xl1990 commited on
Commit
28a65ed
1 Parent(s): 0f998e0

Training in progress, step 200

Browse files
config.json ADDED
@@ -0,0 +1,431 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "dandelin/vilt-b32-mlm",
3
+ "architectures": [
4
+ "ViltForQuestionAnswering"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "hidden_act": "gelu",
8
+ "hidden_dropout_prob": 0.0,
9
+ "hidden_size": 768,
10
+ "id2label": {
11
+ "0": "8",
12
+ "1": "purple",
13
+ "2": "air",
14
+ "3": "bedroom",
15
+ "4": "yellow",
16
+ "5": "forest",
17
+ "6": "person",
18
+ "7": "big ben",
19
+ "8": "ball",
20
+ "9": "doughnut",
21
+ "10": "fence",
22
+ "11": "5",
23
+ "12": "3",
24
+ "13": "1",
25
+ "14": "snowboarder",
26
+ "15": "down",
27
+ "16": "stand",
28
+ "17": "soccer ball",
29
+ "18": "exit",
30
+ "19": "not sure",
31
+ "20": "curtains",
32
+ "21": "2013",
33
+ "22": "security",
34
+ "23": "chopsticks",
35
+ "24": "resting",
36
+ "25": "hawaii",
37
+ "26": "lying down",
38
+ "27": "6",
39
+ "28": "monitor",
40
+ "29": "tower",
41
+ "30": "unknown",
42
+ "31": "bikes",
43
+ "32": "walking",
44
+ "33": "beagle",
45
+ "34": "windows",
46
+ "35": "calico",
47
+ "36": "don't know",
48
+ "37": "large",
49
+ "38": "shadow",
50
+ "39": "style",
51
+ "40": "tabby",
52
+ "41": "station",
53
+ "42": "blue",
54
+ "43": "plate",
55
+ "44": "7",
56
+ "45": "small",
57
+ "46": "4",
58
+ "47": "wine tasting",
59
+ "48": "cloudy",
60
+ "49": "door",
61
+ "50": "dirt",
62
+ "51": "bus",
63
+ "52": "9:35",
64
+ "53": "car",
65
+ "54": "cross",
66
+ "55": "smile",
67
+ "56": "woman",
68
+ "57": "happy",
69
+ "58": "snowboard",
70
+ "59": "skateboard",
71
+ "60": "7:45",
72
+ "61": "wine",
73
+ "62": "chair",
74
+ "63": "tv",
75
+ "64": "solid",
76
+ "65": "bicycle",
77
+ "66": "soccer",
78
+ "67": "beige",
79
+ "68": "screen",
80
+ "69": "giraffes",
81
+ "70": "tired",
82
+ "71": "natural",
83
+ "72": "many",
84
+ "73": "7:35",
85
+ "74": "canopy",
86
+ "75": "cage",
87
+ "76": "clock",
88
+ "77": "curtain",
89
+ "78": "lanyard",
90
+ "79": "church",
91
+ "80": "shade",
92
+ "81": "orange",
93
+ "82": "talking",
94
+ "83": "wall",
95
+ "84": "green",
96
+ "85": "white",
97
+ "86": "arrow",
98
+ "87": "trees",
99
+ "88": "can't tell",
100
+ "89": "low",
101
+ "90": "table",
102
+ "91": "tent",
103
+ "92": "plain",
104
+ "93": "boy",
105
+ "94": "giraffe",
106
+ "95": "women",
107
+ "96": "gray and black",
108
+ "97": "girl",
109
+ "98": "smiling",
110
+ "99": "black and white",
111
+ "100": "sky",
112
+ "101": "leather",
113
+ "102": "bricks",
114
+ "103": "black",
115
+ "104": "on street",
116
+ "105": "woods",
117
+ "106": "2",
118
+ "107": "park",
119
+ "108": "they aren't",
120
+ "109": "out",
121
+ "110": "roof",
122
+ "111": "train",
123
+ "112": "talking on phone",
124
+ "113": "ice cream",
125
+ "114": "platform",
126
+ "115": "zoo",
127
+ "116": "white and black",
128
+ "117": "not there",
129
+ "118": "at table",
130
+ "119": "street",
131
+ "120": "pink",
132
+ "121": "sleeping",
133
+ "122": "desert",
134
+ "123": "wedding",
135
+ "124": "man",
136
+ "125": "shrimp",
137
+ "126": "2000",
138
+ "127": "nothing",
139
+ "128": "8:35",
140
+ "129": "brown",
141
+ "130": "hair",
142
+ "131": "plastic",
143
+ "132": "french",
144
+ "133": "neon",
145
+ "134": "tan",
146
+ "135": "rack",
147
+ "136": "blue and white",
148
+ "137": "shelter",
149
+ "138": "right",
150
+ "139": "sun",
151
+ "140": "queen",
152
+ "141": "yes",
153
+ "142": "king",
154
+ "143": "no",
155
+ "144": "window",
156
+ "145": "gray",
157
+ "146": "full",
158
+ "147": "lg",
159
+ "148": "red and yellow",
160
+ "149": "cat",
161
+ "150": "2010",
162
+ "151": "cup",
163
+ "152": "name tag",
164
+ "153": "necklace",
165
+ "154": "skiing",
166
+ "155": "blonde",
167
+ "156": "white and blue",
168
+ "157": "outside",
169
+ "158": "photographer",
170
+ "159": "human",
171
+ "160": "stripes",
172
+ "161": "africa",
173
+ "162": "little girl",
174
+ "163": "snowboarding",
175
+ "164": "shadows",
176
+ "165": "watching",
177
+ "166": "red",
178
+ "167": "ground",
179
+ "168": "jeep",
180
+ "169": "fashion",
181
+ "170": "camera",
182
+ "171": "in car",
183
+ "172": "hat",
184
+ "173": "backpack",
185
+ "174": "protection",
186
+ "175": "0",
187
+ "176": "crossing",
188
+ "177": "donut",
189
+ "178": "picnic table",
190
+ "179": "double",
191
+ "180": "10",
192
+ "181": "red and blue",
193
+ "182": "skier",
194
+ "183": "crown",
195
+ "184": "lady",
196
+ "185": "clear",
197
+ "186": "sidewalk",
198
+ "187": "net",
199
+ "188": "brick",
200
+ "189": "birthday",
201
+ "190": "clock tower",
202
+ "191": "on road",
203
+ "192": "bike rack",
204
+ "193": "skateboarding",
205
+ "194": "suv",
206
+ "195": "dog",
207
+ "196": "bicycles",
208
+ "197": "snow",
209
+ "198": "laying down"
210
+ },
211
+ "image_size": 384,
212
+ "initializer_range": 0.02,
213
+ "intermediate_size": 3072,
214
+ "label2id": {
215
+ "0": 175,
216
+ "1": 13,
217
+ "10": 180,
218
+ "2": 106,
219
+ "2000": 126,
220
+ "2010": 150,
221
+ "2013": 21,
222
+ "3": 12,
223
+ "4": 46,
224
+ "5": 11,
225
+ "6": 27,
226
+ "7": 44,
227
+ "7:35": 73,
228
+ "7:45": 60,
229
+ "8": 0,
230
+ "8:35": 128,
231
+ "9:35": 52,
232
+ "africa": 161,
233
+ "air": 2,
234
+ "arrow": 86,
235
+ "at table": 118,
236
+ "backpack": 173,
237
+ "ball": 8,
238
+ "beagle": 33,
239
+ "bedroom": 3,
240
+ "beige": 67,
241
+ "bicycle": 65,
242
+ "bicycles": 196,
243
+ "big ben": 7,
244
+ "bike rack": 192,
245
+ "bikes": 31,
246
+ "birthday": 189,
247
+ "black": 103,
248
+ "black and white": 99,
249
+ "blonde": 155,
250
+ "blue": 42,
251
+ "blue and white": 136,
252
+ "boy": 93,
253
+ "brick": 188,
254
+ "bricks": 102,
255
+ "brown": 129,
256
+ "bus": 51,
257
+ "cage": 75,
258
+ "calico": 35,
259
+ "camera": 170,
260
+ "can't tell": 88,
261
+ "canopy": 74,
262
+ "car": 53,
263
+ "cat": 149,
264
+ "chair": 62,
265
+ "chopsticks": 23,
266
+ "church": 79,
267
+ "clear": 185,
268
+ "clock": 76,
269
+ "clock tower": 190,
270
+ "cloudy": 48,
271
+ "cross": 54,
272
+ "crossing": 176,
273
+ "crown": 183,
274
+ "cup": 151,
275
+ "curtain": 77,
276
+ "curtains": 20,
277
+ "desert": 122,
278
+ "dirt": 50,
279
+ "dog": 195,
280
+ "don't know": 36,
281
+ "donut": 177,
282
+ "door": 49,
283
+ "double": 179,
284
+ "doughnut": 9,
285
+ "down": 15,
286
+ "exit": 18,
287
+ "fashion": 169,
288
+ "fence": 10,
289
+ "forest": 5,
290
+ "french": 132,
291
+ "full": 146,
292
+ "giraffe": 94,
293
+ "giraffes": 69,
294
+ "girl": 97,
295
+ "gray": 145,
296
+ "gray and black": 96,
297
+ "green": 84,
298
+ "ground": 167,
299
+ "hair": 130,
300
+ "happy": 57,
301
+ "hat": 172,
302
+ "hawaii": 25,
303
+ "human": 159,
304
+ "ice cream": 113,
305
+ "in car": 171,
306
+ "jeep": 168,
307
+ "king": 142,
308
+ "lady": 184,
309
+ "lanyard": 78,
310
+ "large": 37,
311
+ "laying down": 198,
312
+ "leather": 101,
313
+ "lg": 147,
314
+ "little girl": 162,
315
+ "low": 89,
316
+ "lying down": 26,
317
+ "man": 124,
318
+ "many": 72,
319
+ "monitor": 28,
320
+ "name tag": 152,
321
+ "natural": 71,
322
+ "necklace": 153,
323
+ "neon": 133,
324
+ "net": 187,
325
+ "no": 143,
326
+ "not sure": 19,
327
+ "not there": 117,
328
+ "nothing": 127,
329
+ "on road": 191,
330
+ "on street": 104,
331
+ "orange": 81,
332
+ "out": 109,
333
+ "outside": 157,
334
+ "park": 107,
335
+ "person": 6,
336
+ "photographer": 158,
337
+ "picnic table": 178,
338
+ "pink": 120,
339
+ "plain": 92,
340
+ "plastic": 131,
341
+ "plate": 43,
342
+ "platform": 114,
343
+ "protection": 174,
344
+ "purple": 1,
345
+ "queen": 140,
346
+ "rack": 135,
347
+ "red": 166,
348
+ "red and blue": 181,
349
+ "red and yellow": 148,
350
+ "resting": 24,
351
+ "right": 138,
352
+ "roof": 110,
353
+ "screen": 68,
354
+ "security": 22,
355
+ "shade": 80,
356
+ "shadow": 38,
357
+ "shadows": 164,
358
+ "shelter": 137,
359
+ "shrimp": 125,
360
+ "sidewalk": 186,
361
+ "skateboard": 59,
362
+ "skateboarding": 193,
363
+ "skier": 182,
364
+ "skiing": 154,
365
+ "sky": 100,
366
+ "sleeping": 121,
367
+ "small": 45,
368
+ "smile": 55,
369
+ "smiling": 98,
370
+ "snow": 197,
371
+ "snowboard": 58,
372
+ "snowboarder": 14,
373
+ "snowboarding": 163,
374
+ "soccer": 66,
375
+ "soccer ball": 17,
376
+ "solid": 64,
377
+ "stand": 16,
378
+ "station": 41,
379
+ "street": 119,
380
+ "stripes": 160,
381
+ "style": 39,
382
+ "sun": 139,
383
+ "suv": 194,
384
+ "tabby": 40,
385
+ "table": 90,
386
+ "talking": 82,
387
+ "talking on phone": 112,
388
+ "tan": 134,
389
+ "tent": 91,
390
+ "they aren't": 108,
391
+ "tired": 70,
392
+ "tower": 29,
393
+ "train": 111,
394
+ "trees": 87,
395
+ "tv": 63,
396
+ "unknown": 30,
397
+ "walking": 32,
398
+ "wall": 83,
399
+ "watching": 165,
400
+ "wedding": 123,
401
+ "white": 85,
402
+ "white and black": 116,
403
+ "white and blue": 156,
404
+ "window": 144,
405
+ "windows": 34,
406
+ "wine": 61,
407
+ "wine tasting": 47,
408
+ "woman": 56,
409
+ "women": 95,
410
+ "woods": 105,
411
+ "yellow": 4,
412
+ "yes": 141,
413
+ "zoo": 115
414
+ },
415
+ "layer_norm_eps": 1e-12,
416
+ "max_image_length": -1,
417
+ "max_position_embeddings": 40,
418
+ "modality_type_vocab_size": 2,
419
+ "model_type": "vilt",
420
+ "num_attention_heads": 12,
421
+ "num_channels": 3,
422
+ "num_hidden_layers": 12,
423
+ "num_images": -1,
424
+ "patch_size": 32,
425
+ "qkv_bias": true,
426
+ "tie_word_embeddings": false,
427
+ "torch_dtype": "float32",
428
+ "transformers_version": "4.38.2",
429
+ "type_vocab_size": 2,
430
+ "vocab_size": 30522
431
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dda401260108f166089a552cbe3b4d16a348a65837d69ccd86c1f24ce3c447b7
3
+ size 452365324
preprocessor_config.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_pad": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.5,
8
+ 0.5,
9
+ 0.5
10
+ ],
11
+ "image_processor_type": "ViltImageProcessor",
12
+ "image_std": [
13
+ 0.5,
14
+ 0.5,
15
+ 0.5
16
+ ],
17
+ "processor_class": "ViltProcessor",
18
+ "resample": 3,
19
+ "rescale_factor": 0.00392156862745098,
20
+ "size": {
21
+ "shortest_edge": 384
22
+ },
23
+ "size_divisor": 32
24
+ }
special_tokens_map.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "cls_token": {
3
+ "content": "[CLS]",
4
+ "lstrip": false,
5
+ "normalized": false,
6
+ "rstrip": false,
7
+ "single_word": false
8
+ },
9
+ "mask_token": {
10
+ "content": "[MASK]",
11
+ "lstrip": false,
12
+ "normalized": false,
13
+ "rstrip": false,
14
+ "single_word": false
15
+ },
16
+ "pad_token": {
17
+ "content": "[PAD]",
18
+ "lstrip": false,
19
+ "normalized": false,
20
+ "rstrip": false,
21
+ "single_word": false
22
+ },
23
+ "sep_token": {
24
+ "content": "[SEP]",
25
+ "lstrip": false,
26
+ "normalized": false,
27
+ "rstrip": false,
28
+ "single_word": false
29
+ },
30
+ "unk_token": {
31
+ "content": "[UNK]",
32
+ "lstrip": false,
33
+ "normalized": false,
34
+ "rstrip": false,
35
+ "single_word": false
36
+ }
37
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,56 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "added_tokens_decoder": {
3
+ "0": {
4
+ "content": "[PAD]",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false,
9
+ "special": true
10
+ },
11
+ "100": {
12
+ "content": "[UNK]",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false,
17
+ "special": true
18
+ },
19
+ "101": {
20
+ "content": "[CLS]",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false,
25
+ "special": true
26
+ },
27
+ "102": {
28
+ "content": "[SEP]",
29
+ "lstrip": false,
30
+ "normalized": false,
31
+ "rstrip": false,
32
+ "single_word": false,
33
+ "special": true
34
+ },
35
+ "103": {
36
+ "content": "[MASK]",
37
+ "lstrip": false,
38
+ "normalized": false,
39
+ "rstrip": false,
40
+ "single_word": false,
41
+ "special": true
42
+ }
43
+ },
44
+ "clean_up_tokenization_spaces": true,
45
+ "cls_token": "[CLS]",
46
+ "do_lower_case": true,
47
+ "mask_token": "[MASK]",
48
+ "model_max_length": 40,
49
+ "pad_token": "[PAD]",
50
+ "processor_class": "ViltProcessor",
51
+ "sep_token": "[SEP]",
52
+ "strip_accents": null,
53
+ "tokenize_chinese_chars": true,
54
+ "tokenizer_class": "BertTokenizer",
55
+ "unk_token": "[UNK]"
56
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5226b0203b1f5c4288cc16828cdbf55f3df1127ac8de41e1b3362275c9bee95f
3
+ size 4856
vocab.txt ADDED
The diff for this file is too large to render. See raw diff