File size: 6,695 Bytes
836ffb0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 |
{
"cells": [
{
"cell_type": "code",
"execution_count": null,
"id": "ac7631cc",
"metadata": {},
"outputs": [],
"source": [
"import torch\n",
"import re\n",
"import librosa\n",
"from datasets import load_dataset, load_metric\n",
"from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor\n",
"import warnings\n",
"import os\n",
"\n",
"\n",
"LANG_ID = \"zh-CN\"\n",
"MODEL_ID = \"zh-CN-output-aishell\"\n",
"\n",
"test_dataset = load_dataset(\"common_voice\", LANG_ID, split=\"test\")\n",
"\n",
"wer = load_metric(\"wer\")\n",
"cer = load_metric(\"cer\")\n",
"\n",
"\n",
"\n",
"processor = Wav2Vec2Processor.from_pretrained(MODEL_ID)\n",
"model = Wav2Vec2ForCTC.from_pretrained(MODEL_ID)\n",
"model.to(DEVICE)\n",
"\n",
"# Preprocessing the datasets.\n",
"# We need to read the audio files as arrays\n",
"def speech_file_to_array_fn(batch):\n",
" with warnings.catch_warnings():\n",
" warnings.simplefilter(\"ignore\")\n",
" speech_array, sampling_rate = librosa.load(batch[\"path\"], sr=16_000)\n",
" batch[\"speech\"] = speech_array\n",
" batch[\"sentence\"] = (\n",
" re.sub(\"([^\\u4e00-\\u9fa5\\u0030-\\u0039])\", \"\", batch[\"sentence\"]).lower() + \" \"\n",
" )\n",
" return batch\n",
"\n",
"\n",
"test_dataset = test_dataset.map(\n",
" speech_file_to_array_fn,\n",
" num_proc=15,\n",
" remove_columns=['client_id', 'up_votes', 'down_votes', 'age', 'gender', 'accent', 'locale', 'segment'],\n",
")\n",
"\n",
"# Preprocessing the datasets.\n",
"# We need to read the audio files as arrays\n",
"def evaluate(batch):\n",
" inputs = processor(\n",
" batch[\"speech\"], sampling_rate=16_000, return_tensors=\"pt\", padding=True\n",
" )\n",
"\n",
" with torch.no_grad():\n",
" logits = model(\n",
" inputs.input_values.to(DEVICE),\n",
" attention_mask=inputs.attention_mask.to(DEVICE),\n",
" ).logits\n",
"\n",
" pred_ids = torch.argmax(logits, dim=-1)\n",
" batch[\"pred_strings\"] = processor.batch_decode(pred_ids)\n",
" return batch\n",
"\n",
"\n",
"result = test_dataset.map(evaluate, batched=True, batch_size=8)\n",
"\n",
"predictions = [x.lower() for x in result[\"pred_strings\"]]\n",
"references = [x.lower() for x in result[\"sentence\"]]\n",
"\n",
"print(\n",
" f\"WER: {wer.compute(predictions=predictions, references=references, chunk_size=1000) * 100}\"\n",
")\n",
"print(f\"CER: {cer.compute(predictions=predictions, references=references) * 100}\")"
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "7db04701",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"11/08/2022 09:41:20 - INFO - huggingsound.speech_recognition.model - Loading model...\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"11/08/2022 09:41:23 - WARNING - root - bos_token <s> not in provided tokens. It will be added to the list of tokens\n",
"11/08/2022 09:41:23 - WARNING - root - eos_token </s> not in provided tokens. It will be added to the list of tokens\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"100%|βββββββββββββββββββββββββββββββββββββββββββββββ| 1/1 [00:00<00:00, 2.11it/s]\n"
]
}
],
"source": [
"from huggingsound import SpeechRecognitionModel\n",
"model = SpeechRecognitionModel(\"./wav2vec2-large-xlsr-chinese\")\n",
"audio_paths = [\"1.wav\"]\n",
"transcriptions = model.transcribe(audio_paths)"
]
},
{
"cell_type": "code",
"execution_count": 19,
"id": "23316152",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'δ½ εζ¬’ι₯ε'"
]
},
"execution_count": 19,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# transcriptions[0]['transcription'].replace('[PAD]','')\n",
"transcriptions[0]['transcription']"
]
},
{
"cell_type": "code",
"execution_count": 24,
"id": "730d4afa",
"metadata": {},
"outputs": [],
"source": [
"import torch\n",
"from transformers import Wav2Vec2Processor, HubertForCTC\n",
"from datasets import load_dataset\n",
"\n",
"processor = Wav2Vec2Processor.from_pretrained(\"./english_fine_tune\")\n",
"model = HubertForCTC.from_pretrained(\"./english_fine_tune\")"
]
},
{
"cell_type": "code",
"execution_count": 25,
"id": "f45768e8",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"It is strongly recommended to pass the ``sampling_rate`` argument to this function. Failing to do so can result in silent errors that might be hard to debug.\n"
]
}
],
"source": [
"import librosa\n",
"input_audio, sr = librosa.load('english.wav', sr = 16000)\n",
"input_values = processor(input_audio, return_tensors=\"pt\").input_values # Batch size 1\n",
"logits = model(input_values).logits\n",
"predicted_ids = torch.argmax(logits, dim=-1)\n",
"transcription = processor.decode(predicted_ids[0])"
]
},
{
"cell_type": "code",
"execution_count": 26,
"id": "8bd98a38",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"'WITHOUT THE DATA SET THE ARTICLE IS USELESS'"
]
},
"execution_count": 26,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"transcription"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "db6a5667",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.13"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
|