diff --git a/README.md b/README.md index aaab83d..f323312 100644 --- a/README.md +++ b/README.md @@ -17,10 +17,10 @@ python3 -m pip install --upgrade pip python3 -m pip install -r requirements.txt ``` -Download the LibriTTS generator: +Download the LibriTTS-R generator (exported from [checkpoint](https://huggingface.co/datasets/rhasspy/piper-checkpoints/tree/main/en/en_US/libritts_r/medium)): ``` sh -wget -O models/en-us-libritts-high.pt 'https://github.com/rhasspy/piper-sample-generator/releases/download/v1.0.0/en-us-libritts-high.pt' +wget -O models/en-us-libritts-high.pt 'https://github.com/rhasspy/piper-sample-generator/releases/download/v2.0.0/en_US-libritts_r-medium.pt' ``` @@ -42,7 +42,7 @@ python3 generate_samples.py 'okay, piper.' --max-samples 100 --batch-size 10 --o On an NVidia 2080 Ti with 11GB, a batch size of 100 was possible (generating approximately 100 samples per second). -Setting `--max-speakers` to a value less than 904 (the number if LibriTTS) is recommended. Because very few samples of later speakers were in the original dataset, using them can cause audio artifacts. +Setting `--max-speakers` to a value less than 904 (the number of speakers LibriTTS) is recommended. Because very few samples of later speakers were in the original dataset, using them can cause audio artifacts. See `--help` for more options, including adjust the `--length-scales` (speaking speeds) and `--slerp-weights` (speaker blending) which are cycled per batch. diff --git a/generate_samples.py b/generate_samples.py index f1f098d..1461bbe 100755 --- a/generate_samples.py +++ b/generate_samples.py @@ -10,7 +10,7 @@ from pathlib import Path import numpy as np import torch -from espeak_phonemizer import Phonemizer +from piper_phonemize import phonemize_espeak, phoneme_ids_espeak from piper_train.vits import commons _DIR = Path(__file__).parent @@ -21,7 +21,9 @@ def main() -> None: parser = argparse.ArgumentParser() parser.add_argument("text") parser.add_argument("--max-samples", required=True, type=int) - parser.add_argument("--model", default=_DIR / "models" / "en-us-libritts-high.pt") + parser.add_argument( + "--model", default=_DIR / "models" / "en_US-libritts_r-medium.pt" + ) parser.add_argument("--batch-size", type=int, default=1) parser.add_argument("--slerp-weights", nargs="+", type=float, default=[0.5]) parser.add_argument( @@ -61,9 +63,12 @@ def main() -> None: if args.max_speakers is not None: num_speakers = min(num_speakers, args.max_speakers) - phonemizer = Phonemizer(voice) - phonemes_str = phonemizer.phonemize(args.text) - phonemes = list(unicodedata.normalize("NFD", phonemes_str)) + # Combine all sentences + phonemes = [ + p + for sentence_phonemes in phonemize_espeak(args.text, voice) + for p in sentence_phonemes + ] _LOGGER.debug("Phonemes: %s", phonemes) id_map = config["phoneme_id_map"] diff --git a/models/en_US-libritts_r-medium.pt.json b/models/en_US-libritts_r-medium.pt.json new file mode 100755 index 0000000..e55cfcc --- /dev/null +++ b/models/en_US-libritts_r-medium.pt.json @@ -0,0 +1,1403 @@ +{ + "audio": { + "sample_rate": 22050 + }, + "espeak": { + "voice": "en-us" + }, + "inference": { + "noise_scale": 0.333, + "length_scale": 1, + "noise_w": 0.333 + }, + "phoneme_type": "espeak", + "phoneme_map": {}, + "phoneme_id_map": { + " ": [ + 3 + ], + "!": [ + 4 + ], + "\"": [ + 150 + ], + "#": [ + 149 + ], + "$": [ + 2 + ], + "'": [ + 5 + ], + "(": [ + 6 + ], + ")": [ + 7 + ], + ",": [ + 8 + ], + "-": [ + 9 + ], + ".": [ + 10 + ], + "0": [ + 130 + ], + "1": [ + 131 + ], + "2": [ + 132 + ], + "3": [ + 133 + ], + "4": [ + 134 + ], + "5": [ + 135 + ], + "6": [ + 136 + ], + "7": [ + 137 + ], + "8": [ + 138 + ], + "9": [ + 139 + ], + ":": [ + 11 + ], + ";": [ + 12 + ], + "?": [ + 13 + ], + "X": [ + 156 + ], + "^": [ + 1 + ], + "_": [ + 0 + ], + "a": [ + 14 + ], + "b": [ + 15 + ], + "c": [ + 16 + ], + "d": [ + 17 + ], + "e": [ + 18 + ], + "f": [ + 19 + ], + "g": [ + 154 + ], + "h": [ + 20 + ], + "i": [ + 21 + ], + "j": [ + 22 + ], + "k": [ + 23 + ], + "l": [ + 24 + ], + "m": [ + 25 + ], + "n": [ + 26 + ], + "o": [ + 27 + ], + "p": [ + 28 + ], + "q": [ + 29 + ], + "r": [ + 30 + ], + "s": [ + 31 + ], + "t": [ + 32 + ], + "u": [ + 33 + ], + "v": [ + 34 + ], + "w": [ + 35 + ], + "x": [ + 36 + ], + "y": [ + 37 + ], + "z": [ + 38 + ], + "æ": [ + 39 + ], + "ç": [ + 40 + ], + "ð": [ + 41 + ], + "ø": [ + 42 + ], + "ħ": [ + 43 + ], + "ŋ": [ + 44 + ], + "œ": [ + 45 + ], + "ǀ": [ + 46 + ], + "ǁ": [ + 47 + ], + "ǂ": [ + 48 + ], + "ǃ": [ + 49 + ], + "ɐ": [ + 50 + ], + "ɑ": [ + 51 + ], + "ɒ": [ + 52 + ], + "ɓ": [ + 53 + ], + "ɔ": [ + 54 + ], + "ɕ": [ + 55 + ], + "ɖ": [ + 56 + ], + "ɗ": [ + 57 + ], + "ɘ": [ + 58 + ], + "ə": [ + 59 + ], + "ɚ": [ + 60 + ], + "ɛ": [ + 61 + ], + "ɜ": [ + 62 + ], + "ɞ": [ + 63 + ], + "ɟ": [ + 64 + ], + "ɠ": [ + 65 + ], + "ɡ": [ + 66 + ], + "ɢ": [ + 67 + ], + "ɣ": [ + 68 + ], + "ɤ": [ + 69 + ], + "ɥ": [ + 70 + ], + "ɦ": [ + 71 + ], + "ɧ": [ + 72 + ], + "ɨ": [ + 73 + ], + "ɪ": [ + 74 + ], + "ɫ": [ + 75 + ], + "ɬ": [ + 76 + ], + "ɭ": [ + 77 + ], + "ɮ": [ + 78 + ], + "ɯ": [ + 79 + ], + "ɰ": [ + 80 + ], + "ɱ": [ + 81 + ], + "ɲ": [ + 82 + ], + "ɳ": [ + 83 + ], + "ɴ": [ + 84 + ], + "ɵ": [ + 85 + ], + "ɶ": [ + 86 + ], + "ɸ": [ + 87 + ], + "ɹ": [ + 88 + ], + "ɺ": [ + 89 + ], + "ɻ": [ + 90 + ], + "ɽ": [ + 91 + ], + "ɾ": [ + 92 + ], + "ʀ": [ + 93 + ], + "ʁ": [ + 94 + ], + "ʂ": [ + 95 + ], + "ʃ": [ + 96 + ], + "ʄ": [ + 97 + ], + "ʈ": [ + 98 + ], + "ʉ": [ + 99 + ], + "ʊ": [ + 100 + ], + "ʋ": [ + 101 + ], + "ʌ": [ + 102 + ], + "ʍ": [ + 103 + ], + "ʎ": [ + 104 + ], + "ʏ": [ + 105 + ], + "ʐ": [ + 106 + ], + "ʑ": [ + 107 + ], + "ʒ": [ + 108 + ], + "ʔ": [ + 109 + ], + "ʕ": [ + 110 + ], + "ʘ": [ + 111 + ], + "ʙ": [ + 112 + ], + "ʛ": [ + 113 + ], + "ʜ": [ + 114 + ], + "ʝ": [ + 115 + ], + "ʟ": [ + 116 + ], + "ʡ": [ + 117 + ], + "ʢ": [ + 118 + ], + "ʦ": [ + 155 + ], + "ʰ": [ + 145 + ], + "ʲ": [ + 119 + ], + "ˈ": [ + 120 + ], + "ˌ": [ + 121 + ], + "ː": [ + 122 + ], + "ˑ": [ + 123 + ], + "˞": [ + 124 + ], + "ˤ": [ + 146 + ], + "̃": [ + 141 + ], + "̊": [ + 158 + ], + "̝": [ + 157 + ], + "̧": [ + 140 + ], + "̩": [ + 144 + ], + "̪": [ + 142 + ], + "̯": [ + 143 + ], + "̺": [ + 152 + ], + "̻": [ + 153 + ], + "β": [ + 125 + ], + "ε": [ + 147 + ], + "θ": [ + 126 + ], + "χ": [ + 127 + ], + "ᵻ": [ + 128 + ], + "↑": [ + 151 + ], + "↓": [ + 148 + ], + "ⱱ": [ + 129 + ] + }, + "num_symbols": 256, + "num_speakers": 904, + "speaker_id_map": { + "3922": 0, + "8699": 1, + "4535": 2, + "6701": 3, + "3638": 4, + "922": 5, + "2531": 6, + "1638": 7, + "8848": 8, + "6544": 9, + "3615": 10, + "318": 11, + "6104": 12, + "5400": 13, + "5712": 14, + "1382": 15, + "2769": 16, + "2573": 17, + "1463": 18, + "6458": 19, + "4356": 20, + "3274": 21, + "8498": 22, + "5570": 23, + "176": 24, + "339": 25, + "28": 26, + "5909": 27, + "3869": 28, + "4899": 29, + "64": 30, + "3368": 31, + "3307": 32, + "5618": 33, + "3370": 34, + "4595": 35, + "7704": 36, + "8506": 37, + "8410": 38, + "6904": 39, + "5655": 40, + "2204": 41, + "501": 42, + "7314": 43, + "1027": 44, + "5054": 45, + "534": 46, + "2853": 47, + "5935": 48, + "2404": 49, + "7874": 50, + "816": 51, + "2053": 52, + "8066": 53, + "16": 54, + "4586": 55, + "1923": 56, + "2592": 57, + "1265": 58, + "6189": 59, + "100": 60, + "6371": 61, + "4957": 62, + "4116": 63, + "3003": 64, + "7739": 65, + "1752": 66, + "5717": 67, + "5012": 68, + "5062": 69, + "7481": 70, + "2299": 71, + "7188": 72, + "93": 73, + "4145": 74, + "8684": 75, + "7594": 76, + "2598": 77, + "3540": 78, + "7717": 79, + "6426": 80, + "4148": 81, + "1379": 82, + "2512": 83, + "335": 84, + "242": 85, + "8855": 86, + "8118": 87, + "369": 88, + "6575": 89, + "6694": 90, + "8080": 91, + "1283": 92, + "7434": 93, + "5290": 94, + "1731": 95, + "2401": 96, + "459": 97, + "192": 98, + "7910": 99, + "114": 100, + "5660": 101, + "1313": 102, + "203": 103, + "7460": 104, + "207": 105, + "6497": 106, + "6696": 107, + "7766": 108, + "6233": 109, + "3185": 110, + "2010": 111, + "2056": 112, + "3717": 113, + "5802": 114, + "5622": 115, + "2156": 116, + "4243": 117, + "5039": 118, + "4110": 119, + "1093": 120, + "1776": 121, + "7995": 122, + "6877": 123, + "5635": 124, + "54": 125, + "288": 126, + "4592": 127, + "7276": 128, + "688": 129, + "8388": 130, + "1422": 131, + "8152": 132, + "8194": 133, + "7000": 134, + "8527": 135, + "5126": 136, + "3923": 137, + "1054": 138, + "3927": 139, + "5029": 140, + "4098": 141, + "56": 142, + "7240": 143, + "5538": 144, + "1903": 145, + "6538": 146, + "3380": 147, + "6643": 148, + "7495": 149, + "8718": 150, + "8050": 151, + "1789": 152, + "126": 153, + "7245": 154, + "2517": 155, + "4438": 156, + "7145": 157, + "724": 158, + "4945": 159, + "9022": 160, + "6637": 161, + "6927": 162, + "6937": 163, + "8113": 164, + "5724": 165, + "6006": 166, + "3584": 167, + "2971": 168, + "2230": 169, + "7982": 170, + "1649": 171, + "3994": 172, + "7720": 173, + "6981": 174, + "781": 175, + "4973": 176, + "6206": 177, + "2481": 178, + "3157": 179, + "1509": 180, + "510": 181, + "7540": 182, + "8887": 183, + "7120": 184, + "2882": 185, + "7128": 186, + "8142": 187, + "7229": 188, + "2787": 189, + "8820": 190, + "2368": 191, + "4331": 192, + "4967": 193, + "4427": 194, + "6054": 195, + "3728": 196, + "274": 197, + "7134": 198, + "1603": 199, + "1383": 200, + "1165": 201, + "4363": 202, + "512": 203, + "5985": 204, + "7967": 205, + "2060": 206, + "7752": 207, + "7484": 208, + "8643": 209, + "3549": 210, + "5731": 211, + "7881": 212, + "667": 213, + "6828": 214, + "5740": 215, + "3483": 216, + "718": 217, + "6341": 218, + "1913": 219, + "3228": 220, + "7247": 221, + "7705": 222, + "1018": 223, + "8193": 224, + "6098": 225, + "3989": 226, + "7828": 227, + "5876": 228, + "7754": 229, + "4719": 230, + "8011": 231, + "7939": 232, + "5975": 233, + "2004": 234, + "6139": 235, + "8183": 236, + "3482": 237, + "3361": 238, + "4289": 239, + "231": 240, + "7789": 241, + "4598": 242, + "5239": 243, + "2638": 244, + "6300": 245, + "8474": 246, + "2194": 247, + "7832": 248, + "1079": 249, + "1335": 250, + "188": 251, + "1195": 252, + "5914": 253, + "1401": 254, + "7318": 255, + "5448": 256, + "1392": 257, + "3703": 258, + "2113": 259, + "7783": 260, + "8176": 261, + "6519": 262, + "7933": 263, + "7938": 264, + "7802": 265, + "6120": 266, + "224": 267, + "209": 268, + "5656": 269, + "3032": 270, + "6965": 271, + "258": 272, + "4837": 273, + "5489": 274, + "272": 275, + "3851": 276, + "7140": 277, + "2562": 278, + "1472": 279, + "79": 280, + "2775": 281, + "3046": 282, + "2532": 283, + "8266": 284, + "6099": 285, + "4425": 286, + "5293": 287, + "7981": 288, + "2045": 289, + "920": 290, + "511": 291, + "7416": 292, + "835": 293, + "1289": 294, + "8195": 295, + "7833": 296, + "8772": 297, + "968": 298, + "1641": 299, + "7117": 300, + "1678": 301, + "5809": 302, + "8028": 303, + "500": 304, + "6505": 305, + "7868": 306, + "14": 307, + "2238": 308, + "4744": 309, + "3733": 310, + "7515": 311, + "699": 312, + "5093": 313, + "6388": 314, + "7959": 315, + "98": 316, + "3914": 317, + "5246": 318, + "2570": 319, + "8396": 320, + "3513": 321, + "882": 322, + "7994": 323, + "5968": 324, + "8591": 325, + "806": 326, + "5261": 327, + "1271": 328, + "899": 329, + "3945": 330, + "8404": 331, + "249": 332, + "3008": 333, + "7139": 334, + "6395": 335, + "6215": 336, + "6080": 337, + "4054": 338, + "7825": 339, + "6683": 340, + "8725": 341, + "3230": 342, + "4138": 343, + "6160": 344, + "666": 345, + "6510": 346, + "3551": 347, + "8075": 348, + "225": 349, + "7169": 350, + "1851": 351, + "5984": 352, + "2960": 353, + "8329": 354, + "175": 355, + "6378": 356, + "480": 357, + "7538": 358, + "479": 359, + "5519": 360, + "8534": 361, + "4856": 362, + "101": 363, + "3521": 364, + "2256": 365, + "3083": 366, + "4278": 367, + "8713": 368, + "1226": 369, + "4222": 370, + "8494": 371, + "8776": 372, + "731": 373, + "6574": 374, + "5319": 375, + "8605": 376, + "5583": 377, + "6406": 378, + "4064": 379, + "4806": 380, + "3972": 381, + "7383": 382, + "5133": 383, + "597": 384, + "1025": 385, + "7313": 386, + "5304": 387, + "8758": 388, + "1050": 389, + "6499": 390, + "6956": 391, + "770": 392, + "4108": 393, + "2774": 394, + "3864": 395, + "4490": 396, + "4848": 397, + "1826": 398, + "6294": 399, + "7949": 400, + "1446": 401, + "7867": 402, + "8163": 403, + "953": 404, + "8138": 405, + "353": 406, + "7553": 407, + "8825": 408, + "5189": 409, + "2012": 410, + "948": 411, + "205": 412, + "1535": 413, + "8008": 414, + "1112": 415, + "7926": 416, + "4039": 417, + "716": 418, + "3967": 419, + "7932": 420, + "7525": 421, + "7316": 422, + "2393": 423, + "6788": 424, + "6550": 425, + "7011": 426, + "8791": 427, + "8119": 428, + "1777": 429, + "6014": 430, + "1046": 431, + "6269": 432, + "6188": 433, + "5266": 434, + "3490": 435, + "8786": 436, + "8824": 437, + "3448": 438, + "589": 439, + "576": 440, + "1121": 441, + "1806": 442, + "7294": 443, + "3119": 444, + "2688": 445, + "3446": 446, + "1012": 447, + "4807": 448, + "7498": 449, + "3905": 450, + "7384": 451, + "2992": 452, + "30": 453, + "497": 454, + "227": 455, + "4226": 456, + "5007": 457, + "1066": 458, + "8222": 459, + "7688": 460, + "6865": 461, + "6286": 462, + "8225": 463, + "3224": 464, + "8635": 465, + "1348": 466, + "3645": 467, + "1961": 468, + "8190": 469, + "6032": 470, + "7286": 471, + "5389": 472, + "3105": 473, + "1028": 474, + "6038": 475, + "764": 476, + "7437": 477, + "6555": 478, + "8875": 479, + "2074": 480, + "7809": 481, + "2240": 482, + "2827": 483, + "5386": 484, + "6763": 485, + "3009": 486, + "6339": 487, + "1825": 488, + "7569": 489, + "359": 490, + "7956": 491, + "2137": 492, + "8677": 493, + "4434": 494, + "329": 495, + "3289": 496, + "4290": 497, + "2999": 498, + "2427": 499, + "637": 500, + "2229": 501, + "1874": 502, + "9023": 503, + "3114": 504, + "6235": 505, + "4860": 506, + "4519": 507, + "561": 508, + "70": 509, + "4800": 510, + "2294": 511, + "6115": 512, + "2582": 513, + "8464": 514, + "5139": 515, + "6918": 516, + "337": 517, + "5810": 518, + "8401": 519, + "303": 520, + "5206": 521, + "2589": 522, + "7061": 523, + "2269": 524, + "2758": 525, + "3389": 526, + "4629": 527, + "707": 528, + "5606": 529, + "1513": 530, + "2473": 531, + "664": 532, + "5092": 533, + "5154": 534, + "6288": 535, + "6308": 536, + "4731": 537, + "3328": 538, + "7816": 539, + "3221": 540, + "8687": 541, + "7030": 542, + "476": 543, + "5918": 544, + "6317": 545, + "204": 546, + "8006": 547, + "6895": 548, + "1264": 549, + "112": 550, + "1859": 551, + "398": 552, + "1052": 553, + "3294": 554, + "1460": 555, + "8573": 556, + "5684": 557, + "8421": 558, + "5883": 559, + "7297": 560, + "246": 561, + "8057": 562, + "3835": 563, + "1748": 564, + "3816": 565, + "3357": 566, + "1053": 567, + "409": 568, + "868": 569, + "3118": 570, + "7520": 571, + "2494": 572, + "6686": 573, + "1241": 574, + "5190": 575, + "166": 576, + "4257": 577, + "1482": 578, + "5604": 579, + "1212": 580, + "2741": 581, + "1259": 582, + "984": 583, + "6492": 584, + "6167": 585, + "296": 586, + "6567": 587, + "6924": 588, + "2272": 589, + "7085": 590, + "2388": 591, + "1705": 592, + "1343": 593, + "7241": 594, + "451": 595, + "345": 596, + "5401": 597, + "6446": 598, + "612": 599, + "594": 600, + "7555": 601, + "7069": 602, + "2577": 603, + "5333": 604, + "8742": 605, + "6727": 606, + "1571": 607, + "4734": 608, + "7258": 609, + "3977": 610, + "373": 611, + "5723": 612, + "1365": 613, + "7285": 614, + "580": 615, + "836": 616, + "6782": 617, + "3654": 618, + "1974": 619, + "6258": 620, + "925": 621, + "949": 622, + "2790": 623, + "698": 624, + "6373": 625, + "2785": 626, + "1222": 627, + "2751": 628, + "3825": 629, + "5115": 630, + "1827": 631, + "3171": 632, + "119": 633, + "850": 634, + "3258": 635, + "7909": 636, + "1322": 637, + "8097": 638, + "22": 639, + "7478": 640, + "1349": 641, + "4854": 642, + "2929": 643, + "7335": 644, + "5868": 645, + "454": 646, + "7945": 647, + "2654": 648, + "3493": 649, + "1060": 650, + "8545": 651, + "6509": 652, + "5002": 653, + "7732": 654, + "3082": 655, + "1779": 656, + "2709": 657, + "7398": 658, + "8879": 659, + "639": 660, + "598": 661, + "5672": 662, + "6553": 663, + "4111": 664, + "1417": 665, + "7991": 666, + "380": 667, + "8459": 668, + "8347": 669, + "1769": 670, + "2673": 671, + "3330": 672, + "7051": 673, + "1337": 674, + "4057": 675, + "4839": 676, + "6060": 677, + "7095": 678, + "278": 679, + "1445": 680, + "6518": 681, + "2364": 682, + "1958": 683, + "548": 684, + "4010": 685, + "3072": 686, + "6993": 687, + "8575": 688, + "2149": 689, + "240": 690, + "2920": 691, + "5588": 692, + "1885": 693, + "6082": 694, + "9026": 695, + "340": 696, + "159": 697, + "7730": 698, + "7962": 699, + "1987": 700, + "3876": 701, + "8771": 702, + "5123": 703, + "3866": 704, + "3546": 705, + "7777": 706, + "115": 707, + "5337": 708, + "475": 709, + "1724": 710, + "6359": 711, + "4260": 712, + "2110": 713, + "1845": 714, + "4335": 715, + "4133": 716, + "783": 717, + "8479": 718, + "1448": 719, + "1160": 720, + "7647": 721, + "2618": 722, + "3630": 723, + "4013": 724, + "5242": 725, + "7957": 726, + "3852": 727, + "3889": 728, + "1387": 729, + "439": 730, + "1425": 731, + "2061": 732, + "7395": 733, + "7837": 734, + "5147": 735, + "2319": 736, + "3781": 737, + "1311": 738, + "4733": 739, + "8705": 740, + "3094": 741, + "2823": 742, + "1914": 743, + "954": 744, + "4381": 745, + "4044": 746, + "593": 747, + "8300": 748, + "7558": 749, + "6494": 750, + "6330": 751, + "5940": 752, + "7126": 753, + "1061": 754, + "6352": 755, + "5186": 756, + "1944": 757, + "2285": 758, + "6673": 759, + "5746": 760, + "208": 761, + "492": 762, + "216": 763, + "979": 764, + "1668": 765, + "6620": 766, + "711": 767, + "7733": 768, + "8619": 769, + "5157": 770, + "829": 771, + "3180": 772, + "3979": 773, + "1556": 774, + "3379": 775, + "5727": 776, + "596": 777, + "2127": 778, + "581": 779, + "2652": 780, + "2628": 781, + "1849": 782, + "4238": 783, + "606": 784, + "1224": 785, + "1629": 786, + "1413": 787, + "957": 788, + "8592": 789, + "2254": 790, + "1323": 791, + "122": 792, + "2093": 793, + "1100": 794, + "81": 795, + "323": 796, + "815": 797, + "2581": 798, + "543": 799, + "6037": 800, + "2397": 801, + "5513": 802, + "4495": 803, + "5776": 804, + "17": 805, + "4590": 806, + "8228": 807, + "708": 808, + "3792": 809, + "3790": 810, + "7090": 811, + "1943": 812, + "4246": 813, + "559": 814, + "3738": 815, + "2167": 816, + "1933": 817, + "2162": 818, + "549": 819, + "3025": 820, + "1182": 821, + "4358": 822, + "636": 823, + "986": 824, + "8490": 825, + "3340": 826, + "90": 827, + "1487": 828, + "1639": 829, + "1547": 830, + "4152": 831, + "1498": 832, + "1740": 833, + "6157": 834, + "217": 835, + "2201": 836, + "362": 837, + "2146": 838, + "1801": 839, + "5063": 840, + "7339": 841, + "663": 842, + "38": 843, + "1336": 844, + "3215": 845, + "210": 846, + "6075": 847, + "55": 848, + "2411": 849, + "7445": 850, + "5767": 851, + "2812": 852, + "472": 853, + "803": 854, + "4236": 855, + "7665": 856, + "1607": 857, + "1316": 858, + "7475": 859, + "3001": 860, + "1473": 861, + "3537": 862, + "3070": 863, + "1390": 864, + "1290": 865, + "2499": 866, + "154": 867, + "7518": 868, + "408": 869, + "1811": 870, + "1734": 871, + "7342": 872, + "8722": 873, + "1754": 874, + "7657": 875, + "583": 876, + "830": 877, + "6690": 878, + "1552": 879, + "2498": 880, + "1296": 881, + "3686": 882, + "157": 883, + "487": 884, + "6119": 885, + "4926": 886, + "4846": 887, + "1536": 888, + "2674": 889, + "1645": 890, + "3187": 891, + "1058": 892, + "2039": 893, + "4071": 894, + "4433": 895, + "1175": 896, + "434": 897, + "1001": 898, + "2816": 899, + "820": 900, + "2696": 901, + "4681": 902, + "2085": 903 + }, + "piper_version": "1.0.0" +} diff --git a/requirements.txt b/requirements.txt index fc244b8..d5ff5ff 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,4 +1,4 @@ -audiomentations -espeak-phonemizer -numpy +audiomentations==0.33.0 +piper-phonemize==1.1.0 +numpy<2 torch