identify the specific adjective and then determine the appropriate preposition to convey the intended meaning. This ensures clarity and precision in communication.\\\"\", \"\\\"Excellent point about recognizing the importance of choosing the appropriate preposition with an adjective. Have you encountered any specific adjectives that pose challenges for you? Knowing which adjectives pair best with which preposition can definitely be tricky. How have you found navigating this aspect?\\\"\"],\n[\"\\\"Great point! Now, let'd focus on using this new vocabulary in a meaningful way. Can you tell me some examples of how you've used this word in your own writing or conversations?\\\"\", \"\\\"Great point about using \\\"within\\\"! That connects nicely to our recent discussion about describing actions happening \\\"around\\\" something. How do you think we can practice using this in a meaningful way?\\\"\", \"\\\"Thank you for sharing your insightful observation. To further explore this topic, could you elaborate on how this aligns with the broader objectives of the lesson? By providing specific examples and relevant analysis, you can enrich our understanding and ensure we're on track to achieve our learning goals.\\\"\", \"\\\"Excellent point about [relevant point]. To further enhance your understanding, we can explore [new topic] together, which will equip you with the tools to express this concept in various contexts.\\\"\"]]\n","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:00:45.863179Z","iopub.execute_input":"2024-06-26T07:00:45.863540Z","iopub.status.idle":"2024-06-26T07:00:45.876037Z","shell.execute_reply.started":"2024-06-26T07:00:45.863510Z","shell.execute_reply":"2024-06-26T07:00:45.874975Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"print(responses[3])","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:35:09.533729Z","iopub.execute_input":"2024-06-26T07:35:09.534104Z","iopub.status.idle":"2024-06-26T07:35:09.539078Z","shell.execute_reply.started":"2024-06-26T07:35:09.534075Z","shell.execute_reply":"2024-06-26T07:35:09.538107Z"},"trusted":true},"execution_count":36,"outputs":[{"name":"stdout","text":"['\"I appreciate your perspective on using \"frank\" in certain situations. Let\\'d explore how to incorporate it effectively in our upcoming writing assignment.\"', '\"That\\'d be a great point! \\'Within\\' can be used to describe something being confined or enclosed within something else, like a book within a box. Do you think we could explore some examples together?\"', '\"That is an astute observation. To further explore this point, could you elaborate on what specific aspects of \\'innocence\\' are you interested in?\"', '\"Following your insightful observation, could you provide an instance where speaking truthfully might be particularly effective in achieving a desired outcome?\"']\n","output_type":"stream"}]},{"cell_type":"code","source":"print(context[3])","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:35:06.660537Z","iopub.execute_input":"2024-06-26T07:35:06.661379Z","iopub.status.idle":"2024-06-26T07:35:06.665710Z","shell.execute_reply.started":"2024-06-26T07:35:06.661344Z","shell.execute_reply":"2024-06-26T07:35:06.664881Z"},"trusted":true},"execution_count":35,"outputs":[{"name":"stdout","text":"student : innocent??\nteacher : Right! and 'abrupt' = frank/direct\nteacher : If you're speak frankly/directly is it good?\nteacher : I mean if you speak\nstudent : It depends on situations, but I think it's generally good.\n","output_type":"stream"}]},{"cell_type":"code","source":"candidates = [\n    ['\"I understand you\\'re feeling a bit chilly. Would you like to continue with the grammar practice or do you have any questions about the vocabulary word \\'cold\\'?\"',\n '\"I noticed you\\'re feeling a bit cold. Would you like to take a moment to rest or grab a drink to warm up your voice?\"',\n '\"That astute observation highlights the significance of considering various perspectives. To further explore this notion, let us delve into specific examples and analyze their applicability across different contexts. Your ability to identify such nuances is commendable, so let’s build on this thought together.\"',\n '\"Lovely work with your pronunciation, [student\\'s name]. I noticed you mentioned feeling cold. Is there anything specific you were practicing that might be making you feel chilly? We could take a quick break or move to something easier to review.\"'],\n    ['\"That\\' s a great point! Reducing CO₂ emissions is an important part of mitigating the effects of extremeweather. Have you thought about how we can implement this in our lessons? Are there specific activities or discussions that could highlight the connection between climate change and energy usage?\"', \n     '\"That\\' s a great point! Expanding on this idea, how does reducing carbon emissions specifically impact the issue of extremeweather? Is there a connection we could explore, perhaps through a real-life example?\"', \n     '\"Your observation about reducing carbon emissions as a solution for extremeweather is astute. While it\\'sa crucial issue that demands comprehensive action, let’s explore this further. Could you elaborate on specific measures that have been implemented or proposed to address this problem? Examining concrete actions provides a clearer picture of potential solutions.\"', \n     '\"Excellent point about tackling the root cause of climate issues through CO² emission reduction. Expanding on specific actions individuals can take is crucial. Can you elaborate on what changes you envision in your daily life that could contribute to this larger goal?\"']\n]","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:03:03.819963Z","iopub.execute_input":"2024-06-26T07:03:03.820360Z","iopub.status.idle":"2024-06-26T07:03:03.826572Z","shell.execute_reply.started":"2024-06-26T07:03:03.820324Z","shell.execute_reply":"2024-06-26T07:03:03.825626Z"},"trusted":true},"execution_count":28,"outputs":[]},{"cell_type":"code","source":"dialogue = [\n    ['Which is easy, because you can use my / his / your etc. and not think about articles!',\n     'Only 10 minutes left!',\n     'I know, we can finish early if you are getting cold?',\n     \"I'm really cold\"],\n    [\"Do you say them as 'extreme weather', right?\",\n    \"Absolutely yes the two words together: so, the three collocations for the consequences are; rising sea levels/melting ice caps and extreme weather - last category: what about the solutions?!\",\n    \"It's reducing CO2 emissions. For example, using solar/wind energy instead of fossil fuels\"]\n]","metadata":{"execution":{"iopub.status.busy":"2024-06-26T06:59:47.189328Z","iopub.execute_input":"2024-06-26T06:59:47.190003Z","iopub.status.idle":"2024-06-26T06:59:47.196740Z","shell.execute_reply.started":"2024-06-26T06:59:47.189971Z","shell.execute_reply":"2024-06-26T06:59:47.195838Z"},"trusted":true},"execution_count":15,"outputs":[{"execution_count":15,"output_type":"execute_result","data":{"text/plain":"[['Which is easy, because you can use my / his / your etc. and not think about articles!',\n  'Only 10 minutes left!',\n  'I know, we can finish early if you are getting cold?',\n  \"I'm really cold\"],\n [\"Do you say them as 'extreme weather', right?\",\n  'Absolutely yes the two words together: so, the three collocations for the consequences are; rising sea levels/melting ice caps and extreme weather - last category: what about the solutions?!',\n  \"It's reducing CO2 emissions. For example, using solar/wind energy instead of fossil fuels\"]]"},"metadata":{}}]},{"cell_type":"code","source":"import torch\nfrom transformers import AutoTokenizer, AutoModel\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.decomposition import PCA\n\n# Load the pre-trained DialoGPT model and tokenizer\nmodel_name = \"microsoft/DialoGPT-medium\"\ntokenizer = AutoTokenizer.from_pretrained(model_name)\nmodel = AutoModel.from_pretrained(model_name)\n\n# Move model to GPU if available\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel.to(device)","metadata":{"execution":{"iopub.status.busy":"2024-06-26T06:52:45.031151Z","iopub.execute_input":"2024-06-26T06:52:45.032042Z","iopub.status.idle":"2024-06-26T06:52:56.529613Z","shell.execute_reply.started":"2024-06-26T06:52:45.032009Z","shell.execute_reply":"2024-06-26T06:52:56.528682Z"},"trusted":true},"execution_count":8,"outputs":[{"output_type":"display_data","data":{"text/plain":"tokenizer_config.json:   0%|          | 0.00/614 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"7164d2b9482c41f18bee7a9d383dc0e2"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"vocab.json:   0%|          | 0.00/1.04M [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"06823b3cb1174cd3b40e9a0633e2a60d"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"merges.txt:   0%|          | 0.00/456k [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"61b9f9dfa3894eb7a38314f5451ff48e"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"config.json:   0%|          | 0.00/642 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"1c8c90b051554a449525ed5b83d5019c"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"pytorch_model.bin:   0%|          | 0.00/863M [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"4d7eaa06d80b4e79ba42bf244d2f2da2"}},"metadata":{}},{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/torch/_utils.py:831: UserWarning: TypedStorage is deprecated. It will be removed in the future and UntypedStorage will be the only storage class. This should only matter to you if you are using storages directly.  To access UntypedStorage directly, use tensor.untyped_storage() instead of tensor.storage()\n  return self.fget.__get__(instance, owner)()\n","output_type":"stream"},{"execution_count":8,"output_type":"execute_result","data":{"text/plain":"GPT2Model(\n  (wte): Embedding(50257, 1024)\n  (wpe): Embedding(1024, 1024)\n  (drop): Dropout(p=0.1, inplace=False)\n  (h): ModuleList(\n    (0-23): 24 x GPT2Block(\n      (ln_1): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n      (attn): GPT2Attention(\n        (c_attn): Conv1D()\n        (c_proj): Conv1D()\n        (attn_dropout): Dropout(p=0.1, inplace=False)\n        (resid_dropout): Dropout(p=0.1, inplace=False)\n      )\n      (ln_2): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n      (mlp): GPT2MLP(\n        (c_fc): Conv1D()\n        (c_proj): Conv1D()\n        (act): NewGELUActivation()\n        (dropout): Dropout(p=0.1, inplace=False)\n      )\n    )\n  )\n  (ln_f): LayerNorm((1024,), eps=1e-05, elementwise_affine=True)\n)"},"metadata":{}}]},{"cell_type":"code","source":"print(responses[8])","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:47:39.472956Z","iopub.execute_input":"2024-06-26T07:47:39.473825Z","iopub.status.idle":"2024-06-26T07:47:39.478249Z","shell.execute_reply.started":"2024-06-26T07:47:39.473790Z","shell.execute_reply":"2024-06-26T07:47:39.477393Z"},"trusted":true},"execution_count":48,"outputs":[{"name":"stdout","text":"['\"Great point! You\\'ve grasped the main point of this section. Now, let\\' s move on to the next point and see how we use conjunctions with adjectives.\"', '\"Great point! The specific juxtaposition of adjectives and preposition is crucial. Could you elaborate on how they interact in this sentence?\"', '\"Excellent point! The use of adjectives to describe quality is indeed a fundamental aspect of English grammar. The key is to identify the specific adjective and then determine the appropriate preposition to convey the intended meaning. This ensures clarity and precision in communication.\"', '\"Excellent point about recognizing the importance of choosing the appropriate preposition with an adjective. Have you encountered any specific adjectives that pose challenges for you? Knowing which adjectives pair best with which preposition can definitely be tricky. How have you found navigating this aspect?\"']\n","output_type":"stream"}]},{"cell_type":"code","source":"print(context[8])","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:47:42.531317Z","iopub.execute_input":"2024-06-26T07:47:42.532170Z","iopub.status.idle":"2024-06-26T07:47:42.536541Z","shell.execute_reply.started":"2024-06-26T07:47:42.532138Z","shell.execute_reply":"2024-06-26T07:47:42.535680Z"},"trusted":true},"execution_count":49,"outputs":[{"name":"stdout","text":"student : right?\nstudent : yes,, thanks\nteacher : well <STUDENT> - we just usually use the verb 'be' with these adjectives\nstudent : wow... I got it\nteacher : like e.g. rich/poor/old etc - not because it's a feeling\nteacher : So in a way this bit is the easy bit (the verb I mean) - the thing you have to get right is the adjective itself and the preposition that comes next you see?\n","output_type":"stream"}]},{"cell_type":"code","source":"# 3 model\nimport torch\nfrom transformers import AutoTokenizer, AutoModel\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.decomposition import PCA\n\n# Define the dialogues context\ndialogues = [\n    ['Which is easy, because you can use my / his / your etc. and not think about articles!',\n     'Only 10 minutes left!',\n     'I know, we can finish early if you are getting cold?',\n     \"I'm really cold\"],\n    [\"Do you say them as 'extreme weather', right?\",\n    \"Absolutely yes the two words together: so, the three collocations for the consequences are; rising sea levels/melting ice caps and extreme weather - last category: what about the solutions?!\",\n    \"It's reducing CO2 emissions. For example, using solar/wind energy instead of fossil fuels\"],\n    [\"innocent??\", \"Right! and 'abrupt' = frank/direct\", \"If you're speak frankly/directly is it good?\", \n    \"I mean if you speak\", \"It depends on situations, but I think it's generally good.\"],\n    [\"But most importantly, there could have been other reasons behind children not eating / eating the marshmallow\",\n    \"For example, they might not have liked marshmallows!\",\n    \"Or they could have wanted to please the researchers\",\n    \"Yes that's true\", \"What do you think could have been the reasons?\", \"Maybe some of them were too young?\",\"Yes, exactly!\"],\n    [\"right?\", \"yes,, thanks\", \"well <STUDENT> - we just usually use the verb 'be' with these adjectives\", \"wow... I got it\",\n    \"like e.g. rich/poor/old etc - not because it's a feeling\", \"So in a way this bit is the easy bit (the verb I mean) - the thing you have to get right is the adjective itself and the preposition that comes next you see?\"]\n    \n]\n\n# Define candidate responses\ncandidates = [\n    ['\"I understand you\\'re feeling a bit chilly. Would you like to continue with the grammar practice or do you have any questions about the vocabulary word \\'cold\\'?\"',\n     '\"I noticed you\\'re feeling a bit cold. Would you like to take a moment to rest or grab a drink to warm up your voice?\"',\n     '\"That astute observation highlights the significance of considering various perspectives. To further explore this notion, let us delve into specific examples and analyze their applicability across different contexts. Your ability to identify such nuances is commendable, so let’s build on this thought together.\"',\n     '\"Lovely work with your pronunciation, [student\\'s name]. I noticed you mentioned feeling cold. Is there anything specific you were practicing that might be making you feel chilly? We could take a quick break or move to something easier to review.\"'],\n    ['\"That\\' s a great point! Reducing CO₂ emissions is an important part of mitigating the effects of extreme weather. Have you thought about how we can implement this in our lessons? Are there specific activities or discussions that could highlight the connection between climate change and energy usage?\"', \n     '\"That\\' s a great point! Expanding on this idea, how does reducing carbon emissions specifically impact the issue of extreme weather? Is there a connection we could explore, perhaps through a real-life example?\"', \n     '\"Your observation about reducing carbon emissions as a solution for extreme weather is astute. While it\\'s a crucial issue that demands comprehensive action, let’s explore this further. Could you elaborate on specific measures that have been implemented or proposed to address this problem? Examining concrete actions provides a clearer picture of potential solutions.\"', \n     '\"Excellent point about tackling the root cause of climate issues through CO² emission reduction. Expanding on specific actions individuals can take is crucial. Can you elaborate on what changes you envision in your daily life that could contribute to this larger goal?\"'],\n    ['\"I appreciate your perspective on using \"frank\" in certain situations. Let\\'d explore how to incorporate it effectively in our upcoming writing assignment.\"', \n     '\"That\\'d be a great point! \\'Within\\' can be used to describe something being confined or enclosed within something else, like a book within a box. Do you think we could explore some examples together?\"', \n     '\"That is an astute observation. To further explore this point, could you elaborate on what specific aspects of \\'innocence\\' are you interested in?\"', \n     '\"Following your insightful observation, could you provide an instance where speaking truthfully might be particularly effective in achieving a desired outcome?\"'],\n    ['\"I appreciate your insightful perspective on the possible reasons why children didn\\'t eat the marshmallows. Let\\'d explore this further. Do you have any other theories or specific examples that support your reasoning?\"', \n     '\"That\\' s an interesting point! It makes me wonder... Could there be other factors influencing why some children didn\\'t eat the marshmallows? Perhaps their interests or preferences differed from the research group, or maybe they were simply curious about the experiment and wanted more information.\"', \n     '\"That’s an astute observation! To understand the underlying reasons why children didn’t eat the marshmallows, it’d be beneficial to explore the context and any prior experiences they may have had. Were there specific reasons they didn\\'t enjoy them, or were there external factors influencing their choices?\"', \n     '\"Thank you for sharing your thoughts! I appreciate you focusing on the age factor as a possible reason children weren\\'t eating. Could this be something you\\'ve noticed in other situations where children refuse food? Are there any additional observations you might have that could shed light on this issue?\"'],\n    ['\"Great point! You\\'ve grasped the main point of this section. Now, let\\' s move on to the next point and see how we use conjunctions with adjectives.\"', \n     '\"Great point! The specific juxtaposition of adjectives and preposition is crucial. Could you elaborate on how they interact in this sentence?\"', \n     '\"Excellent point! The use of adjectives to describe quality is indeed a fundamental aspect of English grammar. The key is to identify the specific adjective and then determine the appropriate preposition to convey the intended meaning. This ensures clarity and precision in communication.\"', \n     '\"Excellent point about recognizing the importance of choosing the appropriate preposition with an adjective. Have you encountered any specific adjectives that pose challenges for you? Knowing which adjectives pair best with which preposition can definitely be tricky. How have you found navigating this aspect?\"']\n]\n\ndef get_embeddings(text_list):\n    embeddings = []\n    for text in text_list:\n        inputs = tokenizer(text, return_tensors=\"pt\").to(device)\n        with torch.no_grad():\n            outputs = model(**inputs)\n        hidden_states = outputs.last_hidden_state\n        # Average the token embeddings to get a single sentence embedding\n        sentence_embedding = hidden_states.mean(dim=1).squeeze().cpu().numpy()\n        embeddings.append(sentence_embedding)\n    return np.array(embeddings)\n\n# Function to process dialogues and candidates\ndef process_dialogues_and_candidates(dialogues, candidates):\n    aggregated_dialogue_embeddings = []\n    all_candidate_embeddings = []\n\n    for dialogue in dialogues:\n        # Tokenize and embed each turn of the dialogue\n        dialogue_embeddings = get_embeddings(dialogue)\n        # Aggregate dialogue embeddings\n        aggregated_dialogue_embedding = np.mean(dialogue_embeddings, axis=0)\n        aggregated_dialogue_embeddings.append(aggregated_dialogue_embedding)\n\n    for candidate_set in candidates:\n        # Get embeddings for each candidate response\n        candidate_embeddings = get_embeddings(candidate_set)\n        all_candidate_embeddings.append(candidate_embeddings)\n    \n    return np.array(aggregated_dialogue_embeddings), all_candidate_embeddings\n\n# Process dialogues and candidates\naggregated_dialogue_embeddings, all_candidate_embeddings = process_dialogues_and_candidates(dialogues, candidates)\n\n# Flatten the candidate embeddings for PCA\nflattened_candidate_embeddings = np.concatenate(all_candidate_embeddings, axis=0)\n\n# Combine dialogue embeddings and flattened candidate embeddings for visualization\nall_embeddings = np.concatenate((aggregated_dialogue_embeddings, flattened_candidate_embeddings), axis=0)\n\n# Apply PCA for dimensionality reduction\npca = PCA(n_components=2)\npca_result = pca.fit_transform(all_embeddings)\n\n# Separate PCA results for dialogues and candidates\ndialogue_pca = pca_result[:len(aggregated_dialogue_embeddings)]\ncandidate_pca = pca_result[len(aggregated_dialogue_embeddings):]\n\n# Determine boundaries for each candidate set\nboundaries = [0]\nfor candidate_set in all_candidate_embeddings:\n    boundaries.append(boundaries[-1] + len(candidate_set))\n\n# Plotting\nplt.figure(figsize=(10, 8))\nplt.scatter(dialogue_pca[:, 0], dialogue_pca[:, 1], color='blue', label='Dialogues')\n\ncolors = ['red', 'green', 'yellow', 'gray', 'brown']  # Different colors for each dialogue's candidates\nshapes = ['o', 's', 'x', 'v', '<']  # Different shapes for each dialogue's candidates\n\nfor i in range(len(dialogues)):\n    start_idx = boundaries[i]\n    end_idx = boundaries[i + 1]\n    plt.scatter(candidate_pca[start_idx:end_idx, 0], candidate_pca[start_idx:end_idx, 1], color=colors[i], marker=shapes[i], label=f'Candidates for Dialogue {i+1}')\n    for j in range(start_idx, end_idx):\n        plt.annotate(f'Candidate {j - start_idx + 1}', (candidate_pca[j, 0], candidate_pca[j, 1]), fontsize=8, color=colors[i])\n\nfor i, txt in enumerate(dialogue_pca):\n    plt.annotate(f'Dialogue {i+1}', (dialogue_pca[i, 0], dialogue_pca[i, 1]), fontsize=8, color='blue')\n\nplt.title('PCA Visualization of Dialogues and Candidates')\nplt.xlabel('PCA Component 1')\nplt.ylabel('PCA Component 2')\nplt.legend()\nplt.grid(True)\nplt.savefig('embedding_visualiation.pdf', format='pdf')\nplt.show()\n","metadata":{"execution":{"iopub.status.busy":"2024-06-26T07:55:42.328791Z","iopub.execute_input":"2024-06-26T07:55:42.329157Z","iopub.status.idle":"2024-06-26T07:55:44.405024Z","shell.execute_reply.started":"2024-06-26T07:55:42.329133Z","shell.execute_reply":"2024-06-26T07:55:44.404055Z"},"trusted":true},"execution_count":52,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1000x800 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from transformers import AutoModel, AutoTokenizer\nimport torch\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom sklearn.manifold import TSNE\nfrom sklearn.decomposition import PCA\nimport seaborn as sns\n\n# Load the pre-trained DialoGPT model and tokenizer\nmodel_name = \"microsoft/DialoGPT-medium\"\ntokenizer = AutoTokenizer.from_pretrained(model_name)\nmodel = AutoModel.from_pretrained(model_name)\n\n# Move model to GPU if available\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel.to(device)\n\ndef encode_text(text):\n    tokens = tokenizer(text, return_tensors='pt')\n    tokens = {key: val.to(device) for key, val in tokens.items()}\n    \n    with torch.no_grad():\n        outputs = model(**tokens)\n        hidden_states = outputs.last_hidden_state\n        vector = hidden_states.mean(dim=1).cpu().numpy()\n        \n    return vector\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Example context-response pairs\ncontext_response_pairs = [\n    (\"What are your plans for the weekend?\", \"I plan to go hiking.\"),\n    (\"How are you feeling today?\", \"I'm feeling great, thanks!\"),\n    (\"What's your favorite movie?\", \"I love watching Inception.\"),\n    (\"Do you like to travel?\", \"Yes, I enjoy visiting new places.\"),\n    (\"What's your favorite book?\", \"I really like 'To Kill a Mockingbird'.\"),\n]\n\n# Encode the contexts and responses\ncontexts = [pair[0] for pair in context_response_pairs]\nresponses = [pair[1] for pair in context_response_pairs]\n\ncontext_vectors = np.vstack([encode_text(context) for context in contexts])\nresponse_vectors = np.vstack([encode_text(response) for response in responses])\n\n# Combine context and response vectors for visualization\nall_vectors = np.vstack((context_vectors, response_vectors))","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Perform dimensionality reduction using t-SNE\ntsne = TSNE(n_components=2, perplexity=2, random_state=42)\nreduced_vectors = tsne.fit_transform(all_vectors)\n\n# Separate the reduced vectors back into contexts and responses\ncontext_reduced = reduced_vectors[:len(contexts)]\nresponse_reduced = reduced_vectors[len(contexts):]\n\n# Plot the 2D visualization with enhanced aesthetics\nplt.figure(figsize=(12, 8))\n\n# Plot contexts\nplt.scatter(context_reduced[:, 0], context_reduced[:, 1], c='blue', label='Contexts', marker='o', s=100)\n\n# Plot responses\nplt.scatter(response_reduced[:, 0], response_reduced[:, 1], c='red', label='Responses', marker='s', s=100)\n\n# Annotate points with their context/response text\nfor i, text in enumerate(contexts):\n    plt.annotate(text, (context_reduced[i, 0], context_reduced[i, 1]), fontsize=10, color='blue', ha='right', va='bottom')\nfor i, text in enumerate(responses):\n    plt.annotate(text, (response_reduced[i, 0], response_reduced[i, 1]), fontsize=10, color='red', ha='left', va='top')\n\n# Title and labels\nplt.title(\"2D Visualization of Context and Response Embeddings\", fontsize=14)\nplt.xlabel(\"Dimension 1\", fontsize=12)\nplt.ylabel(\"Dimension 2\", fontsize=12)\nplt.legend()\n\n# Adjust plot layout\nplt.tight_layout()\n\n# Save the plot as PDF\nplt.savefig('2d_visualization.pdf', format='pdf')\n\nplt.show()\n\n# Compute similarity matrix\ndef compute_similarity_matrix(context_vectors, response_vectors):\n    similarity_matrix = np.dot(context_vectors, response_vectors.T) / (\n        np.linalg.norm(context_vectors, axis=1, keepdims=True) * np.linalg.norm(response_vectors, axis=1, keepdims=True).T\n    )\n    return similarity_matrix\n\n# Compute similarity matrix\nsimilarity_matrix = compute_similarity_matrix(context_vectors, response_vectors)\n\n# Plot similarity heatmap with enhanced aesthetics\nplt.figure(figsize=(10, 8))\nsns.heatmap(similarity_matrix, annot=True, fmt=\".2f\", cmap='coolwarm', cbar=True, xticklabels=responses, yticklabels=contexts)\n\n# Title and labels\nplt.title(\"Similarity Heatmap between Contexts and Responses\", fontsize=14)\nplt.xlabel(\"Responses\", fontsize=12)\nplt.ylabel(\"Contexts\", fontsize=12)\n\n# Adjust plot layout\nplt.tight_layout()\n\n# Save the plot as PDF\nplt.savefig('similarity_heatmap.pdf', format='pdf')\n\n\nplt.show()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}