Upload tokenizer
Browse files- special_tokens_map.json +9 -0
- tokenizer.json +691 -0
- tokenizer_config.json +14 -0
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<cls>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "<sep>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,691 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
| 4 |
+
"padding": null,
|
| 5 |
+
"added_tokens": [
|
| 6 |
+
{
|
| 7 |
+
"id": 0,
|
| 8 |
+
"content": "<s>",
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"special": true
|
| 14 |
+
},
|
| 15 |
+
{
|
| 16 |
+
"id": 1,
|
| 17 |
+
"content": "<pad>",
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"normalized": false,
|
| 22 |
+
"special": true
|
| 23 |
+
},
|
| 24 |
+
{
|
| 25 |
+
"id": 2,
|
| 26 |
+
"content": "</s>",
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"special": true
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"id": 3,
|
| 35 |
+
"content": "<unk>",
|
| 36 |
+
"single_word": false,
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"rstrip": false,
|
| 39 |
+
"normalized": false,
|
| 40 |
+
"special": true
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"id": 4,
|
| 44 |
+
"content": "<cls>",
|
| 45 |
+
"single_word": false,
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"id": 5,
|
| 53 |
+
"content": "<sep>",
|
| 54 |
+
"single_word": false,
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"rstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"special": true
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"id": 6,
|
| 62 |
+
"content": "<mask>",
|
| 63 |
+
"single_word": false,
|
| 64 |
+
"lstrip": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"normalized": false,
|
| 67 |
+
"special": true
|
| 68 |
+
}
|
| 69 |
+
],
|
| 70 |
+
"normalizer": null,
|
| 71 |
+
"pre_tokenizer": {
|
| 72 |
+
"type": "WhitespaceSplit"
|
| 73 |
+
},
|
| 74 |
+
"post_processor": null,
|
| 75 |
+
"decoder": null,
|
| 76 |
+
"model": {
|
| 77 |
+
"type": "WordLevel",
|
| 78 |
+
"vocab": {
|
| 79 |
+
"<s>": 0,
|
| 80 |
+
"<pad>": 1,
|
| 81 |
+
"</s>": 2,
|
| 82 |
+
"<unk>": 3,
|
| 83 |
+
"<cls>": 4,
|
| 84 |
+
"<sep>": 5,
|
| 85 |
+
"<mask>": 6,
|
| 86 |
+
"a": 7,
|
| 87 |
+
"aa": 8,
|
| 88 |
+
"0.000": 9,
|
| 89 |
+
"0.001": 10,
|
| 90 |
+
"-0.001": 11,
|
| 91 |
+
"0.002": 12,
|
| 92 |
+
"-0.002": 13,
|
| 93 |
+
"0.003": 14,
|
| 94 |
+
"-0.003": 15,
|
| 95 |
+
"0.004": 16,
|
| 96 |
+
"-0.004": 17,
|
| 97 |
+
"0.005": 18,
|
| 98 |
+
"-0.005": 19,
|
| 99 |
+
"0.006": 20,
|
| 100 |
+
"-0.006": 21,
|
| 101 |
+
"0.007": 22,
|
| 102 |
+
"-0.007": 23,
|
| 103 |
+
"0.008": 24,
|
| 104 |
+
"-0.008": 25,
|
| 105 |
+
"0.009": 26,
|
| 106 |
+
"-0.009": 27,
|
| 107 |
+
"0.010": 28,
|
| 108 |
+
"-0.010": 29,
|
| 109 |
+
"0.011": 30,
|
| 110 |
+
"-0.011": 31,
|
| 111 |
+
"0.012": 32,
|
| 112 |
+
"-0.012": 33,
|
| 113 |
+
"0.013": 34,
|
| 114 |
+
"-0.013": 35,
|
| 115 |
+
"0.014": 36,
|
| 116 |
+
"-0.014": 37,
|
| 117 |
+
"0.015": 38,
|
| 118 |
+
"-0.015": 39,
|
| 119 |
+
"0.016": 40,
|
| 120 |
+
"-0.016": 41,
|
| 121 |
+
"0.017": 42,
|
| 122 |
+
"-0.017": 43,
|
| 123 |
+
"0.018": 44,
|
| 124 |
+
"-0.018": 45,
|
| 125 |
+
"0.019": 46,
|
| 126 |
+
"-0.019": 47,
|
| 127 |
+
"0.020": 48,
|
| 128 |
+
"-0.020": 49,
|
| 129 |
+
"0.021": 50,
|
| 130 |
+
"-0.021": 51,
|
| 131 |
+
"0.022": 52,
|
| 132 |
+
"-0.022": 53,
|
| 133 |
+
"0.023": 54,
|
| 134 |
+
"-0.023": 55,
|
| 135 |
+
"0.024": 56,
|
| 136 |
+
"-0.024": 57,
|
| 137 |
+
"0.025": 58,
|
| 138 |
+
"-0.025": 59,
|
| 139 |
+
"0.026": 60,
|
| 140 |
+
"-0.026": 61,
|
| 141 |
+
"0.027": 62,
|
| 142 |
+
"-0.027": 63,
|
| 143 |
+
"0.028": 64,
|
| 144 |
+
"-0.028": 65,
|
| 145 |
+
"0.029": 66,
|
| 146 |
+
"-0.029": 67,
|
| 147 |
+
"0.030": 68,
|
| 148 |
+
"-0.030": 69,
|
| 149 |
+
"0.031": 70,
|
| 150 |
+
"-0.031": 71,
|
| 151 |
+
"0.032": 72,
|
| 152 |
+
"-0.032": 73,
|
| 153 |
+
"0.033": 74,
|
| 154 |
+
"-0.033": 75,
|
| 155 |
+
"0.034": 76,
|
| 156 |
+
"-0.034": 77,
|
| 157 |
+
"0.035": 78,
|
| 158 |
+
"-0.035": 79,
|
| 159 |
+
"0.036": 80,
|
| 160 |
+
"-0.036": 81,
|
| 161 |
+
"0.037": 82,
|
| 162 |
+
"-0.037": 83,
|
| 163 |
+
"0.038": 84,
|
| 164 |
+
"-0.038": 85,
|
| 165 |
+
"0.039": 86,
|
| 166 |
+
"-0.039": 87,
|
| 167 |
+
"0.040": 88,
|
| 168 |
+
"-0.040": 89,
|
| 169 |
+
"0.041": 90,
|
| 170 |
+
"-0.041": 91,
|
| 171 |
+
"0.042": 92,
|
| 172 |
+
"-0.042": 93,
|
| 173 |
+
"0.043": 94,
|
| 174 |
+
"-0.043": 95,
|
| 175 |
+
"0.044": 96,
|
| 176 |
+
"-0.044": 97,
|
| 177 |
+
"0.045": 98,
|
| 178 |
+
"-0.045": 99,
|
| 179 |
+
"0.046": 100,
|
| 180 |
+
"-0.046": 101,
|
| 181 |
+
"0.047": 102,
|
| 182 |
+
"-0.047": 103,
|
| 183 |
+
"0.048": 104,
|
| 184 |
+
"-0.048": 105,
|
| 185 |
+
"0.049": 106,
|
| 186 |
+
"-0.049": 107,
|
| 187 |
+
"0.050": 108,
|
| 188 |
+
"-0.050": 109,
|
| 189 |
+
"0.051": 110,
|
| 190 |
+
"-0.051": 111,
|
| 191 |
+
"0.052": 112,
|
| 192 |
+
"-0.052": 113,
|
| 193 |
+
"0.053": 114,
|
| 194 |
+
"-0.053": 115,
|
| 195 |
+
"0.054": 116,
|
| 196 |
+
"-0.054": 117,
|
| 197 |
+
"0.055": 118,
|
| 198 |
+
"-0.055": 119,
|
| 199 |
+
"0.056": 120,
|
| 200 |
+
"-0.056": 121,
|
| 201 |
+
"0.057": 122,
|
| 202 |
+
"-0.057": 123,
|
| 203 |
+
"0.058": 124,
|
| 204 |
+
"-0.058": 125,
|
| 205 |
+
"0.059": 126,
|
| 206 |
+
"-0.059": 127,
|
| 207 |
+
"0.060": 128,
|
| 208 |
+
"-0.060": 129,
|
| 209 |
+
"0.061": 130,
|
| 210 |
+
"-0.061": 131,
|
| 211 |
+
"0.062": 132,
|
| 212 |
+
"-0.062": 133,
|
| 213 |
+
"0.063": 134,
|
| 214 |
+
"-0.063": 135,
|
| 215 |
+
"0.064": 136,
|
| 216 |
+
"-0.064": 137,
|
| 217 |
+
"0.065": 138,
|
| 218 |
+
"-0.065": 139,
|
| 219 |
+
"0.066": 140,
|
| 220 |
+
"-0.066": 141,
|
| 221 |
+
"0.067": 142,
|
| 222 |
+
"-0.067": 143,
|
| 223 |
+
"0.068": 144,
|
| 224 |
+
"-0.068": 145,
|
| 225 |
+
"0.069": 146,
|
| 226 |
+
"-0.069": 147,
|
| 227 |
+
"0.070": 148,
|
| 228 |
+
"-0.070": 149,
|
| 229 |
+
"0.071": 150,
|
| 230 |
+
"-0.071": 151,
|
| 231 |
+
"0.072": 152,
|
| 232 |
+
"-0.072": 153,
|
| 233 |
+
"0.073": 154,
|
| 234 |
+
"-0.073": 155,
|
| 235 |
+
"0.074": 156,
|
| 236 |
+
"-0.074": 157,
|
| 237 |
+
"0.075": 158,
|
| 238 |
+
"-0.075": 159,
|
| 239 |
+
"0.076": 160,
|
| 240 |
+
"-0.076": 161,
|
| 241 |
+
"0.077": 162,
|
| 242 |
+
"-0.077": 163,
|
| 243 |
+
"0.078": 164,
|
| 244 |
+
"-0.078": 165,
|
| 245 |
+
"0.079": 166,
|
| 246 |
+
"-0.079": 167,
|
| 247 |
+
"0.080": 168,
|
| 248 |
+
"-0.080": 169,
|
| 249 |
+
"0.081": 170,
|
| 250 |
+
"-0.081": 171,
|
| 251 |
+
"0.082": 172,
|
| 252 |
+
"-0.082": 173,
|
| 253 |
+
"0.083": 174,
|
| 254 |
+
"-0.083": 175,
|
| 255 |
+
"0.084": 176,
|
| 256 |
+
"-0.084": 177,
|
| 257 |
+
"0.085": 178,
|
| 258 |
+
"-0.085": 179,
|
| 259 |
+
"0.086": 180,
|
| 260 |
+
"-0.086": 181,
|
| 261 |
+
"0.087": 182,
|
| 262 |
+
"-0.087": 183,
|
| 263 |
+
"0.088": 184,
|
| 264 |
+
"-0.088": 185,
|
| 265 |
+
"0.089": 186,
|
| 266 |
+
"-0.089": 187,
|
| 267 |
+
"0.090": 188,
|
| 268 |
+
"-0.090": 189,
|
| 269 |
+
"0.091": 190,
|
| 270 |
+
"-0.091": 191,
|
| 271 |
+
"0.092": 192,
|
| 272 |
+
"-0.092": 193,
|
| 273 |
+
"0.093": 194,
|
| 274 |
+
"-0.093": 195,
|
| 275 |
+
"0.094": 196,
|
| 276 |
+
"-0.094": 197,
|
| 277 |
+
"0.095": 198,
|
| 278 |
+
"-0.095": 199,
|
| 279 |
+
"0.096": 200,
|
| 280 |
+
"-0.096": 201,
|
| 281 |
+
"0.097": 202,
|
| 282 |
+
"-0.097": 203,
|
| 283 |
+
"0.098": 204,
|
| 284 |
+
"-0.098": 205,
|
| 285 |
+
"0.099": 206,
|
| 286 |
+
"-0.099": 207,
|
| 287 |
+
"0.100": 208,
|
| 288 |
+
"-0.100": 209,
|
| 289 |
+
"0.101": 210,
|
| 290 |
+
"-0.101": 211,
|
| 291 |
+
"0.102": 212,
|
| 292 |
+
"-0.102": 213,
|
| 293 |
+
"0.103": 214,
|
| 294 |
+
"-0.103": 215,
|
| 295 |
+
"0.104": 216,
|
| 296 |
+
"-0.104": 217,
|
| 297 |
+
"0.105": 218,
|
| 298 |
+
"-0.105": 219,
|
| 299 |
+
"0.106": 220,
|
| 300 |
+
"-0.106": 221,
|
| 301 |
+
"0.107": 222,
|
| 302 |
+
"-0.107": 223,
|
| 303 |
+
"0.108": 224,
|
| 304 |
+
"-0.108": 225,
|
| 305 |
+
"0.109": 226,
|
| 306 |
+
"-0.109": 227,
|
| 307 |
+
"0.110": 228,
|
| 308 |
+
"-0.110": 229,
|
| 309 |
+
"0.111": 230,
|
| 310 |
+
"-0.111": 231,
|
| 311 |
+
"0.112": 232,
|
| 312 |
+
"-0.112": 233,
|
| 313 |
+
"0.113": 234,
|
| 314 |
+
"-0.113": 235,
|
| 315 |
+
"0.114": 236,
|
| 316 |
+
"-0.114": 237,
|
| 317 |
+
"0.115": 238,
|
| 318 |
+
"-0.115": 239,
|
| 319 |
+
"0.116": 240,
|
| 320 |
+
"-0.116": 241,
|
| 321 |
+
"0.117": 242,
|
| 322 |
+
"-0.117": 243,
|
| 323 |
+
"0.118": 244,
|
| 324 |
+
"-0.118": 245,
|
| 325 |
+
"0.119": 246,
|
| 326 |
+
"-0.119": 247,
|
| 327 |
+
"0.120": 248,
|
| 328 |
+
"-0.120": 249,
|
| 329 |
+
"0.121": 250,
|
| 330 |
+
"-0.121": 251,
|
| 331 |
+
"0.122": 252,
|
| 332 |
+
"-0.122": 253,
|
| 333 |
+
"0.123": 254,
|
| 334 |
+
"-0.123": 255,
|
| 335 |
+
"0.124": 256,
|
| 336 |
+
"-0.124": 257,
|
| 337 |
+
"0.125": 258,
|
| 338 |
+
"-0.125": 259,
|
| 339 |
+
"0.126": 260,
|
| 340 |
+
"-0.126": 261,
|
| 341 |
+
"0.127": 262,
|
| 342 |
+
"-0.127": 263,
|
| 343 |
+
"0.128": 264,
|
| 344 |
+
"-0.128": 265,
|
| 345 |
+
"0.129": 266,
|
| 346 |
+
"-0.129": 267,
|
| 347 |
+
"0.130": 268,
|
| 348 |
+
"-0.130": 269,
|
| 349 |
+
"0.131": 270,
|
| 350 |
+
"-0.131": 271,
|
| 351 |
+
"0.132": 272,
|
| 352 |
+
"-0.132": 273,
|
| 353 |
+
"0.133": 274,
|
| 354 |
+
"-0.133": 275,
|
| 355 |
+
"0.134": 276,
|
| 356 |
+
"-0.134": 277,
|
| 357 |
+
"0.135": 278,
|
| 358 |
+
"-0.135": 279,
|
| 359 |
+
"0.136": 280,
|
| 360 |
+
"-0.136": 281,
|
| 361 |
+
"0.137": 282,
|
| 362 |
+
"-0.137": 283,
|
| 363 |
+
"0.138": 284,
|
| 364 |
+
"-0.138": 285,
|
| 365 |
+
"0.139": 286,
|
| 366 |
+
"-0.139": 287,
|
| 367 |
+
"0.140": 288,
|
| 368 |
+
"-0.140": 289,
|
| 369 |
+
"0.141": 290,
|
| 370 |
+
"-0.141": 291,
|
| 371 |
+
"0.142": 292,
|
| 372 |
+
"-0.142": 293,
|
| 373 |
+
"0.143": 294,
|
| 374 |
+
"-0.143": 295,
|
| 375 |
+
"0.144": 296,
|
| 376 |
+
"-0.144": 297,
|
| 377 |
+
"0.145": 298,
|
| 378 |
+
"-0.145": 299,
|
| 379 |
+
"0.146": 300,
|
| 380 |
+
"-0.146": 301,
|
| 381 |
+
"0.147": 302,
|
| 382 |
+
"-0.147": 303,
|
| 383 |
+
"0.148": 304,
|
| 384 |
+
"-0.148": 305,
|
| 385 |
+
"0.149": 306,
|
| 386 |
+
"-0.149": 307,
|
| 387 |
+
"0.150": 308,
|
| 388 |
+
"-0.150": 309,
|
| 389 |
+
"0.151": 310,
|
| 390 |
+
"-0.151": 311,
|
| 391 |
+
"0.152": 312,
|
| 392 |
+
"-0.152": 313,
|
| 393 |
+
"0.153": 314,
|
| 394 |
+
"-0.153": 315,
|
| 395 |
+
"0.154": 316,
|
| 396 |
+
"-0.154": 317,
|
| 397 |
+
"0.155": 318,
|
| 398 |
+
"-0.155": 319,
|
| 399 |
+
"0.156": 320,
|
| 400 |
+
"-0.156": 321,
|
| 401 |
+
"0.157": 322,
|
| 402 |
+
"-0.157": 323,
|
| 403 |
+
"0.158": 324,
|
| 404 |
+
"-0.158": 325,
|
| 405 |
+
"0.159": 326,
|
| 406 |
+
"-0.159": 327,
|
| 407 |
+
"0.160": 328,
|
| 408 |
+
"-0.160": 329,
|
| 409 |
+
"0.161": 330,
|
| 410 |
+
"-0.161": 331,
|
| 411 |
+
"0.162": 332,
|
| 412 |
+
"-0.162": 333,
|
| 413 |
+
"0.163": 334,
|
| 414 |
+
"-0.163": 335,
|
| 415 |
+
"0.164": 336,
|
| 416 |
+
"-0.164": 337,
|
| 417 |
+
"0.165": 338,
|
| 418 |
+
"-0.165": 339,
|
| 419 |
+
"0.166": 340,
|
| 420 |
+
"-0.166": 341,
|
| 421 |
+
"0.167": 342,
|
| 422 |
+
"-0.167": 343,
|
| 423 |
+
"0.168": 344,
|
| 424 |
+
"-0.168": 345,
|
| 425 |
+
"0.169": 346,
|
| 426 |
+
"-0.169": 347,
|
| 427 |
+
"0.170": 348,
|
| 428 |
+
"-0.170": 349,
|
| 429 |
+
"0.171": 350,
|
| 430 |
+
"-0.171": 351,
|
| 431 |
+
"0.172": 352,
|
| 432 |
+
"-0.172": 353,
|
| 433 |
+
"0.173": 354,
|
| 434 |
+
"-0.173": 355,
|
| 435 |
+
"0.174": 356,
|
| 436 |
+
"-0.174": 357,
|
| 437 |
+
"0.175": 358,
|
| 438 |
+
"-0.175": 359,
|
| 439 |
+
"0.176": 360,
|
| 440 |
+
"-0.176": 361,
|
| 441 |
+
"0.177": 362,
|
| 442 |
+
"-0.177": 363,
|
| 443 |
+
"0.178": 364,
|
| 444 |
+
"-0.178": 365,
|
| 445 |
+
"0.179": 366,
|
| 446 |
+
"-0.179": 367,
|
| 447 |
+
"0.180": 368,
|
| 448 |
+
"-0.180": 369,
|
| 449 |
+
"0.181": 370,
|
| 450 |
+
"-0.181": 371,
|
| 451 |
+
"0.182": 372,
|
| 452 |
+
"-0.182": 373,
|
| 453 |
+
"0.183": 374,
|
| 454 |
+
"-0.183": 375,
|
| 455 |
+
"0.184": 376,
|
| 456 |
+
"-0.184": 377,
|
| 457 |
+
"0.185": 378,
|
| 458 |
+
"-0.185": 379,
|
| 459 |
+
"0.186": 380,
|
| 460 |
+
"-0.186": 381,
|
| 461 |
+
"0.187": 382,
|
| 462 |
+
"-0.187": 383,
|
| 463 |
+
"0.188": 384,
|
| 464 |
+
"-0.188": 385,
|
| 465 |
+
"0.189": 386,
|
| 466 |
+
"-0.189": 387,
|
| 467 |
+
"0.190": 388,
|
| 468 |
+
"-0.190": 389,
|
| 469 |
+
"0.191": 390,
|
| 470 |
+
"-0.191": 391,
|
| 471 |
+
"0.192": 392,
|
| 472 |
+
"-0.192": 393,
|
| 473 |
+
"0.193": 394,
|
| 474 |
+
"-0.193": 395,
|
| 475 |
+
"0.194": 396,
|
| 476 |
+
"-0.194": 397,
|
| 477 |
+
"0.195": 398,
|
| 478 |
+
"-0.195": 399,
|
| 479 |
+
"0.196": 400,
|
| 480 |
+
"-0.196": 401,
|
| 481 |
+
"0.197": 402,
|
| 482 |
+
"-0.197": 403,
|
| 483 |
+
"0.198": 404,
|
| 484 |
+
"-0.198": 405,
|
| 485 |
+
"0.199": 406,
|
| 486 |
+
"-0.199": 407,
|
| 487 |
+
"0.200": 408,
|
| 488 |
+
"-0.200": 409,
|
| 489 |
+
"0.201": 410,
|
| 490 |
+
"-0.201": 411,
|
| 491 |
+
"0.202": 412,
|
| 492 |
+
"-0.202": 413,
|
| 493 |
+
"0.203": 414,
|
| 494 |
+
"-0.203": 415,
|
| 495 |
+
"0.204": 416,
|
| 496 |
+
"-0.204": 417,
|
| 497 |
+
"0.205": 418,
|
| 498 |
+
"-0.205": 419,
|
| 499 |
+
"0.206": 420,
|
| 500 |
+
"-0.206": 421,
|
| 501 |
+
"0.207": 422,
|
| 502 |
+
"-0.207": 423,
|
| 503 |
+
"0.208": 424,
|
| 504 |
+
"-0.208": 425,
|
| 505 |
+
"0.209": 426,
|
| 506 |
+
"-0.209": 427,
|
| 507 |
+
"0.210": 428,
|
| 508 |
+
"-0.210": 429,
|
| 509 |
+
"0.211": 430,
|
| 510 |
+
"-0.211": 431,
|
| 511 |
+
"0.212": 432,
|
| 512 |
+
"-0.212": 433,
|
| 513 |
+
"0.213": 434,
|
| 514 |
+
"-0.213": 435,
|
| 515 |
+
"0.214": 436,
|
| 516 |
+
"-0.214": 437,
|
| 517 |
+
"0.215": 438,
|
| 518 |
+
"-0.215": 439,
|
| 519 |
+
"0.216": 440,
|
| 520 |
+
"-0.216": 441,
|
| 521 |
+
"0.217": 442,
|
| 522 |
+
"-0.217": 443,
|
| 523 |
+
"0.218": 444,
|
| 524 |
+
"-0.218": 445,
|
| 525 |
+
"0.219": 446,
|
| 526 |
+
"-0.219": 447,
|
| 527 |
+
"0.220": 448,
|
| 528 |
+
"-0.220": 449,
|
| 529 |
+
"0.221": 450,
|
| 530 |
+
"-0.221": 451,
|
| 531 |
+
"0.222": 452,
|
| 532 |
+
"-0.222": 453,
|
| 533 |
+
"0.223": 454,
|
| 534 |
+
"-0.223": 455,
|
| 535 |
+
"0.224": 456,
|
| 536 |
+
"-0.224": 457,
|
| 537 |
+
"0.225": 458,
|
| 538 |
+
"-0.225": 459,
|
| 539 |
+
"0.226": 460,
|
| 540 |
+
"-0.226": 461,
|
| 541 |
+
"0.227": 462,
|
| 542 |
+
"-0.227": 463,
|
| 543 |
+
"0.228": 464,
|
| 544 |
+
"-0.228": 465,
|
| 545 |
+
"0.229": 466,
|
| 546 |
+
"-0.229": 467,
|
| 547 |
+
"0.230": 468,
|
| 548 |
+
"-0.230": 469,
|
| 549 |
+
"0.231": 470,
|
| 550 |
+
"-0.231": 471,
|
| 551 |
+
"0.232": 472,
|
| 552 |
+
"-0.232": 473,
|
| 553 |
+
"0.233": 474,
|
| 554 |
+
"-0.233": 475,
|
| 555 |
+
"0.234": 476,
|
| 556 |
+
"-0.234": 477,
|
| 557 |
+
"0.235": 478,
|
| 558 |
+
"-0.235": 479,
|
| 559 |
+
"0.236": 480,
|
| 560 |
+
"-0.236": 481,
|
| 561 |
+
"0.237": 482,
|
| 562 |
+
"-0.237": 483,
|
| 563 |
+
"0.238": 484,
|
| 564 |
+
"-0.238": 485,
|
| 565 |
+
"0.239": 486,
|
| 566 |
+
"-0.239": 487,
|
| 567 |
+
"0.240": 488,
|
| 568 |
+
"-0.240": 489,
|
| 569 |
+
"0.241": 490,
|
| 570 |
+
"-0.241": 491,
|
| 571 |
+
"0.242": 492,
|
| 572 |
+
"-0.242": 493,
|
| 573 |
+
"0.243": 494,
|
| 574 |
+
"-0.243": 495,
|
| 575 |
+
"0.244": 496,
|
| 576 |
+
"-0.244": 497,
|
| 577 |
+
"0.245": 498,
|
| 578 |
+
"-0.245": 499,
|
| 579 |
+
"0.246": 500,
|
| 580 |
+
"-0.246": 501,
|
| 581 |
+
"0.247": 502,
|
| 582 |
+
"-0.247": 503,
|
| 583 |
+
"0.248": 504,
|
| 584 |
+
"-0.248": 505,
|
| 585 |
+
"0.249": 506,
|
| 586 |
+
"-0.249": 507,
|
| 587 |
+
"0.250": 508,
|
| 588 |
+
"-0.250": 509,
|
| 589 |
+
"0.251": 510,
|
| 590 |
+
"-0.251": 511,
|
| 591 |
+
"0.252": 512,
|
| 592 |
+
"-0.252": 513,
|
| 593 |
+
"0.253": 514,
|
| 594 |
+
"-0.253": 515,
|
| 595 |
+
"0.254": 516,
|
| 596 |
+
"-0.254": 517,
|
| 597 |
+
"0.255": 518,
|
| 598 |
+
"-0.255": 519,
|
| 599 |
+
"0.256": 520,
|
| 600 |
+
"-0.256": 521,
|
| 601 |
+
"0.257": 522,
|
| 602 |
+
"-0.257": 523,
|
| 603 |
+
"0.258": 524,
|
| 604 |
+
"-0.258": 525,
|
| 605 |
+
"0.259": 526,
|
| 606 |
+
"-0.259": 527,
|
| 607 |
+
"0.260": 528,
|
| 608 |
+
"-0.260": 529,
|
| 609 |
+
"0.261": 530,
|
| 610 |
+
"-0.261": 531,
|
| 611 |
+
"0.262": 532,
|
| 612 |
+
"-0.262": 533,
|
| 613 |
+
"0.263": 534,
|
| 614 |
+
"-0.263": 535,
|
| 615 |
+
"0.264": 536,
|
| 616 |
+
"-0.264": 537,
|
| 617 |
+
"0.265": 538,
|
| 618 |
+
"-0.265": 539,
|
| 619 |
+
"0.266": 540,
|
| 620 |
+
"-0.266": 541,
|
| 621 |
+
"0.267": 542,
|
| 622 |
+
"-0.267": 543,
|
| 623 |
+
"0.268": 544,
|
| 624 |
+
"-0.268": 545,
|
| 625 |
+
"0.269": 546,
|
| 626 |
+
"-0.269": 547,
|
| 627 |
+
"0.270": 548,
|
| 628 |
+
"-0.270": 549,
|
| 629 |
+
"0.271": 550,
|
| 630 |
+
"-0.271": 551,
|
| 631 |
+
"0.272": 552,
|
| 632 |
+
"-0.272": 553,
|
| 633 |
+
"0.273": 554,
|
| 634 |
+
"-0.273": 555,
|
| 635 |
+
"0.274": 556,
|
| 636 |
+
"-0.274": 557,
|
| 637 |
+
"0.275": 558,
|
| 638 |
+
"-0.275": 559,
|
| 639 |
+
"0.276": 560,
|
| 640 |
+
"-0.276": 561,
|
| 641 |
+
"0.277": 562,
|
| 642 |
+
"-0.277": 563,
|
| 643 |
+
"0.278": 564,
|
| 644 |
+
"-0.278": 565,
|
| 645 |
+
"0.279": 566,
|
| 646 |
+
"-0.279": 567,
|
| 647 |
+
"0.280": 568,
|
| 648 |
+
"-0.280": 569,
|
| 649 |
+
"0.281": 570,
|
| 650 |
+
"-0.281": 571,
|
| 651 |
+
"0.282": 572,
|
| 652 |
+
"-0.282": 573,
|
| 653 |
+
"0.283": 574,
|
| 654 |
+
"-0.283": 575,
|
| 655 |
+
"0.284": 576,
|
| 656 |
+
"-0.284": 577,
|
| 657 |
+
"0.285": 578,
|
| 658 |
+
"-0.285": 579,
|
| 659 |
+
"0.286": 580,
|
| 660 |
+
"-0.286": 581,
|
| 661 |
+
"0.287": 582,
|
| 662 |
+
"-0.287": 583,
|
| 663 |
+
"0.288": 584,
|
| 664 |
+
"-0.288": 585,
|
| 665 |
+
"0.289": 586,
|
| 666 |
+
"-0.289": 587,
|
| 667 |
+
"0.290": 588,
|
| 668 |
+
"-0.290": 589,
|
| 669 |
+
"0.291": 590,
|
| 670 |
+
"-0.291": 591,
|
| 671 |
+
"0.292": 592,
|
| 672 |
+
"-0.292": 593,
|
| 673 |
+
"0.293": 594,
|
| 674 |
+
"-0.293": 595,
|
| 675 |
+
"0.294": 596,
|
| 676 |
+
"-0.294": 597,
|
| 677 |
+
"0.295": 598,
|
| 678 |
+
"-0.295": 599,
|
| 679 |
+
"0.296": 600,
|
| 680 |
+
"-0.296": 601,
|
| 681 |
+
"0.297": 602,
|
| 682 |
+
"-0.297": 603,
|
| 683 |
+
"0.298": 604,
|
| 684 |
+
"-0.298": 605,
|
| 685 |
+
"0.299": 606,
|
| 686 |
+
"-0.299": 607,
|
| 687 |
+
"0.300": 608
|
| 688 |
+
},
|
| 689 |
+
"unk_token": "<unk>"
|
| 690 |
+
}
|
| 691 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"clean_up_tokenization_spaces": true,
|
| 3 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 4 |
+
"special_tokens": [
|
| 5 |
+
"<s>",
|
| 6 |
+
"<pad>",
|
| 7 |
+
"</s>",
|
| 8 |
+
"<unk>",
|
| 9 |
+
"<cls>",
|
| 10 |
+
"<sep>",
|
| 11 |
+
"<mask>"
|
| 12 |
+
],
|
| 13 |
+
"tokenizer_class": "PreTrainedTokenizerFast"
|
| 14 |
+
}
|