xlsr_gpt.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "xlsr tw.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZXXYTItj9pDQ"
      },
      "source": [
        "!pip install git+https://github.com/huggingface/datasets.git\n",
        "!pip install git+https://github.com/huggingface/transformers.git\n",
        "!pip install torchaudio\n",
        "!pip install librosa\n",
        "!pip install jiwer\n",
        "!pip install ffmpeg-python"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rtw5lIYx9rRc"
      },
      "source": [
        "AUDIO_HTML = \"\"\"\n",
        "<script>\n",
        "var my_div = document.createElement(\"DIV\");\n",
        "var my_p = document.createElement(\"P\");\n",
        "var my_btn = document.createElement(\"BUTTON\");\n",
        "var t = document.createTextNode(\"Press to start recording\");\n",
        "\n",
        "my_btn.appendChild(t);\n",
        "//my_p.appendChild(my_btn);\n",
        "my_div.appendChild(my_btn);\n",
        "document.body.appendChild(my_div);\n",
        "\n",
        "var base64data = 0;\n",
        "var reader;\n",
        "var recorder, gumStream;\n",
        "var recordButton = my_btn;\n",
        "\n",
        "var handleSuccess = function(stream) {\n",
        "  gumStream = stream;\n",
        "  var options = {\n",
        "    //bitsPerSecond: 8000, //chrome seems to ignore, always 48k\n",
        "    mimeType : 'audio/webm;codecs=opus'\n",
        "    //mimeType : 'audio/webm;codecs=pcm'\n",
        "  };            \n",
        "  //recorder = new MediaRecorder(stream, options);\n",
        "  recorder = new MediaRecorder(stream);\n",
        "  recorder.ondataavailable = function(e) {            \n",
        "    var url = URL.createObjectURL(e.data);\n",
        "    var preview = document.createElement('audio');\n",
        "    preview.controls = true;\n",
        "    preview.src = url;\n",
        "    document.body.appendChild(preview);\n",
        "\n",
        "    reader = new FileReader();\n",
        "    reader.readAsDataURL(e.data); \n",
        "    reader.onloadend = function() {\n",
        "      base64data = reader.result;\n",
        "      //console.log(\"Inside FileReader:\" + base64data);\n",
        "    }\n",
        "  };\n",
        "  recorder.start();\n",
        "  };\n",
        "\n",
        "recordButton.innerText = \"Recording... press to stop\";\n",
        "\n",
        "navigator.mediaDevices.getUserMedia({audio: true}).then(handleSuccess);\n",
        "\n",
        "function toggleRecording() {\n",
        "  if (recorder && recorder.state == \"recording\") {\n",
        "      recorder.stop();\n",
        "      gumStream.getAudioTracks()[0].stop();\n",
        "      recordButton.innerText = \"Saving the recording... pls wait!\"\n",
        "  }\n",
        "}\n",
        "\n",
        "// https://stackoverflow.com/a/951057\n",
        "function sleep(ms) {\n",
        "  return new Promise(resolve => setTimeout(resolve, ms));\n",
        "}\n",
        "\n",
        "var data = new Promise(resolve=>{\n",
        "//recordButton.addEventListener(\"click\", toggleRecording);\n",
        "recordButton.onclick = ()=>{\n",
        "toggleRecording()\n",
        "\n",
        "sleep(2000).then(() => {\n",
        "  // wait 2000ms for the data to be available...\n",
        "  // ideally this should use something like await...\n",
        "  //console.log(\"Inside data:\" + base64data)\n",
        "  resolve(base64data.toString())\n",
        "\n",
        "});\n",
        "\n",
        "}\n",
        "});\n",
        "      \n",
        "</script>\n",
        "\"\"\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LpS8AZbP9tld"
      },
      "source": [
        "from IPython.display import HTML, Audio\n",
        "from google.colab.output import eval_js\n",
        "from base64 import b64decode\n",
        "import wave\n",
        "from scipy.io.wavfile import read as wav_read\n",
        "import io\n",
        "import numpy as np\n",
        "import ffmpeg\n",
        "\n",
        "def write_wav(f, sr, x, normalized=False):\n",
        "    f = wave.open(f, \"wb\")\n",
        "    f.setnchannels(1)\n",
        "    f.setsampwidth(2)\n",
        "    f.setframerate(sr)\n",
        "    \n",
        "    wave_data = x.astype(np.short)\n",
        "    f.writeframes(wave_data.tobytes())\n",
        "    f.close()\n",
        "\n",
        "def get_audio():\n",
        "  global hnum\n",
        "\n",
        "  # call microphone\n",
        "  display(HTML(AUDIO_HTML))\n",
        "  data = eval_js(\"data\")\n",
        "  binary = b64decode(data.split(',')[1])\n",
        "\n",
        "  process = (ffmpeg\n",
        "      .input('pipe:0')\n",
        "      .output('pipe:1', format='wav')\n",
        "      .run_async(pipe_stdin=True, pipe_stdout=True, pipe_stderr=True, quiet=True, overwrite_output=True)\n",
        "  )\n",
        "  output, err = process.communicate(input=binary)\n",
        "\n",
        "  riff_chunk_size = len(output) - 8\n",
        "  # Break up the chunk size into four bytes, held in b.\n",
        "  q = riff_chunk_size\n",
        "  b = []\n",
        "  for i in range(4):\n",
        "      q, r = divmod(q, 256)\n",
        "      b.append(r)\n",
        "\n",
        "  # Replace bytes 4:8 in proc.stdout with the actual size of the RIFF chunk.\n",
        "  riff = output[:4] + bytes(b) + output[8:]\n",
        "  sr, audio = wav_read(io.BytesIO(riff))\n",
        "  # save\n",
        "  human_sound_file = \"demo.wav\"\n",
        "  write_wav(human_sound_file, sr, audio)\n",
        "\n",
        "  return human_sound_file"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T6QGhFeL9vaK"
      },
      "source": [
        "import torchaudio\n",
        "from datasets import load_dataset, load_metric\n",
        "from transformers import (\n",
        "    Wav2Vec2ForCTC,\n",
        "    Wav2Vec2Processor,\n",
        "    AutoTokenizer, \n",
        "    AutoModelWithLMHead \n",
        ")\n",
        "import torch\n",
        "import re\n",
        "import sys\n",
        "\n",
        "model_name = \"voidful/wav2vec2-large-xlsr-53-tw-gpt\"\n",
        "device = \"cuda\"\n",
        "processor_name = \"voidful/wav2vec2-large-xlsr-53-tw-gpt\"\n",
        "\n",
        "chars_to_ignore_regex = r\"[¥•＂＃＄％＆＇（）＊＋，－／：；＜＝＞＠［＼］＾＿｀｛｜｝～｟｠｢｣､　、〃〈〉《》「」『』【】〔〕〖〗〘〙〚〛〜〝〞〟〰〾〿–—‘’‛“”„‟…‧﹏﹑﹔·'℃°•·．﹑︰〈〉─《﹖﹣﹂﹁﹔！？｡。＂＃＄％＆＇（）＊＋，﹐－／：；＜＝＞＠［＼］＾＿｀｛｜｝～｟｠｢｣､、〃》「」『』【】〔〕〖〗〘〙〚〛〜〝〞〟〰〾〿–—‘’‛“”„‟…‧﹏.．!\\\"#$%&()*+,\\-.\\:;<=>?@\\[\\]\\\\\\/^_`{|}~]\"\n",
        "\n",
        "model = Wav2Vec2ForCTC.from_pretrained(model_name).to(device)\n",
        "processor = Wav2Vec2Processor.from_pretrained(processor_name)\n",
        "\n",
        "tokenizer = AutoTokenizer.from_pretrained(\"ckiplab/gpt2-base-chinese\")  \n",
        "gpt_model = AutoModelWithLMHead.from_pretrained(\"ckiplab/gpt2-base-chinese\").to(device)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pfl4sIUIHLVm"
      },
      "source": [
        "Decoding with language model adjustment\n",
        "\n",
        "![CTC LOSS WITH GPT.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xR32Almw9xlI"
      },
      "source": [
        "resampler = torchaudio.transforms.Resample(orig_freq=48_000, new_freq=16_000)\n",
        "\n",
        "def load_file_to_data(file):\n",
        "    batch = {}\n",
        "    speech, _ = torchaudio.load(file)\n",
        "    batch[\"speech\"] = resampler.forward(speech.squeeze(0)).numpy()\n",
        "    batch[\"sampling_rate\"] = resampler.new_freq\n",
        "    return batch\n",
        "\n",
        "def predict(data):\n",
        "    features = processor(data[\"speech\"], sampling_rate=data[\"sampling_rate\"], padding=True, return_tensors=\"pt\")\n",
        "    input_values = features.input_values.to(device)\n",
        "    attention_mask = features.attention_mask.to(device)\n",
        "    with torch.no_grad():\n",
        "        logits = model(input_values, attention_mask=attention_mask).logits\n",
        "    \n",
        "    decoded_results = []\n",
        "    for logit in logits:\n",
        "        pred_ids = torch.argmax(logit, dim=-1)\n",
        "        mask = pred_ids.ge(1).unsqueeze(-1).expand(logit.size())\n",
        "        vocab_size = logit.size()[-1]\n",
        "        voice_prob = torch.nn.functional.softmax((torch.masked_select(logit, mask).view(-1,vocab_size)),dim=-1)\n",
        "        gpt_input = torch.cat((torch.tensor([tokenizer.cls_token_id]).to(device),pred_ids[pred_ids>0]), 0)\n",
        "        gpt_prob = torch.nn.functional.softmax(gpt_model(gpt_input).logits, dim=-1)[:voice_prob.size()[0],:]\n",
        "        comb_pred_ids = torch.argmax(gpt_prob*voice_prob, dim=-1)\n",
        "        decoded_results.append(processor.decode(comb_pred_ids))\n",
        "\n",
        "    return decoded_results\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KvHNnlMw9wwc"
      },
      "source": [
        "get_audio()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pVdIIDl69zFv"
      },
      "source": [
        "predict(load_file_to_data('./demo.wav'))"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}