Upload tokenizer.json with huggingface_hub
Browse files- .gitattributes +1 -0
- tokenizer.json +3 -176
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
tokenizer.json
CHANGED
|
@@ -1,176 +1,3 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
"padding": null,
|
| 5 |
-
"added_tokens": [],
|
| 6 |
-
"normalizer": null,
|
| 7 |
-
"pre_tokenizer": {
|
| 8 |
-
"type": "ByteLevel",
|
| 9 |
-
"add_prefix_space": false,
|
| 10 |
-
"trim_offsets": true,
|
| 11 |
-
"use_regex": true
|
| 12 |
-
},
|
| 13 |
-
"post_processor": null,
|
| 14 |
-
"decoder": {
|
| 15 |
-
"type": "ByteLevel",
|
| 16 |
-
"add_prefix_space": true,
|
| 17 |
-
"trim_offsets": true,
|
| 18 |
-
"use_regex": true
|
| 19 |
-
},
|
| 20 |
-
"model": {
|
| 21 |
-
"type": "WordLevel",
|
| 22 |
-
"vocab": {
|
| 23 |
-
"<unk>": 0,
|
| 24 |
-
"<|endoftext|>": 1,
|
| 25 |
-
"<|im_start|>": 2,
|
| 26 |
-
"<|im_end|>": 3,
|
| 27 |
-
"<|object_ref_start|>": 4,
|
| 28 |
-
"<|object_ref_end|>": 5,
|
| 29 |
-
"<|box_start|>": 6,
|
| 30 |
-
"<|box_end|>": 7,
|
| 31 |
-
"<|quad_start|>": 8,
|
| 32 |
-
"<|quad_end|>": 9,
|
| 33 |
-
"<|vision_start|>": 10,
|
| 34 |
-
"<|vision_end|>": 11,
|
| 35 |
-
"<|vision_pad|>": 12,
|
| 36 |
-
"<|image_pad|>": 13,
|
| 37 |
-
"<|video_pad|>": 14,
|
| 38 |
-
"<tool_call>": 15,
|
| 39 |
-
"</tool_call>": 16,
|
| 40 |
-
"<|fim_prefix|>": 17,
|
| 41 |
-
"<|fim_middle|>": 18,
|
| 42 |
-
"<|fim_suffix|>": 19,
|
| 43 |
-
"<|fim_pad|>": 20,
|
| 44 |
-
"<|repo_name|>": 21,
|
| 45 |
-
"<|file_sep|>": 22,
|
| 46 |
-
"<image>": 23,
|
| 47 |
-
"</image>": 24,
|
| 48 |
-
"<ref>": 25,
|
| 49 |
-
"</ref>": 26,
|
| 50 |
-
"<box>": 27,
|
| 51 |
-
"</box>": 28,
|
| 52 |
-
"<quad>": 29,
|
| 53 |
-
"</quad>": 30,
|
| 54 |
-
"<point>": 31,
|
| 55 |
-
"</point>": 32,
|
| 56 |
-
"<slice>": 33,
|
| 57 |
-
"</slice>": 34,
|
| 58 |
-
"<image_id>": 35,
|
| 59 |
-
"</image_id>": 36,
|
| 60 |
-
"<unit>": 37,
|
| 61 |
-
"</unit>": 38,
|
| 62 |
-
"<asr>": 39,
|
| 63 |
-
"</asr>": 40,
|
| 64 |
-
"<query>": 41,
|
| 65 |
-
"</query>": 42,
|
| 66 |
-
"<|audio_start|>": 43,
|
| 67 |
-
"<|audio|>": 44,
|
| 68 |
-
"<|audio_end|>": 45,
|
| 69 |
-
"<|spk_bos|>": 46,
|
| 70 |
-
"<|spk|>": 47,
|
| 71 |
-
"<|spk_eos|>": 48,
|
| 72 |
-
"<|tts_bos|>": 49,
|
| 73 |
-
"<|tts_eos|>": 50,
|
| 74 |
-
"<|listen|>": 51,
|
| 75 |
-
"<|speak|>": 52,
|
| 76 |
-
"<|interrupt|>": 53,
|
| 77 |
-
"<|vad_start|>": 54,
|
| 78 |
-
"<|vad_end|>": 55,
|
| 79 |
-
"<reserved_43>": 56,
|
| 80 |
-
"<reserved_53>": 57,
|
| 81 |
-
"!": 58,
|
| 82 |
-
"\"": 59,
|
| 83 |
-
"#": 60,
|
| 84 |
-
"$": 61,
|
| 85 |
-
"%": 62,
|
| 86 |
-
"&": 63,
|
| 87 |
-
"'": 64,
|
| 88 |
-
"(": 65,
|
| 89 |
-
")": 66,
|
| 90 |
-
"*": 67,
|
| 91 |
-
"+": 68,
|
| 92 |
-
",": 69,
|
| 93 |
-
"-": 70,
|
| 94 |
-
".": 71,
|
| 95 |
-
"/": 72,
|
| 96 |
-
"0": 73,
|
| 97 |
-
"1": 74,
|
| 98 |
-
"2": 75,
|
| 99 |
-
"3": 76,
|
| 100 |
-
"4": 77,
|
| 101 |
-
"5": 78,
|
| 102 |
-
"6": 79,
|
| 103 |
-
"7": 80,
|
| 104 |
-
"8": 81,
|
| 105 |
-
"9": 82,
|
| 106 |
-
":": 83,
|
| 107 |
-
";": 84,
|
| 108 |
-
"<": 85,
|
| 109 |
-
"=": 86,
|
| 110 |
-
">": 87,
|
| 111 |
-
"?": 88,
|
| 112 |
-
"@": 89,
|
| 113 |
-
"A": 90,
|
| 114 |
-
"B": 91,
|
| 115 |
-
"C": 92,
|
| 116 |
-
"D": 93,
|
| 117 |
-
"E": 94,
|
| 118 |
-
"F": 95,
|
| 119 |
-
"G": 96,
|
| 120 |
-
"H": 97,
|
| 121 |
-
"I": 98,
|
| 122 |
-
"J": 99,
|
| 123 |
-
"K": 100,
|
| 124 |
-
"L": 101,
|
| 125 |
-
"M": 102,
|
| 126 |
-
"N": 103,
|
| 127 |
-
"O": 104,
|
| 128 |
-
"P": 105,
|
| 129 |
-
"Q": 106,
|
| 130 |
-
"R": 107,
|
| 131 |
-
"S": 108,
|
| 132 |
-
"T": 109,
|
| 133 |
-
"U": 110,
|
| 134 |
-
"V": 111,
|
| 135 |
-
"W": 112,
|
| 136 |
-
"X": 113,
|
| 137 |
-
"Y": 114,
|
| 138 |
-
"Z": 115,
|
| 139 |
-
"[": 116,
|
| 140 |
-
"\\": 117,
|
| 141 |
-
"]": 118,
|
| 142 |
-
"^": 119,
|
| 143 |
-
"_": 120,
|
| 144 |
-
"`": 121,
|
| 145 |
-
"a": 122,
|
| 146 |
-
"b": 123,
|
| 147 |
-
"c": 124,
|
| 148 |
-
"d": 125,
|
| 149 |
-
"e": 126,
|
| 150 |
-
"f": 127,
|
| 151 |
-
"g": 128,
|
| 152 |
-
"h": 129,
|
| 153 |
-
"i": 130,
|
| 154 |
-
"j": 131,
|
| 155 |
-
"k": 132,
|
| 156 |
-
"l": 133,
|
| 157 |
-
"m": 134,
|
| 158 |
-
"n": 135,
|
| 159 |
-
"o": 136,
|
| 160 |
-
"p": 137,
|
| 161 |
-
"q": 138,
|
| 162 |
-
"r": 139,
|
| 163 |
-
"s": 140,
|
| 164 |
-
"t": 141,
|
| 165 |
-
"u": 142,
|
| 166 |
-
"v": 143,
|
| 167 |
-
"w": 144,
|
| 168 |
-
"x": 145,
|
| 169 |
-
"y": 146,
|
| 170 |
-
"z": 147,
|
| 171 |
-
"{": 148,
|
| 172 |
-
"|": 149
|
| 173 |
-
},
|
| 174 |
-
"unk_token": "<unk>"
|
| 175 |
-
}
|
| 176 |
-
}
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e43057e8380937b8edf2c9ea6ad34a3875d89c390e97024dfac307f3a4a11321
|
| 3 |
+
size 11428583
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|