Spaces:
Paused
Paused
File size: 11,706 Bytes
ee6e328 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 |
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"This is a notebook to generate mel-spectrograms from a TTS model to be used in a Vocoder training."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"%load_ext autoreload\n",
"%autoreload 2\n",
"import os\n",
"import sys\n",
"import torch\n",
"import importlib\n",
"import numpy as np\n",
"from tqdm import tqdm as tqdm\n",
"from torch.utils.data import DataLoader\n",
"from TTS.tts.datasets.dataset import TTSDataset\n",
"from TTS.tts.layers.losses import L1LossMasked\n",
"from TTS.utils.audio import AudioProcessor\n",
"from TTS.config import load_config\n",
"from TTS.tts.utils.visual import plot_spectrogram\n",
"from TTS.tts.utils.helpers import sequence_mask\n",
"from TTS.tts.models import setup_model\n",
"from TTS.tts.utils.text.symbols import make_symbols, symbols, phonemes\n",
"\n",
"%matplotlib inline\n",
"\n",
"import os\n",
"os.environ['CUDA_VISIBLE_DEVICES']='2'"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def set_filename(wav_path, out_path):\n",
" wav_file = os.path.basename(wav_path)\n",
" file_name = wav_file.split('.')[0]\n",
" os.makedirs(os.path.join(out_path, \"quant\"), exist_ok=True)\n",
" os.makedirs(os.path.join(out_path, \"mel\"), exist_ok=True)\n",
" os.makedirs(os.path.join(out_path, \"wav_gl\"), exist_ok=True)\n",
" wavq_path = os.path.join(out_path, \"quant\", file_name)\n",
" mel_path = os.path.join(out_path, \"mel\", file_name)\n",
" wav_path = os.path.join(out_path, \"wav_gl\", file_name)\n",
" return file_name, wavq_path, mel_path, wav_path"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"OUT_PATH = \"/home/ubuntu/TTS/recipes/ljspeech/LJSpeech-1.1/specs2/\"\n",
"DATA_PATH = \"/home/ubuntu/TTS/recipes/ljspeech/LJSpeech-1.1/\"\n",
"DATASET = \"ljspeech\"\n",
"METADATA_FILE = \"metadata.csv\"\n",
"CONFIG_PATH = \"/home/ubuntu/.local/share/tts/tts_models--en--ljspeech--tacotron2-DDC_ph/config.json\"\n",
"MODEL_FILE = \"/home/ubuntu/.local/share/tts/tts_models--en--ljspeech--tacotron2-DDC_ph/model_file.pth\"\n",
"BATCH_SIZE = 32\n",
"\n",
"QUANTIZED_WAV = False\n",
"QUANTIZE_BIT = None\n",
"DRY_RUN = False # if False, does not generate output files, only computes loss and visuals.\n",
"\n",
"use_cuda = torch.cuda.is_available()\n",
"print(\" > CUDA enabled: \", use_cuda)\n",
"\n",
"C = load_config(CONFIG_PATH)\n",
"C.audio['do_trim_silence'] = False # IMPORTANT!!!!!!!!!!!!!!! disable to align mel specs with the wav files\n",
"ap = AudioProcessor(bits=QUANTIZE_BIT, **C.audio)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"print(C['r'])\n",
"# if the vocabulary was passed, replace the default\n",
"if 'characters' in C and C['characters']:\n",
" symbols, phonemes = make_symbols(**C.characters)\n",
"\n",
"# load the model\n",
"num_chars = len(phonemes) if C.use_phonemes else len(symbols)\n",
"# TODO: multiple speaker\n",
"model = setup_model(C)\n",
"model.load_checkpoint(C, MODEL_FILE, eval=True)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"preprocessor = importlib.import_module(\"TTS.tts.datasets.formatters\")\n",
"preprocessor = getattr(preprocessor, DATASET.lower())\n",
"meta_data = preprocessor(DATA_PATH, METADATA_FILE)\n",
"dataset = TTSDataset(\n",
" checkpoint[\"config\"][\"r\"],\n",
" C.text_cleaner,\n",
" False,\n",
" ap,\n",
" meta_data,\n",
" characters=C.get('characters', None),\n",
" use_phonemes=C.use_phonemes,\n",
" phoneme_cache_path=C.phoneme_cache_path,\n",
" enable_eos_bos=C.enable_eos_bos_chars,\n",
")\n",
"loader = DataLoader(\n",
" dataset, batch_size=BATCH_SIZE, num_workers=4, collate_fn=dataset.collate_fn, shuffle=False, drop_last=False\n",
")\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Generate model outputs "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import pickle\n",
"\n",
"file_idxs = []\n",
"metadata = []\n",
"losses = []\n",
"postnet_losses = []\n",
"criterion = L1LossMasked(seq_len_norm=C.seq_len_norm)\n",
"with torch.no_grad():\n",
" for data in tqdm(loader):\n",
" # setup input data\n",
" text_input = data[0]\n",
" text_lengths = data[1]\n",
" linear_input = data[3]\n",
" mel_input = data[4]\n",
" mel_lengths = data[5]\n",
" stop_targets = data[6]\n",
" item_idx = data[7]\n",
"\n",
" # dispatch data to GPU\n",
" if use_cuda:\n",
" text_input = text_input.cuda()\n",
" text_lengths = text_lengths.cuda()\n",
" mel_input = mel_input.cuda()\n",
" mel_lengths = mel_lengths.cuda()\n",
"\n",
" mask = sequence_mask(text_lengths)\n",
" mel_outputs, postnet_outputs, alignments, stop_tokens = model.forward(text_input, text_lengths, mel_input)\n",
" \n",
" # compute loss\n",
" loss = criterion(mel_outputs, mel_input, mel_lengths)\n",
" loss_postnet = criterion(postnet_outputs, mel_input, mel_lengths)\n",
" losses.append(loss.item())\n",
" postnet_losses.append(loss_postnet.item())\n",
"\n",
" # compute mel specs from linear spec if model is Tacotron\n",
" if C.model == \"Tacotron\":\n",
" mel_specs = []\n",
" postnet_outputs = postnet_outputs.data.cpu().numpy()\n",
" for b in range(postnet_outputs.shape[0]):\n",
" postnet_output = postnet_outputs[b]\n",
" mel_specs.append(torch.FloatTensor(ap.out_linear_to_mel(postnet_output.T).T).cuda())\n",
" postnet_outputs = torch.stack(mel_specs)\n",
" elif C.model == \"Tacotron2\":\n",
" postnet_outputs = postnet_outputs.detach().cpu().numpy()\n",
" alignments = alignments.detach().cpu().numpy()\n",
"\n",
" if not DRY_RUN:\n",
" for idx in range(text_input.shape[0]):\n",
" wav_file_path = item_idx[idx]\n",
" wav = ap.load_wav(wav_file_path)\n",
" file_name, wavq_path, mel_path, wav_path = set_filename(wav_file_path, OUT_PATH)\n",
" file_idxs.append(file_name)\n",
"\n",
" # quantize and save wav\n",
" if QUANTIZED_WAV:\n",
" wavq = ap.quantize(wav)\n",
" np.save(wavq_path, wavq)\n",
"\n",
" # save TTS mel\n",
" mel = postnet_outputs[idx]\n",
" mel_length = mel_lengths[idx]\n",
" mel = mel[:mel_length, :].T\n",
" np.save(mel_path, mel)\n",
"\n",
" metadata.append([wav_file_path, mel_path])\n",
"\n",
" # for wavernn\n",
" if not DRY_RUN:\n",
" pickle.dump(file_idxs, open(OUT_PATH+\"/dataset_ids.pkl\", \"wb\")) \n",
" \n",
" # for pwgan\n",
" with open(os.path.join(OUT_PATH, \"metadata.txt\"), \"w\") as f:\n",
" for data in metadata:\n",
" f.write(f\"{data[0]}|{data[1]+'.npy'}\\n\")\n",
"\n",
" print(np.mean(losses))\n",
" print(np.mean(postnet_losses))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# for pwgan\n",
"with open(os.path.join(OUT_PATH, \"metadata.txt\"), \"w\") as f:\n",
" for data in metadata:\n",
" f.write(f\"{data[0]}|{data[1]+'.npy'}\\n\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Sanity Check"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"idx = 1\n",
"ap.melspectrogram(ap.load_wav(item_idx[idx])).shape"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import soundfile as sf\n",
"wav, sr = sf.read(item_idx[idx])\n",
"mel_postnet = postnet_outputs[idx][:mel_lengths[idx], :]\n",
"mel_decoder = mel_outputs[idx][:mel_lengths[idx], :].detach().cpu().numpy()\n",
"mel_truth = ap.melspectrogram(wav)\n",
"print(mel_truth.shape)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# plot posnet output\n",
"print(mel_postnet[:mel_lengths[idx], :].shape)\n",
"plot_spectrogram(mel_postnet, ap)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# plot decoder output\n",
"print(mel_decoder.shape)\n",
"plot_spectrogram(mel_decoder, ap)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# plot GT specgrogram\n",
"print(mel_truth.shape)\n",
"plot_spectrogram(mel_truth.T, ap)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# postnet, decoder diff\n",
"from matplotlib import pylab as plt\n",
"mel_diff = mel_decoder - mel_postnet\n",
"plt.figure(figsize=(16, 10))\n",
"plt.imshow(abs(mel_diff[:mel_lengths[idx],:]).T,aspect=\"auto\", origin=\"lower\");\n",
"plt.colorbar()\n",
"plt.tight_layout()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# PLOT GT SPECTROGRAM diff\n",
"from matplotlib import pylab as plt\n",
"mel_diff2 = mel_truth.T - mel_decoder\n",
"plt.figure(figsize=(16, 10))\n",
"plt.imshow(abs(mel_diff2).T,aspect=\"auto\", origin=\"lower\");\n",
"plt.colorbar()\n",
"plt.tight_layout()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# PLOT GT SPECTROGRAM diff\n",
"from matplotlib import pylab as plt\n",
"mel = postnet_outputs[idx]\n",
"mel_diff2 = mel_truth.T - mel[:mel_truth.shape[1]]\n",
"plt.figure(figsize=(16, 10))\n",
"plt.imshow(abs(mel_diff2).T,aspect=\"auto\", origin=\"lower\");\n",
"plt.colorbar()\n",
"plt.tight_layout()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"interpreter": {
"hash": "822ce188d9bce5372c4adbb11364eeb49293228c2224eb55307f4664778e7f56"
},
"kernelspec": {
"display_name": "Python 3.9.7 64-bit ('base': conda)",
"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": "3.9.7"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
|