Upload tokenizer

Browse files

Files changed (5) hide show

added_tokens.json +4 -0
special_tokens_map.json +51 -0
tokenizer.json +1079 -0
tokenizer_config.json +78 -0
vocab.txt +911 -0

added_tokens.json ADDED Viewed

	@@ -0,0 +1,4 @@

+{
+  "</s>": 912,
+  "<s>": 911
+}

special_tokens_map.json ADDED Viewed

	@@ -0,0 +1,51 @@

+{
+  "bos_token": {
+    "content": "<s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "cls_token": {
+    "content": "[CLS]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "eos_token": {
+    "content": "</s>",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "mask_token": {
+    "content": "[MASK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "pad_token": {
+    "content": "[PAD]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "sep_token": {
+    "content": "[SEP]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  },
+  "unk_token": {
+    "content": "[UNK]",
+    "lstrip": false,
+    "normalized": false,
+    "rstrip": false,
+    "single_word": false
+  }
+}

tokenizer.json ADDED Viewed

	@@ -0,0 +1,1079 @@

+{
+  "version": "1.0",
+  "truncation": null,
+  "padding": null,
+  "added_tokens": [
+    {
+      "id": 0,
+      "content": "[MASK]",
+      "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": "[UNK]",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 911,
+      "content": "<s>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    },
+    {
+      "id": 912,
+      "content": "</s>",
+      "single_word": false,
+      "lstrip": false,
+      "rstrip": false,
+      "normalized": false,
+      "special": true
+    }
+  ],
+  "normalizer": {
+    "type": "BertNormalizer",
+    "clean_text": true,
+    "handle_chinese_chars": true,
+    "strip_accents": null,
+    "lowercase": false
+  },
+  "pre_tokenizer": {
+    "type": "BertPreTokenizer"
+  },
+  "post_processor": {
+    "type": "TemplateProcessing",
+    "single": [
+      {
+        "SpecialToken": {
+          "id": "[CLS]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 0
+        }
+      }
+    ],
+    "pair": [
+      {
+        "SpecialToken": {
+          "id": "[CLS]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "A",
+          "type_id": 0
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 0
+        }
+      },
+      {
+        "Sequence": {
+          "id": "B",
+          "type_id": 1
+        }
+      },
+      {
+        "SpecialToken": {
+          "id": "[SEP]",
+          "type_id": 1
+        }
+      }
+    ],
+    "special_tokens": {
+      "[CLS]": {
+        "id": "[CLS]",
+        "ids": [
+          2
+        ],
+        "tokens": [
+          "[CLS]"
+        ]
+      },
+      "[SEP]": {
+        "id": "[SEP]",
+        "ids": [
+          3
+        ],
+        "tokens": [
+          "[SEP]"
+        ]
+      }
+    }
+  },
+  "decoder": {
+    "type": "WordPiece",
+    "prefix": "##",
+    "cleanup": true
+  },
+  "model": {
+    "type": "WordPiece",
+    "unk_token": "[UNK]",
+    "continuing_subword_prefix": "##",
+    "max_input_chars_per_word": 100,
+    "vocab": {
+      "[MASK]": 0,
+      "[PAD]": 1,
+      "[CLS]": 2,
+      "[SEP]": 3,
+      "[UNK]": 4,
+      "G31": 5,
+      "F02": 6,
+      "E78": 7,
+      "G20": 8,
+      "K31": 9,
+      "K21": 10,
+      "K44": 11,
+      "F41": 12,
+      "I34": 13,
+      "M81": 14,
+      "G43": 15,
+      "I13": 16,
+      "I50": 17,
+      "E87": 18,
+      "E11": 19,
+      "I25": 20,
+      "D64": 21,
+      "G47": 22,
+      "M1A": 23,
+      "E66": 24,
+      "D50": 25,
+      "K29": 26,
+      "K25": 27,
+      "K57": 28,
+      "I12": 29,
+      "I70": 30,
+      "K22": 31,
+      "D63": 32,
+      "C67": 33,
+      "I10": 34,
+      "D25": 35,
+      "E89": 36,
+      "I26": 37,
+      "I82": 38,
+      "I47": 39,
+      "I87": 40,
+      "B95": 41,
+      "E03": 42,
+      "E83": 43,
+      "K59": 44,
+      "K65": 45,
+      "B96": 46,
+      "F17": 47,
+      "D72": 48,
+      "A41": 49,
+      "D62": 50,
+      "I27": 51,
+      "K43": 52,
+      "E43": 53,
+      "E53": 54,
+      "I48": 55,
+      "I11": 56,
+      "I45": 57,
+      "J44": 58,
+      "G56": 59,
+      "M19": 60,
+      "J45": 61,
+      "I73": 62,
+      "K72": 63,
+      "J96": 64,
+      "I44": 65,
+      "D69": 66,
+      "I46": 67,
+      "M47": 68,
+      "J01": 69,
+      "M79": 70,
+      "K92": 71,
+      "H40": 72,
+      "H69": 73,
+      "I24": 74,
+      "I49": 75,
+      "C16": 76,
+      "C78": 77,
+      "K91": 78,
+      "M86": 79,
+      "L97": 80,
+      "F32": 81,
+      "G25": 82,
+      "L03": 83,
+      "J98": 84,
+      "I96": 85,
+      "G89": 86,
+      "M25": 87,
+      "E46": 88,
+      "G30": 89,
+      "H91": 90,
+      "M17": 91,
+      "D68": 92,
+      "M46": 93,
+      "M70": 94,
+      "E55": 95,
+      "C50": 96,
+      "I83": 97,
+      "L76": 98,
+      "K76": 99,
+      "I67": 100,
+      "M51": 101,
+      "M48": 102,
+      "M35": 103,
+      "C79": 104,
+      "G93": 105,
+      "G91": 106,
+      "C34": 107,
+      "F05": 108,
+      "G51": 109,
+      "M21": 110,
+      "H53": 111,
+      "H04": 112,
+      "H35": 113,
+      "M50": 114,
+      "E04": 115,
+      "M80": 116,
+      "F03": 117,
+      "I28": 118,
+      "J43": 119,
+      "I07": 120,
+      "C43": 121,
+      "I65": 122,
+      "J18": 123,
+      "J90": 124,
+      "K40": 125,
+      "I30": 126,
+      "I31": 127,
+      "M06": 128,
+      "I21": 129,
+      "A04": 130,
+      "F44": 131,
+      "I95": 132,
+      "E86": 133,
+      "D51": 134,
+      "F43": 135,
+      "F50": 136,
+      "K02": 137,
+      "L73": 138,
+      "K85": 139,
+      "F39": 140,
+      "M54": 141,
+      "B20": 142,
+      "B19": 143,
+      "F10": 144,
+      "E16": 145,
+      "F19": 146,
+      "F14": 147,
+      "F20": 148,
+      "F29": 149,
+      "F11": 150,
+      "E10": 151,
+      "D86": 152,
+      "J02": 153,
+      "K20": 154,
+      "E21": 155,
+      "E51": 156,
+      "I69": 157,
+      "I63": 158,
+      "G81": 159,
+      "M85": 160,
+      "D49": 161,
+      "K74": 162,
+      "K66": 163,
+      "K62": 164,
+      "L91": 165,
+      "D57": 166,
+      "C90": 167,
+      "J91": 168,
+      "K56": 169,
+      "C21": 170,
+      "B37": 171,
+      "E88": 172,
+      "I35": 173,
+      "I42": 174,
+      "K61": 175,
+      "K60": 176,
+      "G92": 177,
+      "D46": 178,
+      "J84": 179,
+      "M84": 180,
+      "D18": 181,
+      "I61": 182,
+      "G40": 183,
+      "L02": 184,
+      "F88": 185,
+      "I08": 186,
+      "G62": 187,
+      "F31": 188,
+      "E44": 189,
+      "I15": 190,
+      "K80": 191,
+      "I16": 192,
+      "J69": 193,
+      "J95": 194,
+      "H26": 195,
+      "L40": 196,
+      "L57": 197,
+      "M72": 198,
+      "J30": 199,
+      "H61": 200,
+      "C56": 201,
+      "C77": 202,
+      "I81": 203,
+      "I85": 204,
+      "I86": 205,
+      "K70": 206,
+      "G96": 207,
+      "F12": 208,
+      "M53": 209,
+      "C7B": 210,
+      "C7A": 211,
+      "K58": 212,
+      "K26": 213,
+      "J81": 214,
+      "B18": 215,
+      "A08": 216,
+      "M96": 217,
+      "I97": 218,
+      "C22": 219,
+      "C25": 220,
+      "K86": 221,
+      "K52": 222,
+      "K11": 223,
+      "B00": 224,
+      "A40": 225,
+      "K83": 226,
+      "C24": 227,
+      "D53": 228,
+      "C20": 229,
+      "H90": 230,
+      "D47": 231,
+      "L29": 232,
+      "K81": 233,
+      "K82": 234,
+      "D70": 235,
+      "M10": 236,
+      "F90": 237,
+      "D13": 238,
+      "I71": 239,
+      "F33": 240,
+      "K75": 241,
+      "F22": 242,
+      "F15": 243,
+      "F72": 244,
+      "F13": 245,
+      "J36": 246,
+      "L27": 247,
+      "D61": 248,
+      "C92": 249,
+      "E85": 250,
+      "K12": 251,
+      "L99": 252,
+      "L72": 253,
+      "G58": 254,
+      "F45": 255,
+      "H66": 256,
+      "J11": 257,
+      "F60": 258,
+      "L30": 259,
+      "J86": 260,
+      "J15": 261,
+      "A81": 262,
+      "M40": 263,
+      "L08": 264,
+      "L60": 265,
+      "L82": 266,
+      "M62": 267,
+      "H93": 268,
+      "I72": 269,
+      "F42": 270,
+      "K04": 271,
+      "E80": 272,
+      "L53": 273,
+      "L89": 274,
+      "L98": 275,
+      "H81": 276,
+      "D80": 277,
+      "B35": 278,
+      "C73": 279,
+      "E05": 280,
+      "J39": 281,
+      "I22": 282,
+      "I51": 283,
+      "C18": 284,
+      "C91": 285,
+      "C62": 286,
+      "K64": 287,
+      "K63": 288,
+      "E08": 289,
+      "F09": 290,
+      "K28": 291,
+      "E22": 292,
+      "F25": 293,
+      "M94": 294,
+      "M71": 295,
+      "J20": 296,
+      "K95": 297,
+      "M16": 298,
+      "L85": 299,
+      "B34": 300,
+      "J04": 301,
+      "J47": 302,
+      "M31": 303,
+      "M41": 304,
+      "L71": 305,
+      "M00": 306,
+      "M89": 307,
+      "J94": 308,
+      "C61": 309,
+      "C80": 310,
+      "J93": 311,
+      "K50": 312,
+      "B02": 313,
+      "J06": 314,
+      "I09": 315,
+      "D32": 316,
+      "I33": 317,
+      "L28": 318,
+      "F23": 319,
+      "G12": 320,
+      "E02": 321,
+      "I60": 322,
+      "L95": 323,
+      "H33": 324,
+      "K08": 325,
+      "D73": 326,
+      "I20": 327,
+      "D59": 328,
+      "E79": 329,
+      "J60": 330,
+      "C95": 331,
+      "C94": 332,
+      "F84": 333,
+      "L05": 334,
+      "M30": 335,
+      "J32": 336,
+      "D23": 337,
+      "K42": 338,
+      "D35": 339,
+      "J34": 340,
+      "E26": 341,
+      "I80": 342,
+      "K51": 343,
+      "I36": 344,
+      "K55": 345,
+      "E27": 346,
+      "D75": 347,
+      "D55": 348,
+      "J40": 349,
+      "G90": 350,
+      "D12": 351,
+      "E56": 352,
+      "F51": 353,
+      "G35": 354,
+      "M87": 355,
+      "E06": 356,
+      "I43": 357,
+      "G44": 358,
+      "G95": 359,
+      "C02": 360,
+      "G83": 361,
+      "F79": 362,
+      "I77": 363,
+      "E73": 364,
+      "H05": 365,
+      "H54": 366,
+      "H46": 367,
+      "M08": 368,
+      "M11": 369,
+      "H11": 370,
+      "F28": 371,
+      "M97": 372,
+      "C60": 373,
+      "M99": 374,
+      "M34": 375,
+      "F01": 376,
+      "I66": 377,
+      "H43": 378,
+      "H25": 379,
+      "D89": 380,
+      "G60": 381,
+      "I89": 382,
+      "M20": 383,
+      "E13": 384,
+      "L93": 385,
+      "E61": 386,
+      "A09": 387,
+      "G24": 388,
+      "G21": 389,
+      "M75": 390,
+      "M05": 391,
+      "D60": 392,
+      "B44": 393,
+      "B59": 394,
+      "A49": 395,
+      "A18": 396,
+      "M43": 397,
+      "M23": 398,
+      "K94": 399,
+      "C81": 400,
+      "K35": 401,
+      "F07": 402,
+      "G82": 403,
+      "E74": 404,
+      "C45": 405,
+      "E59": 406,
+      "E60": 407,
+      "A48": 408,
+      "G45": 409,
+      "F30": 410,
+      "H34": 411,
+      "C85": 412,
+      "M32": 413,
+      "I05": 414,
+      "D28": 415,
+      "M24": 416,
+      "E23": 417,
+      "C71": 418,
+      "F40": 419,
+      "D38": 420,
+      "D48": 421,
+      "C64": 422,
+      "H10": 423,
+      "D65": 424,
+      "B49": 425,
+      "H16": 426,
+      "I99": 427,
+      "L70": 428,
+      "C48": 429,
+      "D43": 430,
+      "E07": 431,
+      "L10": 432,
+      "C83": 433,
+      "J70": 434,
+      "I68": 435,
+      "J99": 436,
+      "H15": 437,
+      "B36": 438,
+      "D58": 439,
+      "L23": 440,
+      "I23": 441,
+      "J09": 442,
+      "B94": 443,
+      "I74": 444,
+      "G54": 445,
+      "B99": 446,
+      "G72": 447,
+      "G50": 448,
+      "G97": 449,
+      "D17": 450,
+      "E09": 451,
+      "L13": 452,
+      "G61": 453,
+      "A43": 454,
+      "H02": 455,
+      "H57": 456,
+      "H49": 457,
+      "I38": 458,
+      "G37": 459,
+      "F21": 460,
+      "D56": 461,
+      "K90": 462,
+      "J92": 463,
+      "J12": 464,
+      "B97": 465,
+      "J38": 466,
+      "L80": 467,
+      "H31": 468,
+      "A42": 469,
+      "M13": 470,
+      "M77": 471,
+      "M65": 472,
+      "L88": 473,
+      "C44": 474,
+      "G06": 475,
+      "L25": 476,
+      "D34": 477,
+      "K46": 478,
+      "M60": 479,
+      "I62": 480,
+      "D83": 481,
+      "F64": 482,
+      "D41": 483,
+      "E72": 484,
+      "B60": 485,
+      "J61": 486,
+      "J85": 487,
+      "C10": 488,
+      "A63": 489,
+      "C14": 490,
+      "C09": 491,
+      "F55": 492,
+      "F99": 493,
+      "M76": 494,
+      "A60": 495,
+      "C54": 496,
+      "F91": 497,
+      "F63": 498,
+      "J14": 499,
+      "F06": 500,
+      "C88": 501,
+      "H83": 502,
+      "J31": 503,
+      "I37": 504,
+      "K68": 505,
+      "B45": 506,
+      "G04": 507,
+      "J10": 508,
+      "B85": 509,
+      "C46": 510,
+      "B48": 511,
+      "C40": 512,
+      "J80": 513,
+      "H51": 514,
+      "G46": 515,
+      "M15": 516,
+      "K06": 517,
+      "A69": 518,
+      "J13": 519,
+      "I78": 520,
+      "J00": 521,
+      "M27": 522,
+      "G08": 523,
+      "H92": 524,
+      "E63": 525,
+      "B33": 526,
+      "C19": 527,
+      "E34": 528,
+      "L50": 529,
+      "I06": 530,
+      "K71": 531,
+      "K36": 532,
+      "K14": 533,
+      "H18": 534,
+      "E29": 535,
+      "E54": 536,
+      "D30": 537,
+      "E28": 538,
+      "K30": 539,
+      "B86": 540,
+      "G57": 541,
+      "H21": 542,
+      "H44": 543,
+      "H20": 544,
+      "C15": 545,
+      "G80": 546,
+      "D36": 547,
+      "F34": 548,
+      "D27": 549,
+      "C51": 550,
+      "L90": 551,
+      "D52": 552,
+      "H55": 553,
+      "J62": 554,
+      "C93": 555,
+      "C57": 556,
+      "D3A": 557,
+      "A03": 558,
+      "E84": 559,
+      "D45": 560,
+      "K03": 561,
+      "A31": 562,
+      "B27": 563,
+      "L81": 564,
+      "G63": 565,
+      "D67": 566,
+      "F16": 567,
+      "K41": 568,
+      "K05": 569,
+      "G23": 570,
+      "C41": 571,
+      "K73": 572,
+      "J33": 573,
+      "L21": 574,
+      "D76": 575,
+      "E31": 576,
+      "C82": 577,
+      "B25": 578,
+      "H30": 579,
+      "A07": 580,
+      "B17": 581,
+      "L64": 582,
+      "K27": 583,
+      "L24": 584,
+      "M45": 585,
+      "F48": 586,
+      "A77": 587,
+      "C49": 588,
+      "G71": 589,
+      "A15": 590,
+      "L54": 591,
+      "C23": 592,
+      "G99": 593,
+      "H60": 594,
+      "M12": 595,
+      "F81": 596,
+      "M90": 597,
+      "H36": 598,
+      "D15": 599,
+      "C68": 600,
+      "L74": 601,
+      "C65": 602,
+      "A53": 603,
+      "M95": 604,
+      "E65": 605,
+      "C4A": 606,
+      "H42": 607,
+      "H50": 608,
+      "M88": 609,
+      "H72": 610,
+      "L94": 611,
+      "A19": 612,
+      "C76": 613,
+      "L55": 614,
+      "D71": 615,
+      "D40": 616,
+      "G70": 617,
+      "G52": 618,
+      "L43": 619,
+      "E24": 620,
+      "C52": 621,
+      "L20": 622,
+      "B82": 623,
+      "C37": 624,
+      "D05": 625,
+      "J42": 626,
+      "C53": 627,
+      "J82": 628,
+      "H70": 629,
+      "C74": 630,
+      "D22": 631,
+      "B15": 632,
+      "D02": 633,
+      "B30": 634,
+      "C17": 635,
+      "I76": 636,
+      "B91": 637,
+      "C84": 638,
+      "B78": 639,
+      "L01": 640,
+      "E20": 641,
+      "M92": 642,
+      "J67": 643,
+      "J05": 644,
+      "K13": 645,
+      "A46": 646,
+      "G03": 647,
+      "B09": 648,
+      "F95": 649,
+      "C26": 650,
+      "F53": 651,
+      "C06": 652,
+      "F18": 653,
+      "C72": 654,
+      "D33": 655,
+      "A87": 656,
+      "H47": 657,
+      "I32": 658,
+      "G59": 659,
+      "B16": 660,
+      "K37": 661,
+      "H01": 662,
+      "B01": 663,
+      "G98": 664,
+      "D44": 665,
+      "E50": 666,
+      "H80": 667,
+      "G36": 668,
+      "D74": 669,
+      "M66": 670,
+      "C86": 671,
+      "C32": 672,
+      "C66": 673,
+      "L92": 674,
+      "G65": 675,
+      "D26": 676,
+      "F70": 677,
+      "M26": 678,
+      "C33": 679,
+      "K01": 680,
+      "M18": 681,
+      "D00": 682,
+      "B46": 683,
+      "D37": 684,
+      "A02": 685,
+      "C07": 686,
+      "F71": 687,
+      "G10": 688,
+      "H27": 689,
+      "L63": 690,
+      "K38": 691,
+      "C04": 692,
+      "C08": 693,
+      "J16": 694,
+      "H52": 695,
+      "L65": 696,
+      "L84": 697,
+      "D16": 698,
+      "J03": 699,
+      "L52": 700,
+      "G00": 701,
+      "I75": 702,
+      "H00": 703,
+      "L11": 704,
+      "E01": 705,
+      "H59": 706,
+      "M22": 707,
+      "C96": 708,
+      "B07": 709,
+      "D84": 710,
+      "E32": 711,
+      "A54": 712,
+      "A56": 713,
+      "D39": 714,
+      "E70": 715,
+      "M14": 716,
+      "A52": 717,
+      "F54": 718,
+      "F98": 719,
+      "F68": 720,
+      "A86": 721,
+      "A64": 722,
+      "M61": 723,
+      "I88": 724,
+      "D29": 725,
+      "J22": 726,
+      "F80": 727,
+      "D24": 728,
+      "B53": 729,
+      "J66": 730,
+      "F78": 731,
+      "E75": 732,
+      "M83": 733,
+      "B90": 734,
+      "A44": 735,
+      "E41": 736,
+      "M33": 737,
+      "L59": 738,
+      "M67": 739,
+      "D04": 740,
+      "D09": 741,
+      "K45": 742,
+      "D07": 743,
+      "D01": 744,
+      "M93": 745,
+      "D78": 746,
+      "L12": 747,
+      "I40": 748,
+      "J68": 749,
+      "G11": 750,
+      "L68": 751,
+      "M07": 752,
+      "A36": 753,
+      "F82": 754,
+      "C05": 755,
+      "J35": 756,
+      "F89": 757,
+      "J21": 758,
+      "E71": 759,
+      "G14": 760,
+      "C70": 761,
+      "B10": 762,
+      "C58": 763,
+      "C01": 764,
+      "L44": 765,
+      "C55": 766,
+      "G13": 767,
+      "C13": 768,
+      "A28": 769,
+      "D66": 770,
+      "M36": 771,
+      "B88": 772,
+      "B47": 773,
+      "A74": 774,
+      "B58": 775,
+      "I00": 776,
+      "A85": 777,
+      "L56": 778,
+      "F04": 779,
+      "G94": 780,
+      "A39": 781,
+      "D21": 782,
+      "B57": 783,
+      "B69": 784,
+      "L04": 785,
+      "L51": 786,
+      "M02": 787,
+      "J17": 788,
+      "D82": 789,
+      "C47": 790,
+      "K00": 791,
+      "H73": 792,
+      "J63": 793,
+      "A51": 794,
+      "F65": 795,
+      "F69": 796,
+      "F94": 797,
+      "J41": 798,
+      "D31": 799,
+      "B67": 800,
+      "M01": 801,
+      "G73": 802,
+      "D19": 803,
+      "C12": 804,
+      "D11": 805,
+      "A93": 806,
+      "B65": 807,
+      "L86": 808,
+      "B87": 809,
+      "C75": 810,
+      "C03": 811,
+      "E00": 812,
+      "H71": 813,
+      "C31": 814,
+      "G32": 815,
+      "L83": 816,
+      "H74": 817,
+      "A05": 818,
+      "B50": 819,
+      "D03": 820,
+      "G09": 821,
+      "B54": 822,
+      "E25": 823,
+      "D06": 824,
+      "C11": 825,
+      "A21": 826,
+      "M04": 827,
+      "B08": 828,
+      "D14": 829,
+      "A59": 830,
+      "D10": 831,
+      "A75": 832,
+      "A32": 833,
+      "C00": 834,
+      "M63": 835,
+      "I39": 836,
+      "B83": 837,
+      "B26": 838,
+      "E64": 839,
+      "G53": 840,
+      "A90": 841,
+      "E77": 842,
+      "A70": 843,
+      "G55": 844,
+      "C38": 845,
+      "A84": 846,
+      "E52": 847,
+      "F73": 848,
+      "I01": 849,
+      "A22": 850,
+      "H95": 851,
+      "F52": 852,
+      "G64": 853,
+      "J37": 854,
+      "A37": 855,
+      "D81": 856,
+      "E58": 857,
+      "G05": 858,
+      "H65": 859,
+      "A01": 860,
+      "A26": 861,
+      "E67": 862,
+      "B38": 863,
+      "L87": 864,
+      "M42": 865,
+      "H62": 866,
+      "L75": 867,
+      "A35": 868,
+      "L22": 869,
+      "C63": 870,
+      "A79": 871,
+      "K87": 872,
+      "M91": 873,
+      "L66": 874,
+      "B43": 875,
+      "B40": 876,
+      "B42": 877,
+      "L42": 878,
+      "A24": 879,
+      "K77": 880,
+      "G26": 881,
+      "E68": 882,
+      "C69": 883,
+      "F59": 884,
+      "H75": 885,
+      "H94": 886,
+      "H67": 887,
+      "L26": 888,
+      "C30": 889,
+      "H68": 890,
+      "A55": 891,
+      "A57": 892,
+      "J64": 893,
+      "A50": 894,
+      "B76": 895,
+      "H17": 896,
+      "A06": 897,
+      "E76": 898,
+      "A17": 899,
+      "L41": 900,
+      "B39": 901,
+      "D42": 902,
+      "A92": 903,
+      "A83": 904,
+      "D20": 905,
+      "L58": 906,
+      "B56": 907,
+      "A27": 908,
+      "A80": 909,
+      "A68": 910
+    }
+  }
+}

tokenizer_config.json ADDED Viewed

	@@ -0,0 +1,78 @@

+{
+  "add_prefix_space": false,
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "1": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "2": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "3": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "4": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "911": {
+      "content": "<s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "912": {
+      "content": "</s>",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "bos_token": "<s>",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": false,
+  "eos_token": "</s>",
+  "errors": "replace",
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "RobertaTokenizer",
+  "trim_offsets": true,
+  "unk_token": "[UNK]"
+}

vocab.txt ADDED Viewed

	@@ -0,0 +1,911 @@

+[MASK]
+[PAD]
+[CLS]
+[SEP]
+[UNK]
+G31
+F02
+E78
+G20
+K31
+K21
+K44
+F41
+I34
+M81
+G43
+I13
+I50
+E87
+E11
+I25
+D64
+G47
+M1A
+E66
+D50
+K29
+K25
+K57
+I12
+I70
+K22
+D63
+C67
+I10
+D25
+E89
+I26
+I82
+I47
+I87
+B95
+E03
+E83
+K59
+K65
+B96
+F17
+D72
+A41
+D62
+I27
+K43
+E43
+E53
+I48
+I11
+I45
+J44
+G56
+M19
+J45
+I73
+K72
+J96
+I44
+D69
+I46
+M47
+J01
+M79
+K92
+H40
+H69
+I24
+I49
+C16
+C78
+K91
+M86
+L97
+F32
+G25
+L03
+J98
+I96
+G89
+M25
+E46
+G30
+H91
+M17
+D68
+M46
+M70
+E55
+C50
+I83
+L76
+K76
+I67
+M51
+M48
+M35
+C79
+G93
+G91
+C34
+F05
+G51
+M21
+H53
+H04
+H35
+M50
+E04
+M80
+F03
+I28
+J43
+I07
+C43
+I65
+J18
+J90
+K40
+I30
+I31
+M06
+I21
+A04
+F44
+I95
+E86
+D51
+F43
+F50
+K02
+L73
+K85
+F39
+M54
+B20
+B19
+F10
+E16
+F19
+F14
+F20
+F29
+F11
+E10
+D86
+J02
+K20
+E21
+E51
+I69
+I63
+G81
+M85
+D49
+K74
+K66
+K62
+L91
+D57
+C90
+J91
+K56
+C21
+B37
+E88
+I35
+I42
+K61
+K60
+G92
+D46
+J84
+M84
+D18
+I61
+G40
+L02
+F88
+I08
+G62
+F31
+E44
+I15
+K80
+I16
+J69
+J95
+H26
+L40
+L57
+M72
+J30
+H61
+C56
+C77
+I81
+I85
+I86
+K70
+G96
+F12
+M53
+C7B
+C7A
+K58
+K26
+J81
+B18
+A08
+M96
+I97
+C22
+C25
+K86
+K52
+K11
+B00
+A40
+K83
+C24
+D53
+C20
+H90
+D47
+L29
+K81
+K82
+D70
+M10
+F90
+D13
+I71
+F33
+K75
+F22
+F15
+F72
+F13
+J36
+L27
+D61
+C92
+E85
+K12
+L99
+L72
+G58
+F45
+H66
+J11
+F60
+L30
+J86
+J15
+A81
+M40
+L08
+L60
+L82
+M62
+H93
+I72
+F42
+K04
+E80
+L53
+L89
+L98
+H81
+D80
+B35
+C73
+E05
+J39
+I22
+I51
+C18
+C91
+C62
+K64
+K63
+E08
+F09
+K28
+E22
+F25
+M94
+M71
+J20
+K95
+M16
+L85
+B34
+J04
+J47
+M31
+M41
+L71
+M00
+M89
+J94
+C61
+C80
+J93
+K50
+B02
+J06
+I09
+D32
+I33
+L28
+F23
+G12
+E02
+I60
+L95
+H33
+K08
+D73
+I20
+D59
+E79
+J60
+C95
+C94
+F84
+L05
+M30
+J32
+D23
+K42
+D35
+J34
+E26
+I80
+K51
+I36
+K55
+E27
+D75
+D55
+J40
+G90
+D12
+E56
+F51
+G35
+M87
+E06
+I43
+G44
+G95
+C02
+G83
+F79
+I77
+E73
+H05
+H54
+H46
+M08
+M11
+H11
+F28
+M97
+C60
+M99
+M34
+F01
+I66
+H43
+H25
+D89
+G60
+I89
+M20
+E13
+L93
+E61
+A09
+G24
+G21
+M75
+M05
+D60
+B44
+B59
+A49
+A18
+M43
+M23
+K94
+C81
+K35
+F07
+G82
+E74
+C45
+E59
+E60
+A48
+G45
+F30
+H34
+C85
+M32
+I05
+D28
+M24
+E23
+C71
+F40
+D38
+D48
+C64
+H10
+D65
+B49
+H16
+I99
+L70
+C48
+D43
+E07
+L10
+C83
+J70
+I68
+J99
+H15
+B36
+D58
+L23
+I23
+J09
+B94
+I74
+G54
+B99
+G72
+G50
+G97
+D17
+E09
+L13
+G61
+A43
+H02
+H57
+H49
+I38
+G37
+F21
+D56
+K90
+J92
+J12
+B97
+J38
+L80
+H31
+A42
+M13
+M77
+M65
+L88
+C44
+G06
+L25
+D34
+K46
+M60
+I62
+D83
+F64
+D41
+E72
+B60
+J61
+J85
+C10
+A63
+C14
+C09
+F55
+F99
+M76
+A60
+C54
+F91
+F63
+J14
+F06
+C88
+H83
+J31
+I37
+K68
+B45
+G04
+J10
+B85
+C46
+B48
+C40
+J80
+H51
+G46
+M15
+K06
+A69
+J13
+I78
+J00
+M27
+G08
+H92
+E63
+B33
+C19
+E34
+L50
+I06
+K71
+K36
+K14
+H18
+E29
+E54
+D30
+E28
+K30
+B86
+G57
+H21
+H44
+H20
+C15
+G80
+D36
+F34
+D27
+C51
+L90
+D52
+H55
+J62
+C93
+C57
+D3A
+A03
+E84
+D45
+K03
+A31
+B27
+L81
+G63
+D67
+F16
+K41
+K05
+G23
+C41
+K73
+J33
+L21
+D76
+E31
+C82
+B25
+H30
+A07
+B17
+L64
+K27
+L24
+M45
+F48
+A77
+C49
+G71
+A15
+L54
+C23
+G99
+H60
+M12
+F81
+M90
+H36
+D15
+C68
+L74
+C65
+A53
+M95
+E65
+C4A
+H42
+H50
+M88
+H72
+L94
+A19
+C76
+L55
+D71
+D40
+G70
+G52
+L43
+E24
+C52
+L20
+B82
+C37
+D05
+J42
+C53
+J82
+H70
+C74
+D22
+B15
+D02
+B30
+C17
+I76
+B91
+C84
+B78
+L01
+E20
+M92
+J67
+J05
+K13
+A46
+G03
+B09
+F95
+C26
+F53
+C06
+F18
+C72
+D33
+A87
+H47
+I32
+G59
+B16
+K37
+H01
+B01
+G98
+D44
+E50
+H80
+G36
+D74
+M66
+C86
+C32
+C66
+L92
+G65
+D26
+F70
+M26
+C33
+K01
+M18
+D00
+B46
+D37
+A02
+C07
+F71
+G10
+H27
+L63
+K38
+C04
+C08
+J16
+H52
+L65
+L84
+D16
+J03
+L52
+G00
+I75
+H00
+L11
+E01
+H59
+M22
+C96
+B07
+D84
+E32
+A54
+A56
+D39
+E70
+M14
+A52
+F54
+F98
+F68
+A86
+A64
+M61
+I88
+D29
+J22
+F80
+D24
+B53
+J66
+F78
+E75
+M83
+B90
+A44
+E41
+M33
+L59
+M67
+D04
+D09
+K45
+D07
+D01
+M93
+D78
+L12
+I40
+J68
+G11
+L68
+M07
+A36
+F82
+C05
+J35
+F89
+J21
+E71
+G14
+C70
+B10
+C58
+C01
+L44
+C55
+G13
+C13
+A28
+D66
+M36
+B88
+B47
+A74
+B58
+I00
+A85
+L56
+F04
+G94
+A39
+D21
+B57
+B69
+L04
+L51
+M02
+J17
+D82
+C47
+K00
+H73
+J63
+A51
+F65
+F69
+F94
+J41
+D31
+B67
+M01
+G73
+D19
+C12
+D11
+A93
+B65
+L86
+B87
+C75
+C03
+E00
+H71
+C31
+G32
+L83
+H74
+A05
+B50
+D03
+G09
+B54
+E25
+D06
+C11
+A21
+M04
+B08
+D14
+A59
+D10
+A75
+A32
+C00
+M63
+I39
+B83
+B26
+E64
+G53
+A90
+E77
+A70
+G55
+C38
+A84
+E52
+F73
+I01
+A22
+H95
+F52
+G64
+J37
+A37
+D81
+E58
+G05
+H65
+A01
+A26
+E67
+B38
+L87
+M42
+H62
+L75
+A35
+L22
+C63
+A79
+K87
+M91
+L66
+B43
+B40
+B42
+L42
+A24
+K77
+G26
+E68
+C69
+F59
+H75
+H94
+H67
+L26
+C30
+H68
+A55
+A57
+J64
+A50
+B76
+H17
+A06
+E76
+A17
+L41
+B39
+D42
+A92
+A83
+D20
+L58
+B56
+A27
+A80
+A68