models.json
1 { 2 "tts_models": { 3 "multilingual":{ 4 "multi-dataset":{ 5 "your_tts":{ 6 "description": "Your TTS model accompanying the paper https://arxiv.org/abs/2112.02418", 7 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.10.1_models/tts_models--multilingual--multi-dataset--your_tts.zip", 8 "default_vocoder": null, 9 "commit": "e9a1953e", 10 "license": "CC BY-NC-ND 4.0", 11 "contact": "egolge@coqui.ai" 12 } 13 } 14 }, 15 "bg": { 16 "cv": { 17 "vits":{ 18 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--bg--cv--vits.zip", 19 "default_vocoder": null, 20 "commit": null, 21 "author": "@NeonGeckoCom", 22 "license": "bsd-3-clause" 23 } 24 } 25 }, 26 "cs": { 27 "cv": { 28 "vits":{ 29 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--cs--cv--vits.zip", 30 "default_vocoder": null, 31 "commit": null, 32 "author": "@NeonGeckoCom", 33 "license": "bsd-3-clause" 34 } 35 } 36 }, 37 "da": { 38 "cv": { 39 "vits":{ 40 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--da--cv--vits.zip", 41 "default_vocoder": null, 42 "commit": null, 43 "author": "@NeonGeckoCom", 44 "license": "bsd-3-clause" 45 } 46 } 47 }, 48 "et": { 49 "cv": { 50 "vits":{ 51 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--et--cv--vits.zip", 52 "default_vocoder": null, 53 "commit": null, 54 "author": "@NeonGeckoCom", 55 "license": "bsd-3-clause" 56 } 57 } 58 }, 59 "ga": { 60 "cv": { 61 "vits":{ 62 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--ga--cv--vits.zip", 63 "default_vocoder": null, 64 "commit": null, 65 "author": "@NeonGeckoCom", 66 "license": "bsd-3-clause" 67 } 68 } 69 }, 70 "en": { 71 "ek1": { 72 "tacotron2": { 73 "description": "EK1 en-rp tacotron2 by NMStoker", 74 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ek1--tacotron2.zip", 75 "default_vocoder": "vocoder_models/en/ek1/wavegrad", 76 "commit": "c802255", 77 "license": "apache 2.0" 78 } 79 }, 80 "ljspeech": { 81 "tacotron2-DDC": { 82 "description": "Tacotron2 with Double Decoder Consistency.", 83 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--tacotron2-DDC.zip", 84 "default_vocoder": "vocoder_models/en/ljspeech/hifigan_v2", 85 "commit": "bae2ad0f", 86 "author": "Eren Gölge @erogol", 87 "license": "apache 2.0", 88 "contact": "egolge@coqui.com" 89 }, 90 "tacotron2-DDC_ph": { 91 "description": "Tacotron2 with Double Decoder Consistency with phonemes.", 92 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--tacotron2-DDC_ph.zip", 93 "default_vocoder": "vocoder_models/en/ljspeech/univnet", 94 "commit": "3900448", 95 "author": "Eren Gölge @erogol", 96 "license": "apache 2.0", 97 "contact": "egolge@coqui.com" 98 }, 99 "glow-tts": { 100 "description": "", 101 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--glow-tts.zip", 102 "stats_file": null, 103 "default_vocoder": "vocoder_models/en/ljspeech/multiband-melgan", 104 "commit": "", 105 "author": "Eren Gölge @erogol", 106 "license": "MPL", 107 "contact": "egolge@coqui.com" 108 }, 109 "speedy-speech": { 110 "description": "Speedy Speech model trained on LJSpeech dataset using the Alignment Network for learning the durations.", 111 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--speedy-speech.zip", 112 "stats_file": null, 113 "default_vocoder": "vocoder_models/en/ljspeech/hifigan_v2", 114 "commit": "4581e3d", 115 "author": "Eren Gölge @erogol", 116 "license": "apache 2.0", 117 "contact": "egolge@coqui.com" 118 }, 119 "tacotron2-DCA": { 120 "description": "", 121 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--tacotron2-DCA.zip", 122 "default_vocoder": "vocoder_models/en/ljspeech/multiband-melgan", 123 "commit": "", 124 "author": "Eren Gölge @erogol", 125 "license": "MPL", 126 "contact": "egolge@coqui.com" 127 }, 128 "vits": { 129 "description": "VITS is an End2End TTS model trained on LJSpeech dataset with phonemes.", 130 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--vits.zip", 131 "default_vocoder": null, 132 "commit": "3900448", 133 "author": "Eren Gölge @erogol", 134 "license": "apache 2.0", 135 "contact": "egolge@coqui.com" 136 }, 137 "vits--neon": { 138 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--en--ljspeech--vits.zip", 139 "default_vocoder": null, 140 "author": "@NeonGeckoCom", 141 "license": "bsd-3-clause", 142 "contact": null, 143 "commit": null 144 }, 145 "fast_pitch": { 146 "description": "FastPitch model trained on LJSpeech using the Aligner Network", 147 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--ljspeech--fast_pitch.zip", 148 "default_vocoder": "vocoder_models/en/ljspeech/hifigan_v2", 149 "commit": "b27b3ba", 150 "author": "Eren Gölge @erogol", 151 "license": "apache 2.0", 152 "contact": "egolge@coqui.com" 153 }, 154 "overflow": { 155 "description": "Overflow model trained on LJSpeech", 156 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.10.0_models/tts_models--en--ljspeech--overflow.zip", 157 "default_vocoder": "vocoder_models/en/ljspeech/hifigan_v2", 158 "commit": "3b1a28f", 159 "author": "Eren Gölge @erogol", 160 "license": "apache 2.0", 161 "contact": "egolge@coqui.ai" 162 }, 163 "neural_hmm": { 164 "description": "Neural HMM model trained on LJSpeech", 165 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.11.0_models/tts_models--en--ljspeech--neural_hmm.zip", 166 "default_vocoder": "vocoder_models/en/ljspeech/hifigan_v2", 167 "commit": "3b1a28f", 168 "author": "Shivam Metha @shivammehta25", 169 "license": "apache 2.0", 170 "contact": "d83ee8fe45e3c0d776d4a865aca21d7c2ac324c4" 171 } 172 }, 173 "vctk": { 174 "vits": { 175 "description": "VITS End2End TTS model trained on VCTK dataset with 109 different speakers with EN accent.", 176 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--vctk--vits.zip", 177 "default_vocoder": null, 178 "commit": "3900448", 179 "author": "Eren @erogol", 180 "license": "apache 2.0", 181 "contact": "egolge@coqui.ai" 182 }, 183 "fast_pitch":{ 184 "description": "FastPitch model trained on VCTK dataseset.", 185 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--vctk--fast_pitch.zip", 186 "default_vocoder": null, 187 "commit": "bdab788d", 188 "author": "Eren @erogol", 189 "license": "CC BY-NC-ND 4.0", 190 "contact": "egolge@coqui.ai" 191 } 192 }, 193 "sam": { 194 "tacotron-DDC": { 195 "description": "Tacotron2 with Double Decoder Consistency trained with Aceenture's Sam dataset.", 196 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--en--sam--tacotron-DDC.zip", 197 "default_vocoder": "vocoder_models/en/sam/hifigan_v2", 198 "commit": "bae2ad0f", 199 "author": "Eren Gölge @erogol", 200 "license": "apache 2.0", 201 "contact": "egolge@coqui.com" 202 } 203 }, 204 "blizzard2013": { 205 "capacitron-t2-c50": { 206 "description": "Capacitron additions to Tacotron 2 with Capacity at 50 as in https://arxiv.org/pdf/1906.03402.pdf", 207 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.7.0_models/tts_models--en--blizzard2013--capacitron-t2-c50.zip", 208 "commit": "d6284e7", 209 "default_vocoder": "vocoder_models/en/blizzard2013/hifigan_v2", 210 "author": "Adam Froghyar @a-froghyar", 211 "license": "apache 2.0", 212 "contact": "adamfroghyar@gmail.com" 213 }, 214 "capacitron-t2-c150_v2": { 215 "description": "Capacitron additions to Tacotron 2 with Capacity at 150 as in https://arxiv.org/pdf/1906.03402.pdf", 216 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.7.1_models/tts_models--en--blizzard2013--capacitron-t2-c150_v2.zip", 217 "commit": "a67039d", 218 "default_vocoder": "vocoder_models/en/blizzard2013/hifigan_v2", 219 "author": "Adam Froghyar @a-froghyar", 220 "license": "apache 2.0", 221 "contact": "adamfroghyar@gmail.com" 222 } 223 224 }, 225 "multi-dataset":{ 226 "tortoise-v2":{ 227 "description": "Tortoise tts model https://github.com/neonbjb/tortoise-tts", 228 "github_rls_url": ["https://coqui.gateway.scarf.sh/v0.14.1_models/autoregressive.pth", 229 "https://coqui.gateway.scarf.sh/v0.14.1_models/clvp2.pth", 230 "https://coqui.gateway.scarf.sh/v0.14.1_models/cvvp.pth", 231 "https://coqui.gateway.scarf.sh/v0.14.1_models/diffusion_decoder.pth", 232 "https://coqui.gateway.scarf.sh/v0.14.1_models/rlg_auto.pth", 233 "https://coqui.gateway.scarf.sh/v0.14.1_models/rlg_diffuser.pth", 234 "https://coqui.gateway.scarf.sh/v0.14.1_models/vocoder.pth", 235 "https://coqui.gateway.scarf.sh/v0.14.1_models/mel_norms.pth", 236 "https://coqui.gateway.scarf.sh/v0.14.1_models/config.json" 237 ], 238 "commit": "c1875f6", 239 "default_vocoder": null, 240 "author": "@neonbjb - James Betker, @manmay-nakhashi Manmay Nakhashi", 241 "license": "apache 2.0" 242 } 243 }, 244 "jenny": { 245 "jenny":{ 246 "description": "VITS model trained with Jenny(Dioco) dataset. Named as Jenny as demanded by the license. Original URL for the model https://www.kaggle.com/datasets/noml4u/tts-models--en--jenny-dioco--vits", 247 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.14.0_models/tts_models--en--jenny--jenny.zip", 248 "default_vocoder": null, 249 "commit": "ba40a1c", 250 "license": "custom - see https://github.com/dioco-group/jenny-tts-dataset#important", 251 "author": "@noml4u" 252 } 253 } 254 }, 255 "es": { 256 "mai": { 257 "tacotron2-DDC": { 258 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--es--mai--tacotron2-DDC.zip", 259 "default_vocoder": "vocoder_models/universal/libri-tts/fullband-melgan", 260 "commit": "", 261 "author": "Eren Gölge @erogol", 262 "license": "MPL", 263 "contact": "egolge@coqui.com" 264 } 265 }, 266 "css10":{ 267 "vits":{ 268 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--es--css10--vits.zip", 269 "default_vocoder": null, 270 "commit": null, 271 "author": "@NeonGeckoCom", 272 "license": "bsd-3-clause" 273 } 274 } 275 }, 276 "fr": { 277 "mai": { 278 "tacotron2-DDC": { 279 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--fr--mai--tacotron2-DDC.zip", 280 "default_vocoder": "vocoder_models/universal/libri-tts/fullband-melgan", 281 "commit": null, 282 "author": "Eren Gölge @erogol", 283 "license": "MPL", 284 "contact": "egolge@coqui.com" 285 } 286 }, 287 "css10":{ 288 "vits":{ 289 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--fr--css10--vits.zip", 290 "default_vocoder": null, 291 "commit": null, 292 "author": "@NeonGeckoCom", 293 "license": "bsd-3-clause" 294 } 295 } 296 }, 297 "uk":{ 298 "mai": { 299 "glow-tts": { 300 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--uk--mai--glow-tts.zip", 301 "author":"@robinhad", 302 "commit": "bdab788d", 303 "license": "MIT", 304 "contact": "", 305 "default_vocoder": "vocoder_models/uk/mai/multiband-melgan" 306 }, 307 "vits":{ 308 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--uk--mai--vits.zip", 309 "default_vocoder": null, 310 "commit": null, 311 "author": "@NeonGeckoCom", 312 "license": "bsd-3-clause" 313 } 314 } 315 }, 316 "zh-CN": { 317 "baker": { 318 "tacotron2-DDC-GST": { 319 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--zh-CN--baker--tacotron2-DDC-GST.zip", 320 "commit": "unknown", 321 "author": "@kirianguiller", 322 "license": "apache 2.0", 323 "default_vocoder": null 324 } 325 } 326 }, 327 "nl": { 328 "mai": { 329 "tacotron2-DDC": { 330 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--nl--mai--tacotron2-DDC.zip", 331 "author": "@r-dh", 332 "license": "apache 2.0", 333 "default_vocoder": "vocoder_models/nl/mai/parallel-wavegan", 334 "stats_file": null, 335 "commit": "540d811" 336 } 337 }, 338 "css10":{ 339 "vits":{ 340 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--nl--css10--vits.zip", 341 "default_vocoder": null, 342 "commit": null, 343 "author": "@NeonGeckoCom", 344 "license": "bsd-3-clause" 345 } 346 } 347 }, 348 "de": { 349 "thorsten": { 350 "tacotron2-DCA": { 351 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--de--thorsten--tacotron2-DCA.zip", 352 "default_vocoder": "vocoder_models/de/thorsten/fullband-melgan", 353 "author": "@thorstenMueller", 354 "license": "apache 2.0", 355 "commit": "unknown" 356 }, 357 "vits": { 358 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.7.0_models/tts_models--de--thorsten--vits.zip", 359 "default_vocoder": null, 360 "author": "@thorstenMueller", 361 "license": "apache 2.0", 362 "commit": "unknown" 363 }, 364 "tacotron2-DDC": { 365 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--de--thorsten--tacotron2-DDC.zip", 366 "default_vocoder": "vocoder_models/de/thorsten/hifigan_v1", 367 "description": "Thorsten-Dec2021-22k-DDC", 368 "author": "@thorstenMueller", 369 "license": "apache 2.0", 370 "commit": "unknown" 371 } 372 }, 373 "css10": { 374 "vits-neon":{ 375 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--de--css10--vits.zip", 376 "default_vocoder": null, 377 "author": "@NeonGeckoCom", 378 "license": "bsd-3-clause", 379 "commit": null 380 } 381 } 382 }, 383 "ja": { 384 "kokoro": { 385 "tacotron2-DDC": { 386 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--ja--kokoro--tacotron2-DDC.zip", 387 "default_vocoder": "vocoder_models/ja/kokoro/hifigan_v1", 388 "description": "Tacotron2 with Double Decoder Consistency trained with Kokoro Speech Dataset.", 389 "author": "@kaiidams", 390 "license": "apache 2.0", 391 "commit": "401fbd89" 392 } 393 } 394 }, 395 "tr":{ 396 "common-voice": { 397 "glow-tts":{ 398 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--tr--common-voice--glow-tts.zip", 399 "default_vocoder": "vocoder_models/tr/common-voice/hifigan", 400 "license": "MIT", 401 "description": "Turkish GlowTTS model using an unknown speaker from the Common-Voice dataset.", 402 "author": "Fatih Akademi", 403 "commit": null 404 } 405 } 406 }, 407 "it": { 408 "mai_female": { 409 "glow-tts":{ 410 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_female--glow-tts.zip", 411 "default_vocoder": null, 412 "description": "GlowTTS model as explained on https://github.com/coqui-ai/TTS/issues/1148.", 413 "author": "@nicolalandro", 414 "license": "apache 2.0", 415 "commit": null 416 }, 417 "vits":{ 418 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_female--vits.zip", 419 "default_vocoder": null, 420 "description": "GlowTTS model as explained on https://github.com/coqui-ai/TTS/issues/1148.", 421 "author": "@nicolalandro", 422 "license": "apache 2.0", 423 "commit": null 424 } 425 }, 426 "mai_male": { 427 "glow-tts":{ 428 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_male--glow-tts.zip", 429 "default_vocoder": null, 430 "description": "GlowTTS model as explained on https://github.com/coqui-ai/TTS/issues/1148.", 431 "author": "@nicolalandro", 432 "license": "apache 2.0", 433 "commit": null 434 }, 435 "vits":{ 436 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/tts_models--it--mai_male--vits.zip", 437 "default_vocoder": null, 438 "description": "GlowTTS model as explained on https://github.com/coqui-ai/TTS/issues/1148.", 439 "author": "@nicolalandro", 440 "license": "apache 2.0", 441 "commit": null 442 } 443 } 444 }, 445 "ewe": { 446 "openbible": { 447 "vits":{ 448 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--ewe--openbible--vits.zip", 449 "default_vocoder": null, 450 "license": "CC-BY-SA 4.0", 451 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 452 "author": "@coqui_ai", 453 "commit": "1b22f03" 454 } 455 } 456 }, 457 "hau": { 458 "openbible": { 459 "vits":{ 460 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--hau--openbible--vits.zip", 461 "default_vocoder": null, 462 "license": "CC-BY-SA 4.0", 463 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 464 "author": "@coqui_ai", 465 "commit": "1b22f03" 466 } 467 } 468 }, 469 "lin": { 470 "openbible": { 471 "vits":{ 472 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--lin--openbible--vits.zip", 473 "default_vocoder": null, 474 "license": "CC-BY-SA 4.0", 475 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 476 "author": "@coqui_ai", 477 "commit": "1b22f03" 478 } 479 } 480 }, 481 "tw_akuapem": { 482 "openbible": { 483 "vits":{ 484 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--tw_akuapem--openbible--vits.zip", 485 "default_vocoder": null, 486 "license": "CC-BY-SA 4.0", 487 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 488 "author": "@coqui_ai", 489 "commit": "1b22f03" 490 } 491 } 492 }, 493 "tw_asante": { 494 "openbible": { 495 "vits":{ 496 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--tw_asante--openbible--vits.zip", 497 "default_vocoder": null, 498 "license": "CC-BY-SA 4.0", 499 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 500 "author": "@coqui_ai", 501 "commit": "1b22f03" 502 } 503 } 504 }, 505 "yor": { 506 "openbible": { 507 "vits":{ 508 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.2_models/tts_models--yor--openbible--vits.zip", 509 "default_vocoder": null, 510 "license": "CC-BY-SA 4.0", 511 "description": "Original work (audio and text) by Biblica available for free at www.biblica.com and open.bible.", 512 "author": "@coqui_ai", 513 "commit": "1b22f03" 514 } 515 } 516 }, 517 "hu": { 518 "css10": { 519 "vits": { 520 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--hu--css10--vits.zip", 521 "default_vocoder": null, 522 "commit": null, 523 "author": "@NeonGeckoCom", 524 "license": "bsd-3-clause" 525 } 526 } 527 }, 528 "el": { 529 "cv": { 530 "vits": { 531 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--el--cv--vits.zip", 532 "default_vocoder": null, 533 "commit": null, 534 "author": "@NeonGeckoCom", 535 "license": "bsd-3-clause" 536 } 537 } 538 }, 539 "fi": { 540 "css10": { 541 "vits":{ 542 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--fi--css10--vits.zip", 543 "default_vocoder": null, 544 "commit": null, 545 "author": "@NeonGeckoCom", 546 "license": "bsd-3-clause" 547 } 548 } 549 }, 550 "hr": { 551 "cv": { 552 "vits":{ 553 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--hr--cv--vits.zip", 554 "default_vocoder": null, 555 "commit": null, 556 "author": "@NeonGeckoCom", 557 "license": "bsd-3-clause" 558 } 559 } 560 }, 561 "lt": { 562 "cv": { 563 "vits":{ 564 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--lt--cv--vits.zip", 565 "default_vocoder": null, 566 "commit": null, 567 "author": "@NeonGeckoCom", 568 "license": "bsd-3-clause" 569 } 570 } 571 }, 572 "lv": { 573 "cv": { 574 "vits":{ 575 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--lv--cv--vits.zip", 576 "default_vocoder": null, 577 "commit": null, 578 "author": "@NeonGeckoCom", 579 "license": "bsd-3-clause" 580 } 581 } 582 }, 583 "mt": { 584 "cv": { 585 "vits":{ 586 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--mt--cv--vits.zip", 587 "default_vocoder": null, 588 "commit": null, 589 "author": "@NeonGeckoCom", 590 "license": "bsd-3-clause" 591 } 592 } 593 }, 594 "pl": { 595 "mai_female": { 596 "vits":{ 597 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--pl--mai_female--vits.zip", 598 "default_vocoder": null, 599 "commit": null, 600 "author": "@NeonGeckoCom", 601 "license": "bsd-3-clause" 602 } 603 } 604 }, 605 "pt": { 606 "cv": { 607 "vits":{ 608 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--pt--cv--vits.zip", 609 "default_vocoder": null, 610 "commit": null, 611 "author": "@NeonGeckoCom", 612 "license": "bsd-3-clause" 613 } 614 } 615 }, 616 "ro": { 617 "cv": { 618 "vits":{ 619 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--ro--cv--vits.zip", 620 "default_vocoder": null, 621 "commit": null, 622 "author": "@NeonGeckoCom", 623 "license": "bsd-3-clause" 624 } 625 } 626 }, 627 "sk": { 628 "cv": { 629 "vits":{ 630 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sk--cv--vits.zip", 631 "default_vocoder": null, 632 "commit": null, 633 "author": "@NeonGeckoCom", 634 "license": "bsd-3-clause" 635 } 636 } 637 }, 638 "sl": { 639 "cv": { 640 "vits":{ 641 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sl--cv--vits.zip", 642 "default_vocoder": null, 643 "commit": null, 644 "author": "@NeonGeckoCom", 645 "license": "bsd-3-clause" 646 } 647 } 648 }, 649 "sv": { 650 "cv": { 651 "vits":{ 652 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/tts_models--sv--cv--vits.zip", 653 "default_vocoder": null, 654 "commit": null, 655 "author": "@NeonGeckoCom", 656 "license": "bsd-3-clause" 657 } 658 } 659 }, 660 "ca": { 661 "custom": { 662 "vits":{ 663 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.10.1_models/tts_models--ca--custom--vits.zip", 664 "default_vocoder": null, 665 "commit": null, 666 "description": " It is trained from zero with 101460 utterances consisting of 257 speakers, approx 138 hours of speech. We used three datasets;\nFestcat and Google Catalan TTS (both TTS datasets) and also a part of Common Voice 8. It is trained with TTS v0.8.0.\nhttps://github.com/coqui-ai/TTS/discussions/930#discussioncomment-4466345", 667 "author": "@gullabi", 668 "license": "CC-BY-4.0" 669 } 670 } 671 }, 672 "fa":{ 673 "custom":{ 674 "glow-tts": { 675 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.10.1_models/tts_models--fa--custom--glow-tts.zip", 676 "default_vocoder": null, 677 "commit": null, 678 "description": "persian-tts-female-glow_tts model for text to speech purposes. Single-speaker female voice Trained on persian-tts-dataset-famale. \nThis model has no compatible vocoder thus the output quality is not very good. \nDataset: https://www.kaggle.com/datasets/magnoliasis/persian-tts-dataset-famale.", 679 "author": "@karim23657", 680 "license": "CC-BY-4.0" 681 } 682 } 683 }, 684 "bn":{ 685 "custom":{ 686 "vits-male":{ 687 "github_rls_url":"https://coqui.gateway.scarf.sh/v0.13.3_models/tts_models--bn--custom--vits_male.zip", 688 "default_vocoder": null, 689 "commit": null, 690 "description": "Single speaker Bangla male model. For more information -> https://github.com/mobassir94/comprehensive-bangla-tts", 691 "author": "@mobassir94", 692 "license": "Apache 2.0" 693 }, 694 "vits-female":{ 695 "github_rls_url":"https://coqui.gateway.scarf.sh/v0.13.3_models/tts_models--bn--custom--vits_female.zip", 696 "default_vocoder": null, 697 "commit": null, 698 "description": "Single speaker Bangla female model. For more information -> https://github.com/mobassir94/comprehensive-bangla-tts", 699 "author": "@mobassir94", 700 "license": "Apache 2.0" 701 } 702 } 703 } 704 }, 705 "vocoder_models": { 706 "universal": { 707 "libri-tts": { 708 "wavegrad": { 709 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--universal--libri-tts--wavegrad.zip", 710 "commit": "ea976b0", 711 "author": "Eren Gölge @erogol", 712 "license": "MPL", 713 "contact": "egolge@coqui.com" 714 }, 715 "fullband-melgan": { 716 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--universal--libri-tts--fullband-melgan.zip", 717 "commit": "4132240", 718 "author": "Eren Gölge @erogol", 719 "license": "MPL", 720 "contact": "egolge@coqui.com" 721 } 722 } 723 }, 724 "en": { 725 "ek1": { 726 "wavegrad": { 727 "description": "EK1 en-rp wavegrad by NMStoker", 728 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--ek1--wavegrad.zip", 729 "commit": "c802255", 730 "license": "apache 2.0" 731 } 732 }, 733 "ljspeech": { 734 "multiband-melgan": { 735 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--ljspeech--multiband-melgan.zip", 736 "commit": "ea976b0", 737 "author": "Eren Gölge @erogol", 738 "license": "MPL", 739 "contact": "egolge@coqui.com" 740 }, 741 "hifigan_v2": { 742 "description": "HiFiGAN_v2 LJSpeech vocoder from https://arxiv.org/abs/2010.05646.", 743 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--ljspeech--hifigan_v2.zip", 744 "commit": "bae2ad0f", 745 "author": "@erogol", 746 "license": "apache 2.0", 747 "contact": "egolge@coqui.ai" 748 }, 749 "univnet": { 750 "description": "UnivNet model finetuned on TacotronDDC_ph spectrograms for better compatibility.", 751 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--ljspeech--univnet_v2.zip", 752 "commit": "4581e3d", 753 "author": "Eren @erogol", 754 "license": "apache 2.0", 755 "contact": "egolge@coqui.ai" 756 } 757 }, 758 "blizzard2013": { 759 "hifigan_v2": { 760 "description": "HiFiGAN_v2 LJSpeech vocoder from https://arxiv.org/abs/2010.05646.", 761 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.7.0_models/vocoder_models--en--blizzard2013--hifigan_v2.zip", 762 "commit": "d6284e7", 763 "author": "Adam Froghyar @a-froghyar", 764 "license": "apache 2.0", 765 "contact": "adamfroghyar@gmail.com" 766 } 767 }, 768 "vctk": { 769 "hifigan_v2": { 770 "description": "Finetuned and intended to be used with tts_models/en/vctk/sc-glow-tts", 771 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--vctk--hifigan_v2.zip", 772 "commit": "2f07160", 773 "author": "Edresson Casanova", 774 "license": "apache 2.0", 775 "contact": "" 776 } 777 }, 778 "sam": { 779 "hifigan_v2": { 780 "description": "Finetuned and intended to be used with tts_models/en/sam/tacotron_DDC", 781 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--en--sam--hifigan_v2.zip", 782 "commit": "2f07160", 783 "author": "Eren Gölge @erogol", 784 "license": "apache 2.0", 785 "contact": "egolge@coqui.ai" 786 } 787 } 788 }, 789 "nl": { 790 "mai": { 791 "parallel-wavegan": { 792 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--nl--mai--parallel-wavegan.zip", 793 "author": "@r-dh", 794 "license": "apache 2.0", 795 "commit": "unknown" 796 } 797 } 798 }, 799 "de": { 800 "thorsten": { 801 "wavegrad": { 802 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--de--thorsten--wavegrad.zip", 803 "author": "@thorstenMueller", 804 "license": "apache 2.0", 805 "commit": "unknown" 806 }, 807 "fullband-melgan": { 808 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--de--thorsten--fullband-melgan.zip", 809 "author": "@thorstenMueller", 810 "license": "apache 2.0", 811 "commit": "unknown" 812 }, 813 "hifigan_v1": { 814 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.8.0_models/vocoder_models--de--thorsten--hifigan_v1.zip", 815 "description": "HifiGAN vocoder model for Thorsten Neutral Dec2021 22k Samplerate Tacotron2 DDC model", 816 "author": "@thorstenMueller", 817 "license": "apache 2.0", 818 "commit": "unknown" 819 } 820 } 821 }, 822 "ja": { 823 "kokoro": { 824 "hifigan_v1": { 825 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--ja--kokoro--hifigan_v1.zip", 826 "description": "HifiGAN model trained for kokoro dataset by @kaiidams", 827 "author": "@kaiidams", 828 "license": "apache 2.0", 829 "commit": "3900448" 830 } 831 } 832 }, 833 "uk": { 834 "mai": { 835 "multiband-melgan": { 836 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--uk--mai--multiband-melgan.zip", 837 "author":"@robinhad", 838 "commit": "bdab788d", 839 "license": "MIT", 840 "contact": "" 841 } 842 } 843 }, 844 "tr":{ 845 "common-voice": { 846 "hifigan":{ 847 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.6.1_models/vocoder_models--tr--common-voice--hifigan.zip", 848 "description": "HifiGAN model using an unknown speaker from the Common-Voice dataset.", 849 "author": "Fatih Akademi", 850 "license": "MIT", 851 "commit": null 852 } 853 } 854 } 855 }, 856 "voice_conversion_models":{ 857 "multilingual":{ 858 "vctk":{ 859 "freevc24":{ 860 "github_rls_url": "https://coqui.gateway.scarf.sh/v0.13.0_models/voice_conversion_models--multilingual--vctk--freevc24.zip", 861 "description": "FreeVC model trained on VCTK dataset from https://github.com/OlaWod/FreeVC", 862 "author": "Jing-Yi Li @OlaWod", 863 "license": "MIT", 864 "commit": null 865 } 866 } 867 } 868 } 869 }