emigomez commited on
Commit
9266df2
verified
1 Parent(s): 62f5475

馃嵒 cheers

Browse files
README.md CHANGED
@@ -3,6 +3,7 @@ library_name: transformers
3
  license: apache-2.0
4
  base_model: google/vit-base-patch16-224-in21k
5
  tags:
 
6
  - generated_from_trainer
7
  metrics:
8
  - accuracy
@@ -16,9 +17,9 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  # vit-cropped-faces
18
 
19
- This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
20
  It achieves the following results on the evaluation set:
21
- - Loss: 0.0120
22
  - Accuracy: 1.0
23
 
24
  ## Model description
 
3
  license: apache-2.0
4
  base_model: google/vit-base-patch16-224-in21k
5
  tags:
6
+ - image-classification
7
  - generated_from_trainer
8
  metrics:
9
  - accuracy
 
17
 
18
  # vit-cropped-faces
19
 
20
+ This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the emigomez/vit-cropped-faces dataset.
21
  It achieves the following results on the evaluation set:
22
+ - Loss: 0.0109
23
  - Accuracy: 1.0
24
 
25
  ## Model description
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 1.0,
4
+ "eval_loss": 0.010909353382885456,
5
+ "eval_runtime": 1.6904,
6
+ "eval_samples_per_second": 33.127,
7
+ "eval_steps_per_second": 4.141,
8
+ "total_flos": 7.764766951908557e+17,
9
+ "train_loss": 0.024429786371183582,
10
+ "train_runtime": 337.6815,
11
+ "train_samples_per_second": 29.673,
12
+ "train_steps_per_second": 1.895
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 1.0,
4
+ "eval_loss": 0.010909353382885456,
5
+ "eval_runtime": 1.6904,
6
+ "eval_samples_per_second": 33.127,
7
+ "eval_steps_per_second": 4.141
8
+ }
runs/Mar14_11-15-01_74ea26e313a5/events.out.tfevents.1741951269.74ea26e313a5.1507.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:43e43cdafb3745963c91b589c978aba2b41f88ecc3a0609f64e4e846f2cbdafb
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "total_flos": 7.764766951908557e+17,
4
+ "train_loss": 0.024429786371183582,
5
+ "train_runtime": 337.6815,
6
+ "train_samples_per_second": 29.673,
7
+ "train_steps_per_second": 1.895
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,544 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.010909353382885456,
3
+ "best_model_checkpoint": "emigomez/vit-cropped-faces/checkpoint-200",
4
+ "epoch": 20.0,
5
+ "eval_steps": 100,
6
+ "global_step": 640,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.3125,
13
+ "grad_norm": 1.6401095390319824,
14
+ "learning_rate": 0.000196875,
15
+ "loss": 0.692,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.625,
20
+ "grad_norm": 1.1231616735458374,
21
+ "learning_rate": 0.00019375000000000002,
22
+ "loss": 0.2743,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.9375,
27
+ "grad_norm": 0.7256351709365845,
28
+ "learning_rate": 0.000190625,
29
+ "loss": 0.0795,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 1.25,
34
+ "grad_norm": 0.17406541109085083,
35
+ "learning_rate": 0.0001875,
36
+ "loss": 0.0414,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 1.5625,
41
+ "grad_norm": 6.439935684204102,
42
+ "learning_rate": 0.000184375,
43
+ "loss": 0.0465,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 1.875,
48
+ "grad_norm": 2.296980857849121,
49
+ "learning_rate": 0.00018125000000000001,
50
+ "loss": 0.0526,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 2.1875,
55
+ "grad_norm": 0.09786612540483475,
56
+ "learning_rate": 0.000178125,
57
+ "loss": 0.0372,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 2.5,
62
+ "grad_norm": 0.07251214236021042,
63
+ "learning_rate": 0.000175,
64
+ "loss": 0.0158,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 2.8125,
69
+ "grad_norm": 0.06504912674427032,
70
+ "learning_rate": 0.00017187500000000002,
71
+ "loss": 0.1044,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 3.125,
76
+ "grad_norm": 0.08557004481554031,
77
+ "learning_rate": 0.00016875,
78
+ "loss": 0.0254,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 3.125,
83
+ "eval_accuracy": 1.0,
84
+ "eval_loss": 0.013552392832934856,
85
+ "eval_runtime": 1.5523,
86
+ "eval_samples_per_second": 36.077,
87
+ "eval_steps_per_second": 4.51,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 3.4375,
92
+ "grad_norm": 0.06062212213873863,
93
+ "learning_rate": 0.000165625,
94
+ "loss": 0.0122,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 3.75,
99
+ "grad_norm": 0.050220146775245667,
100
+ "learning_rate": 0.00016250000000000002,
101
+ "loss": 0.0098,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 4.0625,
106
+ "grad_norm": 0.04843802750110626,
107
+ "learning_rate": 0.000159375,
108
+ "loss": 0.0091,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 4.375,
113
+ "grad_norm": 0.04158018156886101,
114
+ "learning_rate": 0.00015625,
115
+ "loss": 0.0081,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 4.6875,
120
+ "grad_norm": 0.03899329528212547,
121
+ "learning_rate": 0.000153125,
122
+ "loss": 0.0075,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 5.0,
127
+ "grad_norm": 0.05264412984251976,
128
+ "learning_rate": 0.00015000000000000001,
129
+ "loss": 0.0069,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 5.3125,
134
+ "grad_norm": 0.036546722054481506,
135
+ "learning_rate": 0.000146875,
136
+ "loss": 0.0064,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 5.625,
141
+ "grad_norm": 0.035343725234270096,
142
+ "learning_rate": 0.00014375,
143
+ "loss": 0.006,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 5.9375,
148
+ "grad_norm": 0.03636905550956726,
149
+ "learning_rate": 0.00014062500000000002,
150
+ "loss": 0.0056,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 6.25,
155
+ "grad_norm": 0.033068690448999405,
156
+ "learning_rate": 0.0001375,
157
+ "loss": 0.0053,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 6.25,
162
+ "eval_accuracy": 1.0,
163
+ "eval_loss": 0.010909353382885456,
164
+ "eval_runtime": 1.5116,
165
+ "eval_samples_per_second": 37.048,
166
+ "eval_steps_per_second": 4.631,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 6.5625,
171
+ "grad_norm": 0.027593158185482025,
172
+ "learning_rate": 0.000134375,
173
+ "loss": 0.0049,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 6.875,
178
+ "grad_norm": 0.02486700937151909,
179
+ "learning_rate": 0.00013125000000000002,
180
+ "loss": 0.0047,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 7.1875,
185
+ "grad_norm": 0.025286702439188957,
186
+ "learning_rate": 0.000128125,
187
+ "loss": 0.0045,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 7.5,
192
+ "grad_norm": 0.02420409955084324,
193
+ "learning_rate": 0.000125,
194
+ "loss": 0.0042,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 7.8125,
199
+ "grad_norm": 0.02255989797413349,
200
+ "learning_rate": 0.00012187500000000001,
201
+ "loss": 0.0041,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 8.125,
206
+ "grad_norm": 0.02108701318502426,
207
+ "learning_rate": 0.00011875,
208
+ "loss": 0.0038,
209
+ "step": 260
210
+ },
211
+ {
212
+ "epoch": 8.4375,
213
+ "grad_norm": 0.020098580047488213,
214
+ "learning_rate": 0.000115625,
215
+ "loss": 0.0037,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 8.75,
220
+ "grad_norm": 0.019474145025014877,
221
+ "learning_rate": 0.00011250000000000001,
222
+ "loss": 0.0036,
223
+ "step": 280
224
+ },
225
+ {
226
+ "epoch": 9.0625,
227
+ "grad_norm": 0.018637843430042267,
228
+ "learning_rate": 0.000109375,
229
+ "loss": 0.0035,
230
+ "step": 290
231
+ },
232
+ {
233
+ "epoch": 9.375,
234
+ "grad_norm": 0.018013661727309227,
235
+ "learning_rate": 0.00010625000000000001,
236
+ "loss": 0.0033,
237
+ "step": 300
238
+ },
239
+ {
240
+ "epoch": 9.375,
241
+ "eval_accuracy": 1.0,
242
+ "eval_loss": 0.013854571618139744,
243
+ "eval_runtime": 1.54,
244
+ "eval_samples_per_second": 36.363,
245
+ "eval_steps_per_second": 4.545,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 9.6875,
250
+ "grad_norm": 0.01775149628520012,
251
+ "learning_rate": 0.000103125,
252
+ "loss": 0.0032,
253
+ "step": 310
254
+ },
255
+ {
256
+ "epoch": 10.0,
257
+ "grad_norm": 0.01715986803174019,
258
+ "learning_rate": 0.0001,
259
+ "loss": 0.0031,
260
+ "step": 320
261
+ },
262
+ {
263
+ "epoch": 10.3125,
264
+ "grad_norm": 0.016615115106105804,
265
+ "learning_rate": 9.687500000000001e-05,
266
+ "loss": 0.003,
267
+ "step": 330
268
+ },
269
+ {
270
+ "epoch": 10.625,
271
+ "grad_norm": 0.016151374205946922,
272
+ "learning_rate": 9.375e-05,
273
+ "loss": 0.0029,
274
+ "step": 340
275
+ },
276
+ {
277
+ "epoch": 10.9375,
278
+ "grad_norm": 0.01536835078150034,
279
+ "learning_rate": 9.062500000000001e-05,
280
+ "loss": 0.0028,
281
+ "step": 350
282
+ },
283
+ {
284
+ "epoch": 11.25,
285
+ "grad_norm": 0.014975355938076973,
286
+ "learning_rate": 8.75e-05,
287
+ "loss": 0.0027,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 11.5625,
292
+ "grad_norm": 0.014888996258378029,
293
+ "learning_rate": 8.4375e-05,
294
+ "loss": 0.0027,
295
+ "step": 370
296
+ },
297
+ {
298
+ "epoch": 11.875,
299
+ "grad_norm": 0.014734145253896713,
300
+ "learning_rate": 8.125000000000001e-05,
301
+ "loss": 0.0026,
302
+ "step": 380
303
+ },
304
+ {
305
+ "epoch": 12.1875,
306
+ "grad_norm": 0.014760036021471024,
307
+ "learning_rate": 7.8125e-05,
308
+ "loss": 0.0025,
309
+ "step": 390
310
+ },
311
+ {
312
+ "epoch": 12.5,
313
+ "grad_norm": 0.013639590702950954,
314
+ "learning_rate": 7.500000000000001e-05,
315
+ "loss": 0.0025,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 12.5,
320
+ "eval_accuracy": 1.0,
321
+ "eval_loss": 0.012760639190673828,
322
+ "eval_runtime": 1.5727,
323
+ "eval_samples_per_second": 35.608,
324
+ "eval_steps_per_second": 4.451,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 12.8125,
329
+ "grad_norm": 0.013434410095214844,
330
+ "learning_rate": 7.1875e-05,
331
+ "loss": 0.0024,
332
+ "step": 410
333
+ },
334
+ {
335
+ "epoch": 13.125,
336
+ "grad_norm": 0.013357667252421379,
337
+ "learning_rate": 6.875e-05,
338
+ "loss": 0.0024,
339
+ "step": 420
340
+ },
341
+ {
342
+ "epoch": 13.4375,
343
+ "grad_norm": 0.01299892459064722,
344
+ "learning_rate": 6.562500000000001e-05,
345
+ "loss": 0.0023,
346
+ "step": 430
347
+ },
348
+ {
349
+ "epoch": 13.75,
350
+ "grad_norm": 0.01252039521932602,
351
+ "learning_rate": 6.25e-05,
352
+ "loss": 0.0023,
353
+ "step": 440
354
+ },
355
+ {
356
+ "epoch": 14.0625,
357
+ "grad_norm": 0.012350697070360184,
358
+ "learning_rate": 5.9375e-05,
359
+ "loss": 0.0022,
360
+ "step": 450
361
+ },
362
+ {
363
+ "epoch": 14.375,
364
+ "grad_norm": 0.012504469603300095,
365
+ "learning_rate": 5.6250000000000005e-05,
366
+ "loss": 0.0022,
367
+ "step": 460
368
+ },
369
+ {
370
+ "epoch": 14.6875,
371
+ "grad_norm": 0.01206807978451252,
372
+ "learning_rate": 5.3125000000000004e-05,
373
+ "loss": 0.0022,
374
+ "step": 470
375
+ },
376
+ {
377
+ "epoch": 15.0,
378
+ "grad_norm": 0.013769702985882759,
379
+ "learning_rate": 5e-05,
380
+ "loss": 0.0021,
381
+ "step": 480
382
+ },
383
+ {
384
+ "epoch": 15.3125,
385
+ "grad_norm": 0.01167486421763897,
386
+ "learning_rate": 4.6875e-05,
387
+ "loss": 0.0021,
388
+ "step": 490
389
+ },
390
+ {
391
+ "epoch": 15.625,
392
+ "grad_norm": 0.011528071016073227,
393
+ "learning_rate": 4.375e-05,
394
+ "loss": 0.0021,
395
+ "step": 500
396
+ },
397
+ {
398
+ "epoch": 15.625,
399
+ "eval_accuracy": 1.0,
400
+ "eval_loss": 0.012214643880724907,
401
+ "eval_runtime": 1.5413,
402
+ "eval_samples_per_second": 36.334,
403
+ "eval_steps_per_second": 4.542,
404
+ "step": 500
405
+ },
406
+ {
407
+ "epoch": 15.9375,
408
+ "grad_norm": 0.011460980400443077,
409
+ "learning_rate": 4.0625000000000005e-05,
410
+ "loss": 0.002,
411
+ "step": 510
412
+ },
413
+ {
414
+ "epoch": 16.25,
415
+ "grad_norm": 0.011635087430477142,
416
+ "learning_rate": 3.7500000000000003e-05,
417
+ "loss": 0.002,
418
+ "step": 520
419
+ },
420
+ {
421
+ "epoch": 16.5625,
422
+ "grad_norm": 0.012200203724205494,
423
+ "learning_rate": 3.4375e-05,
424
+ "loss": 0.002,
425
+ "step": 530
426
+ },
427
+ {
428
+ "epoch": 16.875,
429
+ "grad_norm": 0.011691519059240818,
430
+ "learning_rate": 3.125e-05,
431
+ "loss": 0.002,
432
+ "step": 540
433
+ },
434
+ {
435
+ "epoch": 17.1875,
436
+ "grad_norm": 0.010916861705482006,
437
+ "learning_rate": 2.8125000000000003e-05,
438
+ "loss": 0.002,
439
+ "step": 550
440
+ },
441
+ {
442
+ "epoch": 17.5,
443
+ "grad_norm": 0.010833012871444225,
444
+ "learning_rate": 2.5e-05,
445
+ "loss": 0.0019,
446
+ "step": 560
447
+ },
448
+ {
449
+ "epoch": 17.8125,
450
+ "grad_norm": 0.010946971364319324,
451
+ "learning_rate": 2.1875e-05,
452
+ "loss": 0.0019,
453
+ "step": 570
454
+ },
455
+ {
456
+ "epoch": 18.125,
457
+ "grad_norm": 0.01121476013213396,
458
+ "learning_rate": 1.8750000000000002e-05,
459
+ "loss": 0.0019,
460
+ "step": 580
461
+ },
462
+ {
463
+ "epoch": 18.4375,
464
+ "grad_norm": 0.010572544299066067,
465
+ "learning_rate": 1.5625e-05,
466
+ "loss": 0.0019,
467
+ "step": 590
468
+ },
469
+ {
470
+ "epoch": 18.75,
471
+ "grad_norm": 0.01052548922598362,
472
+ "learning_rate": 1.25e-05,
473
+ "loss": 0.0019,
474
+ "step": 600
475
+ },
476
+ {
477
+ "epoch": 18.75,
478
+ "eval_accuracy": 1.0,
479
+ "eval_loss": 0.011984586715698242,
480
+ "eval_runtime": 1.5235,
481
+ "eval_samples_per_second": 36.758,
482
+ "eval_steps_per_second": 4.595,
483
+ "step": 600
484
+ },
485
+ {
486
+ "epoch": 19.0625,
487
+ "grad_norm": 0.011718451976776123,
488
+ "learning_rate": 9.375000000000001e-06,
489
+ "loss": 0.0019,
490
+ "step": 610
491
+ },
492
+ {
493
+ "epoch": 19.375,
494
+ "grad_norm": 0.011368807405233383,
495
+ "learning_rate": 6.25e-06,
496
+ "loss": 0.0019,
497
+ "step": 620
498
+ },
499
+ {
500
+ "epoch": 19.6875,
501
+ "grad_norm": 0.01071301568299532,
502
+ "learning_rate": 3.125e-06,
503
+ "loss": 0.0019,
504
+ "step": 630
505
+ },
506
+ {
507
+ "epoch": 20.0,
508
+ "grad_norm": 0.011912377551198006,
509
+ "learning_rate": 0.0,
510
+ "loss": 0.0019,
511
+ "step": 640
512
+ },
513
+ {
514
+ "epoch": 20.0,
515
+ "step": 640,
516
+ "total_flos": 7.764766951908557e+17,
517
+ "train_loss": 0.024429786371183582,
518
+ "train_runtime": 337.6815,
519
+ "train_samples_per_second": 29.673,
520
+ "train_steps_per_second": 1.895
521
+ }
522
+ ],
523
+ "logging_steps": 10,
524
+ "max_steps": 640,
525
+ "num_input_tokens_seen": 0,
526
+ "num_train_epochs": 20,
527
+ "save_steps": 100,
528
+ "stateful_callbacks": {
529
+ "TrainerControl": {
530
+ "args": {
531
+ "should_epoch_stop": false,
532
+ "should_evaluate": false,
533
+ "should_log": false,
534
+ "should_save": true,
535
+ "should_training_stop": true
536
+ },
537
+ "attributes": {}
538
+ }
539
+ },
540
+ "total_flos": 7.764766951908557e+17,
541
+ "train_batch_size": 16,
542
+ "trial_name": null,
543
+ "trial_params": null
544
+ }