| { | |
| "version": "1.0", | |
| "truncation": null, | |
| "padding": null, | |
| "added_tokens": [ | |
| { | |
| "id": 0, | |
| "content": "<|endoftext|>", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": true, | |
| "special": true | |
| }, | |
| { | |
| "id": 1, | |
| "content": "a", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 2, | |
| "content": "E", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 3, | |
| "content": "e", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 4, | |
| "content": "Ɛ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 5, | |
| "content": "ɛ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 6, | |
| "content": "Ə", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 7, | |
| "content": "ə", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 8, | |
| "content": "I", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 9, | |
| "content": "i", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 10, | |
| "content": "O", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 11, | |
| "content": "o", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 12, | |
| "content": "Ɔ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 13, | |
| "content": "ɔ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 14, | |
| "content": "U", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 15, | |
| "content": "u", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 16, | |
| "content": "ã", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 17, | |
| "content": "ẽ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 18, | |
| "content": "ĩ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 19, | |
| "content": "õ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 20, | |
| "content": "ũ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 21, | |
| "content": "B", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 22, | |
| "content": "b", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 23, | |
| "content": "D", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 24, | |
| "content": "d", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 25, | |
| "content": "Đ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 26, | |
| "content": "ɖ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 27, | |
| "content": "F", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 28, | |
| "content": "f", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 29, | |
| "content": "Ƒ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 30, | |
| "content": "ƒ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 31, | |
| "content": "G", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 32, | |
| "content": "g", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 33, | |
| "content": "Ɣ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 34, | |
| "content": "ɣ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 35, | |
| "content": "H", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 36, | |
| "content": "h", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 37, | |
| "content": "K", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 38, | |
| "content": "k", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 39, | |
| "content": "L", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 40, | |
| "content": "l", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 41, | |
| "content": "M", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 42, | |
| "content": "m", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 43, | |
| "content": "N", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 44, | |
| "content": "n", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 45, | |
| "content": "Ŋ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 46, | |
| "content": "ŋ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 47, | |
| "content": "P", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 48, | |
| "content": "p", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 49, | |
| "content": "R", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 50, | |
| "content": "r", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 51, | |
| "content": "S", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 52, | |
| "content": "s", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 53, | |
| "content": "T", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 54, | |
| "content": "t", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 55, | |
| "content": "V", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 56, | |
| "content": "v", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 57, | |
| "content": "Ʋ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 58, | |
| "content": "ʋ", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 59, | |
| "content": "W", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 60, | |
| "content": "w", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 61, | |
| "content": "X", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 62, | |
| "content": "x", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 63, | |
| "content": "Y", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 64, | |
| "content": "y", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 65, | |
| "content": "Z", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 66, | |
| "content": "z", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 67, | |
| "content": "̃", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 68, | |
| "content": "ó", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 69, | |
| "content": "À", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 70, | |
| "content": "é", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 71, | |
| "content": "È", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 72, | |
| "content": "ò", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 73, | |
| "content": "à", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| }, | |
| { | |
| "id": 74, | |
| "content": "í", | |
| "single_word": false, | |
| "lstrip": false, | |
| "rstrip": false, | |
| "normalized": false, | |
| "special": true | |
| } | |
| ], | |
| "normalizer": null, | |
| "pre_tokenizer": { | |
| "type": "ByteLevel", | |
| "add_prefix_space": false, | |
| "trim_offsets": true, | |
| "use_regex": true | |
| }, | |
| "post_processor": { | |
| "type": "ByteLevel", | |
| "add_prefix_space": true, | |
| "trim_offsets": false, | |
| "use_regex": true | |
| }, | |
| "decoder": { | |
| "type": "ByteLevel", | |
| "add_prefix_space": true, | |
| "trim_offsets": true, | |
| "use_regex": true | |
| }, | |
| "model": { | |
| "type": "BPE", | |
| "dropout": null, | |
| "unk_token": null, | |
| "continuing_subword_prefix": "", | |
| "end_of_word_suffix": "", | |
| "fuse_unk": false, | |
| "byte_fallback": false, | |
| "ignore_merges": false, | |
| "vocab": { | |
| "<|endoftext|>": 0, | |
| "a": 1, | |
| "E": 2, | |
| "e": 3, | |
| "Ɛ": 4, | |
| "ɛ": 5, | |
| "Ə": 6, | |
| "ə": 7, | |
| "I": 8, | |
| "i": 9, | |
| "O": 10, | |
| "o": 11, | |
| "Ɔ": 12, | |
| "ɔ": 13, | |
| "U": 14, | |
| "u": 15, | |
| "ã": 16, | |
| "ẽ": 17, | |
| "ĩ": 18, | |
| "õ": 19, | |
| "ũ": 20, | |
| "B": 21, | |
| "b": 22, | |
| "D": 23, | |
| "d": 24, | |
| "Đ": 25, | |
| "ɖ": 26, | |
| "F": 27, | |
| "f": 28, | |
| "Ƒ": 29, | |
| "ƒ": 30, | |
| "G": 31, | |
| "g": 32, | |
| "Ɣ": 33, | |
| "ɣ": 34, | |
| "H": 35, | |
| "h": 36, | |
| "K": 37, | |
| "k": 38, | |
| "L": 39, | |
| "l": 40, | |
| "M": 41, | |
| "m": 42, | |
| "N": 43, | |
| "n": 44, | |
| "Ŋ": 45, | |
| "ŋ": 46, | |
| "P": 47, | |
| "p": 48, | |
| "R": 49, | |
| "r": 50, | |
| "S": 51, | |
| "s": 52, | |
| "T": 53, | |
| "t": 54, | |
| "V": 55, | |
| "v": 56, | |
| "Ʋ": 57, | |
| "ʋ": 58, | |
| "W": 59, | |
| "w": 60, | |
| "X": 61, | |
| "x": 62, | |
| "Y": 63, | |
| "y": 64, | |
| "Z": 65, | |
| "z": 66, | |
| "̃": 67, | |
| "ó": 68, | |
| "À": 69, | |
| "é": 70, | |
| "È": 71, | |
| "ò": 72, | |
| "à": 73, | |
| "í": 74, | |
| "!": 75, | |
| "\"": 76, | |
| "#": 77, | |
| "$": 78, | |
| "%": 79, | |
| "&": 80, | |
| "'": 81, | |
| "(": 82, | |
| ")": 83, | |
| "*": 84, | |
| "+": 85, | |
| ",": 86, | |
| "-": 87, | |
| ".": 88, | |
| "/": 89, | |
| "0": 90, | |
| "1": 91, | |
| "2": 92, | |
| "3": 93, | |
| "4": 94, | |
| "5": 95, | |
| "6": 96, | |
| "7": 97, | |
| "8": 98, | |
| "9": 99, | |
| ":": 100, | |
| ";": 101, | |
| "<": 102, | |
| "=": 103, | |
| ">": 104, | |
| "?": 105, | |
| "@": 106, | |
| "A": 107, | |
| "C": 108, | |
| "J": 109, | |
| "Q": 110, | |
| "[": 111, | |
| "\\": 112, | |
| "]": 113, | |
| "^": 114, | |
| "_": 115, | |
| "`": 116, | |
| "c": 117, | |
| "j": 118, | |
| "q": 119, | |
| "{": 120, | |
| "|": 121, | |
| "}": 122, | |
| "~": 123, | |
| "¡": 124, | |
| "¢": 125, | |
| "£": 126, | |
| "¤": 127, | |
| "¥": 128, | |
| "¦": 129, | |
| "§": 130, | |
| "¨": 131, | |
| "©": 132, | |
| "ª": 133, | |
| "«": 134, | |
| "¬": 135, | |
| "®": 136, | |
| "¯": 137, | |
| "°": 138, | |
| "±": 139, | |
| "²": 140, | |
| "³": 141, | |
| "´": 142, | |
| "µ": 143, | |
| "¶": 144, | |
| "·": 145, | |
| "¸": 146, | |
| "¹": 147, | |
| "º": 148, | |
| "»": 149, | |
| "¼": 150, | |
| "½": 151, | |
| "¾": 152, | |
| "¿": 153, | |
| "Á": 154, | |
| "Â": 155, | |
| "Ã": 156, | |
| "Ä": 157, | |
| "Å": 158, | |
| "Æ": 159, | |
| "Ç": 160, | |
| "É": 161, | |
| "Ê": 162, | |
| "Ë": 163, | |
| "Ì": 164, | |
| "Í": 165, | |
| "Î": 166, | |
| "Ï": 167, | |
| "Ð": 168, | |
| "Ñ": 169, | |
| "Ò": 170, | |
| "Ó": 171, | |
| "Ô": 172, | |
| "Õ": 173, | |
| "Ö": 174, | |
| "×": 175, | |
| "Ø": 176, | |
| "Ù": 177, | |
| "Ú": 178, | |
| "Û": 179, | |
| "Ü": 180, | |
| "Ý": 181, | |
| "Þ": 182, | |
| "ß": 183, | |
| "á": 184, | |
| "â": 185, | |
| "ä": 186, | |
| "å": 187, | |
| "æ": 188, | |
| "ç": 189, | |
| "è": 190, | |
| "ê": 191, | |
| "ë": 192, | |
| "ì": 193, | |
| "î": 194, | |
| "ï": 195, | |
| "ð": 196, | |
| "ñ": 197, | |
| "ô": 198, | |
| "ö": 199, | |
| "÷": 200, | |
| "ø": 201, | |
| "ù": 202, | |
| "ú": 203, | |
| "û": 204, | |
| "ü": 205, | |
| "ý": 206, | |
| "þ": 207, | |
| "ÿ": 208, | |
| "Ā": 209, | |
| "ā": 210, | |
| "Ă": 211, | |
| "ă": 212, | |
| "Ą": 213, | |
| "ą": 214, | |
| "Ć": 215, | |
| "ć": 216, | |
| "Ĉ": 217, | |
| "ĉ": 218, | |
| "Ċ": 219, | |
| "ċ": 220, | |
| "Č": 221, | |
| "č": 222, | |
| "Ď": 223, | |
| "ď": 224, | |
| "đ": 225, | |
| "Ē": 226, | |
| "ē": 227, | |
| "Ĕ": 228, | |
| "ĕ": 229, | |
| "Ė": 230, | |
| "ė": 231, | |
| "Ę": 232, | |
| "ę": 233, | |
| "Ě": 234, | |
| "ě": 235, | |
| "Ĝ": 236, | |
| "ĝ": 237, | |
| "Ğ": 238, | |
| "ğ": 239, | |
| "Ġ": 240, | |
| "ġ": 241, | |
| "Ģ": 242, | |
| "ģ": 243, | |
| "Ĥ": 244, | |
| "ĥ": 245, | |
| "Ħ": 246, | |
| "ħ": 247, | |
| "Ĩ": 248, | |
| "Ī": 249, | |
| "ī": 250, | |
| "Ĭ": 251, | |
| "ĭ": 252, | |
| "Į": 253, | |
| "į": 254, | |
| "İ": 255, | |
| "ı": 256, | |
| "IJ": 257, | |
| "ij": 258, | |
| "Ĵ": 259, | |
| "ĵ": 260, | |
| "Ķ": 261, | |
| "ķ": 262, | |
| "ĸ": 263, | |
| "Ĺ": 264, | |
| "ĺ": 265, | |
| "Ļ": 266, | |
| "ļ": 267, | |
| "Ľ": 268, | |
| "ľ": 269, | |
| "Ŀ": 270, | |
| "ŀ": 271, | |
| "Ł": 272, | |
| "ł": 273, | |
| "Ń": 274 | |
| }, | |
| "merges": [] | |
| } | |
| } |