Upload tokenizer.json with huggingface_hub
Browse files- tokenizer.json +161 -0
tokenizer.json
ADDED
|
@@ -0,0 +1,161 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
| 4 |
+
"padding": null,
|
| 5 |
+
"added_tokens": [
|
| 6 |
+
{
|
| 7 |
+
"id": 0,
|
| 8 |
+
"content": "[UNK]",
|
| 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": "[EOS]",
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"lstrip": false,
|
| 29 |
+
"rstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"special": true
|
| 32 |
+
}
|
| 33 |
+
],
|
| 34 |
+
"normalizer": null,
|
| 35 |
+
"pre_tokenizer": {
|
| 36 |
+
"type": "WhitespaceSplit"
|
| 37 |
+
},
|
| 38 |
+
"post_processor": null,
|
| 39 |
+
"decoder": null,
|
| 40 |
+
"model": {
|
| 41 |
+
"type": "WordLevel",
|
| 42 |
+
"vocab": {
|
| 43 |
+
"[UNK]": 0,
|
| 44 |
+
"[PAD]": 1,
|
| 45 |
+
"[EOS]": 2,
|
| 46 |
+
"TIME_DELTA=8": 3,
|
| 47 |
+
"TIME_DELTA=4": 4,
|
| 48 |
+
"BAR_END": 5,
|
| 49 |
+
"BAR_START": 6,
|
| 50 |
+
"DENSITY=2": 7,
|
| 51 |
+
"DENSITY=3": 8,
|
| 52 |
+
"NEXT": 9,
|
| 53 |
+
"NOTE_OFF=62": 10,
|
| 54 |
+
"NOTE_ON=62": 11,
|
| 55 |
+
"NOTE_OFF=67": 12,
|
| 56 |
+
"NOTE_ON=67": 13,
|
| 57 |
+
"NOTE_OFF=65": 14,
|
| 58 |
+
"NOTE_ON=65": 15,
|
| 59 |
+
"NOTE_OFF=64": 16,
|
| 60 |
+
"NOTE_ON=64": 17,
|
| 61 |
+
"DENSITY=4": 18,
|
| 62 |
+
"TIME_DELTA=16": 19,
|
| 63 |
+
"NOTE_OFF=60": 20,
|
| 64 |
+
"NOTE_ON=60": 21,
|
| 65 |
+
"NOTE_OFF=66": 22,
|
| 66 |
+
"NOTE_ON=66": 23,
|
| 67 |
+
"DENSITY=0": 24,
|
| 68 |
+
"NOTE_OFF=63": 25,
|
| 69 |
+
"NOTE_ON=63": 26,
|
| 70 |
+
"NOTE_OFF=69": 27,
|
| 71 |
+
"NOTE_ON=69": 28,
|
| 72 |
+
"NOTE_OFF=68": 29,
|
| 73 |
+
"NOTE_ON=68": 30,
|
| 74 |
+
"NOTE_OFF=61": 31,
|
| 75 |
+
"NOTE_ON=61": 32,
|
| 76 |
+
"NOTE_OFF=59": 33,
|
| 77 |
+
"NOTE_ON=59": 34,
|
| 78 |
+
"NOTE_OFF=58": 35,
|
| 79 |
+
"NOTE_ON=58": 36,
|
| 80 |
+
"NOTE_OFF=57": 37,
|
| 81 |
+
"NOTE_ON=57": 38,
|
| 82 |
+
"NOTE_OFF=55": 39,
|
| 83 |
+
"NOTE_ON=55": 40,
|
| 84 |
+
"NOTE_OFF=70": 41,
|
| 85 |
+
"NOTE_ON=70": 42,
|
| 86 |
+
"NOTE_OFF=56": 43,
|
| 87 |
+
"NOTE_ON=56": 44,
|
| 88 |
+
"NOTE_OFF=71": 45,
|
| 89 |
+
"NOTE_ON=71": 46,
|
| 90 |
+
"NOTE_OFF=53": 47,
|
| 91 |
+
"NOTE_ON=53": 48,
|
| 92 |
+
"NOTE_OFF=54": 49,
|
| 93 |
+
"NOTE_ON=54": 50,
|
| 94 |
+
"NOTE_OFF=52": 51,
|
| 95 |
+
"NOTE_ON=52": 52,
|
| 96 |
+
"NOTE_OFF=72": 53,
|
| 97 |
+
"NOTE_ON=72": 54,
|
| 98 |
+
"NOTE_OFF=50": 55,
|
| 99 |
+
"NOTE_ON=50": 56,
|
| 100 |
+
"NOTE_OFF=51": 57,
|
| 101 |
+
"NOTE_ON=51": 58,
|
| 102 |
+
"NOTE_OFF=73": 59,
|
| 103 |
+
"NOTE_ON=73": 60,
|
| 104 |
+
"NOTE_OFF=48": 61,
|
| 105 |
+
"NOTE_ON=48": 62,
|
| 106 |
+
"NOTE_OFF=74": 63,
|
| 107 |
+
"NOTE_ON=74": 64,
|
| 108 |
+
"NOTE_OFF=49": 65,
|
| 109 |
+
"NOTE_ON=49": 66,
|
| 110 |
+
"NOTE_OFF=47": 67,
|
| 111 |
+
"NOTE_ON=47": 68,
|
| 112 |
+
"TIME_DELTA=2": 69,
|
| 113 |
+
"NOTE_OFF=46": 70,
|
| 114 |
+
"NOTE_ON=46": 71,
|
| 115 |
+
"NOTE_OFF=75": 72,
|
| 116 |
+
"NOTE_ON=75": 73,
|
| 117 |
+
"NOTE_OFF=76": 74,
|
| 118 |
+
"NOTE_ON=76": 75,
|
| 119 |
+
"NOTE_OFF=45": 76,
|
| 120 |
+
"NOTE_ON=45": 77,
|
| 121 |
+
"TIME_DELTA=24": 78,
|
| 122 |
+
"TIME_DELTA=12": 79,
|
| 123 |
+
"NOTE_OFF=44": 80,
|
| 124 |
+
"NOTE_ON=44": 81,
|
| 125 |
+
"NOTE_OFF=43": 82,
|
| 126 |
+
"NOTE_ON=43": 83,
|
| 127 |
+
"NOTE_OFF=77": 84,
|
| 128 |
+
"NOTE_ON=77": 85,
|
| 129 |
+
"GARLAND_END": 86,
|
| 130 |
+
"GARLAND_START": 87,
|
| 131 |
+
"NOTE_OFF=78": 88,
|
| 132 |
+
"NOTE_ON=78": 89,
|
| 133 |
+
"NOTE_OFF=41": 90,
|
| 134 |
+
"NOTE_ON=41": 91,
|
| 135 |
+
"NOTE_OFF=42": 92,
|
| 136 |
+
"NOTE_ON=42": 93,
|
| 137 |
+
"NOTE_OFF=79": 94,
|
| 138 |
+
"NOTE_ON=79": 95,
|
| 139 |
+
"NOTE_OFF=40": 96,
|
| 140 |
+
"NOTE_ON=40": 97,
|
| 141 |
+
"NOTE_OFF=39": 98,
|
| 142 |
+
"NOTE_ON=39": 99,
|
| 143 |
+
"TIME_DELTA=32": 100,
|
| 144 |
+
"TIME_DELTA=6": 101,
|
| 145 |
+
"NOTE_OFF=38": 102,
|
| 146 |
+
"NOTE_ON=38": 103,
|
| 147 |
+
"NOTE_OFF=80": 104,
|
| 148 |
+
"NOTE_ON=80": 105,
|
| 149 |
+
"NOTE_OFF=37": 106,
|
| 150 |
+
"NOTE_ON=37": 107,
|
| 151 |
+
"NOTE_OFF=81": 108,
|
| 152 |
+
"NOTE_ON=81": 109,
|
| 153 |
+
"NOTE_OFF=36": 110,
|
| 154 |
+
"NOTE_ON=36": 111,
|
| 155 |
+
"TIME_DELTA=14": 112,
|
| 156 |
+
"TIME_DELTA=20": 113,
|
| 157 |
+
"TIME_DELTA=10": 114
|
| 158 |
+
},
|
| 159 |
+
"unk_token": "[UNK]"
|
| 160 |
+
}
|
| 161 |
+
}
|