Update README.md
Browse files
README.md
CHANGED
|
@@ -3,6 +3,8 @@ license: apache-2.0
|
|
| 3 |
base_model:
|
| 4 |
- microsoft/wavlm-large
|
| 5 |
pipeline_tag: audio-to-audio
|
|
|
|
|
|
|
| 6 |
---
|
| 7 |
|
| 8 |
# FocalCodec
|
|
@@ -25,16 +27,6 @@ See the readme at: https://github.com/lucadellalib/focalcodec
|
|
| 25 |
|
| 26 |
---------------------------------------------------------------------------------------------------------
|
| 27 |
|
| 28 |
-
## 📌 Available Checkpoints
|
| 29 |
-
|
| 30 |
-
| Checkpoint | Token Rate (Hz) | Bitrate (kbps) | Dataset |
|
| 31 |
-
|:-----------------------:|:---------------:|:--------------:|:-----------:|
|
| 32 |
-
| **LibriTTS960_50Hz** | 50.0 | 0.65 | LibriTTS960 |
|
| 33 |
-
| **LibriTTS960_25Hz** | 25.0 | 0.33 | LibriTTS960 |
|
| 34 |
-
| **LibriTTS960_12_5Hz** | 12.5 | 0.16 | LibriTTS960 |
|
| 35 |
-
|
| 36 |
-
---------------------------------------------------------------------------------------------------------
|
| 37 |
-
|
| 38 |
## @ Citing
|
| 39 |
|
| 40 |
```
|
|
@@ -52,326 +44,4 @@ See the readme at: https://github.com/lucadellalib/focalcodec
|
|
| 52 |
|
| 53 |
[[email protected]](mailto:[email protected])
|
| 54 |
|
| 55 |
-
---------------------------------------------------------------------------------------------------------
|
| 56 |
-
|
| 57 |
-
# File information
|
| 58 |
-
|
| 59 |
-
The repository contains the following file information:
|
| 60 |
-
|
| 61 |
-
Filename: LibriTTS960_25Hz.json
|
| 62 |
-
Content: {
|
| 63 |
-
"encoder_name": "WavLM",
|
| 64 |
-
"encoder_config": {
|
| 65 |
-
"hidden_dims": [
|
| 66 |
-
512,
|
| 67 |
-
512,
|
| 68 |
-
512,
|
| 69 |
-
512,
|
| 70 |
-
512,
|
| 71 |
-
512,
|
| 72 |
-
512
|
| 73 |
-
],
|
| 74 |
-
"kernel_sizes": [
|
| 75 |
-
10,
|
| 76 |
-
3,
|
| 77 |
-
3,
|
| 78 |
-
3,
|
| 79 |
-
3,
|
| 80 |
-
2,
|
| 81 |
-
2
|
| 82 |
-
],
|
| 83 |
-
"strides": [
|
| 84 |
-
5,
|
| 85 |
-
2,
|
| 86 |
-
2,
|
| 87 |
-
2,
|
| 88 |
-
2,
|
| 89 |
-
2,
|
| 90 |
-
2
|
| 91 |
-
],
|
| 92 |
-
"num_layers": 6,
|
| 93 |
-
"dim": 1024,
|
| 94 |
-
"ffn_dim": 4096,
|
| 95 |
-
"num_heads": 16,
|
| 96 |
-
"num_buckets": 320,
|
| 97 |
-
"max_distance": 800,
|
| 98 |
-
"dropout": 0.0,
|
| 99 |
-
"conv_pos": 128,
|
| 100 |
-
"conv_pos_groups": 16
|
| 101 |
-
},
|
| 102 |
-
"compressor_name": "FocalEncoder",
|
| 103 |
-
"compressor_config": {
|
| 104 |
-
"input_dim": 1024,
|
| 105 |
-
"output_dim": 13,
|
| 106 |
-
"hidden_dims": [
|
| 107 |
-
1024,
|
| 108 |
-
512,
|
| 109 |
-
256
|
| 110 |
-
],
|
| 111 |
-
"downscale_factors": [
|
| 112 |
-
2,
|
| 113 |
-
1,
|
| 114 |
-
1
|
| 115 |
-
],
|
| 116 |
-
"focal_window": 7,
|
| 117 |
-
"focal_level": 2,
|
| 118 |
-
"focal_factor": 2,
|
| 119 |
-
"dropout": 0.0,
|
| 120 |
-
"use_post_norm": false,
|
| 121 |
-
"use_layerscale": false,
|
| 122 |
-
"layerscale_init": 0.0001,
|
| 123 |
-
"normalize_modulator": false
|
| 124 |
-
},
|
| 125 |
-
"quantizer_name": "BinarySphericalQuantizer",
|
| 126 |
-
"quantizer_config": {
|
| 127 |
-
"codebook_size": 8192
|
| 128 |
-
},
|
| 129 |
-
"decompressor_name": "FocalDecoder",
|
| 130 |
-
"decompressor_config": {
|
| 131 |
-
"input_dim": 13,
|
| 132 |
-
"output_dim": 1024,
|
| 133 |
-
"hidden_dims": [
|
| 134 |
-
256,
|
| 135 |
-
512,
|
| 136 |
-
1024
|
| 137 |
-
],
|
| 138 |
-
"upscale_factors": [
|
| 139 |
-
1,
|
| 140 |
-
1,
|
| 141 |
-
2
|
| 142 |
-
],
|
| 143 |
-
"focal_window": 7,
|
| 144 |
-
"focal_level": 2,
|
| 145 |
-
"focal_factor": 2,
|
| 146 |
-
"dropout": 0.0,
|
| 147 |
-
"use_post_norm": false,
|
| 148 |
-
"use_layerscale": false,
|
| 149 |
-
"layerscale_init": 0.0001,
|
| 150 |
-
"normalize_modulator": false
|
| 151 |
-
},
|
| 152 |
-
"decoder_name": "Vocos",
|
| 153 |
-
"decoder_config": {
|
| 154 |
-
"input_channels": 1024,
|
| 155 |
-
"num_layers": 8,
|
| 156 |
-
"dim": 512,
|
| 157 |
-
"ffn_dim": 1536,
|
| 158 |
-
"kernel_size": 7,
|
| 159 |
-
"padding": 3,
|
| 160 |
-
"layerscale_init": null,
|
| 161 |
-
"n_fft": 1024,
|
| 162 |
-
"hop_length": 320
|
| 163 |
-
}
|
| 164 |
-
}
|
| 165 |
-
|
| 166 |
-
Filename: focalcodec.png
|
| 167 |
-
Content: "Content of the file is larger than 50 KB, too long to display."
|
| 168 |
-
|
| 169 |
-
Filename: LibriTTS960_50Hz.json
|
| 170 |
-
Content: {
|
| 171 |
-
"encoder_name": "WavLM",
|
| 172 |
-
"encoder_config": {
|
| 173 |
-
"hidden_dims": [
|
| 174 |
-
512,
|
| 175 |
-
512,
|
| 176 |
-
512,
|
| 177 |
-
512,
|
| 178 |
-
512,
|
| 179 |
-
512,
|
| 180 |
-
512
|
| 181 |
-
],
|
| 182 |
-
"kernel_sizes": [
|
| 183 |
-
10,
|
| 184 |
-
3,
|
| 185 |
-
3,
|
| 186 |
-
3,
|
| 187 |
-
3,
|
| 188 |
-
2,
|
| 189 |
-
2
|
| 190 |
-
],
|
| 191 |
-
"strides": [
|
| 192 |
-
5,
|
| 193 |
-
2,
|
| 194 |
-
2,
|
| 195 |
-
2,
|
| 196 |
-
2,
|
| 197 |
-
2,
|
| 198 |
-
2
|
| 199 |
-
],
|
| 200 |
-
"num_layers": 6,
|
| 201 |
-
"dim": 1024,
|
| 202 |
-
"ffn_dim": 4096,
|
| 203 |
-
"num_heads": 16,
|
| 204 |
-
"num_buckets": 320,
|
| 205 |
-
"max_distance": 800,
|
| 206 |
-
"dropout": 0.0,
|
| 207 |
-
"conv_pos": 128,
|
| 208 |
-
"conv_pos_groups": 16
|
| 209 |
-
},
|
| 210 |
-
"compressor_name": "FocalEncoder",
|
| 211 |
-
"compressor_config": {
|
| 212 |
-
"input_dim": 1024,
|
| 213 |
-
"output_dim": 13,
|
| 214 |
-
"hidden_dims": [
|
| 215 |
-
1024,
|
| 216 |
-
512,
|
| 217 |
-
256
|
| 218 |
-
],
|
| 219 |
-
"downscale_factors": [
|
| 220 |
-
1,
|
| 221 |
-
1,
|
| 222 |
-
1
|
| 223 |
-
],
|
| 224 |
-
"focal_window": 7,
|
| 225 |
-
"focal_level": 2,
|
| 226 |
-
"focal_factor": 2,
|
| 227 |
-
"dropout": 0.0,
|
| 228 |
-
"use_post_norm": false,
|
| 229 |
-
"use_layerscale": false,
|
| 230 |
-
"layerscale_init": 0.0001,
|
| 231 |
-
"normalize_modulator": false
|
| 232 |
-
},
|
| 233 |
-
"quantizer_name": "BinarySphericalQuantizer",
|
| 234 |
-
"quantizer_config": {
|
| 235 |
-
"codebook_size": 8192
|
| 236 |
-
},
|
| 237 |
-
"decompressor_name": "FocalDecoder",
|
| 238 |
-
"decompressor_config": {
|
| 239 |
-
"input_dim": 13,
|
| 240 |
-
"output_dim": 1024,
|
| 241 |
-
"hidden_dims": [
|
| 242 |
-
256,
|
| 243 |
-
512,
|
| 244 |
-
1024
|
| 245 |
-
],
|
| 246 |
-
"upscale_factors": [
|
| 247 |
-
1,
|
| 248 |
-
1,
|
| 249 |
-
1
|
| 250 |
-
],
|
| 251 |
-
"focal_window": 7,
|
| 252 |
-
"focal_level": 2,
|
| 253 |
-
"focal_factor": 2,
|
| 254 |
-
"dropout": 0.0,
|
| 255 |
-
"use_post_norm": false,
|
| 256 |
-
"use_layerscale": false,
|
| 257 |
-
"layerscale_init": 0.0001,
|
| 258 |
-
"normalize_modulator": false
|
| 259 |
-
},
|
| 260 |
-
"decoder_name": "Vocos",
|
| 261 |
-
"decoder_config": {
|
| 262 |
-
"input_channels": 1024,
|
| 263 |
-
"num_layers": 8,
|
| 264 |
-
"dim": 512,
|
| 265 |
-
"ffn_dim": 1536,
|
| 266 |
-
"kernel_size": 7,
|
| 267 |
-
"padding": 3,
|
| 268 |
-
"layerscale_init": null,
|
| 269 |
-
"n_fft": 1024,
|
| 270 |
-
"hop_length": 320
|
| 271 |
-
}
|
| 272 |
-
}
|
| 273 |
-
|
| 274 |
-
Filename: LibriTTS960_12_5Hz.json
|
| 275 |
-
Content: {
|
| 276 |
-
"encoder_name": "WavLM",
|
| 277 |
-
"encoder_config": {
|
| 278 |
-
"hidden_dims": [
|
| 279 |
-
512,
|
| 280 |
-
512,
|
| 281 |
-
512,
|
| 282 |
-
512,
|
| 283 |
-
512,
|
| 284 |
-
512,
|
| 285 |
-
512
|
| 286 |
-
],
|
| 287 |
-
"kernel_sizes": [
|
| 288 |
-
10,
|
| 289 |
-
3,
|
| 290 |
-
3,
|
| 291 |
-
3,
|
| 292 |
-
3,
|
| 293 |
-
2,
|
| 294 |
-
2
|
| 295 |
-
],
|
| 296 |
-
"strides": [
|
| 297 |
-
5,
|
| 298 |
-
2,
|
| 299 |
-
2,
|
| 300 |
-
2,
|
| 301 |
-
2,
|
| 302 |
-
2,
|
| 303 |
-
2
|
| 304 |
-
],
|
| 305 |
-
"num_layers": 6,
|
| 306 |
-
"dim": 1024,
|
| 307 |
-
"ffn_dim": 4096,
|
| 308 |
-
"num_heads": 16,
|
| 309 |
-
"num_buckets": 320,
|
| 310 |
-
"max_distance": 800,
|
| 311 |
-
"dropout": 0.0,
|
| 312 |
-
"conv_pos": 128,
|
| 313 |
-
"conv_pos_groups": 16
|
| 314 |
-
},
|
| 315 |
-
"compressor_name": "FocalEncoder",
|
| 316 |
-
"compressor_config": {
|
| 317 |
-
"input_dim": 1024,
|
| 318 |
-
"output_dim": 13,
|
| 319 |
-
"hidden_dims": [
|
| 320 |
-
1024,
|
| 321 |
-
512,
|
| 322 |
-
256
|
| 323 |
-
],
|
| 324 |
-
"downscale_factors": [
|
| 325 |
-
2,
|
| 326 |
-
2,
|
| 327 |
-
1
|
| 328 |
-
],
|
| 329 |
-
"focal_window": 7,
|
| 330 |
-
"focal_level": 2,
|
| 331 |
-
"focal_factor": 2,
|
| 332 |
-
"dropout": 0.0,
|
| 333 |
-
"use_post_norm": false,
|
| 334 |
-
"use_layerscale": false,
|
| 335 |
-
"layerscale_init": 0.0001,
|
| 336 |
-
"normalize_modulator": false
|
| 337 |
-
},
|
| 338 |
-
"quantizer_name": "BinarySphericalQuantizer",
|
| 339 |
-
"quantizer_config": {
|
| 340 |
-
"codebook_size": 8192
|
| 341 |
-
},
|
| 342 |
-
"decompressor_name": "FocalDecoder",
|
| 343 |
-
"decompressor_config": {
|
| 344 |
-
"input_dim": 13,
|
| 345 |
-
"output_dim": 1024,
|
| 346 |
-
"hidden_dims": [
|
| 347 |
-
256,
|
| 348 |
-
512,
|
| 349 |
-
1024
|
| 350 |
-
],
|
| 351 |
-
"upscale_factors": [
|
| 352 |
-
1,
|
| 353 |
-
2,
|
| 354 |
-
2
|
| 355 |
-
],
|
| 356 |
-
"focal_window": 7,
|
| 357 |
-
"focal_level": 2,
|
| 358 |
-
"focal_factor": 2,
|
| 359 |
-
"dropout": 0.0,
|
| 360 |
-
"use_post_norm": false,
|
| 361 |
-
"use_layerscale": false,
|
| 362 |
-
"layerscale_init": 0.0001,
|
| 363 |
-
"normalize_modulator": false
|
| 364 |
-
},
|
| 365 |
-
"decoder_name": "Vocos",
|
| 366 |
-
"decoder_config": {
|
| 367 |
-
"input_channels": 1024,
|
| 368 |
-
"num_layers": 8,
|
| 369 |
-
"dim": 512,
|
| 370 |
-
"ffn_dim": 1536,
|
| 371 |
-
"kernel_size": 7,
|
| 372 |
-
"padding": 3,
|
| 373 |
-
"layerscale_init": null,
|
| 374 |
-
"n_fft": 1024,
|
| 375 |
-
"hop_length": 320
|
| 376 |
-
}
|
| 377 |
-
}
|
|
|
|
| 3 |
base_model:
|
| 4 |
- microsoft/wavlm-large
|
| 5 |
pipeline_tag: audio-to-audio
|
| 6 |
+
datasets:
|
| 7 |
+
- mythicinfinity/libritts
|
| 8 |
---
|
| 9 |
|
| 10 |
# FocalCodec
|
|
|
|
| 27 |
|
| 28 |
---------------------------------------------------------------------------------------------------------
|
| 29 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
## @ Citing
|
| 31 |
|
| 32 |
```
|
|
|
|
| 44 |
|
| 45 |
[[email protected]](mailto:[email protected])
|
| 46 |
|
| 47 |
+
---------------------------------------------------------------------------------------------------------
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|