| { | |
| "version": "1.0", | |
| "truncation": null, | |
| "padding": null, | |
| "added_tokens": [ | |
| { | |
| "id": 0, | |
| "content": "[UNK]", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 1, | |
| "content": "[PAD]", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 2, | |
| "content": "[CLS]", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 3, | |
| "content": "[SEP]", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 4, | |
| "content": "[MASK]", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| } | |
| ], | |
| "normalizer": null, | |
| "pre_tokenizer": { | |
| "type": "Sequence", | |
| "pretokenizers": [ | |
| { | |
| "type": "Whitespace" | |
| }, | |
| { | |
| "type": "Punctuation", | |
| "behavior": "Isolated" | |
| } | |
| ] | |
| }, | |
| "post_processor": null, | |
| "decoder": null, | |
| "model": { | |
| "type": "BPE", | |
| "dropout": null, | |
| "unk_token": null, | |
| "continuing_subword_prefix": null, | |
| "end_of_word_suffix": null, | |
| "fuse_unk": false, | |
| "byte_fallback": false, | |
| "ignore_merges": false, | |
| "vocab": { | |
| "[UNK]": 0, | |
| "[PAD]": 1, | |
| "[CLS]": 2, | |
| "[SEP]": 3, | |
| "[MASK]": 4, | |
| "!": 5, | |
| "'": 6, | |
| ",": 7, | |
| "-": 8, | |
| ".": 9, | |
| "0": 10, | |
| "1": 11, | |
| "2": 12, | |
| "?": 13, | |
| "A": 14, | |
| "B": 15, | |
| "C": 16, | |
| "D": 17, | |
| "E": 18, | |
| "F": 19, | |
| "G": 20, | |
| "H": 21, | |
| "I": 22, | |
| "J": 23, | |
| "K": 24, | |
| "L": 25, | |
| "M": 26, | |
| "N": 27, | |
| "O": 28, | |
| "P": 29, | |
| "R": 30, | |
| "S": 31, | |
| "T": 32, | |
| "U": 33, | |
| "V": 34, | |
| "W": 35, | |
| "Y": 36, | |
| "Z": 37, | |
| "H0": 38, | |
| "AH0": 39, | |
| "H1": 40, | |
| "Y1": 41, | |
| "W1": 42, | |
| "IH0": 43, | |
| "AE": 44, | |
| "IH1": 45, | |
| "AH0N": 46, | |
| "ER": 47, | |
| "EH1": 48, | |
| "AE1": 49, | |
| "AA": 50, | |
| "AH1": 51, | |
| "AA1": 52, | |
| "ER0": 53, | |
| "Y0": 54, | |
| "IY0": 55, | |
| "UW1": 56, | |
| "DH": 57, | |
| "AO": 58, | |
| "AY1": 59, | |
| "IY1": 60, | |
| "EY1": 61, | |
| "AH0L": 62, | |
| "AO1": 63, | |
| "IH0N": 64, | |
| "OW1": 65, | |
| "ST": 66, | |
| "HH": 67, | |
| "SH": 68, | |
| "AA1R": 69, | |
| "DHAH0": 70, | |
| "AO1R": 71, | |
| "AE1N": 72, | |
| "IH1Z": 73, | |
| "JH": 74, | |
| "AH1V": 75, | |
| "TS": 76, | |
| "H2": 77, | |
| "Y2": 78, | |
| "AH0ND": 79, | |
| "TUW1": 80, | |
| "IH0NG": 81, | |
| "ER1": 82, | |
| "EH1N": 83, | |
| "SHAH0N": 84, | |
| "CH": 85, | |
| "AH0T": 86, | |
| "PR": 87, | |
| "IH0K": 88, | |
| "AH0M": 89, | |
| "EH1R": 90, | |
| "YUW1": 91, | |
| "AH1M": 92, | |
| "TH": 93, | |
| "HHAE1": 94, | |
| "AH0S": 95, | |
| "AW1": 96, | |
| "AH0NT": 97, | |
| "W0": 98, | |
| "IH1N": 99, | |
| "ER0Z": 100, | |
| "AH0D": 101, | |
| "AH0Z": 102, | |
| "HHAE1V": 103, | |
| "EH1L": 104, | |
| "RIY0": 105, | |
| "AE1T": 106, | |
| "KT": 107, | |
| "EH2": 108, | |
| "KS": 109, | |
| "AA1N": 110, | |
| "FAO1R": 111, | |
| "AH1N": 112, | |
| "LIY0": 113, | |
| "EY1SHAH0N": 114, | |
| "MOW1": 115, | |
| "UH1": 116, | |
| "EY2": 117, | |
| "DZ": 118, | |
| "AY2": 119, | |
| "SAH1M": 120, | |
| "IH2": 121, | |
| "AO1L": 122, | |
| "DIH0": 123, | |
| "WIH1": 124, | |
| "PL": 125, | |
| "AH0K": 126, | |
| "W2": 127, | |
| "OW0": 128, | |
| "TER0": 129, | |
| "GR": 130, | |
| "LAY1": 131, | |
| "TR": 132, | |
| "SP": 133, | |
| "AH0B": 134, | |
| "AH0LZ": 135, | |
| "MOW1ST": 136, | |
| "EY1N": 137, | |
| "KAE1N": 138, | |
| "FR": 139, | |
| "IY0Z": 140, | |
| "MAH0N": 141, | |
| "IH1L": 142, | |
| "DHAE1T": 143, | |
| "AE2": 144, | |
| "AH0NS": 145, | |
| "ND": 146, | |
| "IH1T": 147, | |
| "AA1RT": 148, | |
| "KAA1": 149, | |
| "KR": 150, | |
| "OW1N": 151, | |
| "AA2": 152, | |
| "IH0V": 153, | |
| "AH0P": 154, | |
| "OW2": 155, | |
| "UW0": 156, | |
| "AA1L": 157, | |
| "AH0TIY0": 158, | |
| "IH0Z": 159, | |
| "STR": 160, | |
| "BIY1": 161, | |
| "SK": 162, | |
| "IH1NG": 163, | |
| "EY2T": 164, | |
| "AH0JH": 165, | |
| "PAA1RT": 166, | |
| "NZ": 167, | |
| "EH1K": 168, | |
| "DHEH1R": 169, | |
| "BR": 170, | |
| "IY1Z": 171, | |
| "IY1T": 172, | |
| "AE1Z": 173, | |
| "AH0NZ": 174, | |
| "VER0": 175, | |
| "WIH1DH": 176, | |
| "AH1T": 177, | |
| "BAY1": 178, | |
| "EH1T": 179, | |
| "DR": 180, | |
| "PAH0L": 181, | |
| "OW1L": 182, | |
| "YUW1Z": 183, | |
| "WAY1": 184, | |
| "AH0F": 185, | |
| "IY2": 186, | |
| "AE0": 187, | |
| "EH1S": 188, | |
| "EY1T": 189, | |
| "EH0": 190, | |
| "EH1M": 191, | |
| "AO2": 192, | |
| "FL": 193, | |
| "MEH1N": 194, | |
| "AE1K": 195, | |
| "BL": 196, | |
| "EH1D": 197, | |
| "PER0": 198, | |
| "IH1R": 199, | |
| "EH1KT": 200, | |
| "OY1": 201, | |
| "MZ": 202, | |
| "IH1F": 203, | |
| "EH1ST": 204, | |
| "MEY1": 205, | |
| "WIY1": 206, | |
| "DHER0": 207, | |
| "CHER0": 208, | |
| "LZ": 209, | |
| "AY0": 210, | |
| "KL": 211, | |
| "DIY0": 212, | |
| "GZ": 213, | |
| "IH1ST": 214, | |
| "VAY1": 215, | |
| "EH1RIY0": 216, | |
| "UH1R": 217, | |
| "DER0": 218, | |
| "AE1L": 219, | |
| "AW1T": 220, | |
| "FRAH1M": 221, | |
| "AH1L": 222, | |
| "MP": 223, | |
| "NOW1": 224, | |
| "AH0ST": 225, | |
| "TAY1": 226, | |
| "PIY1": 227, | |
| "AH0LIY0": 228, | |
| "AE1S": 229, | |
| "WER1": 230, | |
| "IH1S": 231, | |
| "PS": 232, | |
| "IH0T": 233, | |
| "IH2N": 234, | |
| "PRAA1": 235, | |
| "IH0L": 236, | |
| "AA1T": 237, | |
| "SAY1": 238, | |
| "KAH0M": 239, | |
| "MIY1": 240, | |
| "WAH1N": 241, | |
| "WAO1": 242, | |
| "AH0NTS": 243, | |
| "PLAE1N": 244, | |
| "SIY1": 245, | |
| "SHAH0NZ": 246, | |
| "UH1D": 247, | |
| "PIY1PAH0L": 248, | |
| "MAY1": 249, | |
| "VZ": 250, | |
| "MEH1NIY0": 251, | |
| "AE1M": 252, | |
| "MAH0NT": 253, | |
| "UW2": 254, | |
| "AH0W": 255, | |
| "HHEH1L": 256, | |
| "YAH0L": 257, | |
| "IH1RIY0": 258, | |
| "WAO1TER0": 259, | |
| "LOW1": 260, | |
| "KAH0NT": 261, | |
| "AH1ST": 262, | |
| "FIY1": 263, | |
| "AH0MZ": 264, | |
| "MAO1R": 265, | |
| "EY1S": 266, | |
| "AE1KT": 267, | |
| "AH2": 268, | |
| "DUW1": 269, | |
| "AH0BAH0L": 270, | |
| "KAA1Z": 271, | |
| "SEH1L": 272, | |
| "AO1RT": 273, | |
| "YUW0": 274, | |
| "LIY1": 275, | |
| "THR": 276, | |
| "EH1ND": 277, | |
| "UW1D": 278, | |
| "BAH1T": 279, | |
| "SER1": 280, | |
| "NT": 281, | |
| "SOW1": 282, | |
| "FAY1": 283, | |
| "AO1LS": 284, | |
| "AH0G": 285, | |
| "NG": 286, | |
| "DHEY1": 287, | |
| "AH1DHER0": 288, | |
| "AO1LSOW0": 289, | |
| "HHAY1": 290, | |
| "LIH1": 291, | |
| "AH1NG": 292, | |
| "EH1NT": 293, | |
| "WEY1": 294, | |
| "AH0V": 295, | |
| "PRAH0D": 296, | |
| "AW1ER0": 297, | |
| "IH0KAH0L": 298, | |
| "SHAH0L": 299, | |
| "AE1NAH0M": 300, | |
| "DHAE1N": 301, | |
| "HHAE1Z": 302, | |
| "EH1V": 303, | |
| "NAH0S": 304, | |
| "AY1T": 305, | |
| "UW1S": 306, | |
| "RIH1": 307, | |
| "KAE1": 308, | |
| "LD": 309, | |
| "VEH1RIY0": 310, | |
| "RIY1": 311, | |
| "OW1Z": 312, | |
| "SER0": 313, | |
| "IH0KS": 314, | |
| "LAY1K": 315, | |
| "AY2Z": 316, | |
| "WEH1N": 317, | |
| "PT": 318, | |
| "KAH0N": 319, | |
| "NUW1": 320, | |
| "AW1ND": 321, | |
| "RIH0": 322, | |
| "FER0": 323, | |
| "AO1NG": 324, | |
| "AE1ND": 325, | |
| "EY1NJH": 326, | |
| "AH0TS": 327, | |
| "EY1SHAH0NZ": 328, | |
| "WAH1T": 329, | |
| "SEH2": 330, | |
| "DIH0Z": 331, | |
| "PLAE1NTS": 332, | |
| "ER1N": 333, | |
| "THIH1NG": 334, | |
| "0R": 335, | |
| "SIH1": 336, | |
| "GROW1": 337, | |
| "EH1NER0": 338, | |
| "AO2R": 339, | |
| "KW": 340, | |
| "WIH1CH": 341, | |
| "AA1KS": 342, | |
| "SAH1": 343, | |
| "IH2Z": 344, | |
| "AO1F": 345, | |
| "EY2TAH0D": 346, | |
| "MAH1": 347, | |
| "WEH1L": 348, | |
| "AE1NAH0MAH0LZ": 349, | |
| "IY1V": 350, | |
| "EH1VER0": 351, | |
| "HHYUW1": 352, | |
| "TEY1": 353, | |
| "IH0MP": 354, | |
| "SEY1": 355, | |
| "JHAH1ST": 356, | |
| "AY1M": 357, | |
| "AH1P": 358, | |
| "KAH1M": 359, | |
| "EH2RIY0": 360, | |
| "IH1SHAH0N": 361, | |
| "IY1N": 362, | |
| "OW1T": 363, | |
| "BAA1": 364, | |
| "DIH1F": 365, | |
| "IH1K": 366, | |
| "BIH0K": 367, | |
| "BIH0": 368, | |
| "AW1N": 369, | |
| "SM": 370, | |
| "EY1Z": 371, | |
| "MAH0T": 372, | |
| "IH0D": 373, | |
| "KIH0NG": 374, | |
| "WAA1": 375, | |
| "TAY1M": 376, | |
| "HHYUW1MAH0N": 377, | |
| "BER1": 378, | |
| "HHIY1": 379, | |
| "DHIH1S": 380, | |
| "AE1NG": 381, | |
| "IH1V": 382, | |
| "FUW1D": 383 | |
| }, | |
| "merges": [ | |
| "H 0", | |
| "A H0", | |
| "H 1", | |
| "Y 1", | |
| "W 1", | |
| "I H0", | |
| "A E", | |
| "I H1", | |
| "AH0 N", | |
| "E R", | |
| "E H1", | |
| "AE 1", | |
| "A A", | |
| "A H1", | |
| "AA 1", | |
| "ER 0", | |
| "Y 0", | |
| "I Y0", | |
| "U W1", | |
| "D H", | |
| "A O", | |
| "A Y1", | |
| "I Y1", | |
| "E Y1", | |
| "AH0 L", | |
| "AO 1", | |
| "IH0 N", | |
| "O W1", | |
| "S T", | |
| "H H", | |
| "S H", | |
| "AA1 R", | |
| "DH AH0", | |
| "AO1 R", | |
| "AE1 N", | |
| "IH1 Z", | |
| "J H", | |
| "AH1 V", | |
| "T S", | |
| "H 2", | |
| "Y 2", | |
| "AH0N D", | |
| "T UW1", | |
| "IH0N G", | |
| "ER 1", | |
| "EH1 N", | |
| "SH AH0N", | |
| "C H", | |
| "AH0 T", | |
| "P R", | |
| "IH0 K", | |
| "AH0 M", | |
| "EH1 R", | |
| "Y UW1", | |
| "AH1 M", | |
| "T H", | |
| "HH AE1", | |
| "AH0 S", | |
| "A W1", | |
| "AH0N T", | |
| "W 0", | |
| "IH1 N", | |
| "ER0 Z", | |
| "AH0 D", | |
| "AH0 Z", | |
| "HHAE1 V", | |
| "EH1 L", | |
| "R IY0", | |
| "AE1 T", | |
| "K T", | |
| "E H2", | |
| "K S", | |
| "AA1 N", | |
| "F AO1R", | |
| "AH1 N", | |
| "L IY0", | |
| "EY1 SHAH0N", | |
| "M OW1", | |
| "U H1", | |
| "E Y2", | |
| "D Z", | |
| "A Y2", | |
| "S AH1M", | |
| "I H2", | |
| "AO1 L", | |
| "D IH0", | |
| "W IH1", | |
| "P L", | |
| "AH0 K", | |
| "W 2", | |
| "O W0", | |
| "T ER0", | |
| "G R", | |
| "L AY1", | |
| "T R", | |
| "S P", | |
| "AH0 B", | |
| "AH0L Z", | |
| "MOW1 ST", | |
| "EY1 N", | |
| "K AE1N", | |
| "F R", | |
| "IY0 Z", | |
| "M AH0N", | |
| "IH1 L", | |
| "DH AE1T", | |
| "AE 2", | |
| "AH0N S", | |
| "N D", | |
| "IH1 T", | |
| "AA1R T", | |
| "K AA1", | |
| "K R", | |
| "OW1 N", | |
| "AA 2", | |
| "IH0 V", | |
| "AH0 P", | |
| "O W2", | |
| "U W0", | |
| "AA1 L", | |
| "AH0T IY0", | |
| "IH0 Z", | |
| "ST R", | |
| "B IY1", | |
| "S K", | |
| "IH1N G", | |
| "EY2 T", | |
| "AH0 JH", | |
| "P AA1RT", | |
| "N Z", | |
| "EH1 K", | |
| "DH EH1R", | |
| "B R", | |
| "IY1 Z", | |
| "IY1 T", | |
| "AE1 Z", | |
| "AH0N Z", | |
| "V ER0", | |
| "WIH1 DH", | |
| "AH1 T", | |
| "B AY1", | |
| "EH1 T", | |
| "D R", | |
| "P AH0L", | |
| "OW1 L", | |
| "YUW1 Z", | |
| "W AY1", | |
| "AH0 F", | |
| "I Y2", | |
| "AE 0", | |
| "EH1 S", | |
| "EY1 T", | |
| "E H0", | |
| "EH1 M", | |
| "AO 2", | |
| "F L", | |
| "M EH1N", | |
| "AE1 K", | |
| "B L", | |
| "EH1 D", | |
| "P ER0", | |
| "IH1 R", | |
| "EH1 KT", | |
| "O Y1", | |
| "M Z", | |
| "IH1 F", | |
| "EH1 ST", | |
| "M EY1", | |
| "W IY1", | |
| "DH ER0", | |
| "CH ER0", | |
| "L Z", | |
| "A Y0", | |
| "K L", | |
| "D IY0", | |
| "G Z", | |
| "IH1 ST", | |
| "V AY1", | |
| "EH1R IY0", | |
| "UH1 R", | |
| "D ER0", | |
| "AE1 L", | |
| "AW1 T", | |
| "FR AH1M", | |
| "AH1 L", | |
| "M P", | |
| "N OW1", | |
| "AH0 ST", | |
| "T AY1", | |
| "P IY1", | |
| "AH0L IY0", | |
| "AE1 S", | |
| "W ER1", | |
| "IH1 S", | |
| "P S", | |
| "IH0 T", | |
| "IH2 N", | |
| "PR AA1", | |
| "IH0 L", | |
| "AA1 T", | |
| "S AY1", | |
| "K AH0M", | |
| "M IY1", | |
| "W AH1N", | |
| "W AO1", | |
| "AH0N TS", | |
| "PL AE1N", | |
| "S IY1", | |
| "SHAH0N Z", | |
| "UH1 D", | |
| "PIY1 PAH0L", | |
| "M AY1", | |
| "V Z", | |
| "MEH1N IY0", | |
| "AE1 M", | |
| "M AH0NT", | |
| "U W2", | |
| "AH0 W", | |
| "HH EH1L", | |
| "Y AH0L", | |
| "IH1 RIY0", | |
| "WAO1 TER0", | |
| "L OW1", | |
| "K AH0NT", | |
| "AH1 ST", | |
| "F IY1", | |
| "AH0M Z", | |
| "M AO1R", | |
| "EY1 S", | |
| "AE1 KT", | |
| "A H2", | |
| "D UW1", | |
| "AH0B AH0L", | |
| "KAA1 Z", | |
| "S EH1L", | |
| "AO1R T", | |
| "Y UW0", | |
| "L IY1", | |
| "TH R", | |
| "EH1N D", | |
| "UW1 D", | |
| "B AH1T", | |
| "S ER1", | |
| "N T", | |
| "S OW1", | |
| "F AY1", | |
| "AO1L S", | |
| "AH0 G", | |
| "N G", | |
| "DH EY1", | |
| "AH1 DHER0", | |
| "AO1LS OW0", | |
| "HH AY1", | |
| "L IH1", | |
| "AH1N G", | |
| "EH1N T", | |
| "W EY1", | |
| "AH0 V", | |
| "PR AH0D", | |
| "AW1 ER0", | |
| "IH0K AH0L", | |
| "SH AH0L", | |
| "AE1N AH0M", | |
| "DH AE1N", | |
| "HHAE1 Z", | |
| "EH1 V", | |
| "N AH0S", | |
| "AY1 T", | |
| "UW1 S", | |
| "R IH1", | |
| "K AE1", | |
| "L D", | |
| "V EH1RIY0", | |
| "R IY1", | |
| "OW1 Z", | |
| "S ER0", | |
| "IH0K S", | |
| "LAY1 K", | |
| "AY2 Z", | |
| "W EH1N", | |
| "P T", | |
| "K AH0N", | |
| "N UW1", | |
| "AW1 ND", | |
| "R IH0", | |
| "F ER0", | |
| "AO1 NG", | |
| "AE1N D", | |
| "EY1N JH", | |
| "AH0 TS", | |
| "EY1SHAH0N Z", | |
| "W AH1T", | |
| "S EH2", | |
| "DIH0 Z", | |
| "PLAE1N TS", | |
| "ER1 N", | |
| "TH IH1NG", | |
| "0 R", | |
| "S IH1", | |
| "GR OW1", | |
| "EH1N ER0", | |
| "AO2 R", | |
| "K W", | |
| "WIH1 CH", | |
| "AA1 KS", | |
| "S AH1", | |
| "IH2 Z", | |
| "AO1 F", | |
| "EY2T AH0D", | |
| "M AH1", | |
| "W EH1L", | |
| "AE1NAH0M AH0LZ", | |
| "IY1 V", | |
| "EH1 VER0", | |
| "HH YUW1", | |
| "T EY1", | |
| "IH0 MP", | |
| "S EY1", | |
| "JH AH1ST", | |
| "AY1 M", | |
| "AH1 P", | |
| "K AH1M", | |
| "EH2 RIY0", | |
| "IH1 SHAH0N", | |
| "IY1 N", | |
| "OW1 T", | |
| "B AA1", | |
| "D IH1F", | |
| "IH1 K", | |
| "B IH0K", | |
| "B IH0", | |
| "AW1 N", | |
| "S M", | |
| "EY1 Z", | |
| "M AH0T", | |
| "IH0 D", | |
| "K IH0NG", | |
| "W AA1", | |
| "TAY1 M", | |
| "HHYUW1 MAH0N", | |
| "B ER1", | |
| "HH IY1", | |
| "DH IH1S", | |
| "AE1N G", | |
| "IH1 V", | |
| "F UW1D" | |
| ] | |
| } | |
| } |