IMDB_Classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "IMDB_Classification.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "machine_shape": "hm",
      "mount_file_id": "1v5avVmMvSjbz4ehHag_Rd67M1jf4Xwh4",
      "authorship_tag": "ABX9TyPeIzb+Okou4MKkDroEL5yv",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "fb83093c59ab4094abb616135f987365": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_f88144cea24146638da0e7b05256fe2e",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_d67c0f177280476e95440977b63e4647",
              "IPY_MODEL_6f73594f5a32496d80d19637ed97b0e4"
            ]
          }
        },
        "f88144cea24146638da0e7b05256fe2e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "d67c0f177280476e95440977b63e4647": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_80b032cdf13244af976a8ad8f4350fd1",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 898823,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 898823,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_dd835071c2c1415597ee179c248dc0dc"
          }
        },
        "6f73594f5a32496d80d19637ed97b0e4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_1c60f4aab7884a949dd59508349857a3",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 899k/899k [00:00&lt;00:00, 1.44MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_db4541fb2a1b4ae9811b131907030eb7"
          }
        },
        "80b032cdf13244af976a8ad8f4350fd1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "dd835071c2c1415597ee179c248dc0dc": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "1c60f4aab7884a949dd59508349857a3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "db4541fb2a1b4ae9811b131907030eb7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "83e08cf4d66f4b2d9ebd1161ec6808e2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_90fd4c0f8a35494584fa836c27ccab49",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_93a26925dc974a6485f57a60166573d1",
              "IPY_MODEL_0e2e608912484a55a04039c6aeafb184"
            ]
          }
        },
        "90fd4c0f8a35494584fa836c27ccab49": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "93a26925dc974a6485f57a60166573d1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_2fdd9c13409b4bf68124bcef8ad6d5d2",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 456318,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 456318,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_84f6b61b10484f9e9df8178432c1d64a"
          }
        },
        "0e2e608912484a55a04039c6aeafb184": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_526ec6c604b54d82b6a0de4b41a17254",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 456k/456k [00:00&lt;00:00, 2.56MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_2e0dc3c8376c4b77b315fa64c9cf6514"
          }
        },
        "2fdd9c13409b4bf68124bcef8ad6d5d2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "84f6b61b10484f9e9df8178432c1d64a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "526ec6c604b54d82b6a0de4b41a17254": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "2e0dc3c8376c4b77b315fa64c9cf6514": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "da78fea9705b4aceaea8acc37a082a6d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_3423a063293e4bbcaeb7f001f8eec7db",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_38abdee533e04ebfa618920e47bae2f0",
              "IPY_MODEL_7240936095ca4e4cb8b2b25d6f42ba5d"
            ]
          }
        },
        "3423a063293e4bbcaeb7f001f8eec7db": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "38abdee533e04ebfa618920e47bae2f0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_cbdde285fb5a43bfa9819e1ba439f4df",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 481,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 481,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_566c0e129a6c40589cc07c318b300bfb"
          }
        },
        "7240936095ca4e4cb8b2b25d6f42ba5d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_2c5c614f27444456aaba7ed5b8a2d6f8",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 481/481 [00:00&lt;00:00, 1.68kB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_745d5685826e4f8f98b4fe26250cbcc8"
          }
        },
        "cbdde285fb5a43bfa9819e1ba439f4df": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "566c0e129a6c40589cc07c318b300bfb": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "2c5c614f27444456aaba7ed5b8a2d6f8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "745d5685826e4f8f98b4fe26250cbcc8": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "b26c9fc5bd644302badb167b83db2c1f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_cb302d8eb9c547c38c02ef4199fcf24b",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_044d872e86f040c68dc047a884a4fd0f",
              "IPY_MODEL_fb820510cb4c40be9bfde0758bf01fa4"
            ]
          }
        },
        "cb302d8eb9c547c38c02ef4199fcf24b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "044d872e86f040c68dc047a884a4fd0f": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_2a511b388c4a4664b4a0f009d421fc98",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 501200538,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 501200538,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6e5859cfa09845f8b28244fcaa283025"
          }
        },
        "fb820510cb4c40be9bfde0758bf01fa4": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_e7bd0d2e2ee74e52a3c599c102cc5ece",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 501M/501M [02:02&lt;00:00, 4.08MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_87da97ff4fa94cf1827e2f11cf36b439"
          }
        },
        "2a511b388c4a4664b4a0f009d421fc98": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "6e5859cfa09845f8b28244fcaa283025": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "e7bd0d2e2ee74e52a3c599c102cc5ece": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "87da97ff4fa94cf1827e2f11cf36b439": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nidharap/Notebooks/blob/master/IMDB_Classification.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zvw06g7ksdxx",
        "colab_type": "text"
      },
      "source": [
        "# This notebook intends to fine-tune the RoBERTa architecture on the IMDB 50K movie reviews dataset.\n",
        "\n",
        "##### Most of the code was inspired from https://www.curiousily.com/posts/sentiment-analysis-with-bert-and-hugging-face-using-pytorch-and-python/"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kIONNP_NwkAK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        },
        "outputId": "29e9c28c-e959-4f60-b965-b896b9e5b627"
      },
      "source": [
        "!pip install -q -U watermark\n",
        "!pip install -qq transformers"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[K     |████████████████████████████████| 890kB 3.6MB/s \n",
            "\u001b[K     |████████████████████████████████| 1.1MB 16.4MB/s \n",
            "\u001b[K     |████████████████████████████████| 890kB 25.6MB/s \n",
            "\u001b[K     |████████████████████████████████| 3.0MB 32.1MB/s \n",
            "\u001b[?25h  Building wheel for sacremoses (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xrf5YSkXHoYT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 138
        },
        "outputId": "be267602-8c4c-4517-e15e-dd8e37b8a48b"
      },
      "source": [
        "%reload_ext watermark\n",
        "%watermark -v -p numpy,pandas,torch,transformers"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPython 3.6.9\n",
            "IPython 5.5.0\n",
            "\n",
            "numpy 1.18.5\n",
            "pandas 1.0.5\n",
            "torch 1.6.0+cu101\n",
            "transformers 3.1.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-SLArSk9H6j0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        },
        "outputId": "ad2177cb-e446-43cc-d51f-688fb6112e14"
      },
      "source": [
        "import transformers\n",
        "from transformers import AdamW, get_linear_schedule_with_warmup\n",
        "from transformers import RobertaModel, RobertaTokenizer\n",
        "import torch\n",
        "import numpy as np\n",
        "import os\n",
        "import time\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from pylab import rcParams\n",
        "import matplotlib.pyplot as plt\n",
        "from matplotlib import rc\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from collections import defaultdict\n",
        "from textwrap import wrap\n",
        "from torch import nn, optim\n",
        "import time\n",
        "from bs4 import BeautifulSoup\n",
        "from torch.utils.data import Dataset, DataLoader\n",
        "import glob\n",
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format='retina'\n",
        "sns.set(style='whitegrid', palette='muted', font_scale=1.2)\n",
        "HAPPY_COLORS_PALETTE = [\"#01BEFE\", \"#FFDD00\", \"#FF7D00\", \"#FF006D\", \"#ADFF02\", \"#8F00FF\"]\n",
        "sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))\n",
        "rcParams['figure.figsize'] = 12, 8\n",
        "RANDOM_SEED = 42\n",
        "np.random.seed(RANDOM_SEED)\n",
        "torch.manual_seed(RANDOM_SEED)\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
            "  import pandas.util.testing as tm\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E5yc7vpY7cGj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 484
        },
        "outputId": "a1749674-94f6-4e7f-e54d-27dd8ddf5c66"
      },
      "source": [
        "from tensorflow.python.client import device_lib\n",
        "device_lib.list_local_devices()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[name: \"/device:CPU:0\"\n",
              " device_type: \"CPU\"\n",
              " memory_limit: 268435456\n",
              " locality {\n",
              " }\n",
              " incarnation: 11781244283972302971, name: \"/device:XLA_CPU:0\"\n",
              " device_type: \"XLA_CPU\"\n",
              " memory_limit: 17179869184\n",
              " locality {\n",
              " }\n",
              " incarnation: 2265414263351774177\n",
              " physical_device_desc: \"device: XLA_CPU device\", name: \"/device:XLA_GPU:0\"\n",
              " device_type: \"XLA_GPU\"\n",
              " memory_limit: 17179869184\n",
              " locality {\n",
              " }\n",
              " incarnation: 5209877246765252818\n",
              " physical_device_desc: \"device: XLA_GPU device\", name: \"/device:GPU:0\"\n",
              " device_type: \"GPU\"\n",
              " memory_limit: 15177091648\n",
              " locality {\n",
              "   bus_id: 1\n",
              "   links {\n",
              "   }\n",
              " }\n",
              " incarnation: 4823337555824421073\n",
              " physical_device_desc: \"device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\"]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "02Uc0VBRqDWx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "4753cacc-7361-457c-88a2-681189289740"
      },
      "source": [
        "#we will load the dataset from google drive\n",
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XDI_Tc6qtGgs",
        "colab_type": "text"
      },
      "source": [
        "#Dataset\n",
        "### The dataset can be downloaded from multiple places , but I used this - https://www.kaggle.com/atulanandjha/imdb-50k-movie-reviews-test-your-bert"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4GJLD33rrLqZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "b84843ce-06c8-407a-dca5-2e7ab214ce13"
      },
      "source": [
        "df = pd.read_csv(\"/content/drive/My Drive/Datasets/IMDBMovieReviews/train.csv\")\n",
        "df_test = pd.read_csv(\"/content/drive/My Drive/Datasets/IMDBMovieReviews/test.csv\")\n",
        "df['label_encoded'] = df['sentiment'].apply(lambda x: 0 if x=='neg' else 1)\n",
        "df_test['label_encoded'] = df_test['sentiment'].apply(lambda x: 0 if x=='neg' else 1)\n",
        "df['review_len'] = df['text'].apply(lambda x : len(x.split()))\n",
        "class_names = ['neg', 'pos']\n",
        "\n",
        "#remove html tags\n",
        "df['text'] = df['text'].apply(lambda x: BeautifulSoup(x, \"lxml\").text)\n",
        "df_test['text'] = df_test['text'].apply(lambda x: BeautifulSoup(x, \"lxml\").text)\n",
        "\n",
        "#separate the train set into train and val\n",
        "df_train, df_val = train_test_split(df, test_size=0.2, random_state=RANDOM_SEED)\n",
        "\n",
        "df.shape, df_train.shape, df_val.shape, df_test.shape"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((25000, 4), (20000, 4), (5000, 4), (25000, 3))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "46M1rHODwcG9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 505
        },
        "outputId": "863db992-8b5b-4902-ac2b-7de7521048cd"
      },
      "source": [
        "df_train['review_len'].hist()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f6ea00bc2e8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 735,
              "height": 471
            }
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "outj5oXKIIkR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "PRE_TRAINED_MODEL_NAME = 'roberta-base' #'bert-base-cased'\n",
        "MAX_LEN = 512     #for the most part, things are below 500, lets set it to max len possible in bert\n",
        "\n",
        "TIMESTR = time.strftime(\"%Y%m%d-%H%M%S\")\n",
        "\n",
        "MODEL_SAVE_NAME = \"imdb_movie_large_roberta_state_{}\".format(TIMESTR)\n",
        "\n",
        "LOAD_PREV_MODEL = False     #set to True to load a trained model from local "
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j2fKuE91H1KG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#TO LOAD MODEL , Lets override\n",
        "LOAD_PREV_MODEL = True\n",
        "MODEL_SAVE_NAME = \"imdb_movie_large_roberta_state_20200905-154517\""
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kDI5pgJmIMHB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 214,
          "referenced_widgets": [
            "fb83093c59ab4094abb616135f987365",
            "f88144cea24146638da0e7b05256fe2e",
            "d67c0f177280476e95440977b63e4647",
            "6f73594f5a32496d80d19637ed97b0e4",
            "80b032cdf13244af976a8ad8f4350fd1",
            "dd835071c2c1415597ee179c248dc0dc",
            "1c60f4aab7884a949dd59508349857a3",
            "db4541fb2a1b4ae9811b131907030eb7",
            "83e08cf4d66f4b2d9ebd1161ec6808e2",
            "90fd4c0f8a35494584fa836c27ccab49",
            "93a26925dc974a6485f57a60166573d1",
            "0e2e608912484a55a04039c6aeafb184",
            "2fdd9c13409b4bf68124bcef8ad6d5d2",
            "84f6b61b10484f9e9df8178432c1d64a",
            "526ec6c604b54d82b6a0de4b41a17254",
            "2e0dc3c8376c4b77b315fa64c9cf6514",
            "da78fea9705b4aceaea8acc37a082a6d",
            "3423a063293e4bbcaeb7f001f8eec7db",
            "38abdee533e04ebfa618920e47bae2f0",
            "7240936095ca4e4cb8b2b25d6f42ba5d",
            "cbdde285fb5a43bfa9819e1ba439f4df",
            "566c0e129a6c40589cc07c318b300bfb",
            "2c5c614f27444456aaba7ed5b8a2d6f8",
            "745d5685826e4f8f98b4fe26250cbcc8",
            "b26c9fc5bd644302badb167b83db2c1f",
            "cb302d8eb9c547c38c02ef4199fcf24b",
            "044d872e86f040c68dc047a884a4fd0f",
            "fb820510cb4c40be9bfde0758bf01fa4",
            "2a511b388c4a4664b4a0f009d421fc98",
            "6e5859cfa09845f8b28244fcaa283025",
            "e7bd0d2e2ee74e52a3c599c102cc5ece",
            "87da97ff4fa94cf1827e2f11cf36b439"
          ]
        },
        "outputId": "1cfdefb3-10c3-44c7-d49a-85054d0f445a"
      },
      "source": [
        "tokenizer = RobertaTokenizer.from_pretrained(PRE_TRAINED_MODEL_NAME)\n",
        "bert_model = RobertaModel.from_pretrained(PRE_TRAINED_MODEL_NAME)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "fb83093c59ab4094abb616135f987365",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=898823.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "83e08cf4d66f4b2d9ebd1161ec6808e2",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=456318.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "da78fea9705b4aceaea8acc37a082a6d",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=481.0, style=ProgressStyle(description_…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "b26c9fc5bd644302badb167b83db2c1f",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=501200538.0, style=ProgressStyle(descri…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "75a-9qLYIMWp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "MAX_LEN = 512"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LdVfm8WeItu3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class ReviewsDataset(Dataset):\n",
        "  '''\n",
        "  custom pytorch dataset class\n",
        "  '''\n",
        "  def __init__(self, reviews, targets, tokenizer, max_len):\n",
        "    self.reviews = reviews\n",
        "    self.targets = targets\n",
        "    self.tokenizer = tokenizer\n",
        "    self.max_len = max_len\n",
        "\n",
        "  def __len__(self):\n",
        "    return len(self.reviews)\n",
        "\n",
        "  def __getitem__(self, item):\n",
        "    review = str(self.reviews[item])\n",
        "    target = self.targets[item]\n",
        "\n",
        "\n",
        "    encoding = self.tokenizer.encode_plus(\n",
        "        review, \n",
        "        add_special_tokens=True,\n",
        "        max_length=self.max_len,\n",
        "        return_token_type_ids=False,\n",
        "        padding='max_length',\n",
        "        return_attention_mask=True,\n",
        "        return_tensors='pt',\n",
        "        truncation=True\n",
        "    )\n",
        "\n",
        "    #return the review and attention masks\n",
        "    return {\n",
        "      'review_text': review,\n",
        "      'input_ids': encoding['input_ids'].flatten(),\n",
        "      'attention_mask': encoding['attention_mask'].flatten(),\n",
        "      'targets': torch.tensor(target, dtype=torch.long)\n",
        "    }"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7Wct3g5WK72I",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def create_data_loader(df, tokenizer, max_len, batch_size):\n",
        "  '''\n",
        "  create a data loader for the dataset\n",
        "  '''\n",
        "  ds = ReviewsDataset(\n",
        "    reviews=df.text.to_numpy(),\n",
        "    targets=df.label_encoded.to_numpy(),\n",
        "    tokenizer=tokenizer,\n",
        "    max_len=MAX_LEN\n",
        "  )\n",
        "\n",
        "  return DataLoader(\n",
        "    ds,\n",
        "    batch_size=batch_size,\n",
        "    num_workers=4\n",
        "  )"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ymqeEe8LK9Nw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#batch size is tricky . Larger batch size lead to memory issue on GPU\n",
        "#I have a colab pro subscription that gives me larger memory and gpus, but for free versions,\n",
        "#I suggest using batch size of 8. It will take longer run train with smaller batch sizes\n",
        "BATCH_SIZE = 16    \n",
        "\n",
        "#initialize data loaders\n",
        "train_data_loader = create_data_loader(df_train, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "val_data_loader = create_data_loader(df_val, tokenizer, MAX_LEN, BATCH_SIZE)\n",
        "test_data_loader = create_data_loader(df_test, tokenizer, MAX_LEN, BATCH_SIZE)"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5woZX0ooLmv-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class SentimentClassifier(nn.Module):\n",
        "  '''\n",
        "  Wrapper class to add dropout and FC layer on top of the \n",
        "  pooled output provided by hugging face transformers\n",
        "  '''\n",
        "  def __init__(self, n_classes, bertmodel, dropout_p=0.3):\n",
        "    super(SentimentClassifier, self).__init__()\n",
        "    self.bert = bertmodel\n",
        "    self.dropout = nn.Dropout(p=dropout_p)\n",
        "    self.out = nn.Linear(self.bert.config.hidden_size, n_classes)\n",
        "\n",
        "  def forward(self, input_ids, attention_mask):\n",
        "    last_hidden_state, pooled_output = self.bert(input_ids=input_ids, attention_mask=attention_mask)\n",
        "    output = self.dropout(pooled_output)\n",
        "    return self.out(output)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mY9-KVijUsa7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#init the model\n",
        "model = SentimentClassifier(len(class_names), bertmodel=bert_model).to(device)"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gtt0LTv6lv0Y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "bc7f9eb8-313d-4746-96f7-5d5c925ed8c0"
      },
      "source": [
        "EPOCHS = 10\n",
        "\n",
        "#adam w is a special version of Adam optimizer that transformers library provides\n",
        "#https://huggingface.co/transformers/main_classes/optimizer_schedules.html\n",
        "\n",
        "optimizer = AdamW(model.parameters(), lr=2e-5, correct_bias=False) \n",
        "loss_fn = nn.CrossEntropyLoss().to(device)\n",
        "print(\"len(train_data_loader): \", len(train_data_loader))\n",
        "total_steps = len(train_data_loader) * EPOCHS\n",
        "print(\"total_steps: \", total_steps)\n",
        "\n",
        "#define LR scheduler\n",
        "scheduler = get_linear_schedule_with_warmup(\n",
        "  optimizer,\n",
        "  num_warmup_steps=0,\n",
        "  num_training_steps=total_steps\n",
        ")"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "len(train_data_loader):  1250\n",
            "total_steps:  12500\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tFiPRtubqL2n",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "history = None\n",
        "best_accuracy = None"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vp790i7ADQEV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 106
        },
        "outputId": "fb12b168-58b5-49c9-8b3a-d930051a1319"
      },
      "source": [
        "if LOAD_PREV_MODEL and len(list(glob.glob(\"/content/drive/My Drive/Datasets/IMDBMovieReviews/{}*\".format(MODEL_SAVE_NAME)))) > 0:\n",
        "  #Load previous model to continue training\n",
        "  state_file_name = sorted(list(glob.glob(\"/content/drive/My Drive/Datasets/IMDBMovieReviews/{}*\".format(MODEL_SAVE_NAME))))[-1]\n",
        "  print(\"Loading : {}\".format(state_file_name))\n",
        "  state = torch.load(state_file_name)\n",
        "  print(state.keys())\n",
        "  \n",
        "\n",
        "  model.load_state_dict(state['model'])\n",
        "\n",
        "  optimizer.load_state_dict(state['optimizer'])\n",
        "  scheduler.load_state_dict(state['scheduler'])\n",
        "  history = state['history']\n",
        "  best_accuracy = state['best_accuracy']\n",
        "\n",
        "  EPOCHS = EPOCHS - state['epoch']-1\n",
        "  state = None"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loading : /content/drive/My Drive/Datasets/IMDBMovieReviews/imdb_movie_large_roberta_state_20200905-154517_0\n",
            "dict_keys(['epoch', 'model', 'optimizer', 'scheduler', 'history', 'best_accuracy'])\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/optim/lr_scheduler.py:218: UserWarning: Please also save or load the state of the optimzer when saving or loading the scheduler.\n",
            "  warnings.warn(SAVE_STATE_WARNING, UserWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "23Gr5HlqWpZ0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "torch.cuda.empty_cache()"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e7Tc5KjDWRos",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train_epoch(model, dataloader, lossfn, opt, device, schedulr , n_examples):\n",
        "  model.train();\n",
        "  losses = []\n",
        "  correct_preds = 0\n",
        "\n",
        "  for ind, d in enumerate(dataloader):\n",
        "    if ind%100 == 0:\n",
        "      print(f\"iter#{ind}\")\n",
        "    input_ids = d[\"input_ids\"].to(device)\n",
        "    targets = d[\"targets\"].to(device)\n",
        "    attention_mask = d['attention_mask'].to(device)\n",
        "\n",
        "    outputs = model(\n",
        "      input_ids=input_ids,\n",
        "      attention_mask=attention_mask\n",
        "    )\n",
        "\n",
        "    max_values, max_indices = torch.max(outputs, dim=1)\n",
        "    loss = lossfn(outputs, targets)\n",
        "    correct_preds+=torch.sum(max_indices == targets)\n",
        "    losses.append(loss.item())\n",
        "    \n",
        "    loss.backward()\n",
        "    nn.utils.clip_grad_norm_(model.parameters(), max_norm=1.0)\n",
        "    opt.step()\n",
        "    schedulr.step()\n",
        "    opt.zero_grad()\n",
        "\n",
        "  return correct_preds.double()/n_examples, np.mean(losses)"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dEtKOmH-ah3u",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def eval_model(model, dataloader, lossfn, device , n_examples):\n",
        "  model.eval();\n",
        "  losses = []\n",
        "  correct_preds = 0\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in dataloader:\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "      attention_mask = d['attention_mask'].to(device)\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask\n",
        "      )\n",
        "\n",
        "      max_values, max_indices = torch.max(outputs, dim=1)\n",
        "      loss = lossfn(outputs, targets)\n",
        "      correct_preds+=torch.sum(max_indices == targets)\n",
        "      losses.append(loss.item())\n",
        "\n",
        "  return correct_preds.double()/n_examples, np.mean(losses)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qCMF2rv6p3gS",
        "colab_type": "text"
      },
      "source": [
        "#Lets begin training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VHXR63m7an4o",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "d5549863-01ca-401b-db49-4faa1fef3d54"
      },
      "source": [
        "if history is None:\n",
        "  history = defaultdict(list)\n",
        "if best_accuracy is None:\n",
        "  best_accuracy = 0\n",
        "\n",
        "\n",
        "for epoch in range(EPOCHS):\n",
        "  start_time = time.time()\n",
        "\n",
        "  print(f'Epoch {epoch + 1}/{EPOCHS}')\n",
        "  print('-' * 10)\n",
        "\n",
        "  train_acc, train_loss = train_epoch(\n",
        "    model,\n",
        "    train_data_loader,    \n",
        "    loss_fn, \n",
        "    optimizer, \n",
        "    device, \n",
        "    scheduler, \n",
        "    len(df_train)\n",
        "  )\n",
        "\n",
        "  print(f'Train loss {train_loss} accuracy {train_acc}')\n",
        "\n",
        "  val_acc, val_loss = eval_model(\n",
        "    model,\n",
        "    val_data_loader,\n",
        "    loss_fn, \n",
        "    device, \n",
        "    len(df_val)\n",
        "  )\n",
        "\n",
        "  print(f'Val   loss {val_loss} accuracy {val_acc}')\n",
        "  print()\n",
        "\n",
        "  history['train_acc'].append(train_acc)\n",
        "  history['train_loss'].append(train_loss)\n",
        "  history['val_acc'].append(val_acc)\n",
        "  history['val_loss'].append(val_loss)\n",
        "  print(\"--- %s seconds ---\" % (time.time() - start_time))\n",
        "\n",
        "  #save only those models that do better on val\n",
        "  if val_acc > best_accuracy:\n",
        "    drive.mount('/content/drive')\n",
        "    state = {\n",
        "    'epoch': epoch,\n",
        "    'model': model.state_dict(),\n",
        "    'optimizer': optimizer.state_dict(),\n",
        "    'scheduler': scheduler.state_dict(),\n",
        "    'history': history,\n",
        "    'best_accuracy': best_accuracy\n",
        "\n",
        "    }\n",
        "    torch.save(state, \"/content/drive/My Drive/Datasets/IMDBMovieReviews/{}_{}\".format(MODEL_SAVE_NAME, epoch))\n",
        "\n",
        "    best_accuracy = val_acc"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.10577935135336593 accuracy 0.9730000000000001\n",
            "Val   loss 0.325952225993313 accuracy 0.9338000000000001\n",
            "\n",
            "--- 1169.7395250797272 seconds ---\n",
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/optim/lr_scheduler.py:200: UserWarning: Please also save or load the state of the optimzer when saving or loading the scheduler.\n",
            "  warnings.warn(SAVE_STATE_WARNING, UserWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 2/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.10723657275061123 accuracy 0.9746\n",
            "Val   loss 0.3054136142689592 accuracy 0.9332\n",
            "\n",
            "--- 1169.6421444416046 seconds ---\n",
            "Epoch 3/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.06332904680930078 accuracy 0.9863500000000001\n",
            "Val   loss 0.3217761030408759 accuracy 0.9390000000000001\n",
            "\n",
            "--- 1169.1732811927795 seconds ---\n",
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n",
            "Epoch 4/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.04372967040722724 accuracy 0.9911500000000001\n",
            "Val   loss 0.35703370585786953 accuracy 0.935\n",
            "\n",
            "--- 1168.878390789032 seconds ---\n",
            "Epoch 5/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.027374642371118533 accuracy 0.9951000000000001\n",
            "Val   loss 0.4327657846772981 accuracy 0.9276000000000001\n",
            "\n",
            "--- 1168.899046421051 seconds ---\n",
            "Epoch 6/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.022806652674335057 accuracy 0.996\n",
            "Val   loss 0.40589323708021935 accuracy 0.9398000000000001\n",
            "\n",
            "--- 1168.4733574390411 seconds ---\n",
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n",
            "Epoch 7/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.013027524928969797 accuracy 0.99775\n",
            "Val   loss 0.4583241831719635 accuracy 0.9372\n",
            "\n",
            "--- 1168.927967786789 seconds ---\n",
            "Epoch 8/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.009591659283504124 accuracy 0.9984500000000001\n",
            "Val   loss 0.4790034453262677 accuracy 0.9372\n",
            "\n",
            "--- 1169.0722229480743 seconds ---\n",
            "Epoch 9/9\n",
            "----------\n",
            "iter#0\n",
            "iter#100\n",
            "iter#200\n",
            "iter#300\n",
            "iter#400\n",
            "iter#500\n",
            "iter#600\n",
            "iter#700\n",
            "iter#800\n",
            "iter#900\n",
            "iter#1000\n",
            "iter#1100\n",
            "iter#1200\n",
            "Train loss 0.005919923061551527 accuracy 0.9991000000000001\n",
            "Val   loss 0.4733588481107139 accuracy 0.9408000000000001\n",
            "\n",
            "--- 1169.633512020111 seconds ---\n",
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8j6y_e-Rwlz0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 523
        },
        "outputId": "e62a115f-0f36-43cd-8a30-317ac5a91649"
      },
      "source": [
        "plt.plot(history['train_acc'], label='train accuracy')\n",
        "plt.plot(history['val_acc'], label='validation accuracy')\n",
        "\n",
        "plt.title('Training history')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epoch')\n",
        "plt.legend()\n",
        "plt.ylim([0, 1]);"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 732,
              "height": 506
            }
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_21Lll17qX9l",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "ed5f6301-4f5e-46d8-ddd9-3530d4089a6a"
      },
      "source": [
        "#Evaluate accuracy on test set (blind)\n",
        "test_acc, _ = eval_model(\n",
        "  model,\n",
        "  test_data_loader,\n",
        "  loss_fn,\n",
        "  device,\n",
        "  len(df_test)\n",
        ")\n",
        "\n",
        "test_acc.item()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-8c90b1b0b3a8>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m test_acc, _ = eval_model(\n\u001b[0m\u001b[1;32m      2\u001b[0m   \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m   \u001b[0mtest_data_loader\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m   \u001b[0mloss_fn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m   \u001b[0mdevice\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'eval_model' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ehfr03J6qYAU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_predictions(model, data_loader):\n",
        "  # This is similar to the evaluation function, \n",
        "  #except that we're storing the text of the reviews and the predicted \n",
        "  #probabilities (by applying the softmax on the model outputs)\n",
        "  model = model.eval()\n",
        "  \n",
        "  review_texts = []\n",
        "  predictions = []\n",
        "  prediction_probs = []\n",
        "  real_values = []\n",
        "\n",
        "  with torch.no_grad():\n",
        "    for d in data_loader:\n",
        "\n",
        "      texts = d[\"review_text\"]\n",
        "      input_ids = d[\"input_ids\"].to(device)\n",
        "      attention_mask = d[\"attention_mask\"].to(device)\n",
        "      targets = d[\"targets\"].to(device)\n",
        "\n",
        "      outputs = model(\n",
        "        input_ids=input_ids,\n",
        "        attention_mask=attention_mask\n",
        "      )\n",
        "      _, preds = torch.max(outputs, dim=1)\n",
        "\n",
        "      probs = nn.functional.softmax(outputs, dim=1)\n",
        "\n",
        "      review_texts.extend(texts)\n",
        "      predictions.extend(preds)\n",
        "      prediction_probs.extend(probs)\n",
        "      real_values.extend(targets)\n",
        "\n",
        "  predictions = torch.stack(predictions).cpu()\n",
        "  prediction_probs = torch.stack(prediction_probs).cpu()\n",
        "  real_values = torch.stack(real_values).cpu()\n",
        "  return review_texts, predictions, prediction_probs, real_values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mgFudrutqtok",
        "colab_type": "text"
      },
      "source": [
        "#Lets look at the confusion matrix on test"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0J44TT6Cqyfn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Get confusion matrix on test\n",
        "def show_confusion_matrix(confusion_matrix):\n",
        "  hmap = sns.heatmap(confusion_matrix, annot=True, fmt=\"d\", cmap=\"Blues\")\n",
        "  hmap.yaxis.set_ticklabels(hmap.yaxis.get_ticklabels(), rotation=0, ha='right')\n",
        "  hmap.xaxis.set_ticklabels(hmap.xaxis.get_ticklabels(), rotation=30, ha='right')\n",
        "  plt.ylabel('True sentiment')\n",
        "  plt.xlabel('Predicted sentiment');\n",
        "\n",
        "#get preds from test set\n",
        "y_review_texts, y_pred, y_pred_probs, y_test = get_predictions(\n",
        "  model,\n",
        "  test_data_loader\n",
        ")\n",
        "\n",
        "#visualize confusion matrix\n",
        "cm = confusion_matrix(y_test, y_pred)\n",
        "df_cm = pd.DataFrame(cm, index=class_names, columns=class_names)\n",
        "show_confusion_matrix(df_cm)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fdCvEceyq9dE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#example to predict on one example\n",
        "review_text = \"Hated it with all my being. Worst movie ever. Mentally- scarred. Help me. It was that bad.TRUST ME!!!\"\n",
        "encoded_review = tokenizer.encode_plus(\n",
        "  review_text,\n",
        "  max_length=MAX_LEN,\n",
        "  add_special_tokens=True,\n",
        "  return_token_type_ids=False,\n",
        "  pad_to_max_length=True,\n",
        "  return_attention_mask=True,\n",
        "  return_tensors='pt',\n",
        "  truncation=True\n",
        ")\n",
        "input_ids = encoded_review['input_ids'].to(device)\n",
        "attention_mask = encoded_review['attention_mask'].to(device)\n",
        "\n",
        "model.eval();\n",
        "output = model(input_ids, attention_mask)\n",
        "_, prediction = torch.max(output, dim=1)\n",
        "\n",
        "print(f'Review text: {review_text}')\n",
        "print(f'Sentiment  : {class_names[prediction]}')"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}