hf-transformers-bot commited on
Commit
fe9a3c0
·
verified ·
1 Parent(s): 0cc8365

Update tiny models for MarkupLMModel

Browse files
config.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "MarkupLMModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 32,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 37,
15
+ "layer_norm_eps": 1e-12,
16
+ "max_depth": 10,
17
+ "max_position_embeddings": 512,
18
+ "max_xpath_subs_unit_embeddings": 30,
19
+ "max_xpath_tag_unit_embeddings": 20,
20
+ "model_type": "markuplm",
21
+ "num_attention_heads": 4,
22
+ "num_hidden_layers": 2,
23
+ "pad_token_id": 1,
24
+ "subs_pad_id": 2,
25
+ "tag_pad_id": 2,
26
+ "transformers_version": "5.16.0.dev0",
27
+ "type_vocab_size": 16,
28
+ "use_cache": true,
29
+ "vocab_size": 50267,
30
+ "xpath_unit_hidden_size": 32
31
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:42e1e14a396be2f3c5d38029a3b39abfc0b46231e76987c2d4faf332e597eccc
3
+ size 6853920
preprocessor_config.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "feature_extractor_type": "MarkupLMFeatureExtractor"
3
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,240 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "backend": "tokenizers",
4
+ "bos_token": "<s>",
5
+ "cls_token": "<s>",
6
+ "eos_token": "</s>",
7
+ "errors": "replace",
8
+ "from_slow": true,
9
+ "is_local": true,
10
+ "local_files_only": false,
11
+ "mask_token": "<mask>",
12
+ "max_depth": 50,
13
+ "max_width": 1000,
14
+ "model_max_length": 512,
15
+ "only_label_first_subword": true,
16
+ "pad_token": "<pad>",
17
+ "pad_token_label": -100,
18
+ "pad_width": 1001,
19
+ "sep_token": "</s>",
20
+ "tags_dict": {
21
+ "a": 0,
22
+ "abbr": 1,
23
+ "acronym": 2,
24
+ "address": 3,
25
+ "altGlyph": 4,
26
+ "altGlyphDef": 5,
27
+ "altGlyphItem": 6,
28
+ "animate": 7,
29
+ "animateColor": 8,
30
+ "animateMotion": 9,
31
+ "animateTransform": 10,
32
+ "applet": 11,
33
+ "area": 12,
34
+ "article": 13,
35
+ "aside": 14,
36
+ "audio": 15,
37
+ "b": 16,
38
+ "base": 17,
39
+ "basefont": 18,
40
+ "bdi": 19,
41
+ "bdo": 20,
42
+ "bgsound": 21,
43
+ "big": 22,
44
+ "blink": 23,
45
+ "blockquote": 24,
46
+ "body": 25,
47
+ "br": 26,
48
+ "button": 27,
49
+ "canvas": 28,
50
+ "caption": 29,
51
+ "center": 30,
52
+ "circle": 31,
53
+ "cite": 32,
54
+ "clipPath": 33,
55
+ "code": 34,
56
+ "col": 35,
57
+ "colgroup": 36,
58
+ "color-profile": 37,
59
+ "content": 38,
60
+ "cursor": 39,
61
+ "data": 40,
62
+ "datalist": 41,
63
+ "dd": 42,
64
+ "defs": 43,
65
+ "del": 44,
66
+ "desc": 45,
67
+ "details": 46,
68
+ "dfn": 47,
69
+ "dialog": 48,
70
+ "dir": 49,
71
+ "div": 50,
72
+ "dl": 51,
73
+ "dt": 52,
74
+ "ellipse": 53,
75
+ "em": 54,
76
+ "embed": 55,
77
+ "feBlend": 56,
78
+ "feColorMatrix": 57,
79
+ "feComponentTransfer": 58,
80
+ "feComposite": 59,
81
+ "feConvolveMatrix": 60,
82
+ "feDiffuseLighting": 61,
83
+ "feDisplacementMap": 62,
84
+ "feDistantLight": 63,
85
+ "feFlood": 64,
86
+ "feFuncA": 65,
87
+ "feFuncB": 66,
88
+ "feFuncG": 67,
89
+ "feFuncR": 68,
90
+ "feGaussianBlur": 69,
91
+ "feImage": 70,
92
+ "feMerge": 71,
93
+ "feMergeNode": 72,
94
+ "feMorphology": 73,
95
+ "feOffset": 74,
96
+ "fePointLight": 75,
97
+ "feSpecularLighting": 76,
98
+ "feSpotLight": 77,
99
+ "feTile": 78,
100
+ "feTurbulence": 79,
101
+ "fieldset": 80,
102
+ "figcaption": 81,
103
+ "figure": 82,
104
+ "filter": 83,
105
+ "font": 89,
106
+ "font-face": 88,
107
+ "font-face-format": 84,
108
+ "font-face-name": 85,
109
+ "font-face-src": 86,
110
+ "font-face-uri": 87,
111
+ "footer": 90,
112
+ "foreignObject": 91,
113
+ "form": 92,
114
+ "frame": 93,
115
+ "frameset": 94,
116
+ "g": 95,
117
+ "glyph": 96,
118
+ "glyphRef": 97,
119
+ "h1": 98,
120
+ "h2": 99,
121
+ "h3": 100,
122
+ "h4": 101,
123
+ "h5": 102,
124
+ "h6": 103,
125
+ "head": 104,
126
+ "header": 105,
127
+ "hgroup": 106,
128
+ "hkern": 107,
129
+ "hr": 108,
130
+ "html": 109,
131
+ "i": 110,
132
+ "iframe": 111,
133
+ "image": 112,
134
+ "img": 113,
135
+ "input": 114,
136
+ "ins": 115,
137
+ "kbd": 116,
138
+ "keygen": 117,
139
+ "label": 118,
140
+ "legend": 119,
141
+ "li": 120,
142
+ "line": 121,
143
+ "linearGradient": 122,
144
+ "link": 123,
145
+ "main": 124,
146
+ "map": 125,
147
+ "mark": 126,
148
+ "marker": 127,
149
+ "marquee": 128,
150
+ "mask": 129,
151
+ "math": 130,
152
+ "menu": 131,
153
+ "menuitem": 132,
154
+ "meta": 133,
155
+ "metadata": 134,
156
+ "meter": 135,
157
+ "missing-glyph": 136,
158
+ "mpath": 137,
159
+ "nav": 138,
160
+ "nobr": 139,
161
+ "noembed": 140,
162
+ "noframes": 141,
163
+ "noscript": 142,
164
+ "object": 143,
165
+ "ol": 144,
166
+ "optgroup": 145,
167
+ "option": 146,
168
+ "output": 147,
169
+ "p": 148,
170
+ "param": 149,
171
+ "path": 150,
172
+ "pattern": 151,
173
+ "picture": 152,
174
+ "plaintext": 153,
175
+ "polygon": 154,
176
+ "polyline": 155,
177
+ "portal": 156,
178
+ "pre": 157,
179
+ "progress": 158,
180
+ "q": 159,
181
+ "radialGradient": 160,
182
+ "rb": 161,
183
+ "rect": 162,
184
+ "rp": 163,
185
+ "rt": 164,
186
+ "rtc": 165,
187
+ "ruby": 166,
188
+ "s": 167,
189
+ "samp": 168,
190
+ "script": 169,
191
+ "section": 170,
192
+ "select": 171,
193
+ "set": 172,
194
+ "shadow": 173,
195
+ "slot": 174,
196
+ "small": 175,
197
+ "source": 176,
198
+ "spacer": 177,
199
+ "span": 178,
200
+ "stop": 179,
201
+ "strike": 180,
202
+ "strong": 181,
203
+ "style": 182,
204
+ "sub": 183,
205
+ "summary": 184,
206
+ "sup": 185,
207
+ "svg": 186,
208
+ "switch": 187,
209
+ "symbol": 188,
210
+ "table": 189,
211
+ "tbody": 190,
212
+ "td": 191,
213
+ "template": 192,
214
+ "text": 193,
215
+ "textPath": 194,
216
+ "textarea": 195,
217
+ "tfoot": 196,
218
+ "th": 197,
219
+ "thead": 198,
220
+ "time": 199,
221
+ "title": 200,
222
+ "tr": 201,
223
+ "track": 202,
224
+ "tref": 203,
225
+ "tspan": 204,
226
+ "tt": 205,
227
+ "u": 206,
228
+ "ul": 207,
229
+ "use": 208,
230
+ "var": 209,
231
+ "video": 210,
232
+ "view": 211,
233
+ "vkern": 212,
234
+ "wbr": 213,
235
+ "xmp": 214
236
+ },
237
+ "tokenizer_class": "MarkupLMTokenizer",
238
+ "trim_offsets": false,
239
+ "unk_token": "<unk>"
240
+ }