Hugging Face's logo

Jaguar7788
/
vilt_finetuned_200

Visual Question Answering
Transformers
TensorBoard
Safetensors
vilt
Inference Endpoints
Model card Files Metrics Community
Jaguar7788 commited on
Commit
8131229
•
1 Parent(s): 8032d0f

Training in progress, step 200

Browse files
Files changed (9) hide show
  1. config.json +431 -0
  2. model.safetensors +3 -0
  3. preprocessor_config.json +40 -0
  4. runs/Jun06_07-01-39_Jacks-MBP/events.out.tfevents.1717614101.Jacks-MBP.16824.0 +3 -0
  5. special_tokens_map.json +37 -0
  6. tokenizer.json +0 -0
  7. tokenizer_config.json +56 -0
  8. training_args.bin +3 -0
  9. vocab.txt +0 -0
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": "skiing",
12
+ "1": "security",
13
+ "2": "bikes",
14
+ "3": "3",
15
+ "4": "stand",
16
+ "5": "church",
17
+ "6": "in car",
18
+ "7": "snowboarder",
19
+ "8": "women",
20
+ "9": "clock",
21
+ "10": "blue and white",
22
+ "11": "little girl",
23
+ "12": "0",
24
+ "13": "soccer",
25
+ "14": "tabby",
26
+ "15": "not sure",
27
+ "16": "sleeping",
28
+ "17": "on road",
29
+ "18": "chopsticks",
30
+ "19": "desert",
31
+ "20": "cup",
32
+ "21": "clear",
33
+ "22": "dirt",
34
+ "23": "monitor",
35
+ "24": "bricks",
36
+ "25": "protection",
37
+ "26": "hawaii",
38
+ "27": "walking",
39
+ "28": "at table",
40
+ "29": "bicycles",
41
+ "30": "resting",
42
+ "31": "sidewalk",
43
+ "32": "wedding",
44
+ "33": "unknown",
45
+ "34": "2010",
46
+ "35": "8:35",
47
+ "36": "jeep",
48
+ "37": "lg",
49
+ "38": "roof",
50
+ "39": "solid",
51
+ "40": "lady",
52
+ "41": "door",
53
+ "42": "neon",
54
+ "43": "suv",
55
+ "44": "crown",
56
+ "45": "wine tasting",
57
+ "46": "hair",
58
+ "47": "king",
59
+ "48": "sun",
60
+ "49": "7:35",
61
+ "50": "human",
62
+ "51": "4",
63
+ "52": "green",
64
+ "53": "10",
65
+ "54": "many",
66
+ "55": "red and blue",
67
+ "56": "africa",
68
+ "57": "giraffes",
69
+ "58": "purple",
70
+ "59": "beagle",
71
+ "60": "skateboarding",
72
+ "61": "blue",
73
+ "62": "tan",
74
+ "63": "5",
75
+ "64": "watching",
76
+ "65": "right",
77
+ "66": "nothing",
78
+ "67": "hat",
79
+ "68": "6",
80
+ "69": "arrow",
81
+ "70": "9:35",
82
+ "71": "stripes",
83
+ "72": "natural",
84
+ "73": "they aren't",
85
+ "74": "8",
86
+ "75": "screen",
87
+ "76": "down",
88
+ "77": "window",
89
+ "78": "don't know",
90
+ "79": "big ben",
91
+ "80": "happy",
92
+ "81": "bedroom",
93
+ "82": "2013",
94
+ "83": "park",
95
+ "84": "shadows",
96
+ "85": "cross",
97
+ "86": "out",
98
+ "87": "cat",
99
+ "88": "plain",
100
+ "89": "photographer",
101
+ "90": "queen",
102
+ "91": "chair",
103
+ "92": "brick",
104
+ "93": "sky",
105
+ "94": "2",
106
+ "95": "net",
107
+ "96": "yellow",
108
+ "97": "dog",
109
+ "98": "person",
110
+ "99": "forest",
111
+ "100": "trees",
112
+ "101": "plate",
113
+ "102": "picnic table",
114
+ "103": "street",
115
+ "104": "snow",
116
+ "105": "table",
117
+ "106": "cage",
118
+ "107": "platform",
119
+ "108": "cloudy",
120
+ "109": "ball",
121
+ "110": "double",
122
+ "111": "zoo",
123
+ "112": "curtain",
124
+ "113": "doughnut",
125
+ "114": "giraffe",
126
+ "115": "white and blue",
127
+ "116": "brown",
128
+ "117": "tower",
129
+ "118": "boy",
130
+ "119": "black and white",
131
+ "120": "skier",
132
+ "121": "tired",
133
+ "122": "7:45",
134
+ "123": "bicycle",
135
+ "124": "outside",
136
+ "125": "shelter",
137
+ "126": "plastic",
138
+ "127": "soccer ball",
139
+ "128": "talking on phone",
140
+ "129": "yes",
141
+ "130": "lying down",
142
+ "131": "bike rack",
143
+ "132": "bus",
144
+ "133": "talking",
145
+ "134": "snowboard",
146
+ "135": "snowboarding",
147
+ "136": "leather",
148
+ "137": "shade",
149
+ "138": "blonde",
150
+ "139": "girl",
151
+ "140": "clock tower",
152
+ "141": "on street",
153
+ "142": "name tag",
154
+ "143": "french",
155
+ "144": "beige",
156
+ "145": "camera",
157
+ "146": "backpack",
158
+ "147": "air",
159
+ "148": "small",
160
+ "149": "style",
161
+ "150": "can't tell",
162
+ "151": "crossing",
163
+ "152": "canopy",
164
+ "153": "train",
165
+ "154": "rack",
166
+ "155": "7",
167
+ "156": "laying down",
168
+ "157": "smiling",
169
+ "158": "shadow",
170
+ "159": "fence",
171
+ "160": "lanyard",
172
+ "161": "white and black",
173
+ "162": "wall",
174
+ "163": "ice cream",
175
+ "164": "red",
176
+ "165": "wine",
177
+ "166": "2000",
178
+ "167": "1",
179
+ "168": "red and yellow",
180
+ "169": "pink",
181
+ "170": "large",
182
+ "171": "birthday",
183
+ "172": "low",
184
+ "173": "tv",
185
+ "174": "car",
186
+ "175": "white",
187
+ "176": "ground",
188
+ "177": "woman",
189
+ "178": "calico",
190
+ "179": "donut",
191
+ "180": "fashion",
192
+ "181": "no",
193
+ "182": "smile",
194
+ "183": "full",
195
+ "184": "skateboard",
196
+ "185": "woods",
197
+ "186": "exit",
198
+ "187": "windows",
199
+ "188": "necklace",
200
+ "189": "gray and black",
201
+ "190": "man",
202
+ "191": "curtains",
203
+ "192": "black",
204
+ "193": "orange",
205
+ "194": "not there",
206
+ "195": "station",
207
+ "196": "gray",
208
+ "197": "shrimp",
209
+ "198": "tent"
210
+ },
211
+ "image_size": 384,
212
+ "initializer_range": 0.02,
213
+ "intermediate_size": 3072,
214
+ "label2id": {
215
+ "0": 12,
216
+ "1": 167,
217
+ "10": 53,
218
+ "2": 94,
219
+ "2000": 166,
220
+ "2010": 34,
221
+ "2013": 82,
222
+ "3": 3,
223
+ "4": 51,
224
+ "5": 63,
225
+ "6": 68,
226
+ "7": 155,
227
+ "7:35": 49,
228
+ "7:45": 122,
229
+ "8": 74,
230
+ "8:35": 35,
231
+ "9:35": 70,
232
+ "africa": 56,
233
+ "air": 147,
234
+ "arrow": 69,
235
+ "at table": 28,
236
+ "backpack": 146,
237
+ "ball": 109,
238
+ "beagle": 59,
239
+ "bedroom": 81,
240
+ "beige": 144,
241
+ "bicycle": 123,
242
+ "bicycles": 29,
243
+ "big ben": 79,
244
+ "bike rack": 131,
245
+ "bikes": 2,
246
+ "birthday": 171,
247
+ "black": 192,
248
+ "black and white": 119,
249
+ "blonde": 138,
250
+ "blue": 61,
251
+ "blue and white": 10,
252
+ "boy": 118,
253
+ "brick": 92,
254
+ "bricks": 24,
255
+ "brown": 116,
256
+ "bus": 132,
257
+ "cage": 106,
258
+ "calico": 178,
259
+ "camera": 145,
260
+ "can't tell": 150,
261
+ "canopy": 152,
262
+ "car": 174,
263
+ "cat": 87,
264
+ "chair": 91,
265
+ "chopsticks": 18,
266
+ "church": 5,
267
+ "clear": 21,
268
+ "clock": 9,
269
+ "clock tower": 140,
270
+ "cloudy": 108,
271
+ "cross": 85,
272
+ "crossing": 151,
273
+ "crown": 44,
274
+ "cup": 20,
275
+ "curtain": 112,
276
+ "curtains": 191,
277
+ "desert": 19,
278
+ "dirt": 22,
279
+ "dog": 97,
280
+ "don't know": 78,
281
+ "donut": 179,
282
+ "door": 41,
283
+ "double": 110,
284
+ "doughnut": 113,
285
+ "down": 76,
286
+ "exit": 186,
287
+ "fashion": 180,
288
+ "fence": 159,
289
+ "forest": 99,
290
+ "french": 143,
291
+ "full": 183,
292
+ "giraffe": 114,
293
+ "giraffes": 57,
294
+ "girl": 139,
295
+ "gray": 196,
296
+ "gray and black": 189,
297
+ "green": 52,
298
+ "ground": 176,
299
+ "hair": 46,
300
+ "happy": 80,
301
+ "hat": 67,
302
+ "hawaii": 26,
303
+ "human": 50,
304
+ "ice cream": 163,
305
+ "in car": 6,
306
+ "jeep": 36,
307
+ "king": 47,
308
+ "lady": 40,
309
+ "lanyard": 160,
310
+ "large": 170,
311
+ "laying down": 156,
312
+ "leather": 136,
313
+ "lg": 37,
314
+ "little girl": 11,
315
+ "low": 172,
316
+ "lying down": 130,
317
+ "man": 190,
318
+ "many": 54,
319
+ "monitor": 23,
320
+ "name tag": 142,
321
+ "natural": 72,
322
+ "necklace": 188,
323
+ "neon": 42,
324
+ "net": 95,
325
+ "no": 181,
326
+ "not sure": 15,
327
+ "not there": 194,
328
+ "nothing": 66,
329
+ "on road": 17,
330
+ "on street": 141,
331
+ "orange": 193,
332
+ "out": 86,
333
+ "outside": 124,
334
+ "park": 83,
335
+ "person": 98,
336
+ "photographer": 89,
337
+ "picnic table": 102,
338
+ "pink": 169,
339
+ "plain": 88,
340
+ "plastic": 126,
341
+ "plate": 101,
342
+ "platform": 107,
343
+ "protection": 25,
344
+ "purple": 58,
345
+ "queen": 90,
346
+ "rack": 154,
347
+ "red": 164,
348
+ "red and blue": 55,
349
+ "red and yellow": 168,
350
+ "resting": 30,
351
+ "right": 65,
352
+ "roof": 38,
353
+ "screen": 75,
354
+ "security": 1,
355
+ "shade": 137,
356
+ "shadow": 158,
357
+ "shadows": 84,
358
+ "shelter": 125,
359
+ "shrimp": 197,
360
+ "sidewalk": 31,
361
+ "skateboard": 184,
362
+ "skateboarding": 60,
363
+ "skier": 120,
364
+ "skiing": 0,
365
+ "sky": 93,
366
+ "sleeping": 16,
367
+ "small": 148,
368
+ "smile": 182,
369
+ "smiling": 157,
370
+ "snow": 104,
371
+ "snowboard": 134,
372
+ "snowboarder": 7,
373
+ "snowboarding": 135,
374
+ "soccer": 13,
375
+ "soccer ball": 127,
376
+ "solid": 39,
377
+ "stand": 4,
378
+ "station": 195,
379
+ "street": 103,
380
+ "stripes": 71,
381
+ "style": 149,
382
+ "sun": 48,
383
+ "suv": 43,
384
+ "tabby": 14,
385
+ "table": 105,
386
+ "talking": 133,
387
+ "talking on phone": 128,
388
+ "tan": 62,
389
+ "tent": 198,
390
+ "they aren't": 73,
391
+ "tired": 121,
392
+ "tower": 117,
393
+ "train": 153,
394
+ "trees": 100,
395
+ "tv": 173,
396
+ "unknown": 33,
397
+ "walking": 27,
398
+ "wall": 162,
399
+ "watching": 64,
400
+ "wedding": 32,
401
+ "white": 175,
402
+ "white and black": 161,
403
+ "white and blue": 115,
404
+ "window": 77,
405
+ "windows": 187,
406
+ "wine": 165,
407
+ "wine tasting": 45,
408
+ "woman": 177,
409
+ "women": 8,
410
+ "woods": 185,
411
+ "yellow": 96,
412
+ "yes": 129,
413
+ "zoo": 111
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": "4.41.2",
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 sha256:02874854b7f500f9057d06e98893a1971d3f43c48bbdaa11c006b18a861076b7
3
+ size 452365324
preprocessor_config.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_valid_processor_keys": [
3
+ "images",
4
+ "do_resize",
5
+ "size",
6
+ "size_divisor",
7
+ "resample",
8
+ "do_rescale",
9
+ "rescale_factor",
10
+ "do_normalize",
11
+ "image_mean",
12
+ "image_std",
13
+ "do_pad",
14
+ "return_tensors",
15
+ "data_format",
16
+ "input_data_format"
17
+ ],
18
+ "do_normalize": true,
19
+ "do_pad": true,
20
+ "do_rescale": true,
21
+ "do_resize": true,
22
+ "image_mean": [
23
+ 0.5,
24
+ 0.5,
25
+ 0.5
26
+ ],
27
+ "image_processor_type": "ViltImageProcessor",
28
+ "image_std": [
29
+ 0.5,
30
+ 0.5,
31
+ 0.5
32
+ ],
33
+ "processor_class": "ViltProcessor",
34
+ "resample": 3,
35
+ "rescale_factor": 0.00392156862745098,
36
+ "size": {
37
+ "shortest_edge": 384
38
+ },
39
+ "size_divisor": 32
40
+ }
runs/Jun06_07-01-39_Jacks-MBP/events.out.tfevents.1717614101.Jacks-MBP.16824.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9f06648b032d811d9c6470775608918690bf7a422b1ff0870b7f3c17295d834c
3
+ size 13517
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 sha256:28be3070702b39be1cf78844fe44fe19904e3a0af6de7b90e056745296ed0e96
3
+ size 5112
vocab.txt ADDED
The diff for this file is too large to render. See raw diff