whisper_vocab_cleanup.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oIPd30CuphDD"
      },
      "source": [
        "# Updating Whisper's default vocabulary\n",
        "\n",
        "Some of the upstream Whisper models have been observed to swear (in English) when given non-speech input. This notebook adjusts the `vocab.json` for the `whisper-tiny` model to make such swearing less likely.\n",
        "\n",
        "More specifically, this notebook:\n",
        "- Fetches the upstream `whisper-tiny` model.\n",
        "- Replaces most full-swearword entries (see below).\n",
        "- Packages the model as GGML.\n",
        "\n",
        "After running this notebook, it will still be nessary to put the model in a `.zip` file in the format supported by Joplin (see the existing models for an example).\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "whisper_mode = 'tiny'"
      ],
      "metadata": {
        "id": "9XTvbHFnIjJ-"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6J7tG1nfpeQP",
        "outputId": "cd6a3c04-d416-4373-88f1-c15ab053111d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: pip in /usr/local/lib/python3.11/dist-packages (24.1.2)\n",
            "Collecting pip\n",
            "  Downloading pip-25.0.1-py3-none-any.whl.metadata (3.7 kB)\n",
            "Downloading pip-25.0.1-py3-none-any.whl (1.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.8/1.8 MB\u001b[0m \u001b[31m10.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hInstalling collected packages: pip\n",
            "  Attempting uninstall: pip\n",
            "    Found existing installation: pip 24.1.2\n",
            "    Uninstalling pip-24.1.2:\n",
            "      Successfully uninstalled pip-24.1.2\n",
            "Successfully installed pip-25.0.1\n",
            "Requirement already satisfied: transformers in /usr/local/lib/python3.11/dist-packages (4.48.3)\n",
            "Collecting transformers\n",
            "  Downloading transformers-4.49.0-py3-none-any.whl.metadata (44 kB)\n",
            "Collecting evaluate\n",
            "  Downloading evaluate-0.4.3-py3-none-any.whl.metadata (9.2 kB)\n",
            "Collecting jiwer\n",
            "  Downloading jiwer-3.1.0-py3-none-any.whl.metadata (2.6 kB)\n",
            "Collecting datasets[audio]\n",
            "  Downloading datasets-3.3.2-py3-none-any.whl.metadata (19 kB)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (3.17.0)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (1.26.4)\n",
            "Requirement already satisfied: pyarrow>=15.0.0 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (18.1.0)\n",
            "Collecting dill<0.3.9,>=0.3.0 (from datasets[audio])\n",
            "  Downloading dill-0.3.8-py3-none-any.whl.metadata (10 kB)\n",
            "Requirement already satisfied: pandas in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (2.2.2)\n",
            "Requirement already satisfied: requests>=2.32.2 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (2.32.3)\n",
            "Requirement already satisfied: tqdm>=4.66.3 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (4.67.1)\n",
            "Collecting xxhash (from datasets[audio])\n",
            "  Downloading xxhash-3.5.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (12 kB)\n",
            "Collecting multiprocess<0.70.17 (from datasets[audio])\n",
            "  Downloading multiprocess-0.70.16-py311-none-any.whl.metadata (7.2 kB)\n",
            "Requirement already satisfied: fsspec<=2024.12.0,>=2023.1.0 in /usr/local/lib/python3.11/dist-packages (from fsspec[http]<=2024.12.0,>=2023.1.0->datasets[audio]) (2024.10.0)\n",
            "Requirement already satisfied: aiohttp in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (3.11.13)\n",
            "Requirement already satisfied: huggingface-hub>=0.24.0 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (0.28.1)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (24.2)\n",
            "Requirement already satisfied: pyyaml>=5.1 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (6.0.2)\n",
            "Requirement already satisfied: soundfile>=0.12.1 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (0.13.1)\n",
            "Requirement already satisfied: librosa in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (0.10.2.post1)\n",
            "Requirement already satisfied: soxr>=0.4.0 in /usr/local/lib/python3.11/dist-packages (from datasets[audio]) (0.5.0.post1)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.11/dist-packages (from transformers) (2024.11.6)\n",
            "Requirement already satisfied: tokenizers<0.22,>=0.21 in /usr/local/lib/python3.11/dist-packages (from transformers) (0.21.0)\n",
            "Requirement already satisfied: safetensors>=0.4.1 in /usr/local/lib/python3.11/dist-packages (from transformers) (0.5.3)\n",
            "Requirement already satisfied: click>=8.1.8 in /usr/local/lib/python3.11/dist-packages (from jiwer) (8.1.8)\n",
            "Collecting rapidfuzz>=3.9.7 (from jiwer)\n",
            "  Downloading rapidfuzz-3.12.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (11 kB)\n",
            "Requirement already satisfied: aiohappyeyeballs>=2.3.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (2.4.6)\n",
            "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (1.3.2)\n",
            "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (25.1.0)\n",
            "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (1.5.0)\n",
            "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (6.1.0)\n",
            "Requirement already satisfied: propcache>=0.2.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (0.3.0)\n",
            "Requirement already satisfied: yarl<2.0,>=1.17.0 in /usr/local/lib/python3.11/dist-packages (from aiohttp->datasets[audio]) (1.18.3)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.11/dist-packages (from huggingface-hub>=0.24.0->datasets[audio]) (4.12.2)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.11/dist-packages (from requests>=2.32.2->datasets[audio]) (3.4.1)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.11/dist-packages (from requests>=2.32.2->datasets[audio]) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.11/dist-packages (from requests>=2.32.2->datasets[audio]) (2.3.0)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.11/dist-packages (from requests>=2.32.2->datasets[audio]) (2025.1.31)\n",
            "Requirement already satisfied: cffi>=1.0 in /usr/local/lib/python3.11/dist-packages (from soundfile>=0.12.1->datasets[audio]) (1.17.1)\n",
            "Requirement already satisfied: audioread>=2.1.9 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (3.0.1)\n",
            "Requirement already satisfied: scipy>=1.2.0 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (1.13.1)\n",
            "Requirement already satisfied: scikit-learn>=0.20.0 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (1.6.1)\n",
            "Requirement already satisfied: joblib>=0.14 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (1.4.2)\n",
            "Requirement already satisfied: decorator>=4.3.0 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (4.4.2)\n",
            "Requirement already satisfied: numba>=0.51.0 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (0.61.0)\n",
            "Requirement already satisfied: pooch>=1.1 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (1.8.2)\n",
            "Requirement already satisfied: lazy-loader>=0.1 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (0.4)\n",
            "Requirement already satisfied: msgpack>=1.0 in /usr/local/lib/python3.11/dist-packages (from librosa->datasets[audio]) (1.1.0)\n",
            "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.11/dist-packages (from pandas->datasets[audio]) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.11/dist-packages (from pandas->datasets[audio]) (2025.1)\n",
            "Requirement already satisfied: tzdata>=2022.7 in /usr/local/lib/python3.11/dist-packages (from pandas->datasets[audio]) (2025.1)\n",
            "Requirement already satisfied: pycparser in /usr/local/lib/python3.11/dist-packages (from cffi>=1.0->soundfile>=0.12.1->datasets[audio]) (2.22)\n",
            "Requirement already satisfied: llvmlite<0.45,>=0.44.0dev0 in /usr/local/lib/python3.11/dist-packages (from numba>=0.51.0->librosa->datasets[audio]) (0.44.0)\n",
            "Requirement already satisfied: platformdirs>=2.5.0 in /usr/local/lib/python3.11/dist-packages (from pooch>=1.1->librosa->datasets[audio]) (4.3.6)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.11/dist-packages (from python-dateutil>=2.8.2->pandas->datasets[audio]) (1.17.0)\n",
            "Requirement already satisfied: threadpoolctl>=3.1.0 in /usr/local/lib/python3.11/dist-packages (from scikit-learn>=0.20.0->librosa->datasets[audio]) (3.5.0)\n",
            "Downloading transformers-4.49.0-py3-none-any.whl (10.0 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m10.0/10.0 MB\u001b[0m \u001b[31m43.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading evaluate-0.4.3-py3-none-any.whl (84 kB)\n",
            "Downloading jiwer-3.1.0-py3-none-any.whl (22 kB)\n",
            "Downloading datasets-3.3.2-py3-none-any.whl (485 kB)\n",
            "Downloading dill-0.3.8-py3-none-any.whl (116 kB)\n",
            "Downloading multiprocess-0.70.16-py311-none-any.whl (143 kB)\n",
            "Downloading rapidfuzz-3.12.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (3.1 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m3.1/3.1 MB\u001b[0m \u001b[31m25.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading xxhash-3.5.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (194 kB)\n",
            "Installing collected packages: xxhash, rapidfuzz, dill, multiprocess, jiwer, transformers, datasets, evaluate\n",
            "  Attempting uninstall: transformers\n",
            "    Found existing installation: transformers 4.48.3\n",
            "    Uninstalling transformers-4.48.3:\n",
            "      Successfully uninstalled transformers-4.48.3\n",
            "Successfully installed datasets-3.3.2 dill-0.3.8 evaluate-0.4.3 jiwer-3.1.0 multiprocess-0.70.16 rapidfuzz-3.12.1 transformers-4.49.0 xxhash-3.5.0\n"
          ]
        }
      ],
      "source": [
        "!pip install --upgrade pip\n",
        "# jiwer is used for the word error rate (WER) metric\n",
        "!pip install --upgrade datasets[audio] transformers evaluate jiwer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 39
        },
        "id": "uCkfg9oqhW97",
        "outputId": "a7b72027-7e1c-4d7d-c2a7-383512e2da87"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<button onClick=\"this.nextSibling.style.display='block';this.style.display='none';\">Display W&B run</button><iframe src='https://wandb.ai/dummy/dummy/runs/zp6n1nzs?jupyter=true' style='border:none;width:100%;height:420px;display:none;'></iframe>"
            ],
            "text/plain": [
              "<wandb.sdk.wandb_run.Run at 0x7e2c8f00cf50>"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ],
      "source": [
        "import wandb\n",
        "# See https://discuss.huggingface.co/t/how-to-turn-wandb-off-in-trainer/6237/10\n",
        "wandb.init(mode='disabled')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "hmdwvSfAmPwo"
      },
      "outputs": [],
      "source": [
        "from pathlib import Path\n",
        "checkpoint_path = Path('./whisper/checkpoints').resolve()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "h24UgNf7mVOB"
      },
      "outputs": [],
      "source": [
        "import shutil\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M1HyDz2cRvVQ"
      },
      "source": [
        "## Create the feature extractor and tokenizer\n",
        "\n",
        "Fetch the feature extractor and tokenizer for this model from Huggingface:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "FPGYcxa7zmO5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 273,
          "referenced_widgets": [
            "51e33c5351a74ed08a3a17523af54aca",
            "8e6923494af245b1b1885baa6f4ece77",
            "45d711f248bf4c81b45c5e6f5ea53037",
            "ceef1fbfad6a41e9a21ee8142bb790f8",
            "efffdb9d52e44381932613bec4038f80",
            "7974a0822a3b4e3a97d9291de4769614",
            "b44b737695d34f1187181709ed82b131",
            "54a1b3498da04043bb32a7524f4cd63f",
            "3ee92336df6a4ca0b3712fa6753bd519",
            "1177d97f2b254bcf9c4cd0f02167732a",
            "287813d864024d358f7f4148220d0791",
            "31691f20a6dc40a98422ec168251dc5d",
            "35b7313901554d46b4fb52198a00340d",
            "e40a68341c3a40e28ccfd2063f46b374",
            "51553b842c2f41b5954122939b1660a3",
            "e8ac88d9de14476a979c9e145dfbc122",
            "2a1f4e1fa85b42eaa0b4a5cbf0b64f8b",
            "d9930f3cd3074d9face7778a8511e2e2",
            "078038fa96644f4e82054970f978c165",
            "88caa5476c414b0189242f165b0f7f50",
            "4de1dc09c2174b34b34a29627f64d82c",
            "a7b986ddd910453f9ff39499b00cf034",
            "14a90640bfde4589bc0ad25d1848266b",
            "77b9055500a041db9a6427e795ca2440",
            "f3b04084b8744cd6b36e82230adfba37",
            "d5e3cc1505324285802ef60eba8167e5",
            "c6461c1dd97a41538e076fc5388ab128",
            "d62cce5d5d3b426c9b347f10d7a1c1a7",
            "978cb2c586b4428281f036daa52b5fcc",
            "d199c5845385448e9a8c592cb7afac17",
            "96ee23a97ba64a518495e015f2943f66",
            "e536ea375cc849669bad66b055d6f2df",
            "129e4b5d228f441ba7bcda7031d73e16",
            "f2f0b564fc7f408d8a858f144d136e12",
            "a61ee831e9a641ed9d053fe2ae247479",
            "145101e9efa7454babbf19483af7d05c",
            "56f1c590824a49e3b1f737863e0d2cef",
            "ccd6078d37f3413f90c590a8b2cabd22",
            "a3e3c77a10dc413e9d0bffec564fec74",
            "afb06c06d562464caa9cb2d214a44c6b",
            "2021afccba0a473ca7d2de62e808927f",
            "390b00c3b6284b3cb26b34965f85f121",
            "e426d69a070a4f05ae6e2d7fd136a0b5",
            "dda824f44e2f46e8bc02ee226f9e15c3",
            "cbe590638e1c4ccf91bc27e483ec4da3",
            "f0f78d395bc44366bd0395e0cbfe491c",
            "a9169039361b49fe8aaf571a028ebe53",
            "374c527161ab4d19af2687c6327449c9",
            "e69ab604b17544d2a8284b9de7b70f1c",
            "fca7440f08f444e38f207bcc28af4e98",
            "87934b97cc5f4fe5aa9eda646c282098",
            "7ed2c7076f614a01bfda4e26aeca086b",
            "3ec75421502e459d97326b4706bf4514",
            "fc153028c0ce4529b6194dd54c8f7549",
            "a189e5ac99c3456cb6c1059ac5188bfc",
            "97b1014caa84403ea551cdd8be05c914",
            "cb99055498334cecad43b136cc70ce72",
            "317fbd3b3dca44c6896a6608dc04b4ac",
            "c68afccfd19f4f67b368a93d01d3d4b9",
            "6b368e5500a04e75a7be4b65004d1ec6",
            "bcb12acd5e564288a2c2a74fbf1a4868",
            "dd3f34e27dba438eb5feb13ac938fee8",
            "e59bde27da1747e4b5f60ec467090a1b",
            "0b60360e788b4636b9d9a683ef1358bd",
            "7091c2f3880b4ed2a7bb1f92c5865847",
            "862fd35e1d984c56b8615a1a6d647672",
            "32bffc2faba844a8b33dfb7ea81effad",
            "8dad60fe67b24411a75476870e86a724",
            "d7a5ca4b2e4546a1bedefe00b3365e55",
            "cd5c485cb4eb4682b45d7042b5afddfe",
            "56d2d0258c7d4b278de3ba42826680c1",
            "2928325954d143608225a5a220f80bfb",
            "c92c1443d92440b3ad6861998693e921",
            "c8d079922c9447a096eb22f139e41d51",
            "aa177ec5c523461081fba9f533e09c73",
            "b90600a633b64293b03dd3d48f8cdc34",
            "6037bc8a0eb24eae9adf7ddd1e4920fb",
            "36237cc2ee7d401e8c74f0b32df25845",
            "abcd74688b674f3282300fe52587ac48",
            "323419bb521348a6862317d6229a1bfc",
            "7ec43ca1b0e64892b6fccef9b123b4af",
            "7de61e234b05499a960ea520fc678a23",
            "5fcf053e08d54910a77610f3e8d6e51d",
            "21f393e2240c488598702104ef9aa584",
            "d466f57c1f8e4eb99aff9d1d1b336890",
            "10905252963449608b4cf2a72d04659b",
            "2ba32ce91480468cb22d44221047259f",
            "c9e35fc216754d878f125877944afd22"
          ]
        },
        "outputId": "094129d9-c5d0-48a5-ea4f-771c046af6b1"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "preprocessor_config.json:   0%|          | 0.00/185k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "51e33c5351a74ed08a3a17523af54aca"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer_config.json:   0%|          | 0.00/283k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "31691f20a6dc40a98422ec168251dc5d"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "vocab.json:   0%|          | 0.00/836k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "14a90640bfde4589bc0ad25d1848266b"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "tokenizer.json:   0%|          | 0.00/2.48M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "f2f0b564fc7f408d8a858f144d136e12"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "merges.txt:   0%|          | 0.00/494k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "cbe590638e1c4ccf91bc27e483ec4da3"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "normalizer.json:   0%|          | 0.00/52.7k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "97b1014caa84403ea551cdd8be05c914"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "added_tokens.json:   0%|          | 0.00/34.6k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "32bffc2faba844a8b33dfb7ea81effad"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "special_tokens_map.json:   0%|          | 0.00/2.19k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "36237cc2ee7d401e8c74f0b32df25845"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "from transformers import WhisperFeatureExtractor, WhisperTokenizer\n",
        "\n",
        "finetune_from_id = 'openai/whisper-' + whisper_mode\n",
        "feature_extractor = WhisperFeatureExtractor.from_pretrained(finetune_from_id,  task='transcribe')\n",
        "tokenizer_original = WhisperTokenizer.from_pretrained(finetune_from_id, task='transcribe')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dAjHVQPCueTO"
      },
      "source": [
        "We'll create a customized tokenizer based on `tokenizer_original` in the next section."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S2-3B1TW3Tc1"
      },
      "source": [
        "## Vocabulary adjustements\n",
        "\n",
        "Next, we remove several unwanted tokens from the vocabulary:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WWsP8YfrcmUC",
        "outputId": "bf15066e-b179-4168-c479-19e3ebc35573"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('whisper-default-tokenizer/tokenizer_config.json',\n",
              " 'whisper-default-tokenizer/special_tokens_map.json',\n",
              " 'whisper-default-tokenizer/vocab.json',\n",
              " 'whisper-default-tokenizer/merges.txt',\n",
              " 'whisper-default-tokenizer/normalizer.json',\n",
              " 'whisper-default-tokenizer/added_tokens.json')"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ],
      "source": [
        "# Step 1: Save the vocabulary to a file\n",
        "tokenizer_directory = Path('whisper-default-tokenizer')\n",
        "tokenizer_original.save_pretrained(tokenizer_directory)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ES0enoY-ezxm"
      },
      "source": [
        "Now that the tokenizer is saved in `tokenizer_directory`, we can load `tokenizer_directory/vocab.json` and modify it:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "jtHNBGVPewT2"
      },
      "outputs": [],
      "source": [
        "# Step 2: Get vocab.json\n",
        "import json\n",
        "\n",
        "def json_from_path(path: Path):\n",
        "    with open(path, 'r', encoding='utf-8') as f:\n",
        "        return json.loads(f.read())\n",
        "\n",
        "vocab = json_from_path(tokenizer_directory / 'vocab.json')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "TBqPdNpLkSYg"
      },
      "outputs": [],
      "source": [
        "import re\n",
        "NONWORD_REGEX = re.compile(r'[ \\t?.,;!()/\\-«»]+')\n",
        "def split_by_word(text: str):\n",
        "    \"\"\" Splits the given `text` into words. Returns a list of those words. \"\"\"\n",
        "    return NONWORD_REGEX.split(text)\n",
        "\n",
        "\n",
        "# This character marks the beginning of a word in vocab.json\n",
        "word_start_char = 'Ġ'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "PPBQrfHmsEdz"
      },
      "outputs": [],
      "source": [
        "# Step 3: Replace!\n",
        "next_replacement_idx = 0\n",
        "new_vocab = {}\n",
        "\n",
        "# Token IDs can be found by inspecting the original vocab.json. These token IDs\n",
        "# are specific to the multilingual whisper-tiny, but may also work for whisper-base. Each remapping should be unique.\n",
        "token_id_remappings = {\n",
        "    19186: \"[swearS1]\", # s***\n",
        "    30748: word_start_char + \"[swearS2]\",\n",
        "    4611: word_start_char + \"[swearS3]\",\n",
        "    19593: word_start_char + \"[swearS4]\", # S***\n",
        "    10965: word_start_char + \"[swearF1]\", # F***\n",
        "    26154: word_start_char + \"[swearF2]\", # F***\n",
        "    33342: word_start_char + \"[swearF3]\",\n",
        "    47069: word_start_char + \"[swearF4]\", # f****\n",
        "    3275: word_start_char + \"[swearF5]\",\n",
        "    22518: word_start_char + \"[swearF6]\",\n",
        "    20022: word_start_char + \"[swearF7]\",\n",
        "    5546: word_start_char + \"[swearF8]\",\n",
        "    47069: word_start_char + \"[swearM1]\",\n",
        "    29537: word_start_char + \"[swearM2]\",\n",
        "    22676: word_start_char + \"[swearB1]\", # bull****\n",
        "    11960: word_start_char + \"[swearB2]\",\n",
        "    42094: word_start_char + \"[swearB3]\",\n",
        "    40678: word_start_char + \"[swearB4]\"\n",
        "}\n",
        "replaced_keys = set()\n",
        "\n",
        "for key in vocab:\n",
        "    token_id = vocab[key]\n",
        "    if token_id in token_id_remappings:\n",
        "        new_key = token_id_remappings[token_id]\n",
        "        new_vocab[new_key] = token_id\n",
        "        replaced_keys.add(key)\n",
        "    else:\n",
        "        new_vocab[key] = token_id\n",
        "\n",
        "new_merges = []\n",
        "with open(tokenizer_directory / 'merges.txt', 'r', encoding='utf-8') as merges:\n",
        "    for line in merges.readlines():\n",
        "        if len(line) == 0:\n",
        "            continue\n",
        "        words = split_by_word(line)\n",
        "        if not (words[0] in replaced_keys):\n",
        "            new_merges.append(line.strip())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EKCuKjtPIRaR"
      },
      "source": [
        "To check for other indexes to replace (keeping in mind that the output should still be multi-lingual), we could do something like this:\n",
        "```python\n",
        "!pip install better_profanity==0.7.0\n",
        "\n",
        "from better_profanity import profanity\n",
        "\n",
        "profanity.load_censor_words()\n",
        "for key in new_vocab:\n",
        "    word = key\n",
        "    if key.startswith(word_start_char):\n",
        "        word = key[1:]\n",
        "    if profanity.contains_profanity(word):\n",
        "        print(\"Consider replacing\", key, new_vocab[key])\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# !pip install better_profanity==0.7.0\n",
        "\n",
        "# from better_profanity import profanity\n",
        "\n",
        "# profanity.load_censor_words()\n",
        "# for key in new_vocab:\n",
        "#     word = key\n",
        "#     if key.startswith(word_start_char):\n",
        "#         word = key[1:]\n",
        "#     if profanity.contains_profanity(word):\n",
        "#         print(\"Consider replacing\", key, new_vocab[key])"
      ],
      "metadata": {
        "id": "mhNhzp98FjFR"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y_lDgRsIuLRY"
      },
      "source": [
        "Great! We now have an updated vocab file!"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "xa4M3M_02GwI"
      },
      "outputs": [],
      "source": [
        "# Write to a file\n",
        "tokenizer_fr_directory = Path('updated-tokenizer')\n",
        "if tokenizer_fr_directory.exists():\n",
        "    shutil.rmtree(tokenizer_fr_directory)\n",
        "shutil.copytree(tokenizer_directory, tokenizer_fr_directory)\n",
        "with open(tokenizer_fr_directory / 'vocab.json', 'w', encoding='utf-8') as f:\n",
        "    json.dump(new_vocab, f, ensure_ascii=False)\n",
        "\n",
        "\n",
        "with open(tokenizer_fr_directory / 'merges.txt', 'w', encoding='utf-8') as f:\n",
        "    f.write('\\n'.join(new_merges))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2oSULQLZz5rQ",
        "outputId": "9097ce7f-4781-44c5-b339-2c350d008263"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "105"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "from transformers import WhisperTokenizer\n",
        "\n",
        "# Use a normal WhisperTokenizer -- WhisperTokenizerFast has trouble with the updated\n",
        "# vocabulary.\n",
        "tokenizer = WhisperTokenizer(\n",
        "    tokenizer_fr_directory / 'vocab.json',\n",
        "    tokenizer_fr_directory / 'merges.txt',\n",
        "    tokenizer_fr_directory / 'normalizer.json',\n",
        "    bos_token='<|startoftranscript|>',\n",
        "    unk_token='',\n",
        "    pad_token='<|endoftext|>',\n",
        ")\n",
        "\n",
        "# See https://discuss.huggingface.co/t/fine-tuning-whisper-on-my-own-dataset-with-a-customized-tokenizer/25903\n",
        "tokenizer.add_special_tokens(tokenizer_original.special_tokens_map)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wl_cU43c7IHR",
        "outputId": "dd124140-5bd6-43b4-ea7a-18582aaa0384"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('updated-tokenizer/tokenizer_config.json',\n",
              " 'updated-tokenizer/special_tokens_map.json',\n",
              " 'updated-tokenizer/vocab.json',\n",
              " 'updated-tokenizer/merges.txt',\n",
              " 'updated-tokenizer/normalizer.json',\n",
              " 'updated-tokenizer/added_tokens.json')"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ],
      "source": [
        "# For debugging, update the output directory\n",
        "shutil.rmtree(tokenizer_fr_directory)\n",
        "tokenizer.save_pretrained(tokenizer_fr_directory)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DEcid4Cj4H_V"
      },
      "source": [
        "## Create the processor\n",
        "\n",
        "Next, load the `WhisperProcessor`, which combines a feature extractor and tokenizer."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "3usf8ghTRHrW"
      },
      "outputs": [],
      "source": [
        "from transformers import WhisperProcessor\n",
        "\n",
        "processor = WhisperProcessor(feature_extractor, tokenizer)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QMpuESBPSRs6"
      },
      "source": [
        "Next, build the model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "X6EHVfw4TRLz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 113,
          "referenced_widgets": [
            "b8b1f114b84a48c8befc607d75d19217",
            "7c043b15c7cb471297f6cfacabce40be",
            "d7a2aa1f36fb40f1a3c9f101da6f1fcc",
            "9cd2de48c15e485492cbc9ac76e21df0",
            "716626235b15484fbe3bdb87dbb80d3b",
            "92d4442fe8b94b3cb6d652b72dc55428",
            "3e3cbed9d80f4d9c87db0e375fa2ee2b",
            "16ab731566804579a6548ddde822518f",
            "d360f8037eba412882d9807b838ae534",
            "26882105686847bdb309af4c0faacbb9",
            "b1171a864e524ffc9378d79e240081e0",
            "b07ce25529044a6081faf108a91b46f9",
            "b74145815a8047aca9da40a2f5170bd2",
            "ee8ae5e130904544bb1963d5b61f66a9",
            "52a58a18fb6b4f5a9fb1ae4aec4dbe98",
            "a9bb8a8adf3443dc9588f886dd17ec4a",
            "7b7bf3be758c4dd59476bc05ad4b4e0f",
            "49d5167b64634d609265c804662c70a9",
            "0e49a94023644036a8feb6e783c77566",
            "1903365501904d11bcfbcf76e9cbd323",
            "84886ef71ecb449aa0d5363131d2c5d2",
            "f64bde590bf14a36b35c66467da00a79",
            "fa9d2724157f4da2b7acc8590e35b981",
            "c9dc2532e8a9404588c31ccaeca42bc6",
            "40aa791f749943d6b6317ee254304d13",
            "c629090bb492462a9e4066ed58a91589",
            "e673672365ad409e9e83af4c1141c0dd",
            "8dbbea5d076548cbabf14a1ef610f71b",
            "b5b0df9d638e4c8a87880c4e7b0c7ca3",
            "4bac8d95267743d9baa4bc4c3292521f",
            "4c0b2925e91b455ea7127a2fdf75d1fa",
            "11689b3e624b4fb69c53a3e3f8503099",
            "87e6900700e54ff7bc409e0c2f2b7728"
          ]
        },
        "outputId": "8801c3aa-72ef-48df-9728-9cb84352d765"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "config.json:   0%|          | 0.00/1.98k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "b8b1f114b84a48c8befc607d75d19217"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "model.safetensors:   0%|          | 0.00/151M [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "b07ce25529044a6081faf108a91b46f9"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "generation_config.json:   0%|          | 0.00/3.75k [00:00<?, ?B/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "fa9d2724157f4da2b7acc8590e35b981"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "from transformers import WhisperForConditionalGeneration\n",
        "\n",
        "model = WhisperForConditionalGeneration.from_pretrained(finetune_from_id)\n",
        "model.generation_config.forced_decoder_ids = None\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "JVAJV11oUs5A"
      },
      "outputs": [],
      "source": [
        "from dataclasses import dataclass\n",
        "from typing import Any\n",
        "import torch\n",
        "# See the linked blog post and https://huggingface.co/docs/transformers/main_classes/data_collator\n",
        "\n",
        "@dataclass\n",
        "class DataCollatorWithPadding:\n",
        "    ''' Converts raw data into a batch ready for the model '''\n",
        "    processor: Any\n",
        "    decoder_start_token_id: int\n",
        "\n",
        "    def __call__(self, features: list) -> dict[str, torch.Tensor]:\n",
        "        input_features = [{'input_features': f['input_features']} for f in features]\n",
        "        label_features = [{'input_ids': f['labels']} for f in features]\n",
        "\n",
        "        # According to the linked blog post, the input and label features need\n",
        "        # to be padded separately (due to different final lengths), then\n",
        "        # recombined:\n",
        "        batch = self.processor.feature_extractor.pad(input_features, return_tensors='pt')\n",
        "\n",
        "        labels_batch = self.processor.tokenizer.pad(label_features, return_tensors='pt')\n",
        "\n",
        "        # transformers uses -100 for masking\n",
        "        labels = labels_batch['input_ids'].masked_fill(labels_batch.attention_mask.ne(1), -100)\n",
        "\n",
        "        # Don't double-prepend the beginning of sequence token:\n",
        "        if (labels[:,0] == self.decoder_start_token_id).all().cpu().item():\n",
        "            labels = labels[:, 1:]\n",
        "\n",
        "        batch['labels'] = labels\n",
        "        return batch\n",
        "\n",
        "data_collator = DataCollatorWithPadding(processor=processor, decoder_start_token_id=model.config.decoder_start_token_id)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jAjMqolA-957",
        "outputId": "89b3f08e-f0f4-4b16-80db-5e817ddd6f87"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.11/dist-packages/transformers/modeling_utils.py:2810: UserWarning: Moving the following attributes in the config to the generation config: {'max_length': 448, 'suppress_tokens': [1, 2, 7, 8, 9, 10, 14, 25, 26, 27, 28, 29, 31, 58, 59, 60, 61, 62, 63, 90, 91, 92, 93, 359, 503, 522, 542, 873, 893, 902, 918, 922, 931, 1350, 1853, 1982, 2460, 2627, 3246, 3253, 3268, 3536, 3846, 3961, 4183, 4667, 6585, 6647, 7273, 9061, 9383, 10428, 10929, 11938, 12033, 12331, 12562, 13793, 14157, 14635, 15265, 15618, 16553, 16604, 18362, 18956, 20075, 21675, 22520, 26130, 26161, 26435, 28279, 29464, 31650, 32302, 32470, 36865, 42863, 47425, 49870, 50254, 50258, 50358, 50359, 50360, 50361, 50362], 'begin_suppress_tokens': [220, 50257]}. You are seeing this warning because you've set generation parameters in the model config, as opposed to in the generation config.\n",
            "  warnings.warn(\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('/content/final-model/tokenizer_config.json',\n",
              " '/content/final-model/special_tokens_map.json',\n",
              " '/content/final-model/vocab.json',\n",
              " '/content/final-model/merges.txt',\n",
              " '/content/final-model/normalizer.json',\n",
              " '/content/final-model/added_tokens.json')"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "model_output_dir = Path('./final-model').resolve()\n",
        "model.save_pretrained(model_output_dir)\n",
        "tokenizer.save_pretrained(model_output_dir)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g0e9GSFSfDsr"
      },
      "source": [
        "# Model conversion\n",
        "\n",
        "Next, we need to convert the model into a format usable by Joplin. This next step converts the model from PyTorch to GGML."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Sz5y6PdM5Qei",
        "outputId": "5329621e-ae11-4138-c92f-f817d26f40f0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cloning into 'whisper-github'...\n",
            "remote: Enumerating objects: 828, done.\u001b[K\n",
            "remote: Counting objects: 100% (370/370), done.\u001b[K\n",
            "remote: Compressing objects: 100% (69/69), done.\u001b[K\n",
            "remote: Total 828 (delta 333), reused 301 (delta 301), pack-reused 458 (from 2)\u001b[K\n",
            "Receiving objects: 100% (828/828), 8.26 MiB | 18.15 MiB/s, done.\n",
            "Resolving deltas: 100% (496/496), done.\n",
            "Cloning into 'whisper.cpp'...\n",
            "remote: Enumerating objects: 15791, done.\u001b[K\n",
            "remote: Counting objects: 100% (2864/2864), done.\u001b[K\n",
            "remote: Compressing objects: 100% (453/453), done.\u001b[K\n",
            "remote: Total 15791 (delta 2513), reused 2411 (delta 2410), pack-reused 12927 (from 4)\u001b[K\n",
            "Receiving objects: 100% (15791/15791), 19.84 MiB | 16.74 MiB/s, done.\n",
            "Resolving deltas: 100% (10873/10873), done.\n",
            "Note: switching to 'v1.7.4'.\n",
            "\n",
            "You are in 'detached HEAD' state. You can look around, make experimental\n",
            "changes and commit them, and you can discard any commits you make in this\n",
            "state without impacting any branches by switching back to a branch.\n",
            "\n",
            "If you want to create a new branch to retain commits you create, you may\n",
            "do so (now or later) by using -c with the switch command. Example:\n",
            "\n",
            "  git switch -c <new-branch-name>\n",
            "\n",
            "Or undo this operation with:\n",
            "\n",
            "  git switch -\n",
            "\n",
            "Turn off this advice by setting config variable advice.detachedHead to false\n",
            "\n",
            "HEAD is now at 8a9ad78 release : v1.7.4\n"
          ]
        }
      ],
      "source": [
        "!git clone /~https://github.com/openai/whisper whisper-github\n",
        "!git clone /~https://github.com/ggerganov/whisper.cpp\n",
        "!cd whisper.cpp && git checkout v1.7.4"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "MUMzKNCmJfq0"
      },
      "outputs": [],
      "source": [
        "# Patch convert-h5-to-ggml to work with more recent model versions\n",
        "conversion_script_path = Path('whisper.cpp/models/convert-h5-to-ggml.py')\n",
        "conversion_script_content = conversion_script_path.read_text()\n",
        "with open(conversion_script_path, 'w') as conversion_script:\n",
        "    bad_if_statement = 'if \"max_length\" not in hparams:'\n",
        "    replaced_if_statement = 'if \"max_length\" not in hparams or hparams[\"max_length\"] == None:'\n",
        "    conversion_script.write(conversion_script_content.replace(bad_if_statement, replaced_if_statement))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "A6eeyWNgdw8f",
        "outputId": "22c44f79-9a7a-4bec-d3bb-4167f26585b5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2025-03-01 00:41:01.294144: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:477] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered\n",
            "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
            "E0000 00:00:1740789661.339763     859 cuda_dnn.cc:8310] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered\n",
            "E0000 00:00:1740789661.353759     859 cuda_blas.cc:1418] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered\n",
            "model.encoder.conv1.weight  ->  encoder.conv1.weight\n",
            "encoder.conv1.weight 3 (384, 80, 3)\n",
            "model.encoder.conv1.bias  ->  encoder.conv1.bias\n",
            "  Reshaped variable:  encoder.conv1.bias  to shape:  (384, 1)\n",
            "encoder.conv1.bias 2 (384, 1)\n",
            "  Converting to float32\n",
            "model.encoder.conv2.weight  ->  encoder.conv2.weight\n",
            "encoder.conv2.weight 3 (384, 384, 3)\n",
            "model.encoder.conv2.bias  ->  encoder.conv2.bias\n",
            "  Reshaped variable:  encoder.conv2.bias  to shape:  (384, 1)\n",
            "encoder.conv2.bias 2 (384, 1)\n",
            "  Converting to float32\n",
            "model.encoder.embed_positions.weight  ->  encoder.positional_embedding\n",
            "encoder.positional_embedding 2 (1500, 384)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.self_attn.k_proj.weight  ->  encoder.blocks.0.attn.key.weight\n",
            "encoder.blocks.0.attn.key.weight 2 (384, 384)\n",
            "model.encoder.layers.0.self_attn.v_proj.weight  ->  encoder.blocks.0.attn.value.weight\n",
            "encoder.blocks.0.attn.value.weight 2 (384, 384)\n",
            "model.encoder.layers.0.self_attn.v_proj.bias  ->  encoder.blocks.0.attn.value.bias\n",
            "encoder.blocks.0.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.self_attn.q_proj.weight  ->  encoder.blocks.0.attn.query.weight\n",
            "encoder.blocks.0.attn.query.weight 2 (384, 384)\n",
            "model.encoder.layers.0.self_attn.q_proj.bias  ->  encoder.blocks.0.attn.query.bias\n",
            "encoder.blocks.0.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.self_attn.out_proj.weight  ->  encoder.blocks.0.attn.out.weight\n",
            "encoder.blocks.0.attn.out.weight 2 (384, 384)\n",
            "model.encoder.layers.0.self_attn.out_proj.bias  ->  encoder.blocks.0.attn.out.bias\n",
            "encoder.blocks.0.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.self_attn_layer_norm.weight  ->  encoder.blocks.0.attn_ln.weight\n",
            "encoder.blocks.0.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.self_attn_layer_norm.bias  ->  encoder.blocks.0.attn_ln.bias\n",
            "encoder.blocks.0.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.fc1.weight  ->  encoder.blocks.0.mlp.0.weight\n",
            "encoder.blocks.0.mlp.0.weight 2 (1536, 384)\n",
            "model.encoder.layers.0.fc1.bias  ->  encoder.blocks.0.mlp.0.bias\n",
            "encoder.blocks.0.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.fc2.weight  ->  encoder.blocks.0.mlp.2.weight\n",
            "encoder.blocks.0.mlp.2.weight 2 (384, 1536)\n",
            "model.encoder.layers.0.fc2.bias  ->  encoder.blocks.0.mlp.2.bias\n",
            "encoder.blocks.0.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.final_layer_norm.weight  ->  encoder.blocks.0.mlp_ln.weight\n",
            "encoder.blocks.0.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.0.final_layer_norm.bias  ->  encoder.blocks.0.mlp_ln.bias\n",
            "encoder.blocks.0.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.self_attn.k_proj.weight  ->  encoder.blocks.1.attn.key.weight\n",
            "encoder.blocks.1.attn.key.weight 2 (384, 384)\n",
            "model.encoder.layers.1.self_attn.v_proj.weight  ->  encoder.blocks.1.attn.value.weight\n",
            "encoder.blocks.1.attn.value.weight 2 (384, 384)\n",
            "model.encoder.layers.1.self_attn.v_proj.bias  ->  encoder.blocks.1.attn.value.bias\n",
            "encoder.blocks.1.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.self_attn.q_proj.weight  ->  encoder.blocks.1.attn.query.weight\n",
            "encoder.blocks.1.attn.query.weight 2 (384, 384)\n",
            "model.encoder.layers.1.self_attn.q_proj.bias  ->  encoder.blocks.1.attn.query.bias\n",
            "encoder.blocks.1.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.self_attn.out_proj.weight  ->  encoder.blocks.1.attn.out.weight\n",
            "encoder.blocks.1.attn.out.weight 2 (384, 384)\n",
            "model.encoder.layers.1.self_attn.out_proj.bias  ->  encoder.blocks.1.attn.out.bias\n",
            "encoder.blocks.1.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.self_attn_layer_norm.weight  ->  encoder.blocks.1.attn_ln.weight\n",
            "encoder.blocks.1.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.self_attn_layer_norm.bias  ->  encoder.blocks.1.attn_ln.bias\n",
            "encoder.blocks.1.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.fc1.weight  ->  encoder.blocks.1.mlp.0.weight\n",
            "encoder.blocks.1.mlp.0.weight 2 (1536, 384)\n",
            "model.encoder.layers.1.fc1.bias  ->  encoder.blocks.1.mlp.0.bias\n",
            "encoder.blocks.1.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.fc2.weight  ->  encoder.blocks.1.mlp.2.weight\n",
            "encoder.blocks.1.mlp.2.weight 2 (384, 1536)\n",
            "model.encoder.layers.1.fc2.bias  ->  encoder.blocks.1.mlp.2.bias\n",
            "encoder.blocks.1.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.final_layer_norm.weight  ->  encoder.blocks.1.mlp_ln.weight\n",
            "encoder.blocks.1.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.1.final_layer_norm.bias  ->  encoder.blocks.1.mlp_ln.bias\n",
            "encoder.blocks.1.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.self_attn.k_proj.weight  ->  encoder.blocks.2.attn.key.weight\n",
            "encoder.blocks.2.attn.key.weight 2 (384, 384)\n",
            "model.encoder.layers.2.self_attn.v_proj.weight  ->  encoder.blocks.2.attn.value.weight\n",
            "encoder.blocks.2.attn.value.weight 2 (384, 384)\n",
            "model.encoder.layers.2.self_attn.v_proj.bias  ->  encoder.blocks.2.attn.value.bias\n",
            "encoder.blocks.2.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.self_attn.q_proj.weight  ->  encoder.blocks.2.attn.query.weight\n",
            "encoder.blocks.2.attn.query.weight 2 (384, 384)\n",
            "model.encoder.layers.2.self_attn.q_proj.bias  ->  encoder.blocks.2.attn.query.bias\n",
            "encoder.blocks.2.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.self_attn.out_proj.weight  ->  encoder.blocks.2.attn.out.weight\n",
            "encoder.blocks.2.attn.out.weight 2 (384, 384)\n",
            "model.encoder.layers.2.self_attn.out_proj.bias  ->  encoder.blocks.2.attn.out.bias\n",
            "encoder.blocks.2.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.self_attn_layer_norm.weight  ->  encoder.blocks.2.attn_ln.weight\n",
            "encoder.blocks.2.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.self_attn_layer_norm.bias  ->  encoder.blocks.2.attn_ln.bias\n",
            "encoder.blocks.2.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.fc1.weight  ->  encoder.blocks.2.mlp.0.weight\n",
            "encoder.blocks.2.mlp.0.weight 2 (1536, 384)\n",
            "model.encoder.layers.2.fc1.bias  ->  encoder.blocks.2.mlp.0.bias\n",
            "encoder.blocks.2.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.fc2.weight  ->  encoder.blocks.2.mlp.2.weight\n",
            "encoder.blocks.2.mlp.2.weight 2 (384, 1536)\n",
            "model.encoder.layers.2.fc2.bias  ->  encoder.blocks.2.mlp.2.bias\n",
            "encoder.blocks.2.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.final_layer_norm.weight  ->  encoder.blocks.2.mlp_ln.weight\n",
            "encoder.blocks.2.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.2.final_layer_norm.bias  ->  encoder.blocks.2.mlp_ln.bias\n",
            "encoder.blocks.2.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.self_attn.k_proj.weight  ->  encoder.blocks.3.attn.key.weight\n",
            "encoder.blocks.3.attn.key.weight 2 (384, 384)\n",
            "model.encoder.layers.3.self_attn.v_proj.weight  ->  encoder.blocks.3.attn.value.weight\n",
            "encoder.blocks.3.attn.value.weight 2 (384, 384)\n",
            "model.encoder.layers.3.self_attn.v_proj.bias  ->  encoder.blocks.3.attn.value.bias\n",
            "encoder.blocks.3.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.self_attn.q_proj.weight  ->  encoder.blocks.3.attn.query.weight\n",
            "encoder.blocks.3.attn.query.weight 2 (384, 384)\n",
            "model.encoder.layers.3.self_attn.q_proj.bias  ->  encoder.blocks.3.attn.query.bias\n",
            "encoder.blocks.3.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.self_attn.out_proj.weight  ->  encoder.blocks.3.attn.out.weight\n",
            "encoder.blocks.3.attn.out.weight 2 (384, 384)\n",
            "model.encoder.layers.3.self_attn.out_proj.bias  ->  encoder.blocks.3.attn.out.bias\n",
            "encoder.blocks.3.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.self_attn_layer_norm.weight  ->  encoder.blocks.3.attn_ln.weight\n",
            "encoder.blocks.3.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.self_attn_layer_norm.bias  ->  encoder.blocks.3.attn_ln.bias\n",
            "encoder.blocks.3.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.fc1.weight  ->  encoder.blocks.3.mlp.0.weight\n",
            "encoder.blocks.3.mlp.0.weight 2 (1536, 384)\n",
            "model.encoder.layers.3.fc1.bias  ->  encoder.blocks.3.mlp.0.bias\n",
            "encoder.blocks.3.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.fc2.weight  ->  encoder.blocks.3.mlp.2.weight\n",
            "encoder.blocks.3.mlp.2.weight 2 (384, 1536)\n",
            "model.encoder.layers.3.fc2.bias  ->  encoder.blocks.3.mlp.2.bias\n",
            "encoder.blocks.3.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.final_layer_norm.weight  ->  encoder.blocks.3.mlp_ln.weight\n",
            "encoder.blocks.3.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layers.3.final_layer_norm.bias  ->  encoder.blocks.3.mlp_ln.bias\n",
            "encoder.blocks.3.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layer_norm.weight  ->  encoder.ln_post.weight\n",
            "encoder.ln_post.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.encoder.layer_norm.bias  ->  encoder.ln_post.bias\n",
            "encoder.ln_post.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.embed_tokens.weight  ->  decoder.token_embedding.weight\n",
            "decoder.token_embedding.weight 2 (51865, 384)\n",
            "model.decoder.embed_positions.weight  ->  decoder.positional_embedding\n",
            "decoder.positional_embedding 2 (448, 384)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.self_attn.k_proj.weight  ->  decoder.blocks.0.attn.key.weight\n",
            "decoder.blocks.0.attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.0.self_attn.v_proj.weight  ->  decoder.blocks.0.attn.value.weight\n",
            "decoder.blocks.0.attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.0.self_attn.v_proj.bias  ->  decoder.blocks.0.attn.value.bias\n",
            "decoder.blocks.0.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.self_attn.q_proj.weight  ->  decoder.blocks.0.attn.query.weight\n",
            "decoder.blocks.0.attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.0.self_attn.q_proj.bias  ->  decoder.blocks.0.attn.query.bias\n",
            "decoder.blocks.0.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.self_attn.out_proj.weight  ->  decoder.blocks.0.attn.out.weight\n",
            "decoder.blocks.0.attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.0.self_attn.out_proj.bias  ->  decoder.blocks.0.attn.out.bias\n",
            "decoder.blocks.0.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.self_attn_layer_norm.weight  ->  decoder.blocks.0.attn_ln.weight\n",
            "decoder.blocks.0.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.self_attn_layer_norm.bias  ->  decoder.blocks.0.attn_ln.bias\n",
            "decoder.blocks.0.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.encoder_attn.k_proj.weight  ->  decoder.blocks.0.cross_attn.key.weight\n",
            "decoder.blocks.0.cross_attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.0.encoder_attn.v_proj.weight  ->  decoder.blocks.0.cross_attn.value.weight\n",
            "decoder.blocks.0.cross_attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.0.encoder_attn.v_proj.bias  ->  decoder.blocks.0.cross_attn.value.bias\n",
            "decoder.blocks.0.cross_attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.encoder_attn.q_proj.weight  ->  decoder.blocks.0.cross_attn.query.weight\n",
            "decoder.blocks.0.cross_attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.0.encoder_attn.q_proj.bias  ->  decoder.blocks.0.cross_attn.query.bias\n",
            "decoder.blocks.0.cross_attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.encoder_attn.out_proj.weight  ->  decoder.blocks.0.cross_attn.out.weight\n",
            "decoder.blocks.0.cross_attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.0.encoder_attn.out_proj.bias  ->  decoder.blocks.0.cross_attn.out.bias\n",
            "decoder.blocks.0.cross_attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.encoder_attn_layer_norm.weight  ->  decoder.blocks.0.cross_attn_ln.weight\n",
            "decoder.blocks.0.cross_attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.encoder_attn_layer_norm.bias  ->  decoder.blocks.0.cross_attn_ln.bias\n",
            "decoder.blocks.0.cross_attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.fc1.weight  ->  decoder.blocks.0.mlp.0.weight\n",
            "decoder.blocks.0.mlp.0.weight 2 (1536, 384)\n",
            "model.decoder.layers.0.fc1.bias  ->  decoder.blocks.0.mlp.0.bias\n",
            "decoder.blocks.0.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.fc2.weight  ->  decoder.blocks.0.mlp.2.weight\n",
            "decoder.blocks.0.mlp.2.weight 2 (384, 1536)\n",
            "model.decoder.layers.0.fc2.bias  ->  decoder.blocks.0.mlp.2.bias\n",
            "decoder.blocks.0.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.final_layer_norm.weight  ->  decoder.blocks.0.mlp_ln.weight\n",
            "decoder.blocks.0.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.0.final_layer_norm.bias  ->  decoder.blocks.0.mlp_ln.bias\n",
            "decoder.blocks.0.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.self_attn.k_proj.weight  ->  decoder.blocks.1.attn.key.weight\n",
            "decoder.blocks.1.attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.1.self_attn.v_proj.weight  ->  decoder.blocks.1.attn.value.weight\n",
            "decoder.blocks.1.attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.1.self_attn.v_proj.bias  ->  decoder.blocks.1.attn.value.bias\n",
            "decoder.blocks.1.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.self_attn.q_proj.weight  ->  decoder.blocks.1.attn.query.weight\n",
            "decoder.blocks.1.attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.1.self_attn.q_proj.bias  ->  decoder.blocks.1.attn.query.bias\n",
            "decoder.blocks.1.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.self_attn.out_proj.weight  ->  decoder.blocks.1.attn.out.weight\n",
            "decoder.blocks.1.attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.1.self_attn.out_proj.bias  ->  decoder.blocks.1.attn.out.bias\n",
            "decoder.blocks.1.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.self_attn_layer_norm.weight  ->  decoder.blocks.1.attn_ln.weight\n",
            "decoder.blocks.1.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.self_attn_layer_norm.bias  ->  decoder.blocks.1.attn_ln.bias\n",
            "decoder.blocks.1.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.encoder_attn.k_proj.weight  ->  decoder.blocks.1.cross_attn.key.weight\n",
            "decoder.blocks.1.cross_attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.1.encoder_attn.v_proj.weight  ->  decoder.blocks.1.cross_attn.value.weight\n",
            "decoder.blocks.1.cross_attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.1.encoder_attn.v_proj.bias  ->  decoder.blocks.1.cross_attn.value.bias\n",
            "decoder.blocks.1.cross_attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.encoder_attn.q_proj.weight  ->  decoder.blocks.1.cross_attn.query.weight\n",
            "decoder.blocks.1.cross_attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.1.encoder_attn.q_proj.bias  ->  decoder.blocks.1.cross_attn.query.bias\n",
            "decoder.blocks.1.cross_attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.encoder_attn.out_proj.weight  ->  decoder.blocks.1.cross_attn.out.weight\n",
            "decoder.blocks.1.cross_attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.1.encoder_attn.out_proj.bias  ->  decoder.blocks.1.cross_attn.out.bias\n",
            "decoder.blocks.1.cross_attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.encoder_attn_layer_norm.weight  ->  decoder.blocks.1.cross_attn_ln.weight\n",
            "decoder.blocks.1.cross_attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.encoder_attn_layer_norm.bias  ->  decoder.blocks.1.cross_attn_ln.bias\n",
            "decoder.blocks.1.cross_attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.fc1.weight  ->  decoder.blocks.1.mlp.0.weight\n",
            "decoder.blocks.1.mlp.0.weight 2 (1536, 384)\n",
            "model.decoder.layers.1.fc1.bias  ->  decoder.blocks.1.mlp.0.bias\n",
            "decoder.blocks.1.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.fc2.weight  ->  decoder.blocks.1.mlp.2.weight\n",
            "decoder.blocks.1.mlp.2.weight 2 (384, 1536)\n",
            "model.decoder.layers.1.fc2.bias  ->  decoder.blocks.1.mlp.2.bias\n",
            "decoder.blocks.1.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.final_layer_norm.weight  ->  decoder.blocks.1.mlp_ln.weight\n",
            "decoder.blocks.1.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.1.final_layer_norm.bias  ->  decoder.blocks.1.mlp_ln.bias\n",
            "decoder.blocks.1.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.self_attn.k_proj.weight  ->  decoder.blocks.2.attn.key.weight\n",
            "decoder.blocks.2.attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.2.self_attn.v_proj.weight  ->  decoder.blocks.2.attn.value.weight\n",
            "decoder.blocks.2.attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.2.self_attn.v_proj.bias  ->  decoder.blocks.2.attn.value.bias\n",
            "decoder.blocks.2.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.self_attn.q_proj.weight  ->  decoder.blocks.2.attn.query.weight\n",
            "decoder.blocks.2.attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.2.self_attn.q_proj.bias  ->  decoder.blocks.2.attn.query.bias\n",
            "decoder.blocks.2.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.self_attn.out_proj.weight  ->  decoder.blocks.2.attn.out.weight\n",
            "decoder.blocks.2.attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.2.self_attn.out_proj.bias  ->  decoder.blocks.2.attn.out.bias\n",
            "decoder.blocks.2.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.self_attn_layer_norm.weight  ->  decoder.blocks.2.attn_ln.weight\n",
            "decoder.blocks.2.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.self_attn_layer_norm.bias  ->  decoder.blocks.2.attn_ln.bias\n",
            "decoder.blocks.2.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.encoder_attn.k_proj.weight  ->  decoder.blocks.2.cross_attn.key.weight\n",
            "decoder.blocks.2.cross_attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.2.encoder_attn.v_proj.weight  ->  decoder.blocks.2.cross_attn.value.weight\n",
            "decoder.blocks.2.cross_attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.2.encoder_attn.v_proj.bias  ->  decoder.blocks.2.cross_attn.value.bias\n",
            "decoder.blocks.2.cross_attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.encoder_attn.q_proj.weight  ->  decoder.blocks.2.cross_attn.query.weight\n",
            "decoder.blocks.2.cross_attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.2.encoder_attn.q_proj.bias  ->  decoder.blocks.2.cross_attn.query.bias\n",
            "decoder.blocks.2.cross_attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.encoder_attn.out_proj.weight  ->  decoder.blocks.2.cross_attn.out.weight\n",
            "decoder.blocks.2.cross_attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.2.encoder_attn.out_proj.bias  ->  decoder.blocks.2.cross_attn.out.bias\n",
            "decoder.blocks.2.cross_attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.encoder_attn_layer_norm.weight  ->  decoder.blocks.2.cross_attn_ln.weight\n",
            "decoder.blocks.2.cross_attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.encoder_attn_layer_norm.bias  ->  decoder.blocks.2.cross_attn_ln.bias\n",
            "decoder.blocks.2.cross_attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.fc1.weight  ->  decoder.blocks.2.mlp.0.weight\n",
            "decoder.blocks.2.mlp.0.weight 2 (1536, 384)\n",
            "model.decoder.layers.2.fc1.bias  ->  decoder.blocks.2.mlp.0.bias\n",
            "decoder.blocks.2.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.fc2.weight  ->  decoder.blocks.2.mlp.2.weight\n",
            "decoder.blocks.2.mlp.2.weight 2 (384, 1536)\n",
            "model.decoder.layers.2.fc2.bias  ->  decoder.blocks.2.mlp.2.bias\n",
            "decoder.blocks.2.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.final_layer_norm.weight  ->  decoder.blocks.2.mlp_ln.weight\n",
            "decoder.blocks.2.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.2.final_layer_norm.bias  ->  decoder.blocks.2.mlp_ln.bias\n",
            "decoder.blocks.2.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.self_attn.k_proj.weight  ->  decoder.blocks.3.attn.key.weight\n",
            "decoder.blocks.3.attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.3.self_attn.v_proj.weight  ->  decoder.blocks.3.attn.value.weight\n",
            "decoder.blocks.3.attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.3.self_attn.v_proj.bias  ->  decoder.blocks.3.attn.value.bias\n",
            "decoder.blocks.3.attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.self_attn.q_proj.weight  ->  decoder.blocks.3.attn.query.weight\n",
            "decoder.blocks.3.attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.3.self_attn.q_proj.bias  ->  decoder.blocks.3.attn.query.bias\n",
            "decoder.blocks.3.attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.self_attn.out_proj.weight  ->  decoder.blocks.3.attn.out.weight\n",
            "decoder.blocks.3.attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.3.self_attn.out_proj.bias  ->  decoder.blocks.3.attn.out.bias\n",
            "decoder.blocks.3.attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.self_attn_layer_norm.weight  ->  decoder.blocks.3.attn_ln.weight\n",
            "decoder.blocks.3.attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.self_attn_layer_norm.bias  ->  decoder.blocks.3.attn_ln.bias\n",
            "decoder.blocks.3.attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.encoder_attn.k_proj.weight  ->  decoder.blocks.3.cross_attn.key.weight\n",
            "decoder.blocks.3.cross_attn.key.weight 2 (384, 384)\n",
            "model.decoder.layers.3.encoder_attn.v_proj.weight  ->  decoder.blocks.3.cross_attn.value.weight\n",
            "decoder.blocks.3.cross_attn.value.weight 2 (384, 384)\n",
            "model.decoder.layers.3.encoder_attn.v_proj.bias  ->  decoder.blocks.3.cross_attn.value.bias\n",
            "decoder.blocks.3.cross_attn.value.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.encoder_attn.q_proj.weight  ->  decoder.blocks.3.cross_attn.query.weight\n",
            "decoder.blocks.3.cross_attn.query.weight 2 (384, 384)\n",
            "model.decoder.layers.3.encoder_attn.q_proj.bias  ->  decoder.blocks.3.cross_attn.query.bias\n",
            "decoder.blocks.3.cross_attn.query.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.encoder_attn.out_proj.weight  ->  decoder.blocks.3.cross_attn.out.weight\n",
            "decoder.blocks.3.cross_attn.out.weight 2 (384, 384)\n",
            "model.decoder.layers.3.encoder_attn.out_proj.bias  ->  decoder.blocks.3.cross_attn.out.bias\n",
            "decoder.blocks.3.cross_attn.out.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.encoder_attn_layer_norm.weight  ->  decoder.blocks.3.cross_attn_ln.weight\n",
            "decoder.blocks.3.cross_attn_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.encoder_attn_layer_norm.bias  ->  decoder.blocks.3.cross_attn_ln.bias\n",
            "decoder.blocks.3.cross_attn_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.fc1.weight  ->  decoder.blocks.3.mlp.0.weight\n",
            "decoder.blocks.3.mlp.0.weight 2 (1536, 384)\n",
            "model.decoder.layers.3.fc1.bias  ->  decoder.blocks.3.mlp.0.bias\n",
            "decoder.blocks.3.mlp.0.bias 1 (1536,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.fc2.weight  ->  decoder.blocks.3.mlp.2.weight\n",
            "decoder.blocks.3.mlp.2.weight 2 (384, 1536)\n",
            "model.decoder.layers.3.fc2.bias  ->  decoder.blocks.3.mlp.2.bias\n",
            "decoder.blocks.3.mlp.2.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.final_layer_norm.weight  ->  decoder.blocks.3.mlp_ln.weight\n",
            "decoder.blocks.3.mlp_ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layers.3.final_layer_norm.bias  ->  decoder.blocks.3.mlp_ln.bias\n",
            "decoder.blocks.3.mlp_ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layer_norm.weight  ->  decoder.ln.weight\n",
            "decoder.ln.weight 1 (384,)\n",
            "  Converting to float32\n",
            "model.decoder.layer_norm.bias  ->  decoder.ln.bias\n",
            "decoder.ln.bias 1 (384,)\n",
            "  Converting to float32\n",
            "Skipping proj_out.weight\n",
            "Done. Output file:  ggml/ggml-model.bin\n",
            "\n"
          ]
        }
      ],
      "source": [
        "!mkdir ./ggml\n",
        "!python whisper.cpp/models/convert-h5-to-ggml.py ./final-model ./whisper-github ./ggml\n",
        "!mv ./ggml/ggml-model.bin ./ggml/ggml-clean.bin"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UOdPpP1fMRr7"
      },
      "source": [
        "For smaller size and better performance, we can also quantize the GGML model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xD4AX-4ZL8xN",
        "outputId": "95ecebc5-8867-46b4-f17e-9ce89e1ec44b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[0mCMake Deprecation Warning at CMakeLists.txt:1 (cmake_minimum_required):\n",
            "  Compatibility with CMake < 3.10 will be removed from a future version of\n",
            "  CMake.\n",
            "\n",
            "  Update the VERSION argument <min> value.  Or, use the <min>...<max> syntax\n",
            "  to tell CMake that the project requires at least <min> but has been updated\n",
            "  to work with policies introduced by <max> or earlier.\n",
            "\n",
            "\u001b[0m\n",
            "-- The C compiler identification is GNU 11.4.0\n",
            "-- The CXX compiler identification is GNU 11.4.0\n",
            "-- Detecting C compiler ABI info\n",
            "-- Detecting C compiler ABI info - done\n",
            "-- Check for working C compiler: /usr/bin/cc - skipped\n",
            "-- Detecting C compile features\n",
            "-- Detecting C compile features - done\n",
            "-- Detecting CXX compiler ABI info\n",
            "-- Detecting CXX compiler ABI info - done\n",
            "-- Check for working CXX compiler: /usr/bin/c++ - skipped\n",
            "-- Detecting CXX compile features\n",
            "-- Detecting CXX compile features - done\n",
            "-- Found Git: /usr/bin/git (found version \"2.34.1\")\n",
            "-- Performing Test CMAKE_HAVE_LIBC_PTHREAD\n",
            "-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Success\n",
            "-- Found Threads: TRUE\n",
            "-- Warning: ccache not found - consider installing it for faster compilation or disable this warning with GGML_CCACHE=OFF\n",
            "-- CMAKE_SYSTEM_PROCESSOR: x86_64\n",
            "-- Including CPU backend\n",
            "-- Found OpenMP_C: -fopenmp (found version \"4.5\")\n",
            "-- Found OpenMP_CXX: -fopenmp (found version \"4.5\")\n",
            "-- Found OpenMP: TRUE (found version \"4.5\")\n",
            "-- x86 detected\n",
            "-- Adding CPU backend variant ggml-cpu: -march=native \n",
            "-- Configuring done (2.4s)\n",
            "-- Generating done (0.1s)\n",
            "-- Build files have been written to: /content/whisper.cpp/build\n",
            "[  2%] \u001b[32mBuilding C object ggml/src/CMakeFiles/ggml-base.dir/ggml.c.o\u001b[0m\n",
            "[  5%] \u001b[32mBuilding C object ggml/src/CMakeFiles/ggml-base.dir/ggml-alloc.c.o\u001b[0m\n",
            "[  7%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-backend.cpp.o\u001b[0m\n",
            "[ 10%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-opt.cpp.o\u001b[0m\n",
            "[ 12%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-base.dir/ggml-threading.cpp.o\u001b[0m\n",
            "[ 15%] \u001b[32mBuilding C object ggml/src/CMakeFiles/ggml-base.dir/ggml-quants.c.o\u001b[0m\n",
            "[ 17%] \u001b[32m\u001b[1mLinking CXX shared library libggml-base.so\u001b[0m\n",
            "[ 17%] Built target ggml-base\n",
            "[ 20%] \u001b[32mBuilding C object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu.c.o\u001b[0m\n",
            "[ 22%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu.cpp.o\u001b[0m\n",
            "[ 25%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-aarch64.cpp.o\u001b[0m\n",
            "[ 27%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-hbm.cpp.o\u001b[0m\n",
            "[ 30%] \u001b[32mBuilding C object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-quants.c.o\u001b[0m\n",
            "[ 32%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/ggml-cpu-traits.cpp.o\u001b[0m\n",
            "[ 35%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/amx.cpp.o\u001b[0m\n",
            "[ 37%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml-cpu.dir/ggml-cpu/amx/mmq.cpp.o\u001b[0m\n",
            "[ 40%] \u001b[32m\u001b[1mLinking CXX shared library libggml-cpu.so\u001b[0m\n",
            "[ 40%] Built target ggml-cpu\n",
            "[ 42%] \u001b[32mBuilding CXX object ggml/src/CMakeFiles/ggml.dir/ggml-backend-reg.cpp.o\u001b[0m\n",
            "[ 45%] \u001b[32m\u001b[1mLinking CXX shared library libggml.so\u001b[0m\n",
            "[ 45%] Built target ggml\n",
            "[ 47%] \u001b[32mBuilding CXX object src/CMakeFiles/whisper.dir/whisper.cpp.o\u001b[0m\n",
            "[ 50%] \u001b[32m\u001b[1mLinking CXX shared library libwhisper.so\u001b[0m\n",
            "[ 50%] Built target whisper\n",
            "[ 52%] \u001b[32mBuilding CXX object examples/CMakeFiles/common.dir/common.cpp.o\u001b[0m\n",
            "[ 55%] \u001b[32mBuilding CXX object examples/CMakeFiles/common.dir/common-ggml.cpp.o\u001b[0m\n",
            "[ 57%] \u001b[32mBuilding CXX object examples/CMakeFiles/common.dir/grammar-parser.cpp.o\u001b[0m\n",
            "[ 60%] \u001b[32m\u001b[1mLinking CXX static library libcommon.a\u001b[0m\n",
            "[ 60%] Built target common\n",
            "[ 62%] \u001b[32mBuilding CXX object examples/cli/CMakeFiles/whisper-cli.dir/cli.cpp.o\u001b[0m\n",
            "[ 65%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/whisper-cli\u001b[0m\n",
            "[ 65%] Built target whisper-cli\n",
            "[ 67%] \u001b[32mBuilding CXX object examples/bench/CMakeFiles/whisper-bench.dir/bench.cpp.o\u001b[0m\n",
            "[ 70%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/whisper-bench\u001b[0m\n",
            "[ 70%] Built target whisper-bench\n",
            "[ 72%] \u001b[32mBuilding CXX object examples/server/CMakeFiles/whisper-server.dir/server.cpp.o\u001b[0m\n",
            "[ 75%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/whisper-server\u001b[0m\n",
            "[ 75%] Built target whisper-server\n",
            "[ 77%] \u001b[32mBuilding CXX object examples/quantize/CMakeFiles/quantize.dir/quantize.cpp.o\u001b[0m\n",
            "[ 80%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/quantize\u001b[0m\n",
            "[ 80%] Built target quantize\n",
            "[ 82%] \u001b[32mBuilding CXX object examples/deprecation-warning/CMakeFiles/main.dir/deprecation-warning.cpp.o\u001b[0m\n",
            "[ 85%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/main\u001b[0m\n",
            "[ 85%] Built target main\n",
            "[ 87%] \u001b[32mBuilding CXX object examples/deprecation-warning/CMakeFiles/bench.dir/deprecation-warning.cpp.o\u001b[0m\n",
            "[ 90%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/bench\u001b[0m\n",
            "[ 90%] Built target bench\n",
            "[ 92%] \u001b[32mBuilding CXX object examples/deprecation-warning/CMakeFiles/stream.dir/deprecation-warning.cpp.o\u001b[0m\n",
            "[ 95%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/stream\u001b[0m\n",
            "[ 95%] Built target stream\n",
            "[ 97%] \u001b[32mBuilding CXX object examples/deprecation-warning/CMakeFiles/command.dir/deprecation-warning.cpp.o\u001b[0m\n",
            "[100%] \u001b[32m\u001b[1mLinking CXX executable ../../bin/command\u001b[0m\n",
            "[100%] Built target command\n",
            "whisper_model_quantize: loading model from './ggml/ggml-clean.bin'\n",
            "whisper_model_quantize: n_vocab       = 51865\n",
            "whisper_model_quantize: n_audio_ctx   = 1500\n",
            "whisper_model_quantize: n_audio_state = 384\n",
            "whisper_model_quantize: n_audio_head  = 6\n",
            "whisper_model_quantize: n_audio_layer = 4\n",
            "whisper_model_quantize: n_text_ctx    = 448\n",
            "whisper_model_quantize: n_text_state  = 384\n",
            "whisper_model_quantize: n_text_head   = 6\n",
            "whisper_model_quantize: n_text_layer  = 4\n",
            "whisper_model_quantize: n_mels        = 80\n",
            "whisper_model_quantize: ftype (src)   = 1\n",
            "whisper_model_quantize: qntvr (src)   = 0\n",
            "whisper_model_quantize: ftype (dst)   = 2007\n",
            "whisper_model_quantize: qntvr (dst)   = 2\n",
            "                                            encoder.conv1.weight - [    3,    80,   384], type =    f16 size =    0.176 MB\n",
            "                                              encoder.conv1.bias - [    1,   384,     1], type =    f32 size =    0.001 MB\n",
            "                                            encoder.conv2.weight - [    3,   384,   384], type =    f16 size =    0.844 MB\n",
            "                                              encoder.conv2.bias - [    1,   384,     1], type =    f32 size =    0.001 MB\n",
            "                                    encoder.positional_embedding - [  384,  1500,     1], type =    f32 size =    2.197 MB\n",
            "                                encoder.blocks.0.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              encoder.blocks.0.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.0.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              encoder.blocks.0.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.0.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.0.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  encoder.blocks.0.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 encoder.blocks.0.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.0.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.0.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.0.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   encoder.blocks.0.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.0.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  encoder.blocks.0.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    encoder.blocks.0.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.1.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              encoder.blocks.1.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.1.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              encoder.blocks.1.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.1.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.1.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  encoder.blocks.1.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 encoder.blocks.1.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.1.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.1.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.1.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   encoder.blocks.1.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.1.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  encoder.blocks.1.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    encoder.blocks.1.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.2.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              encoder.blocks.2.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.2.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              encoder.blocks.2.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.2.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.2.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  encoder.blocks.2.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 encoder.blocks.2.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.2.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.2.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.2.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   encoder.blocks.2.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.2.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  encoder.blocks.2.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    encoder.blocks.2.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.3.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              encoder.blocks.3.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.3.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              encoder.blocks.3.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                encoder.blocks.3.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                encoder.blocks.3.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  encoder.blocks.3.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 encoder.blocks.3.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.3.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   encoder.blocks.3.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.3.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   encoder.blocks.3.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     encoder.blocks.3.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  encoder.blocks.3.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    encoder.blocks.3.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                          encoder.ln_post.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                            encoder.ln_post.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  decoder.token_embedding.weight - [  384, 51865,     1], type =    f16 size =    75.97 MB ->    20.18 MB\n",
            "                                    decoder.positional_embedding - [  384,   448,     1], type =    f32 size =    0.656 MB\n",
            "                                decoder.blocks.0.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              decoder.blocks.0.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.0.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              decoder.blocks.0.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.0.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.0.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  decoder.blocks.0.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 decoder.blocks.0.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.0.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.0.cross_attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                        decoder.blocks.0.cross_attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.0.cross_attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                        decoder.blocks.0.cross_attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.0.cross_attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.0.cross_attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                            decoder.blocks.0.cross_attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                           decoder.blocks.0.cross_attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                             decoder.blocks.0.cross_attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.0.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.0.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   decoder.blocks.0.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.0.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  decoder.blocks.0.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    decoder.blocks.0.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.1.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              decoder.blocks.1.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.1.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              decoder.blocks.1.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.1.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.1.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  decoder.blocks.1.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 decoder.blocks.1.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.1.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.1.cross_attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                        decoder.blocks.1.cross_attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.1.cross_attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                        decoder.blocks.1.cross_attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.1.cross_attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.1.cross_attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                            decoder.blocks.1.cross_attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                           decoder.blocks.1.cross_attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                             decoder.blocks.1.cross_attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.1.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.1.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   decoder.blocks.1.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.1.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  decoder.blocks.1.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    decoder.blocks.1.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.2.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              decoder.blocks.2.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.2.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              decoder.blocks.2.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.2.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.2.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  decoder.blocks.2.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 decoder.blocks.2.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.2.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.2.cross_attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                        decoder.blocks.2.cross_attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.2.cross_attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                        decoder.blocks.2.cross_attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.2.cross_attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.2.cross_attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                            decoder.blocks.2.cross_attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                           decoder.blocks.2.cross_attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                             decoder.blocks.2.cross_attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.2.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.2.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   decoder.blocks.2.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.2.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  decoder.blocks.2.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    decoder.blocks.2.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.3.attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                              decoder.blocks.3.attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.3.attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                              decoder.blocks.3.attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                decoder.blocks.3.attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                decoder.blocks.3.attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                                  decoder.blocks.3.attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                 decoder.blocks.3.attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.3.attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.3.cross_attn.key.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                        decoder.blocks.3.cross_attn.value.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.3.cross_attn.value.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                        decoder.blocks.3.cross_attn.query.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                          decoder.blocks.3.cross_attn.query.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                          decoder.blocks.3.cross_attn.out.weight - [  384,   384,     1], type =    f16 size =     0.56 MB ->     0.15 MB\n",
            "                            decoder.blocks.3.cross_attn.out.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                           decoder.blocks.3.cross_attn_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                             decoder.blocks.3.cross_attn_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                   decoder.blocks.3.mlp.0.weight - [  384,  1536,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.3.mlp.0.bias - [ 1536,     1,     1], type =    f32 size =    0.006 MB\n",
            "                                   decoder.blocks.3.mlp.2.weight - [ 1536,   384,     1], type =    f16 size =     2.25 MB ->     0.60 MB\n",
            "                                     decoder.blocks.3.mlp.2.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                  decoder.blocks.3.mlp_ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                    decoder.blocks.3.mlp_ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                               decoder.ln.weight - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "                                                 decoder.ln.bias - [  384,     1,     1], type =    f32 size =    0.001 MB\n",
            "ggml_common_quantize_0: model size  =   144.05 MB\n",
            "ggml_common_quantize_0: quant size  =    40.97 MB | ftype = 7 (q8_0)\n",
            "\n",
            "main: quantize time =   702.23 ms\n",
            "main:    total time =   702.23 ms\n"
          ]
        }
      ],
      "source": [
        "!cd whisper.cpp && cmake -B build && cmake --build build --config Release\n",
        "!./whisper.cpp/build/bin/quantize ./ggml/ggml-clean.bin ./ggml/ggml-clean-q8_0.bin q8_0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZNH_WV-lMW9n"
      },
      "source": [
        "## Testing it\n",
        "\n",
        "Now, let's make sure that the `.ggml` model works. Start by downloading some test audio:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YFcwkwhXMYKt",
        "outputId": "6a38a55c-bde1-4b32-a928-0a32dac2184d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2025-03-01 00:43:12--  https://www.archive.org/download/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3\n",
            "Resolving www.archive.org (www.archive.org)... 207.241.224.2\n",
            "Connecting to www.archive.org (www.archive.org)|207.241.224.2|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Moved Temporarily\n",
            "Location: https://archive.org/download/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3 [following]\n",
            "--2025-03-01 00:43:12--  https://archive.org/download/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3\n",
            "Resolving archive.org (archive.org)... 207.241.224.2\n",
            "Connecting to archive.org (archive.org)|207.241.224.2|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://ia803201.us.archive.org/25/items/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3 [following]\n",
            "--2025-03-01 00:43:12--  https://ia803201.us.archive.org/25/items/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3\n",
            "Resolving ia803201.us.archive.org (ia803201.us.archive.org)... 207.241.234.121\n",
            "Connecting to ia803201.us.archive.org (ia803201.us.archive.org)|207.241.234.121|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 17754577 (17M) [audio/mpeg]\n",
            "Saving to: ‘./test-audio/aliceaupays_01_carroll_128kb.mp3’\n",
            "\n",
            "aliceaupays_01_carr 100%[===================>]  16.93M  20.4MB/s    in 0.8s    \n",
            "\n",
            "2025-03-01 00:43:14 (20.4 MB/s) - ‘./test-audio/aliceaupays_01_carroll_128kb.mp3’ saved [17754577/17754577]\n",
            "\n",
            "--2025-03-01 00:43:14--  https://www.archive.org/download/alice_in_wonderland_librivox/wonderland_ch_01.mp3\n",
            "Resolving www.archive.org (www.archive.org)... 207.241.224.2\n",
            "Connecting to www.archive.org (www.archive.org)|207.241.224.2|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Moved Temporarily\n",
            "Location: https://archive.org/download/alice_in_wonderland_librivox/wonderland_ch_01.mp3 [following]\n",
            "--2025-03-01 00:43:14--  https://archive.org/download/alice_in_wonderland_librivox/wonderland_ch_01.mp3\n",
            "Resolving archive.org (archive.org)... 207.241.224.2\n",
            "Connecting to archive.org (archive.org)|207.241.224.2|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://ia803401.us.archive.org/8/items/alice_in_wonderland_librivox/wonderland_ch_01.mp3 [following]\n",
            "--2025-03-01 00:43:15--  https://ia803401.us.archive.org/8/items/alice_in_wonderland_librivox/wonderland_ch_01.mp3\n",
            "Resolving ia803401.us.archive.org (ia803401.us.archive.org)... 207.241.232.191\n",
            "Connecting to ia803401.us.archive.org (ia803401.us.archive.org)|207.241.232.191|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 10249859 (9.8M) [audio/mpeg]\n",
            "Saving to: ‘./test-audio/wonderland_ch_01.mp3’\n",
            "\n",
            "wonderland_ch_01.mp 100%[===================>]   9.77M  14.7MB/s    in 0.7s    \n",
            "\n",
            "2025-03-01 00:43:16 (14.7 MB/s) - ‘./test-audio/wonderland_ch_01.mp3’ saved [10249859/10249859]\n",
            "\n",
            "ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers\n",
            "  built with gcc 11 (Ubuntu 11.2.0-19ubuntu1)\n",
            "  configuration: --prefix=/usr --extra-version=0ubuntu0.22.04.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzimg --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-librsvg --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared\n",
            "  libavutil      56. 70.100 / 56. 70.100\n",
            "  libavcodec     58.134.100 / 58.134.100\n",
            "  libavformat    58. 76.100 / 58. 76.100\n",
            "  libavdevice    58. 13.100 / 58. 13.100\n",
            "  libavfilter     7.110.100 /  7.110.100\n",
            "  libswscale      5.  9.100 /  5.  9.100\n",
            "  libswresample   3.  9.100 /  3.  9.100\n",
            "  libpostproc    55.  9.100 / 55.  9.100\n",
            "Input #0, mp3, from './test-audio/aliceaupays_01_carroll_128kb.mp3':\n",
            "  Metadata:\n",
            "    title           : 01 - Au Fond du Terrier\n",
            "    artist          : Lewis Carroll\n",
            "    album           : Aventures d'Alice au Pays des Merveilles\n",
            "    genre           : speech\n",
            "    track           : 2\n",
            "  Duration: 00:18:29.37, start: 0.025057, bitrate: 128 kb/s\n",
            "  Stream #0:0: Audio: mp3, 44100 Hz, mono, fltp, 128 kb/s\n",
            "    Metadata:\n",
            "      encoder         : LAME3.99r\n",
            "Stream mapping:\n",
            "  Stream #0:0 -> #0:0 (mp3 (mp3float) -> pcm_s16le (native))\n",
            "Press [q] to stop, [?] for help\n",
            "Output #0, wav, to './test-audio/recording-fr.wav':\n",
            "  Metadata:\n",
            "    INAM            : 01 - Au Fond du Terrier\n",
            "    IART            : Lewis Carroll\n",
            "    IPRD            : Aventures d'Alice au Pays des Merveilles\n",
            "    IGNR            : speech\n",
            "    IPRT            : 2\n",
            "    ISFT            : Lavf58.76.100\n",
            "  Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s\n",
            "    Metadata:\n",
            "      encoder         : Lavc58.134.100 pcm_s16le\n",
            "size=     938kB time=00:00:29.99 bitrate= 256.1kbits/s speed= 398x    \n",
            "video:0kB audio:938kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.021667%\n",
            "ffmpeg version 4.4.2-0ubuntu0.22.04.1 Copyright (c) 2000-2021 the FFmpeg developers\n",
            "  built with gcc 11 (Ubuntu 11.2.0-19ubuntu1)\n",
            "  configuration: --prefix=/usr --extra-version=0ubuntu0.22.04.1 --toolchain=hardened --libdir=/usr/lib/x86_64-linux-gnu --incdir=/usr/include/x86_64-linux-gnu --arch=amd64 --enable-gpl --disable-stripping --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzimg --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --enable-pocketsphinx --enable-librsvg --enable-libmfx --enable-libdc1394 --enable-libdrm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared\n",
            "  libavutil      56. 70.100 / 56. 70.100\n",
            "  libavcodec     58.134.100 / 58.134.100\n",
            "  libavformat    58. 76.100 / 58. 76.100\n",
            "  libavdevice    58. 13.100 / 58. 13.100\n",
            "  libavfilter     7.110.100 /  7.110.100\n",
            "  libswscale      5.  9.100 /  5.  9.100\n",
            "  libswresample   3.  9.100 /  3.  9.100\n",
            "  libpostproc    55.  9.100 / 55.  9.100\n",
            "\u001b[0;35m[mp3 @ 0x55d22a2ff040] \u001b[0m\u001b[0;33mEstimating duration from bitrate, this may be inaccurate\n",
            "\u001b[0mInput #0, mp3, from './test-audio/wonderland_ch_01.mp3':\n",
            "  Metadata:\n",
            "    title           : wonderland_ch_01\n",
            "    iTunNORM        :  0000024F 000001A7 00000B2D 0000079B 000975F7 000668CE 000059C5 00004487 000975E0 000975E0\n",
            "    artist          : Lewis Carroll\n",
            "    album           : Alice in Wonderland\n",
            "    genre           : Audio Book\n",
            "    comment         : LibriVox.org\n",
            "    track           : 1/12\n",
            "  Duration: 00:10:40.57, start: 0.000000, bitrate: 128 kb/s\n",
            "  Stream #0:0: Audio: mp3, 44100 Hz, stereo, fltp, 128 kb/s\n",
            "Stream mapping:\n",
            "  Stream #0:0 -> #0:0 (mp3 (mp3float) -> pcm_s16le (native))\n",
            "Press [q] to stop, [?] for help\n",
            "Output #0, wav, to './test-audio/recording-en.wav':\n",
            "  Metadata:\n",
            "    INAM            : wonderland_ch_01\n",
            "    iTunNORM        :  0000024F 000001A7 00000B2D 0000079B 000975F7 000668CE 000059C5 00004487 000975E0 000975E0\n",
            "    IART            : Lewis Carroll\n",
            "    IPRD            : Alice in Wonderland\n",
            "    IGNR            : Audio Book\n",
            "    ICMT            : LibriVox.org\n",
            "    IPRT            : 1/12\n",
            "    ISFT            : Lavf58.76.100\n",
            "  Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s\n",
            "    Metadata:\n",
            "      encoder         : Lavc58.134.100 pcm_s16le\n",
            "size=     938kB time=00:00:29.99 bitrate= 256.1kbits/s speed= 264x    \n",
            "video:0kB audio:938kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.021875%\n"
          ]
        }
      ],
      "source": [
        "!mkdir ./test-audio\n",
        "# Download the first chapter of Alice in Wonderland (in French)\n",
        "!wget -P ./test-audio/ https://www.archive.org/download/alice_au_pays_des_merveilles_1811_librivox/aliceaupays_01_carroll_128kb.mp3\n",
        "!wget -P ./test-audio/ https://www.archive.org/download/alice_in_wonderland_librivox/wonderland_ch_01.mp3\n",
        "# Convert it to a format that's understandable by whisper.cpp:\n",
        "# -t 30                 Take the first 30s\n",
        "# -i ...                Input path\n",
        "# -ar 16000             Sample rate of 16000 HZ\n",
        "# -ac 1                 1 audio channel\n",
        "# -codec:a pcm_s16le    Audio codec\n",
        "!ffmpeg -t 30 -i ./test-audio/aliceaupays_01_carroll_128kb.mp3 -ar 16000 -ac 1 -codec:a pcm_s16le ./test-audio/recording-fr.wav\n",
        "!ffmpeg -t 30 -i ./test-audio/wonderland_ch_01.mp3 -ar 16000 -ac 1 -codec:a pcm_s16le ./test-audio/recording-en.wav"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kg5hkQpyq1io"
      },
      "source": [
        "Next, use the `whisper-cli` command to transcribe the audio using our GGML model:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ur0wEF6SULr4",
        "outputId": "1cbe5046-b40b-4d1e-d1f3-8cf8f1733935"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "whisper_init_from_file_with_params_no_state: loading model from './ggml/ggml-clean.bin'\n",
            "whisper_init_with_params_no_state: use gpu    = 1\n",
            "whisper_init_with_params_no_state: flash attn = 0\n",
            "whisper_init_with_params_no_state: gpu_device = 0\n",
            "whisper_init_with_params_no_state: dtw        = 0\n",
            "whisper_init_with_params_no_state: devices    = 1\n",
            "whisper_init_with_params_no_state: backends   = 1\n",
            "whisper_model_load: loading model\n",
            "whisper_model_load: n_vocab       = 51865\n",
            "whisper_model_load: n_audio_ctx   = 1500\n",
            "whisper_model_load: n_audio_state = 384\n",
            "whisper_model_load: n_audio_head  = 6\n",
            "whisper_model_load: n_audio_layer = 4\n",
            "whisper_model_load: n_text_ctx    = 448\n",
            "whisper_model_load: n_text_state  = 384\n",
            "whisper_model_load: n_text_head   = 6\n",
            "whisper_model_load: n_text_layer  = 4\n",
            "whisper_model_load: n_mels        = 80\n",
            "whisper_model_load: ftype         = 1\n",
            "whisper_model_load: qntvr         = 0\n",
            "whisper_model_load: type          = 1 (tiny)\n",
            "whisper_model_load: adding 1607 extra tokens\n",
            "whisper_model_load: n_langs       = 99\n",
            "whisper_model_load:      CPU total size =    77.11 MB\n",
            "whisper_model_load: model size    =   77.11 MB\n",
            "whisper_init_state: kv self size  =    3.15 MB\n",
            "whisper_init_state: kv cross size =    9.44 MB\n",
            "whisper_init_state: kv pad  size  =    2.36 MB\n",
            "whisper_init_state: compute buffer (conv)   =   13.19 MB\n",
            "whisper_init_state: compute buffer (encode) =   64.79 MB\n",
            "whisper_init_state: compute buffer (cross)  =    3.88 MB\n",
            "whisper_init_state: compute buffer (decode) =   95.89 MB\n",
            "\n",
            "system_info: n_threads = 2 / 2 | AVX = 1 | AVX2 = 1 | AVX512 = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | COREML = 0 | OPENVINO = 0 | \n",
            "\n",
            "main: processing './test-audio/recording-fr.wav' (480000 samples, 30.0 sec), 2 threads, 1 processors, 5 beams + best of 5, lang = fr, task = transcribe, timestamps = 0 ...\n",
            "\n",
            "\n",
            " … J'appêtre premier de aventures d'Alice au pays d'Emerveille par Louis Carl. Praduit par Henri-Bué, c'est un registrement libre-vox fait partie du domaine public. Enregistré par Linda Olsen-Fightak, Los Angeles. Au fond du terrier, Alice, assise au prête s'assure.\n",
            "\n",
            "whisper_print_timings:     load time =   141.68 ms\n",
            "whisper_print_timings:     fallbacks =   0 p /   0 h\n",
            "whisper_print_timings:      mel time =   109.07 ms\n",
            "whisper_print_timings:   sample time =   707.69 ms /   342 runs (    2.07 ms per run)\n",
            "whisper_print_timings:   encode time =  3276.56 ms /     1 runs ( 3276.56 ms per run)\n",
            "whisper_print_timings:   decode time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:   batchd time =  2004.00 ms /   341 runs (    5.88 ms per run)\n",
            "whisper_print_timings:   prompt time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:    total time =  6287.07 ms\n",
            "whisper_init_from_file_with_params_no_state: loading model from './ggml/ggml-clean.bin'\n",
            "whisper_init_with_params_no_state: use gpu    = 1\n",
            "whisper_init_with_params_no_state: flash attn = 0\n",
            "whisper_init_with_params_no_state: gpu_device = 0\n",
            "whisper_init_with_params_no_state: dtw        = 0\n",
            "whisper_init_with_params_no_state: devices    = 1\n",
            "whisper_init_with_params_no_state: backends   = 1\n",
            "whisper_model_load: loading model\n",
            "whisper_model_load: n_vocab       = 51865\n",
            "whisper_model_load: n_audio_ctx   = 1500\n",
            "whisper_model_load: n_audio_state = 384\n",
            "whisper_model_load: n_audio_head  = 6\n",
            "whisper_model_load: n_audio_layer = 4\n",
            "whisper_model_load: n_text_ctx    = 448\n",
            "whisper_model_load: n_text_state  = 384\n",
            "whisper_model_load: n_text_head   = 6\n",
            "whisper_model_load: n_text_layer  = 4\n",
            "whisper_model_load: n_mels        = 80\n",
            "whisper_model_load: ftype         = 1\n",
            "whisper_model_load: qntvr         = 0\n",
            "whisper_model_load: type          = 1 (tiny)\n",
            "whisper_model_load: adding 1607 extra tokens\n",
            "whisper_model_load: n_langs       = 99\n",
            "whisper_model_load:      CPU total size =    77.11 MB\n",
            "whisper_model_load: model size    =   77.11 MB\n",
            "whisper_init_state: kv self size  =    3.15 MB\n",
            "whisper_init_state: kv cross size =    9.44 MB\n",
            "whisper_init_state: kv pad  size  =    2.36 MB\n",
            "whisper_init_state: compute buffer (conv)   =   13.19 MB\n",
            "whisper_init_state: compute buffer (encode) =   64.79 MB\n",
            "whisper_init_state: compute buffer (cross)  =    3.88 MB\n",
            "whisper_init_state: compute buffer (decode) =   95.89 MB\n",
            "\n",
            "system_info: n_threads = 2 / 2 | AVX = 1 | AVX2 = 1 | AVX512 = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | COREML = 0 | OPENVINO = 0 | \n",
            "\n",
            "main: processing './test-audio/recording-en.wav' (480000 samples, 30.0 sec), 2 threads, 1 processors, 5 beams + best of 5, lang = en, task = transcribe, timestamps = 0 ...\n",
            "\n",
            "\n",
            " This is a Librivox recording. All Librivox recordings are in the public domain for more information or to volunteer visit Librivox.org. Alice's adventures in Wonderland by Lewis Carroll. Chapter 1, down the rabbit hole. Alice was beginning to get very tired of sitting by her sister on the bank and of having nothing to do. Once or twice, she peeped into the book her sister was reading, but had no pictures or conversations in it. And what's the use of a book?\n",
            "\n",
            "whisper_print_timings:     load time =   139.36 ms\n",
            "whisper_print_timings:     fallbacks =   0 p /   0 h\n",
            "whisper_print_timings:      mel time =   110.43 ms\n",
            "whisper_print_timings:   sample time =  1066.73 ms /   550 runs (    1.94 ms per run)\n",
            "whisper_print_timings:   encode time =  4792.14 ms /     1 runs ( 4792.14 ms per run)\n",
            "whisper_print_timings:   decode time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:   batchd time =  2872.72 ms /   549 runs (    5.23 ms per run)\n",
            "whisper_print_timings:   prompt time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:    total time =  9034.25 ms\n"
          ]
        }
      ],
      "source": [
        "# Test converting the WAV file to text using the GGML file that we built\n",
        "!./whisper.cpp/build/bin/whisper-cli --language fr --no-timestamps -m ./ggml/ggml-clean.bin ./test-audio/recording-fr.wav\n",
        "!./whisper.cpp/build/bin/whisper-cli --language en --no-timestamps -m ./ggml/ggml-clean.bin ./test-audio/recording-en.wav"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kT2qwsdQL9sP",
        "outputId": "714c03bf-5f06-44c4-c082-bb5fe1720ead"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading ggml model tiny from 'https://huggingface.co/ggerganov/whisper.cpp' ...\n",
            "ggml-tiny.bin       100%[===================>]  74.09M  74.6MB/s    in 1.0s    \n",
            "Done! Model 'tiny' saved in './ggml-upstream//ggml-tiny.bin'\n",
            "You can now use it like this:\n",
            "\n",
            "  $ ./main -m ./ggml-upstream//ggml-tiny.bin -f samples/jfk.wav\n",
            "\n",
            "whisper_init_from_file_with_params_no_state: loading model from './ggml-upstream/ggml-tiny.bin'\n",
            "whisper_init_with_params_no_state: use gpu    = 1\n",
            "whisper_init_with_params_no_state: flash attn = 0\n",
            "whisper_init_with_params_no_state: gpu_device = 0\n",
            "whisper_init_with_params_no_state: dtw        = 0\n",
            "whisper_init_with_params_no_state: devices    = 1\n",
            "whisper_init_with_params_no_state: backends   = 1\n",
            "whisper_model_load: loading model\n",
            "whisper_model_load: n_vocab       = 51865\n",
            "whisper_model_load: n_audio_ctx   = 1500\n",
            "whisper_model_load: n_audio_state = 384\n",
            "whisper_model_load: n_audio_head  = 6\n",
            "whisper_model_load: n_audio_layer = 4\n",
            "whisper_model_load: n_text_ctx    = 448\n",
            "whisper_model_load: n_text_state  = 384\n",
            "whisper_model_load: n_text_head   = 6\n",
            "whisper_model_load: n_text_layer  = 4\n",
            "whisper_model_load: n_mels        = 80\n",
            "whisper_model_load: ftype         = 1\n",
            "whisper_model_load: qntvr         = 0\n",
            "whisper_model_load: type          = 1 (tiny)\n",
            "whisper_model_load: adding 1608 extra tokens\n",
            "whisper_model_load: n_langs       = 99\n",
            "whisper_model_load:      CPU total size =    77.11 MB\n",
            "whisper_model_load: model size    =   77.11 MB\n",
            "whisper_init_state: kv self size  =    3.15 MB\n",
            "whisper_init_state: kv cross size =    9.44 MB\n",
            "whisper_init_state: kv pad  size  =    2.36 MB\n",
            "whisper_init_state: compute buffer (conv)   =   13.19 MB\n",
            "whisper_init_state: compute buffer (encode) =   64.79 MB\n",
            "whisper_init_state: compute buffer (cross)  =    3.88 MB\n",
            "whisper_init_state: compute buffer (decode) =   95.89 MB\n",
            "\n",
            "system_info: n_threads = 2 / 2 | AVX = 1 | AVX2 = 1 | AVX512 = 0 | FMA = 1 | NEON = 0 | ARM_FMA = 0 | F16C = 1 | FP16_VA = 0 | WASM_SIMD = 0 | SSE3 = 1 | SSSE3 = 1 | VSX = 0 | COREML = 0 | OPENVINO = 0 | \n",
            "\n",
            "main: processing './test-audio/recording-fr.wav' (480000 samples, 30.0 sec), 2 threads, 1 processors, 5 beams + best of 5, lang = fr, task = transcribe, timestamps = 0 ...\n",
            "\n",
            "\n",
            " … J'appêtre premier de aventures d'Alice au pays d'Emerveille par Louis Carl. Praduit par Henri-Bué, c'est un registrement libre-vox fait partie du domaine public. Enregistré par Linda Olsen-Fightak, Los Angeles. Au fond du terrier, Alice, assise au prête s'assure.\n",
            "\n",
            "whisper_print_timings:     load time =   135.71 ms\n",
            "whisper_print_timings:     fallbacks =   0 p /   0 h\n",
            "whisper_print_timings:      mel time =   109.31 ms\n",
            "whisper_print_timings:   sample time =   896.58 ms /   342 runs (    2.62 ms per run)\n",
            "whisper_print_timings:   encode time =  3244.05 ms /     1 runs ( 3244.05 ms per run)\n",
            "whisper_print_timings:   decode time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:   batchd time =  3859.78 ms /   341 runs (   11.32 ms per run)\n",
            "whisper_print_timings:   prompt time =     0.00 ms /     1 runs (    0.00 ms per run)\n",
            "whisper_print_timings:    total time =  8305.80 ms\n"
          ]
        }
      ],
      "source": [
        "# Compare with the upstream model\n",
        "!mkdir ./ggml-upstream/\n",
        "!sh ./whisper.cpp/models/download-ggml-model.sh tiny ./ggml-upstream/\n",
        "!./whisper.cpp/build/bin/whisper-cli --language fr --no-timestamps -m ./ggml-upstream/ggml-tiny.bin ./test-audio/recording-fr.wav"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Building the Joplin-compatible model\n",
        "\n",
        "Next, we need to convert the model to a format compatible with Joplin. A `.zip` file is created with the following structure:\n",
        "```\n",
        "model_name.zip/\n",
        "| model_name/\n",
        "| | README.md\n",
        "| | model.bin\n",
        "| | config.json\n",
        "```"
      ],
      "metadata": {
        "id": "bNKGqVn0zRg6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from pathlib import Path\n",
        "import shutil, json, zipfile\n",
        "\n",
        "def package_output(source_model: Path, output_dir: Path, output_filename: str):\n",
        "    if not output_dir.exists():\n",
        "        output_dir.mkdir()\n",
        "    unzipped_dir = output_dir / output_filename\n",
        "    if unzipped_dir.exists():\n",
        "        shutil.rmtree(unzipped_dir)\n",
        "    unzipped_dir.mkdir()\n",
        "\n",
        "    shutil.copyfile(source_model, unzipped_dir / 'model.bin')\n",
        "    # config.json\n",
        "    config_filepath = unzipped_dir / 'config.json'\n",
        "    config_filepath.write_text(json.dumps({\n",
        "        'prompts': {\n",
        "            # Custom prompts can improve accuracy.\n",
        "            'en': 'Joplin is a note-taking application. This is a Joplin note.'\n",
        "        },\n",
        "        'output': {\n",
        "            '//': 'Each of the replacements is in the form [ original, replaceWith ]. For example, [\"test\", \"\"] replaces all instances of \"test\" with the empty string.',\n",
        "            'stringReplacements': [\n",
        "                [ '[BLANK_AUDIO]', '' ],\n",
        "            ],\n",
        "            'regexReplacements': [\n",
        "                [ r'^\\([^(),.?]+\\)$', ''],\n",
        "                [ r'^\\[[^(),.?]+\\]$', ''],\n",
        "                [ r'^[.,?!]$', '' ],\n",
        "                [ r'\\[swearB1\\]', 'BS' ],\n",
        "                [ r'\\[swear[A-Z][0-9]+\\]', '****' ],\n",
        "            ],\n",
        "        }\n",
        "    }, indent='\\t'))\n",
        "    # README.md\n",
        "    readme_filepath = unzipped_dir / 'README.md'\n",
        "    readme_filepath.write_text('\\n'.join([\n",
        "        '# {}'.format(output_filename),\n",
        "        '',\n",
        "        'This model is a version of `whisper-' + whisper_mode + '` with an [adjusted vocab.json](/~https://github.com/personalizedrefrigerator/joplin-voice-typing-test/blob/main/whisper_vocab_cleanup.ipynb) to reduce the probability of profanity when given noisy non-speech input.',\n",
        "        '',\n",
        "        '## License',\n",
        "        '',\n",
        "        'The Whisper model from which this is modified has the following license:',\n",
        "        '''\n",
        "        MIT License\n",
        "\n",
        "        Copyright (c) 2022 OpenAI\n",
        "\n",
        "        Permission is hereby granted, free of charge, to any person obtaining a copy\n",
        "        of this software and associated documentation files (the \"Software\"), to deal\n",
        "        in the Software without restriction, including without limitation the rights\n",
        "        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell\n",
        "        copies of the Software, and to permit persons to whom the Software is\n",
        "        furnished to do so, subject to the following conditions:\n",
        "\n",
        "        The above copyright notice and this permission notice shall be included in all\n",
        "        copies or substantial portions of the Software.\n",
        "\n",
        "        THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n",
        "        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n",
        "        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE\n",
        "        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n",
        "        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,\n",
        "        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE\n",
        "        SOFTWARE.\n",
        "        '''\n",
        "    ]))\n",
        "\n",
        "    # Make the .zip file\n",
        "    # See https://docs.python.org/3/library/shutil.html\n",
        "    shutil.make_archive(\n",
        "        output_dir / output_filename,\n",
        "        'zip',\n",
        "        root_dir=output_dir,\n",
        "        base_dir=output_filename,\n",
        "    )\n",
        "\n",
        "package_output(\n",
        "    Path('./ggml/ggml-clean.bin'),\n",
        "    Path('./joplin-model'),\n",
        "    'whisper-{}'.format(whisper_mode)\n",
        ")\n",
        "package_output(\n",
        "    Path('./ggml/ggml-clean-q8_0.bin'),\n",
        "    Path('./joplin-model-q8_0'),\n",
        "    'whisper-{}-q8_0'.format(whisper_mode)\n",
        ")"
      ],
      "metadata": {
        "id": "ucsOUrTqzRSz"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "The models are now built! They're stored in the `./joplin-model` and `./joplin-model-q8_0` directories."
      ],
      "metadata": {
        "id": "aCEyuCYi3Phx"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "8iX0Q0ui26p1"
      },
      "execution_count": 26,
      "outputs": []
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "51e33c5351a74ed08a3a17523af54aca": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_8e6923494af245b1b1885baa6f4ece77",
              "IPY_MODEL_45d711f248bf4c81b45c5e6f5ea53037",
              "IPY_MODEL_ceef1fbfad6a41e9a21ee8142bb790f8"
            ],
            "layout": "IPY_MODEL_efffdb9d52e44381932613bec4038f80"
          }
        },
        "8e6923494af245b1b1885baa6f4ece77": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7974a0822a3b4e3a97d9291de4769614",
            "placeholder": "​",
            "style": "IPY_MODEL_b44b737695d34f1187181709ed82b131",
            "value": "preprocessor_config.json: 100%"
          }
        },
        "45d711f248bf4c81b45c5e6f5ea53037": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_54a1b3498da04043bb32a7524f4cd63f",
            "max": 184990,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_3ee92336df6a4ca0b3712fa6753bd519",
            "value": 184990
          }
        },
        "ceef1fbfad6a41e9a21ee8142bb790f8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_1177d97f2b254bcf9c4cd0f02167732a",
            "placeholder": "​",
            "style": "IPY_MODEL_287813d864024d358f7f4148220d0791",
            "value": " 185k/185k [00:00&lt;00:00, 2.10MB/s]"
          }
        },
        "efffdb9d52e44381932613bec4038f80": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7974a0822a3b4e3a97d9291de4769614": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b44b737695d34f1187181709ed82b131": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "54a1b3498da04043bb32a7524f4cd63f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3ee92336df6a4ca0b3712fa6753bd519": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "1177d97f2b254bcf9c4cd0f02167732a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "287813d864024d358f7f4148220d0791": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "31691f20a6dc40a98422ec168251dc5d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_35b7313901554d46b4fb52198a00340d",
              "IPY_MODEL_e40a68341c3a40e28ccfd2063f46b374",
              "IPY_MODEL_51553b842c2f41b5954122939b1660a3"
            ],
            "layout": "IPY_MODEL_e8ac88d9de14476a979c9e145dfbc122"
          }
        },
        "35b7313901554d46b4fb52198a00340d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2a1f4e1fa85b42eaa0b4a5cbf0b64f8b",
            "placeholder": "​",
            "style": "IPY_MODEL_d9930f3cd3074d9face7778a8511e2e2",
            "value": "tokenizer_config.json: 100%"
          }
        },
        "e40a68341c3a40e28ccfd2063f46b374": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_078038fa96644f4e82054970f978c165",
            "max": 282683,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_88caa5476c414b0189242f165b0f7f50",
            "value": 282683
          }
        },
        "51553b842c2f41b5954122939b1660a3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_4de1dc09c2174b34b34a29627f64d82c",
            "placeholder": "​",
            "style": "IPY_MODEL_a7b986ddd910453f9ff39499b00cf034",
            "value": " 283k/283k [00:00&lt;00:00, 3.72MB/s]"
          }
        },
        "e8ac88d9de14476a979c9e145dfbc122": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "2a1f4e1fa85b42eaa0b4a5cbf0b64f8b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "d9930f3cd3074d9face7778a8511e2e2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "078038fa96644f4e82054970f978c165": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "88caa5476c414b0189242f165b0f7f50": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "4de1dc09c2174b34b34a29627f64d82c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a7b986ddd910453f9ff39499b00cf034": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "14a90640bfde4589bc0ad25d1848266b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_77b9055500a041db9a6427e795ca2440",
              "IPY_MODEL_f3b04084b8744cd6b36e82230adfba37",
              "IPY_MODEL_d5e3cc1505324285802ef60eba8167e5"
            ],
            "layout": "IPY_MODEL_c6461c1dd97a41538e076fc5388ab128"
          }
        },
        "77b9055500a041db9a6427e795ca2440": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d62cce5d5d3b426c9b347f10d7a1c1a7",
            "placeholder": "​",
            "style": "IPY_MODEL_978cb2c586b4428281f036daa52b5fcc",
            "value": "vocab.json: 100%"
          }
        },
        "f3b04084b8744cd6b36e82230adfba37": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d199c5845385448e9a8c592cb7afac17",
            "max": 835550,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_96ee23a97ba64a518495e015f2943f66",
            "value": 835550
          }
        },
        "d5e3cc1505324285802ef60eba8167e5": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_e536ea375cc849669bad66b055d6f2df",
            "placeholder": "​",
            "style": "IPY_MODEL_129e4b5d228f441ba7bcda7031d73e16",
            "value": " 836k/836k [00:00&lt;00:00, 7.94MB/s]"
          }
        },
        "c6461c1dd97a41538e076fc5388ab128": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "d62cce5d5d3b426c9b347f10d7a1c1a7": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "978cb2c586b4428281f036daa52b5fcc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "d199c5845385448e9a8c592cb7afac17": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "96ee23a97ba64a518495e015f2943f66": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "e536ea375cc849669bad66b055d6f2df": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "129e4b5d228f441ba7bcda7031d73e16": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "f2f0b564fc7f408d8a858f144d136e12": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_a61ee831e9a641ed9d053fe2ae247479",
              "IPY_MODEL_145101e9efa7454babbf19483af7d05c",
              "IPY_MODEL_56f1c590824a49e3b1f737863e0d2cef"
            ],
            "layout": "IPY_MODEL_ccd6078d37f3413f90c590a8b2cabd22"
          }
        },
        "a61ee831e9a641ed9d053fe2ae247479": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_a3e3c77a10dc413e9d0bffec564fec74",
            "placeholder": "​",
            "style": "IPY_MODEL_afb06c06d562464caa9cb2d214a44c6b",
            "value": "tokenizer.json: 100%"
          }
        },
        "145101e9efa7454babbf19483af7d05c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2021afccba0a473ca7d2de62e808927f",
            "max": 2480466,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_390b00c3b6284b3cb26b34965f85f121",
            "value": 2480466
          }
        },
        "56f1c590824a49e3b1f737863e0d2cef": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_e426d69a070a4f05ae6e2d7fd136a0b5",
            "placeholder": "​",
            "style": "IPY_MODEL_dda824f44e2f46e8bc02ee226f9e15c3",
            "value": " 2.48M/2.48M [00:00&lt;00:00, 11.0MB/s]"
          }
        },
        "ccd6078d37f3413f90c590a8b2cabd22": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a3e3c77a10dc413e9d0bffec564fec74": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "afb06c06d562464caa9cb2d214a44c6b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "2021afccba0a473ca7d2de62e808927f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "390b00c3b6284b3cb26b34965f85f121": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "e426d69a070a4f05ae6e2d7fd136a0b5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "dda824f44e2f46e8bc02ee226f9e15c3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "cbe590638e1c4ccf91bc27e483ec4da3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_f0f78d395bc44366bd0395e0cbfe491c",
              "IPY_MODEL_a9169039361b49fe8aaf571a028ebe53",
              "IPY_MODEL_374c527161ab4d19af2687c6327449c9"
            ],
            "layout": "IPY_MODEL_e69ab604b17544d2a8284b9de7b70f1c"
          }
        },
        "f0f78d395bc44366bd0395e0cbfe491c": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_fca7440f08f444e38f207bcc28af4e98",
            "placeholder": "​",
            "style": "IPY_MODEL_87934b97cc5f4fe5aa9eda646c282098",
            "value": "merges.txt: 100%"
          }
        },
        "a9169039361b49fe8aaf571a028ebe53": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7ed2c7076f614a01bfda4e26aeca086b",
            "max": 493869,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_3ec75421502e459d97326b4706bf4514",
            "value": 493869
          }
        },
        "374c527161ab4d19af2687c6327449c9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_fc153028c0ce4529b6194dd54c8f7549",
            "placeholder": "​",
            "style": "IPY_MODEL_a189e5ac99c3456cb6c1059ac5188bfc",
            "value": " 494k/494k [00:00&lt;00:00, 10.2MB/s]"
          }
        },
        "e69ab604b17544d2a8284b9de7b70f1c": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "fca7440f08f444e38f207bcc28af4e98": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "87934b97cc5f4fe5aa9eda646c282098": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "7ed2c7076f614a01bfda4e26aeca086b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3ec75421502e459d97326b4706bf4514": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "fc153028c0ce4529b6194dd54c8f7549": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "a189e5ac99c3456cb6c1059ac5188bfc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "97b1014caa84403ea551cdd8be05c914": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_cb99055498334cecad43b136cc70ce72",
              "IPY_MODEL_317fbd3b3dca44c6896a6608dc04b4ac",
              "IPY_MODEL_c68afccfd19f4f67b368a93d01d3d4b9"
            ],
            "layout": "IPY_MODEL_6b368e5500a04e75a7be4b65004d1ec6"
          }
        },
        "cb99055498334cecad43b136cc70ce72": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_bcb12acd5e564288a2c2a74fbf1a4868",
            "placeholder": "​",
            "style": "IPY_MODEL_dd3f34e27dba438eb5feb13ac938fee8",
            "value": "normalizer.json: 100%"
          }
        },
        "317fbd3b3dca44c6896a6608dc04b4ac": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_e59bde27da1747e4b5f60ec467090a1b",
            "max": 52666,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_0b60360e788b4636b9d9a683ef1358bd",
            "value": 52666
          }
        },
        "c68afccfd19f4f67b368a93d01d3d4b9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7091c2f3880b4ed2a7bb1f92c5865847",
            "placeholder": "​",
            "style": "IPY_MODEL_862fd35e1d984c56b8615a1a6d647672",
            "value": " 52.7k/52.7k [00:00&lt;00:00, 593kB/s]"
          }
        },
        "6b368e5500a04e75a7be4b65004d1ec6": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "bcb12acd5e564288a2c2a74fbf1a4868": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "dd3f34e27dba438eb5feb13ac938fee8": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "e59bde27da1747e4b5f60ec467090a1b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "0b60360e788b4636b9d9a683ef1358bd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "7091c2f3880b4ed2a7bb1f92c5865847": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "862fd35e1d984c56b8615a1a6d647672": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "32bffc2faba844a8b33dfb7ea81effad": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_8dad60fe67b24411a75476870e86a724",
              "IPY_MODEL_d7a5ca4b2e4546a1bedefe00b3365e55",
              "IPY_MODEL_cd5c485cb4eb4682b45d7042b5afddfe"
            ],
            "layout": "IPY_MODEL_56d2d0258c7d4b278de3ba42826680c1"
          }
        },
        "8dad60fe67b24411a75476870e86a724": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2928325954d143608225a5a220f80bfb",
            "placeholder": "​",
            "style": "IPY_MODEL_c92c1443d92440b3ad6861998693e921",
            "value": "added_tokens.json: 100%"
          }
        },
        "d7a5ca4b2e4546a1bedefe00b3365e55": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_c8d079922c9447a096eb22f139e41d51",
            "max": 34604,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_aa177ec5c523461081fba9f533e09c73",
            "value": 34604
          }
        },
        "cd5c485cb4eb4682b45d7042b5afddfe": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_b90600a633b64293b03dd3d48f8cdc34",
            "placeholder": "​",
            "style": "IPY_MODEL_6037bc8a0eb24eae9adf7ddd1e4920fb",
            "value": " 34.6k/34.6k [00:00&lt;00:00, 578kB/s]"
          }
        },
        "56d2d0258c7d4b278de3ba42826680c1": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "2928325954d143608225a5a220f80bfb": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c92c1443d92440b3ad6861998693e921": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "c8d079922c9447a096eb22f139e41d51": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "aa177ec5c523461081fba9f533e09c73": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "b90600a633b64293b03dd3d48f8cdc34": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "6037bc8a0eb24eae9adf7ddd1e4920fb": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "36237cc2ee7d401e8c74f0b32df25845": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_abcd74688b674f3282300fe52587ac48",
              "IPY_MODEL_323419bb521348a6862317d6229a1bfc",
              "IPY_MODEL_7ec43ca1b0e64892b6fccef9b123b4af"
            ],
            "layout": "IPY_MODEL_7de61e234b05499a960ea520fc678a23"
          }
        },
        "abcd74688b674f3282300fe52587ac48": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_5fcf053e08d54910a77610f3e8d6e51d",
            "placeholder": "​",
            "style": "IPY_MODEL_21f393e2240c488598702104ef9aa584",
            "value": "special_tokens_map.json: 100%"
          }
        },
        "323419bb521348a6862317d6229a1bfc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_d466f57c1f8e4eb99aff9d1d1b336890",
            "max": 2194,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_10905252963449608b4cf2a72d04659b",
            "value": 2194
          }
        },
        "7ec43ca1b0e64892b6fccef9b123b4af": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_2ba32ce91480468cb22d44221047259f",
            "placeholder": "​",
            "style": "IPY_MODEL_c9e35fc216754d878f125877944afd22",
            "value": " 2.19k/2.19k [00:00&lt;00:00, 35.2kB/s]"
          }
        },
        "7de61e234b05499a960ea520fc678a23": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "5fcf053e08d54910a77610f3e8d6e51d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "21f393e2240c488598702104ef9aa584": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "d466f57c1f8e4eb99aff9d1d1b336890": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "10905252963449608b4cf2a72d04659b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "2ba32ce91480468cb22d44221047259f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "c9e35fc216754d878f125877944afd22": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "b8b1f114b84a48c8befc607d75d19217": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_7c043b15c7cb471297f6cfacabce40be",
              "IPY_MODEL_d7a2aa1f36fb40f1a3c9f101da6f1fcc",
              "IPY_MODEL_9cd2de48c15e485492cbc9ac76e21df0"
            ],
            "layout": "IPY_MODEL_716626235b15484fbe3bdb87dbb80d3b"
          }
        },
        "7c043b15c7cb471297f6cfacabce40be": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_92d4442fe8b94b3cb6d652b72dc55428",
            "placeholder": "​",
            "style": "IPY_MODEL_3e3cbed9d80f4d9c87db0e375fa2ee2b",
            "value": "config.json: 100%"
          }
        },
        "d7a2aa1f36fb40f1a3c9f101da6f1fcc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_16ab731566804579a6548ddde822518f",
            "max": 1983,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_d360f8037eba412882d9807b838ae534",
            "value": 1983
          }
        },
        "9cd2de48c15e485492cbc9ac76e21df0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_26882105686847bdb309af4c0faacbb9",
            "placeholder": "​",
            "style": "IPY_MODEL_b1171a864e524ffc9378d79e240081e0",
            "value": " 1.98k/1.98k [00:00&lt;00:00, 97.1kB/s]"
          }
        },
        "716626235b15484fbe3bdb87dbb80d3b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "92d4442fe8b94b3cb6d652b72dc55428": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3e3cbed9d80f4d9c87db0e375fa2ee2b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "16ab731566804579a6548ddde822518f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "d360f8037eba412882d9807b838ae534": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "26882105686847bdb309af4c0faacbb9": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b1171a864e524ffc9378d79e240081e0": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "b07ce25529044a6081faf108a91b46f9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_b74145815a8047aca9da40a2f5170bd2",
              "IPY_MODEL_ee8ae5e130904544bb1963d5b61f66a9",
              "IPY_MODEL_52a58a18fb6b4f5a9fb1ae4aec4dbe98"
            ],
            "layout": "IPY_MODEL_a9bb8a8adf3443dc9588f886dd17ec4a"
          }
        },
        "b74145815a8047aca9da40a2f5170bd2": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_7b7bf3be758c4dd59476bc05ad4b4e0f",
            "placeholder": "​",
            "style": "IPY_MODEL_49d5167b64634d609265c804662c70a9",
            "value": "model.safetensors: 100%"
          }
        },
        "ee8ae5e130904544bb1963d5b61f66a9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_0e49a94023644036a8feb6e783c77566",
            "max": 151061672,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_1903365501904d11bcfbcf76e9cbd323",
            "value": 151061672
          }
        },
        "52a58a18fb6b4f5a9fb1ae4aec4dbe98": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_84886ef71ecb449aa0d5363131d2c5d2",
            "placeholder": "​",
            "style": "IPY_MODEL_f64bde590bf14a36b35c66467da00a79",
            "value": " 151M/151M [00:02&lt;00:00, 75.8MB/s]"
          }
        },
        "a9bb8a8adf3443dc9588f886dd17ec4a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "7b7bf3be758c4dd59476bc05ad4b4e0f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "49d5167b64634d609265c804662c70a9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "0e49a94023644036a8feb6e783c77566": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "1903365501904d11bcfbcf76e9cbd323": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "84886ef71ecb449aa0d5363131d2c5d2": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "f64bde590bf14a36b35c66467da00a79": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "fa9d2724157f4da2b7acc8590e35b981": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_c9dc2532e8a9404588c31ccaeca42bc6",
              "IPY_MODEL_40aa791f749943d6b6317ee254304d13",
              "IPY_MODEL_c629090bb492462a9e4066ed58a91589"
            ],
            "layout": "IPY_MODEL_e673672365ad409e9e83af4c1141c0dd"
          }
        },
        "c9dc2532e8a9404588c31ccaeca42bc6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_8dbbea5d076548cbabf14a1ef610f71b",
            "placeholder": "​",
            "style": "IPY_MODEL_b5b0df9d638e4c8a87880c4e7b0c7ca3",
            "value": "generation_config.json: 100%"
          }
        },
        "40aa791f749943d6b6317ee254304d13": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_4bac8d95267743d9baa4bc4c3292521f",
            "max": 3747,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_4c0b2925e91b455ea7127a2fdf75d1fa",
            "value": 3747
          }
        },
        "c629090bb492462a9e4066ed58a91589": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_11689b3e624b4fb69c53a3e3f8503099",
            "placeholder": "​",
            "style": "IPY_MODEL_87e6900700e54ff7bc409e0c2f2b7728",
            "value": " 3.75k/3.75k [00:00&lt;00:00, 196kB/s]"
          }
        },
        "e673672365ad409e9e83af4c1141c0dd": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "8dbbea5d076548cbabf14a1ef610f71b": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "b5b0df9d638e4c8a87880c4e7b0c7ca3": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "4bac8d95267743d9baa4bc4c3292521f": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "4c0b2925e91b455ea7127a2fdf75d1fa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "11689b3e624b4fb69c53a3e3f8503099": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "87e6900700e54ff7bc409e0c2f2b7728": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}