Upload tokenizer.json with huggingface_hub
Browse files- tokenizer.json +139 -0
tokenizer.json
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"vocab": {
|
| 3 |
+
"<PAD>": 0,
|
| 4 |
+
"N": 1,
|
| 5 |
+
"5": 2,
|
| 6 |
+
"b": 3,
|
| 7 |
+
"g": 4,
|
| 8 |
+
"6": 5,
|
| 9 |
+
"e": 6,
|
| 10 |
+
"h": 7,
|
| 11 |
+
"2": 8,
|
| 12 |
+
"a": 9,
|
| 13 |
+
"1": 10,
|
| 14 |
+
"7": 11,
|
| 15 |
+
"R": 12,
|
| 16 |
+
"d": 13,
|
| 17 |
+
"B": 14,
|
| 18 |
+
"3": 15,
|
| 19 |
+
"K": 16,
|
| 20 |
+
"8": 17,
|
| 21 |
+
"Q": 18,
|
| 22 |
+
"f": 19,
|
| 23 |
+
"c": 20,
|
| 24 |
+
"P": 21,
|
| 25 |
+
"4": 22,
|
| 26 |
+
"-": 23,
|
| 27 |
+
".": 24,
|
| 28 |
+
"+": 25,
|
| 29 |
+
"#": 26,
|
| 30 |
+
"9": 27,
|
| 31 |
+
"x": 28,
|
| 32 |
+
"0": 29,
|
| 33 |
+
"O": 30,
|
| 34 |
+
"=": 31,
|
| 35 |
+
" ": 32,
|
| 36 |
+
"\n": 33,
|
| 37 |
+
"\t": 34,
|
| 38 |
+
",": 35,
|
| 39 |
+
"'": 36,
|
| 40 |
+
"[": 37,
|
| 41 |
+
"\"": 38,
|
| 42 |
+
"]": 39,
|
| 43 |
+
")": 40,
|
| 44 |
+
"!": 41,
|
| 45 |
+
"/": 42,
|
| 46 |
+
"(": 43
|
| 47 |
+
},
|
| 48 |
+
"id2token": {
|
| 49 |
+
"0": "<PAD>",
|
| 50 |
+
"1": "N",
|
| 51 |
+
"2": "5",
|
| 52 |
+
"3": "b",
|
| 53 |
+
"4": "g",
|
| 54 |
+
"5": "6",
|
| 55 |
+
"6": "e",
|
| 56 |
+
"7": "h",
|
| 57 |
+
"8": "2",
|
| 58 |
+
"9": "a",
|
| 59 |
+
"10": "1",
|
| 60 |
+
"11": "7",
|
| 61 |
+
"12": "R",
|
| 62 |
+
"13": "d",
|
| 63 |
+
"14": "B",
|
| 64 |
+
"15": "3",
|
| 65 |
+
"16": "K",
|
| 66 |
+
"17": "8",
|
| 67 |
+
"18": "Q",
|
| 68 |
+
"19": "f",
|
| 69 |
+
"20": "c",
|
| 70 |
+
"21": "P",
|
| 71 |
+
"22": "4",
|
| 72 |
+
"23": "-",
|
| 73 |
+
"24": ".",
|
| 74 |
+
"25": "+",
|
| 75 |
+
"26": "#",
|
| 76 |
+
"27": "9",
|
| 77 |
+
"28": "x",
|
| 78 |
+
"29": "0",
|
| 79 |
+
"30": "O",
|
| 80 |
+
"31": "=",
|
| 81 |
+
"32": " ",
|
| 82 |
+
"33": "\n",
|
| 83 |
+
"34": "\t",
|
| 84 |
+
"35": ",",
|
| 85 |
+
"36": "'",
|
| 86 |
+
"37": "[",
|
| 87 |
+
"38": "\"",
|
| 88 |
+
"39": "]",
|
| 89 |
+
"40": ")",
|
| 90 |
+
"41": "!",
|
| 91 |
+
"42": "/",
|
| 92 |
+
"43": "("
|
| 93 |
+
},
|
| 94 |
+
"valid_chars": [
|
| 95 |
+
"N",
|
| 96 |
+
"5",
|
| 97 |
+
".",
|
| 98 |
+
"b",
|
| 99 |
+
"9",
|
| 100 |
+
"[",
|
| 101 |
+
"g",
|
| 102 |
+
"=",
|
| 103 |
+
"6",
|
| 104 |
+
"e",
|
| 105 |
+
" ",
|
| 106 |
+
"-",
|
| 107 |
+
"h",
|
| 108 |
+
"2",
|
| 109 |
+
"\"",
|
| 110 |
+
"]",
|
| 111 |
+
"a",
|
| 112 |
+
"1",
|
| 113 |
+
"7",
|
| 114 |
+
"R",
|
| 115 |
+
"'",
|
| 116 |
+
"d",
|
| 117 |
+
"B",
|
| 118 |
+
"x",
|
| 119 |
+
"3",
|
| 120 |
+
"0",
|
| 121 |
+
"O",
|
| 122 |
+
")",
|
| 123 |
+
"(",
|
| 124 |
+
"\n",
|
| 125 |
+
"K",
|
| 126 |
+
"\t",
|
| 127 |
+
"+",
|
| 128 |
+
",",
|
| 129 |
+
"8",
|
| 130 |
+
"#",
|
| 131 |
+
"Q",
|
| 132 |
+
"f",
|
| 133 |
+
"c",
|
| 134 |
+
"P",
|
| 135 |
+
"!",
|
| 136 |
+
"4",
|
| 137 |
+
"/"
|
| 138 |
+
]
|
| 139 |
+
}
|