indic-whisper-jax-tpu-kaggle.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.8.16","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"tpu1vmV38","dataSources":[{"sourceId":7723170,"sourceType":"datasetVersion","datasetId":4511660}],"dockerImageVersionId":30459,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"## Whisper JAX ⚡️\n\nThis Kaggle notebook demonstratese how to run Indic Whisper JAX on a TPU v3-8. Indic Whisper JAX is a highly optimised JAX implementation of the Indic Whisper model by AI4Bharat. Compared to OpenAI's PyTorch code, Whisper JAX runs over **70x faster**, making it the fastest Whisper implementation available.\n\nYou can find the code [here](https://github.com/sanchit-gandhi/whisper-jax).","metadata":{}},{"cell_type":"markdown","source":"## Let's get started!\n\nThe first thing we need to do is connect to a TPU. Kaggle offers 20 hours of TPU v3-8 usage per month for free, which we'll make use of for this notebook. Refer to the guide [Introducing TPUs to Kaggle](https://www.kaggle.com/product-feedback/129828) for more information on TPU quotas in Kaggle.\n\nYou will need to register a Kaggle account and verify your phone number if you haven't done so already. Once verified, open up the settings menu in the Notebook editor (the small arrow in the bottom right). Then under _Notebook options_, select ‘TPU VM v3-8’ from the _Accelerator_ menu. You will also need to toggle the internet switch so that it is set to \"on\".\n\nOnce we've got a TPU allocated (there might be a queue to get one!), we can run the following to see the TPU devices we have available:","metadata":{}},{"cell_type":"code","source":"import jax\njax.devices()","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:35:34.084039Z","iopub.execute_input":"2024-03-01T10:35:34.084365Z","iopub.status.idle":"2024-03-01T10:35:42.028904Z","shell.execute_reply.started":"2024-03-01T10:35:34.084337Z","shell.execute_reply":"2024-03-01T10:35:42.028004Z"},"trusted":true},"execution_count":1,"outputs":[{"execution_count":1,"output_type":"execute_result","data":{"text/plain":"[TpuDevice(id=0, process_index=0, coords=(0,0,0), core_on_chip=0),\n TpuDevice(id=1, process_index=0, coords=(0,0,0), core_on_chip=1),\n TpuDevice(id=2, process_index=0, coords=(1,0,0), core_on_chip=0),\n TpuDevice(id=3, process_index=0, coords=(1,0,0), core_on_chip=1),\n TpuDevice(id=4, process_index=0, coords=(0,1,0), core_on_chip=0),\n TpuDevice(id=5, process_index=0, coords=(0,1,0), core_on_chip=1),\n TpuDevice(id=6, process_index=0, coords=(1,1,0), core_on_chip=0),\n TpuDevice(id=7, process_index=0, coords=(1,1,0), core_on_chip=1)]"},"metadata":{}}]},{"cell_type":"markdown","source":"Cool! We've got 8 TPU devices packaged into one accelerator.\n\nKaggle TPUs come with JAX pre-installed, so we can directly install the remaining Python packages. If you're running the notebook on a Cloud TPU, ensure you have installed JAX according to the official [installation guide](https://github.com/google/jax#pip-installation-google-cloud-tpu). \n\nWe'll install [Whisper JAX](https://github.com/sanchit-gandhi/whisper-jax) from main, as well as `datasets` and `librosa` for loading audio files:","metadata":{}},{"cell_type":"code","source":"!pip install --quiet --upgrade pip\n!pip install --quiet git+https://github.com/sanchit-gandhi/whisper-jax.git datasets librosa tqdm","metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","tags":[],"execution":{"iopub.status.busy":"2024-03-01T10:35:42.030386Z","iopub.execute_input":"2024-03-01T10:35:42.030841Z","iopub.status.idle":"2024-03-01T10:36:32.690121Z","shell.execute_reply.started":"2024-03-01T10:35:42.030813Z","shell.execute_reply":"2024-03-01T10:36:32.688948Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stdout","text":"\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"markdown","source":"## Loading the Pipeline\n\nThe recommended way of running Whisper JAX is through the [`FlaxWhisperPipline`](https://github.com/sanchit-gandhi/whisper-jax/blob/main/whisper_jax/pipeline.py#L57) class. This class handles all the necessary pre- and post-processing for the model, as well as wrapping the generate method for data parallelism across all available accelerator devices\n\n\nLet's load the large-v2 model in bfloat16 (half-precision). Using half-precision will speed-up the computation quite considerably by storing intermediate tensors in half-precision. There is no change to the precision of the model weights.\n\nWe'll also make use of _batching_ for single audio inputs: the audio is first chunked into 30 second segments, and then chunks dispatched to the model to be transcribed in parallel. By batching an audio input and transcribing it in parallel, we get a ~10x speed-up compared to transcribing the audio samples sequentially.","metadata":{}},{"cell_type":"code","source":"from whisper_jax import FlaxWhisperPipline\nimport jax.numpy as jnp\n\npipeline = FlaxWhisperPipline(\"parthiv11/indic_whisper_hi_multi_gpu\", dtype=jnp.bfloat16, batch_size=16)","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2024-03-01T10:36:32.691632Z","iopub.execute_input":"2024-03-01T10:36:32.691963Z","iopub.status.idle":"2024-03-01T10:37:32.714517Z","shell.execute_reply.started":"2024-03-01T10:36:32.691935Z","shell.execute_reply":"2024-03-01T10:37:32.713578Z"},"trusted":true},"execution_count":3,"outputs":[{"name":"stderr","text":"/usr/local/lib/python3.8/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n  from .autonotebook import tqdm as notebook_tqdm\nSome of the weights of FlaxWhisperForConditionalGeneration were initialized in float16 precision from the model checkpoint at parthiv11/indic_whisper_hi_multi_gpu:\n[('model', 'decoder', 'embed_positions', 'embedding'), ('model', 'decoder', 'embed_tokens', 'embedding'), ('model', 'decoder', 'layer_norm', 'bias'), ('model', 'decoder', 'layer_norm', 'scale'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '0', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '0', 'fc1', 'bias'), ('model', 'decoder', 'layers', '0', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '0', 'fc2', 'bias'), ('model', 'decoder', 'layers', '0', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '0', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '0', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '0', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '0', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '0', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '0', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '1', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '1', 'fc1', 'bias'), ('model', 'decoder', 'layers', '1', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '1', 'fc2', 'bias'), ('model', 'decoder', 'layers', '1', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '1', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '1', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '1', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '1', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '1', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '1', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '10', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '10', 'fc1', 'bias'), ('model', 'decoder', 'layers', '10', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '10', 'fc2', 'bias'), ('model', 'decoder', 'layers', '10', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '10', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '10', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '10', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '10', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '10', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '10', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '11', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '11', 'fc1', 'bias'), ('model', 'decoder', 'layers', '11', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '11', 'fc2', 'bias'), ('model', 'decoder', 'layers', '11', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '11', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '11', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '11', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '11', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '11', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '11', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '12', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '12', 'fc1', 'bias'), ('model', 'decoder', 'layers', '12', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '12', 'fc2', 'bias'), ('model', 'decoder', 'layers', '12', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '12', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '12', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '12', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '12', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '12', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '12', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '13', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '13', 'fc1', 'bias'), ('model', 'decoder', 'layers', '13', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '13', 'fc2', 'bias'), ('model', 'decoder', 'layers', '13', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '13', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '13', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '13', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '13', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '13', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '13', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '14', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '14', 'fc1', 'bias'), ('model', 'decoder', 'layers', '14', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '14', 'fc2', 'bias'), ('model', 'decoder', 'layers', '14', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '14', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '14', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '14', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '14', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '14', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '14', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '15', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '15', 'fc1', 'bias'), ('model', 'decoder', 'layers', '15', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '15', 'fc2', 'bias'), ('model', 'decoder', 'layers', '15', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '15', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '15', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '15', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '15', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '15', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '15', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '16', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '16', 'fc1', 'bias'), ('model', 'decoder', 'layers', '16', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '16', 'fc2', 'bias'), ('model', 'decoder', 'layers', '16', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '16', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '16', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '16', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '16', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '16', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '16', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '17', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '17', 'fc1', 'bias'), ('model', 'decoder', 'layers', '17', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '17', 'fc2', 'bias'), ('model', 'decoder', 'layers', '17', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '17', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '17', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '17', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '17', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '17', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '17', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '18', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '18', 'fc1', 'bias'), ('model', 'decoder', 'layers', '18', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '18', 'fc2', 'bias'), ('model', 'decoder', 'layers', '18', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '18', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '18', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '18', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '18', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '18', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '18', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '19', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '19', 'fc1', 'bias'), ('model', 'decoder', 'layers', '19', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '19', 'fc2', 'bias'), ('model', 'decoder', 'layers', '19', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '19', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '19', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '19', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '19', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '19', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '19', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '2', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '2', 'fc1', 'bias'), ('model', 'decoder', 'layers', '2', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '2', 'fc2', 'bias'), ('model', 'decoder', 'layers', '2', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '2', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '2', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '2', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '2', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '2', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '2', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '20', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '20', 'fc1', 'bias'), ('model', 'decoder', 'layers', '20', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '20', 'fc2', 'bias'), ('model', 'decoder', 'layers', '20', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '20', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '20', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '20', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '20', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '20', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '20', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '21', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '21', 'fc1', 'bias'), ('model', 'decoder', 'layers', '21', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '21', 'fc2', 'bias'), ('model', 'decoder', 'layers', '21', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '21', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '21', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '21', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '21', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '21', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '21', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '22', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '22', 'fc1', 'bias'), ('model', 'decoder', 'layers', '22', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '22', 'fc2', 'bias'), ('model', 'decoder', 'layers', '22', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '22', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '22', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '22', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '22', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '22', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '22', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '23', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '23', 'fc1', 'bias'), ('model', 'decoder', 'layers', '23', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '23', 'fc2', 'bias'), ('model', 'decoder', 'layers', '23', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '23', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '23', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '23', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '23', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '23', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '23', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '3', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '3', 'fc1', 'bias'), ('model', 'decoder', 'layers', '3', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '3', 'fc2', 'bias'), ('model', 'decoder', 'layers', '3', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '3', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '3', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '3', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '3', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '3', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '3', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '4', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '4', 'fc1', 'bias'), ('model', 'decoder', 'layers', '4', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '4', 'fc2', 'bias'), ('model', 'decoder', 'layers', '4', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '4', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '4', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '4', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '4', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '4', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '4', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '5', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '5', 'fc1', 'bias'), ('model', 'decoder', 'layers', '5', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '5', 'fc2', 'bias'), ('model', 'decoder', 'layers', '5', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '5', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '5', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '5', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '5', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '5', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '5', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '6', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '6', 'fc1', 'bias'), ('model', 'decoder', 'layers', '6', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '6', 'fc2', 'bias'), ('model', 'decoder', 'layers', '6', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '6', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '6', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '6', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '6', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '6', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '6', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '7', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '7', 'fc1', 'bias'), ('model', 'decoder', 'layers', '7', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '7', 'fc2', 'bias'), ('model', 'decoder', 'layers', '7', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '7', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '7', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '7', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '7', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '7', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '7', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '8', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '8', 'fc1', 'bias'), ('model', 'decoder', 'layers', '8', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '8', 'fc2', 'bias'), ('model', 'decoder', 'layers', '8', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '8', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '8', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '8', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '8', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '8', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '8', 'self_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'encoder_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'encoder_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '9', 'encoder_attn_layer_norm', 'scale'), ('model', 'decoder', 'layers', '9', 'fc1', 'bias'), ('model', 'decoder', 'layers', '9', 'fc1', 'kernel'), ('model', 'decoder', 'layers', '9', 'fc2', 'bias'), ('model', 'decoder', 'layers', '9', 'fc2', 'kernel'), ('model', 'decoder', 'layers', '9', 'final_layer_norm', 'bias'), ('model', 'decoder', 'layers', '9', 'final_layer_norm', 'scale'), ('model', 'decoder', 'layers', '9', 'self_attn', 'k_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'self_attn', 'out_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'self_attn', 'out_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'self_attn', 'q_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'self_attn', 'q_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'self_attn', 'v_proj', 'bias'), ('model', 'decoder', 'layers', '9', 'self_attn', 'v_proj', 'kernel'), ('model', 'decoder', 'layers', '9', 'self_attn_layer_norm', 'bias'), ('model', 'decoder', 'layers', '9', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'conv1', 'bias'), ('model', 'encoder', 'conv1', 'kernel'), ('model', 'encoder', 'conv2', 'bias'), ('model', 'encoder', 'conv2', 'kernel'), ('model', 'encoder', 'embed_positions', 'embedding'), ('model', 'encoder', 'layer_norm', 'bias'), ('model', 'encoder', 'layer_norm', 'scale'), ('model', 'encoder', 'layers', '0', 'fc1', 'bias'), ('model', 'encoder', 'layers', '0', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '0', 'fc2', 'bias'), ('model', 'encoder', 'layers', '0', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '0', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '0', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '0', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '0', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '0', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '0', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '0', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '0', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '0', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '0', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '0', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '1', 'fc1', 'bias'), ('model', 'encoder', 'layers', '1', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '1', 'fc2', 'bias'), ('model', 'encoder', 'layers', '1', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '1', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '1', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '1', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '1', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '1', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '1', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '1', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '1', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '1', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '1', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '1', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '10', 'fc1', 'bias'), ('model', 'encoder', 'layers', '10', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '10', 'fc2', 'bias'), ('model', 'encoder', 'layers', '10', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '10', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '10', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '10', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '10', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '10', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '10', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '10', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '10', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '10', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '10', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '10', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '11', 'fc1', 'bias'), ('model', 'encoder', 'layers', '11', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '11', 'fc2', 'bias'), ('model', 'encoder', 'layers', '11', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '11', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '11', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '11', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '11', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '11', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '11', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '11', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '11', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '11', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '11', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '11', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '12', 'fc1', 'bias'), ('model', 'encoder', 'layers', '12', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '12', 'fc2', 'bias'), ('model', 'encoder', 'layers', '12', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '12', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '12', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '12', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '12', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '12', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '12', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '12', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '12', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '12', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '12', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '12', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '13', 'fc1', 'bias'), ('model', 'encoder', 'layers', '13', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '13', 'fc2', 'bias'), ('model', 'encoder', 'layers', '13', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '13', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '13', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '13', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '13', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '13', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '13', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '13', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '13', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '13', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '13', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '13', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '14', 'fc1', 'bias'), ('model', 'encoder', 'layers', '14', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '14', 'fc2', 'bias'), ('model', 'encoder', 'layers', '14', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '14', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '14', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '14', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '14', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '14', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '14', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '14', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '14', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '14', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '14', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '14', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '15', 'fc1', 'bias'), ('model', 'encoder', 'layers', '15', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '15', 'fc2', 'bias'), ('model', 'encoder', 'layers', '15', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '15', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '15', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '15', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '15', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '15', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '15', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '15', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '15', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '15', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '15', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '15', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '16', 'fc1', 'bias'), ('model', 'encoder', 'layers', '16', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '16', 'fc2', 'bias'), ('model', 'encoder', 'layers', '16', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '16', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '16', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '16', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '16', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '16', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '16', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '16', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '16', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '16', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '16', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '16', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '17', 'fc1', 'bias'), ('model', 'encoder', 'layers', '17', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '17', 'fc2', 'bias'), ('model', 'encoder', 'layers', '17', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '17', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '17', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '17', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '17', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '17', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '17', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '17', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '17', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '17', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '17', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '17', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '18', 'fc1', 'bias'), ('model', 'encoder', 'layers', '18', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '18', 'fc2', 'bias'), ('model', 'encoder', 'layers', '18', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '18', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '18', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '18', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '18', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '18', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '18', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '18', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '18', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '18', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '18', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '18', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '19', 'fc1', 'bias'), ('model', 'encoder', 'layers', '19', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '19', 'fc2', 'bias'), ('model', 'encoder', 'layers', '19', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '19', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '19', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '19', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '19', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '19', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '19', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '19', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '19', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '19', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '19', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '19', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '2', 'fc1', 'bias'), ('model', 'encoder', 'layers', '2', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '2', 'fc2', 'bias'), ('model', 'encoder', 'layers', '2', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '2', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '2', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '2', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '2', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '2', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '2', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '2', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '2', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '2', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '2', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '2', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '20', 'fc1', 'bias'), ('model', 'encoder', 'layers', '20', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '20', 'fc2', 'bias'), ('model', 'encoder', 'layers', '20', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '20', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '20', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '20', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '20', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '20', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '20', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '20', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '20', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '20', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '20', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '20', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '21', 'fc1', 'bias'), ('model', 'encoder', 'layers', '21', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '21', 'fc2', 'bias'), ('model', 'encoder', 'layers', '21', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '21', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '21', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '21', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '21', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '21', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '21', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '21', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '21', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '21', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '21', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '21', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '22', 'fc1', 'bias'), ('model', 'encoder', 'layers', '22', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '22', 'fc2', 'bias'), ('model', 'encoder', 'layers', '22', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '22', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '22', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '22', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '22', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '22', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '22', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '22', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '22', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '22', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '22', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '22', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '23', 'fc1', 'bias'), ('model', 'encoder', 'layers', '23', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '23', 'fc2', 'bias'), ('model', 'encoder', 'layers', '23', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '23', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '23', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '23', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '23', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '23', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '23', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '23', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '23', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '23', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '23', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '23', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '3', 'fc1', 'bias'), ('model', 'encoder', 'layers', '3', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '3', 'fc2', 'bias'), ('model', 'encoder', 'layers', '3', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '3', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '3', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '3', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '3', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '3', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '3', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '3', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '3', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '3', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '3', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '3', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '4', 'fc1', 'bias'), ('model', 'encoder', 'layers', '4', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '4', 'fc2', 'bias'), ('model', 'encoder', 'layers', '4', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '4', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '4', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '4', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '4', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '4', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '4', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '4', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '4', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '4', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '4', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '4', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '5', 'fc1', 'bias'), ('model', 'encoder', 'layers', '5', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '5', 'fc2', 'bias'), ('model', 'encoder', 'layers', '5', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '5', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '5', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '5', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '5', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '5', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '5', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '5', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '5', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '5', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '5', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '5', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '6', 'fc1', 'bias'), ('model', 'encoder', 'layers', '6', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '6', 'fc2', 'bias'), ('model', 'encoder', 'layers', '6', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '6', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '6', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '6', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '6', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '6', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '6', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '6', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '6', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '6', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '6', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '6', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '7', 'fc1', 'bias'), ('model', 'encoder', 'layers', '7', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '7', 'fc2', 'bias'), ('model', 'encoder', 'layers', '7', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '7', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '7', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '7', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '7', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '7', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '7', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '7', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '7', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '7', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '7', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '7', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '8', 'fc1', 'bias'), ('model', 'encoder', 'layers', '8', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '8', 'fc2', 'bias'), ('model', 'encoder', 'layers', '8', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '8', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '8', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '8', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '8', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '8', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '8', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '8', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '8', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '8', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '8', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '8', 'self_attn_layer_norm', 'scale'), ('model', 'encoder', 'layers', '9', 'fc1', 'bias'), ('model', 'encoder', 'layers', '9', 'fc1', 'kernel'), ('model', 'encoder', 'layers', '9', 'fc2', 'bias'), ('model', 'encoder', 'layers', '9', 'fc2', 'kernel'), ('model', 'encoder', 'layers', '9', 'final_layer_norm', 'bias'), ('model', 'encoder', 'layers', '9', 'final_layer_norm', 'scale'), ('model', 'encoder', 'layers', '9', 'self_attn', 'k_proj', 'kernel'), ('model', 'encoder', 'layers', '9', 'self_attn', 'out_proj', 'bias'), ('model', 'encoder', 'layers', '9', 'self_attn', 'out_proj', 'kernel'), ('model', 'encoder', 'layers', '9', 'self_attn', 'q_proj', 'bias'), ('model', 'encoder', 'layers', '9', 'self_attn', 'q_proj', 'kernel'), ('model', 'encoder', 'layers', '9', 'self_attn', 'v_proj', 'bias'), ('model', 'encoder', 'layers', '9', 'self_attn', 'v_proj', 'kernel'), ('model', 'encoder', 'layers', '9', 'self_attn_layer_norm', 'bias'), ('model', 'encoder', 'layers', '9', 'self_attn_layer_norm', 'scale')]\nYou should probably UPCAST the model weights to float32 if this was not intended. See [`~FlaxPreTrainedModel.to_fp32`] for further information on how to do this.\n","output_type":"stream"}]},{"cell_type":"markdown","source":"We'll then initialise a compilation cache, which will speed-up the compilation time if we close our kernel and want to compile the model again:\n\n","metadata":{}},{"cell_type":"code","source":"%cd /kaggle/working\nfrom jax.experimental.compilation_cache import compilation_cache as cc\n\ncc.initialize_cache(\"./jax_cache\")","metadata":{"tags":[],"execution":{"iopub.status.busy":"2024-03-01T10:37:32.717488Z","iopub.execute_input":"2024-03-01T10:37:32.717771Z","iopub.status.idle":"2024-03-01T10:37:32.724173Z","shell.execute_reply.started":"2024-03-01T10:37:32.717748Z","shell.execute_reply":"2024-03-01T10:37:32.723435Z"},"trusted":true},"execution_count":4,"outputs":[{"name":"stderr","text":"WARNING:jax.experimental.compilation_cache.compilation_cache:Initialized persistent compilation cache at ./jax_cache\n","output_type":"stream"},{"name":"stdout","text":"/kaggle/working\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## 🎶 Load an audio file\n\nLet's load up a long audio file for our tests.Note that you can also pass in any `.mp3`, `.wav` or `.flac` audio file directly to the Whisper JAX pipeline, and it will take care of loading the audio file for you.","metadata":{}},{"cell_type":"markdown","source":"We can take a listen to the audio file that we've loaded - we'll see that it's approximately 5 mins long:","metadata":{}},{"cell_type":"code","source":"# 1 min audio\nfile_path='/kaggle/input/cpgram-audios/7020587774.mp3'","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:38:31.407112Z","iopub.execute_input":"2024-03-01T10:38:31.407375Z","iopub.status.idle":"2024-03-01T10:38:31.410757Z","shell.execute_reply.started":"2024-03-01T10:38:31.407352Z","shell.execute_reply":"2024-03-01T10:38:31.409922Z"},"trusted":true},"execution_count":9,"outputs":[]},{"cell_type":"code","source":"from IPython.display import Audio\n\nAudio(file_path)","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:38:31.411781Z","iopub.execute_input":"2024-03-01T10:38:31.412046Z","iopub.status.idle":"2024-03-01T10:38:31.425997Z","shell.execute_reply.started":"2024-03-01T10:38:31.412025Z","shell.execute_reply":"2024-03-01T10:38:31.425237Z"},"trusted":true},"execution_count":10,"outputs":[{"execution_count":10,"output_type":"execute_result","data":{"text/plain":"<IPython.lib.display.Audio object>","text/html":"\n                <audio  controls=\"controls\" >\n                    <source src=\"data:audio/mpeg;base64,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\" type=\"audio/mpeg\" />\n                    Your browser does not support the audio element.\n                </audio>\n              "},"metadata":{}}]},{"cell_type":"code","source":"import librosa\n\n# Load the audio file\nsignal, sample_rate = librosa.load(file_path, sr=16000)\n\n# Print waveform shape and sample rate\nprint(\"Waveform shape:\", signal.shape)\nprint(\"Sample rate:\", sample_rate)\n","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:38:31.426878Z","iopub.execute_input":"2024-03-01T10:38:31.427109Z","iopub.status.idle":"2024-03-01T10:38:31.455736Z","shell.execute_reply.started":"2024-03-01T10:38:31.427089Z","shell.execute_reply":"2024-03-01T10:38:31.454635Z"},"trusted":true},"execution_count":11,"outputs":[{"name":"stdout","text":"Waveform shape: (975744,)\nSample rate: 16000\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## Run the model\n\nNow we're ready to transcribe! We'll need to compile the `pmap` function the first time we use it. You can expect compilation to take ~2 minutes on a TPU v3-8 with a batch size of 16. Enough time to grab a coffee ☕️\n\nThereafter, we can use our cached `pmap` function, which you'll see is amazingly fast.","metadata":{}},{"cell_type":"code","source":"# JIT compile the forward call - slow, but we only do once\n%time text= pipeline(signal)","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:37:41.990717Z","iopub.execute_input":"2024-03-01T10:37:41.991317Z","iopub.status.idle":"2024-03-01T10:38:31.405269Z","shell.execute_reply.started":"2024-03-01T10:37:41.991290Z","shell.execute_reply":"2024-03-01T10:38:31.404488Z"},"trusted":true},"execution_count":8,"outputs":[{"name":"stderr","text":"WARNING:whisper_jax.pipeline:Numpy array passed as input - no sampling rate checks will be performed.It is strongly recommended to pass the input as a dictionary with an 'array' key containing the numpy array representing the audio, and a 'sampling_rate' key containing the sampling rate associated with the audio array.Failing to do so can result in silent errors that might be hard to debug.\n","output_type":"stream"},{"name":"stdout","text":"CPU times: user 57.6 s, sys: 20.7 s, total: 1min 18s\nWall time: 49.4 s\n","output_type":"stream"}]},{"cell_type":"code","source":"# used cached function thereafter - super fast!\n%time text= pipeline(signal)","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:11:50.574436Z","iopub.execute_input":"2024-03-01T10:11:50.574848Z","iopub.status.idle":"2024-03-01T10:11:53.910240Z","shell.execute_reply.started":"2024-03-01T10:11:50.574817Z","shell.execute_reply":"2024-03-01T10:11:53.909281Z"},"trusted":true},"execution_count":63,"outputs":[{"name":"stderr","text":"WARNING:whisper_jax.pipeline:Numpy array passed as input - no sampling rate checks will be performed.It is strongly recommended to pass the input as a dictionary with an 'array' key containing the numpy array representing the audio, and a 'sampling_rate' key containing the sampling rate associated with the audio array.Failing to do so can result in silent errors that might be hard to debug.\n","output_type":"stream"},{"name":"stdout","text":"CPU times: user 12 s, sys: 20.5 s, total: 32.5 s\nWall time: 3.33 s\n","output_type":"stream"}]},{"cell_type":"code","source":"# let's check our transcription - looks spot on!\nprint(text)","metadata":{"tags":[],"execution":{"iopub.status.busy":"2024-03-01T10:11:57.416291Z","iopub.execute_input":"2024-03-01T10:11:57.416689Z","iopub.status.idle":"2024-03-01T10:11:57.421479Z","shell.execute_reply.started":"2024-03-01T10:11:57.416658Z","shell.execute_reply":"2024-03-01T10:11:57.420478Z"},"trusted":true},"execution_count":64,"outputs":[{"name":"stdout","text":"{'text': 'नमस्कार मैं विथापनी ऑफ़ एडमिनिस्ट्रेटिव फॉर्म फॉइन पब्लिक के लिए अब सब की तरफ की प्रश्ना बात करती हूँ बोली बोली मैं आपका यह कॉल आपकी ग्राह ऐसी न ने आपको सब का कॉल किया गया है जैसे एम सरकार प्रायः आपकी शिकायत दर जब तक दो अगस्त दो हज़ार बाईस तेईस को आपका फॉसिएशिएटिव दू बाई है मिनिस्ट्री ऑफ़ लेबर एंड इम्प्लॉयेंट अथपर जैसे हम क्या कर प्रकार हैं की अपनी शिकायत को बंद कर दिया जैसे इम्प्लायमेंट फण्ड ऑफ़ अर्गनाइजेशन की तरफ ऐसी इसमें आपको जो रिप्लाई दिया गया आपने चेक किया था तो सिर के आपको क्या जवाब दिया गया क्या समाधान मिला हाँ हाँ ठीक किया न ब्लैंड मैंने जी तो आपको जो रिप्लाई दिया गया आपने ऑफिसर सतिस्फाइड है और नॉट सतिस्फाइपको जो है रिप्लाई दिया की आपने आपसे सतिफाईड है और नॉट सतिफाईड है आप आपसे सतिफाईड है थैंक यू बेरिमस्टर कोई अब फीडबाइट गैस है इससे कोई सजेशन नहीं ब्रह्म दो मौली रूपए दो मिला था अच्छा है मैं रख रहा हूँ थैंक यू बेरिमस्टर कितना कीमती समय देने आरोप आपको जो हो'}\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## Run it again!","metadata":{}},{"cell_type":"markdown","source":"Now let's step it up a notch. Let's try transcribing 81 minutes of audio.","metadata":{}},{"cell_type":"code","source":"!pip install y2mate-api","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:40:35.743525Z","iopub.execute_input":"2024-03-01T10:40:35.744752Z","iopub.status.idle":"2024-03-01T10:40:40.380224Z","shell.execute_reply.started":"2024-03-01T10:40:35.744713Z","shell.execute_reply":"2024-03-01T10:40:40.378809Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stdout","text":"Requirement already satisfied: y2mate-api in /usr/local/lib/python3.8/site-packages (1.0.4)\nCollecting argparse>=1.1 (from y2mate-api)\n  Using cached argparse-1.4.0-py2.py3-none-any.whl (23 kB)\nRequirement already satisfied: tqdm==4.65.0 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (4.65.0)\nRequirement already satisfied: requests==2.28.2 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (2.28.2)\nRequirement already satisfied: colorama==0.4.6 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (0.4.6)\nRequirement already satisfied: appdirs==1.4.4 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (1.4.4)\nRequirement already satisfied: click==8.1.3 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (8.1.3)\nRequirement already satisfied: brotli==1.1.0 in /usr/local/lib/python3.8/site-packages (from y2mate-api) (1.1.0)\nRequirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.8/site-packages (from requests==2.28.2->y2mate-api) (3.1.0)\nRequirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.8/site-packages (from requests==2.28.2->y2mate-api) (3.4)\nRequirement already satisfied: urllib3<1.27,>=1.21.1 in /usr/local/lib/python3.8/site-packages (from requests==2.28.2->y2mate-api) (1.26.15)\nRequirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.8/site-packages (from requests==2.28.2->y2mate-api) (2022.12.7)\nInstalling collected packages: argparse\nSuccessfully installed argparse-1.4.0\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"code","source":"from y2mate_api import Handler\napi = Handler(\"CcY4GWjyl2U\") # youtube id for audio i have downloaded from\napi.save({'size': '75.4 MB', 'f': 'mp3', 'q': '128kbps', 'q_text': 'MP3 - 128kbps', 'k': 'joBHWYWzwJKgasL176uEqMDxAl7j4vImhNgqlAxyU/NQ9dhhhr32bZgdeelenN4=', 'status': 'ok', 'mess': '', 'c_status': 'CONVERTED', 'vid': 'CcY4GWjyl2U', 'title': \"India's #1 of the BEST Life Changing Seminar by Mohammad Shakeel (Motivational Speech)\", 'ftype': 'mp3', 'fquality': '128', 'dlink': 'https://dl204.filemate22.shop/?file=M3R4SUNiN3JsOHJ6WWQ2a3NQS1Y5ZGlxVlZIOCtyZ05sdEV6eGdOb1VPQkRvTVk3MytIckFleDhHdXdqeEl5bld2MWM5RERmZU42TWV6eU11NUVvUTJISzljOTJsRERIOG9NdFdNMDZaUjc1a09PbW1ucGJoaFA4YU5uWkhMZFliSHN3a0ZCbTFpV2JndnpFNmxQK3VuR29tRjJDZUN4WDkya3RKUHJGNVlwSzBDeVpTZnIwZ05WWGkzRGFzTHhxMmNuSmt6YjgycjVzalpKNFRoY3lkSlZTaEsvR21mWFV0VTRMaFkwVjEwajUrTFgyWDh4dFJmekxLbkZqTnpnTXRMaTJERWhHbTNaSHZuK29xNnNudWc9PQ%3D%3D'})\n","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:42:15.704988Z","iopub.execute_input":"2024-03-01T10:42:15.705390Z","iopub.status.idle":"2024-03-01T10:42:21.171440Z","shell.execute_reply.started":"2024-03-01T10:42:15.705361Z","shell.execute_reply":"2024-03-01T10:42:21.170297Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stdout","text":"10:42:17 - ERROR : __init__() got an unexpected keyword argument 'exit_on_error'\nIndia's #1 of the BEST Life Changing Seminar by Mohammad Shakeel (Motivational Speech) CcY4GWjyl2U_128.mp3\n","output_type":"stream"},{"name":"stderr","text":"\u001b[32m78 MB \u001b[36m           \u001b[33m|\u001b[39m    \n","output_type":"stream"},{"execution_count":17,"output_type":"execute_result","data":{"text/plain":"\"India's #1 of the BEST Life Changing Seminar by Mohammad Shakeel (Motivational Speech) CcY4GWjyl2U_128.mp3\""},"metadata":{}}]},{"cell_type":"code","source":"!mv \"/kaggle/working/India's #1 of the BEST Life Changing Seminar by Mohammad Shakeel (Motivational Speech) CcY4GWjyl2U_128.mp3\" /kaggle/working/81min.mp3","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:42:57.801751Z","iopub.execute_input":"2024-03-01T10:42:57.802158Z","iopub.status.idle":"2024-03-01T10:42:59.082689Z","shell.execute_reply.started":"2024-03-01T10:42:57.802126Z","shell.execute_reply":"2024-03-01T10:42:59.081263Z"},"trusted":true},"execution_count":18,"outputs":[]},{"cell_type":"code","source":"signal, sample_rate = librosa.load('/kaggle/working/81min.mp3', sr=16000)\n","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:46:25.897632Z","iopub.execute_input":"2024-03-01T10:46:25.898050Z","iopub.status.idle":"2024-03-01T10:46:39.112813Z","shell.execute_reply.started":"2024-03-01T10:46:25.898021Z","shell.execute_reply":"2024-03-01T10:46:39.111625Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"# transcribe using cached function\n%time text = pipeline(signal) #81 min long audio transcribing\n# 1min 19s means 80 sec for 81 min audio, that super fast, right?","metadata":{"execution":{"iopub.status.busy":"2024-03-01T10:47:19.211338Z","iopub.execute_input":"2024-03-01T10:47:19.211738Z","iopub.status.idle":"2024-03-01T10:48:38.244726Z","shell.execute_reply.started":"2024-03-01T10:47:19.211710Z","shell.execute_reply":"2024-03-01T10:48:38.243603Z"},"trusted":true},"execution_count":22,"outputs":[{"name":"stderr","text":"WARNING:whisper_jax.pipeline:Numpy array passed as input - no sampling rate checks will be performed.It is strongly recommended to pass the input as a dictionary with an 'array' key containing the numpy array representing the audio, and a 'sampling_rate' key containing the sampling rate associated with the audio array.Failing to do so can result in silent errors that might be hard to debug.\n","output_type":"stream"},{"name":"stdout","text":"CPU times: user 11min 6s, sys: 18min 51s, total: 29min 57s\nWall time: 1min 19s\n","output_type":"stream"}]},{"cell_type":"markdown","source":"Just **74s** to transcribe for **81 mins** of audio! That means you could transcribe an entire **2 hour movie in under 2.5 minutes 🤯** By increasing the batch size, we could also reduce the transcription time for long audio files further: increasing the batch size by 2x roughly decreases the transcription time by 2x, provided the overall batch size is less than the total audio time.\n\nIf you're fortunate enough to have access to a TPU v4, you'll find that the transcription times a factor of 2 faster than on a v3 - you can quickly see how we can get super fast transcription times using Whisper JAX on TPU!","metadata":{}},{"cell_type":"markdown","source":"## ⏰ Timestamps and more\n\nWe can also get timestamps from the model by passing `return_timestamps=True`, but this will require a recompilation since we change the signature of the forward pass. \n\nThe timestamps compilation takes longer than the non-timestamps one. Luckily, because we initialised our compilation cache above, we're not starting from scratch in compiling this time. This is the last compilation we need to do!","metadata":{}},{"cell_type":"code","source":"# compile the forward call with timestamps - slow but we only do once\n%time outputs = pipeline(signal, return_timestamps=True)\ntext = outputs[\"text\"]  # transcription\nchunks = outputs[\"chunks\"]  # transcription + timestamps","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# use cached timestamps function - super fast!\n%time outputs = pipeline(signal, return_timestamps=True)\ntext = outputs[\"text\"] \nchunks = outputs[\"chunks\"]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"We've shown how you can transcibe an audio file in English. The pipeline is also compatible with two further arguments that you can use to control the generation process. It's perfectly fine to omit these if you want speech transcription and the Whisper model to automatically detect which language the audio is in. Otherwise, you can change them depending on your task/language:\n\n\n* `task`: task to use for generation, either `\"transcribe\"` or `\"translate\"`. Defaults to `\"transcribe\"`.\n* `language`: language token to use for generation, can be either in the form of `\"<|en|>\"`, `\"en\"` or `\"english\"`. Defaults to `None`, meaning the language is automatically inferred from the audio input. Optional, and only relevant if the source audio language is known a-priori.","metadata":{}}]}