Hugging Face's logo

Manasee27
/
locater_ai1

Image Classification
Transformers
PyTorch
vit
Model card Files
xet
 Community
5

Upload folder using huggingface_hub

#1
by Manasee27 - opened
base: refs/heads/main
←
from: refs/pr/1
Discussion Files changed
+6433
-0
Files changed (23) hide show
  1. all_results.json +8 -0
  2. checkpoint-1200/config.json +40 -0
  3. checkpoint-1200/optimizer.pt +3 -0
  4. checkpoint-1200/preprocessor_config.json +22 -0
  5. checkpoint-1200/pytorch_model.bin +3 -0
  6. checkpoint-1200/rng_state.pth +3 -0
  7. checkpoint-1200/scheduler.pt +3 -0
  8. checkpoint-1200/trainer_state.json +1006 -0
  9. checkpoint-1200/training_args.bin +3 -0
  10. checkpoint-3120/config.json +40 -0
  11. checkpoint-3120/optimizer.pt +3 -0
  12. checkpoint-3120/preprocessor_config.json +22 -0
  13. checkpoint-3120/pytorch_model.bin +3 -0
  14. checkpoint-3120/rng_state.pth +3 -0
  15. checkpoint-3120/scheduler.pt +3 -0
  16. checkpoint-3120/trainer_state.json +2590 -0
  17. checkpoint-3120/training_args.bin +3 -0
  18. config.json +40 -0
  19. preprocessor_config.json +22 -0
  20. pytorch_model.bin +3 -0
  21. train_results.json +8 -0
  22. trainer_state.json +2599 -0
  23. training_args.bin +3 -0
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "total_flos": 3.8629591832685773e+18,
4
+ "train_loss": 0.2533179052472592,
5
+ "train_runtime": 4957.8804,
6
+ "train_samples_per_second": 10.069,
7
+ "train_steps_per_second": 0.629
8
+ }
checkpoint-1200/config.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Brazil",
13
+ "1": "Canada",
14
+ "2": "Finland",
15
+ "3": "Japan",
16
+ "4": "United-Kingdom",
17
+ "5": "United_States"
18
+ },
19
+ "image_size": 224,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Brazil": "0",
24
+ "Canada": "1",
25
+ "Finland": "2",
26
+ "Japan": "3",
27
+ "United-Kingdom": "4",
28
+ "United_States": "5"
29
+ },
30
+ "layer_norm_eps": 1e-12,
31
+ "model_type": "vit",
32
+ "num_attention_heads": 12,
33
+ "num_channels": 3,
34
+ "num_hidden_layers": 12,
35
+ "patch_size": 16,
36
+ "problem_type": "single_label_classification",
37
+ "qkv_bias": true,
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.31.0"
40
+ }
checkpoint-1200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b53cf19e0fe68219c3cde14ec10a9e31bbfe37f2bddc8174b04e0adce4f1221e
3
+ size 686537861
checkpoint-1200/preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
checkpoint-1200/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e4cf7738c400c63d307d24c20ed9c77e1a7e56df0099b73936659a2e13d9670
3
+ size 343278253
checkpoint-1200/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:15eae6ef31a72124390c6c3b9dc58b9cb83b5171321842da3e78c939836919a1
3
+ size 13617
checkpoint-1200/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5a81ad569a08b812a8c2c69f31db173295038be7ba83b72561fb876fd57769fc
3
+ size 627
checkpoint-1200/trainer_state.json ADDED
@@ -0,0 +1,1006 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.6381103992462158,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-1200",
4
+ "epoch": 3.076923076923077,
5
+ "global_step": 1200,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.03,
12
+ "learning_rate": 0.00019935897435897437,
13
+ "loss": 1.7678,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 0.00019871794871794874,
19
+ "loss": 1.7014,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.08,
24
+ "learning_rate": 0.0001980769230769231,
25
+ "loss": 1.4957,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 0.00019743589743589744,
31
+ "loss": 1.5043,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.1,
36
+ "eval_accuracy": 0.44556677890011226,
37
+ "eval_loss": 1.493240237236023,
38
+ "eval_runtime": 59.3866,
39
+ "eval_samples_per_second": 60.081,
40
+ "eval_steps_per_second": 7.51,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.13,
45
+ "learning_rate": 0.00019679487179487178,
46
+ "loss": 1.3931,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.15,
51
+ "learning_rate": 0.00019615384615384615,
52
+ "loss": 1.3029,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.18,
57
+ "learning_rate": 0.0001955128205128205,
58
+ "loss": 1.1753,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.21,
63
+ "learning_rate": 0.00019487179487179487,
64
+ "loss": 1.3894,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.21,
69
+ "eval_accuracy": 0.5297418630751964,
70
+ "eval_loss": 1.2430940866470337,
71
+ "eval_runtime": 38.1635,
72
+ "eval_samples_per_second": 93.492,
73
+ "eval_steps_per_second": 11.687,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.23,
78
+ "learning_rate": 0.00019423076923076924,
79
+ "loss": 1.118,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.26,
84
+ "learning_rate": 0.0001935897435897436,
85
+ "loss": 1.1395,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.28,
90
+ "learning_rate": 0.00019294871794871797,
91
+ "loss": 1.0946,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.31,
96
+ "learning_rate": 0.00019230769230769233,
97
+ "loss": 1.208,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.31,
102
+ "eval_accuracy": 0.6464646464646465,
103
+ "eval_loss": 0.9881709218025208,
104
+ "eval_runtime": 34.6718,
105
+ "eval_samples_per_second": 102.908,
106
+ "eval_steps_per_second": 12.863,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.33,
111
+ "learning_rate": 0.00019166666666666667,
112
+ "loss": 0.9775,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.36,
117
+ "learning_rate": 0.00019102564102564104,
118
+ "loss": 1.0216,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.38,
123
+ "learning_rate": 0.00019038461538461538,
124
+ "loss": 1.1942,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.41,
129
+ "learning_rate": 0.00018974358974358974,
130
+ "loss": 0.9847,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.41,
135
+ "eval_accuracy": 0.6599326599326599,
136
+ "eval_loss": 0.950343132019043,
137
+ "eval_runtime": 40.3536,
138
+ "eval_samples_per_second": 88.418,
139
+ "eval_steps_per_second": 11.052,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.44,
144
+ "learning_rate": 0.0001891025641025641,
145
+ "loss": 0.9707,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.46,
150
+ "learning_rate": 0.00018846153846153847,
151
+ "loss": 1.1941,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.49,
156
+ "learning_rate": 0.00018782051282051283,
157
+ "loss": 1.0309,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 0.51,
162
+ "learning_rate": 0.0001871794871794872,
163
+ "loss": 1.102,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 0.51,
168
+ "eval_accuracy": 0.6374859708193041,
169
+ "eval_loss": 0.9868837594985962,
170
+ "eval_runtime": 32.1858,
171
+ "eval_samples_per_second": 110.856,
172
+ "eval_steps_per_second": 13.857,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 0.54,
177
+ "learning_rate": 0.00018653846153846154,
178
+ "loss": 0.9899,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 0.56,
183
+ "learning_rate": 0.0001858974358974359,
184
+ "loss": 0.8926,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 0.59,
189
+ "learning_rate": 0.00018525641025641027,
190
+ "loss": 0.9552,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 0.62,
195
+ "learning_rate": 0.00018461538461538463,
196
+ "loss": 0.9109,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 0.62,
201
+ "eval_accuracy": 0.6380471380471381,
202
+ "eval_loss": 0.9887320399284363,
203
+ "eval_runtime": 32.5312,
204
+ "eval_samples_per_second": 109.679,
205
+ "eval_steps_per_second": 13.71,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 0.64,
210
+ "learning_rate": 0.00018397435897435897,
211
+ "loss": 0.832,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 0.67,
216
+ "learning_rate": 0.00018333333333333334,
217
+ "loss": 0.9323,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 0.69,
222
+ "learning_rate": 0.0001826923076923077,
223
+ "loss": 1.0061,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 0.72,
228
+ "learning_rate": 0.00018205128205128207,
229
+ "loss": 0.9627,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 0.72,
234
+ "eval_accuracy": 0.6156004489337823,
235
+ "eval_loss": 1.057395100593567,
236
+ "eval_runtime": 34.1152,
237
+ "eval_samples_per_second": 104.587,
238
+ "eval_steps_per_second": 13.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 0.74,
243
+ "learning_rate": 0.00018141025641025643,
244
+ "loss": 0.9613,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 0.77,
249
+ "learning_rate": 0.00018076923076923077,
250
+ "loss": 0.8986,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 0.79,
255
+ "learning_rate": 0.00018012820512820513,
256
+ "loss": 0.9344,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 0.82,
261
+ "learning_rate": 0.0001794871794871795,
262
+ "loss": 0.7031,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 0.82,
267
+ "eval_accuracy": 0.7087542087542088,
268
+ "eval_loss": 0.813530683517456,
269
+ "eval_runtime": 39.1342,
270
+ "eval_samples_per_second": 91.173,
271
+ "eval_steps_per_second": 11.397,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 0.85,
276
+ "learning_rate": 0.00017884615384615386,
277
+ "loss": 0.7596,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 0.87,
282
+ "learning_rate": 0.00017820512820512823,
283
+ "loss": 0.7207,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 0.9,
288
+ "learning_rate": 0.00017756410256410257,
289
+ "loss": 0.8904,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 0.92,
294
+ "learning_rate": 0.00017692307692307693,
295
+ "loss": 0.7605,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 0.92,
300
+ "eval_accuracy": 0.7272727272727273,
301
+ "eval_loss": 0.7620912790298462,
302
+ "eval_runtime": 33.0054,
303
+ "eval_samples_per_second": 108.103,
304
+ "eval_steps_per_second": 13.513,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 0.95,
309
+ "learning_rate": 0.0001762820512820513,
310
+ "loss": 0.8093,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 0.97,
315
+ "learning_rate": 0.00017564102564102566,
316
+ "loss": 0.8198,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.0,
321
+ "learning_rate": 0.000175,
322
+ "loss": 0.9515,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 1.03,
327
+ "learning_rate": 0.00017435897435897436,
328
+ "loss": 0.8467,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 1.03,
333
+ "eval_accuracy": 0.7244668911335578,
334
+ "eval_loss": 0.7974632978439331,
335
+ "eval_runtime": 33.3512,
336
+ "eval_samples_per_second": 106.983,
337
+ "eval_steps_per_second": 13.373,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 1.05,
342
+ "learning_rate": 0.00017371794871794873,
343
+ "loss": 0.5829,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 1.08,
348
+ "learning_rate": 0.0001730769230769231,
349
+ "loss": 0.5576,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 1.1,
354
+ "learning_rate": 0.00017243589743589746,
355
+ "loss": 0.6035,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 1.13,
360
+ "learning_rate": 0.0001717948717948718,
361
+ "loss": 0.6653,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 1.13,
366
+ "eval_accuracy": 0.7317620650953984,
367
+ "eval_loss": 0.7574185132980347,
368
+ "eval_runtime": 32.4855,
369
+ "eval_samples_per_second": 109.834,
370
+ "eval_steps_per_second": 13.729,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 1.15,
375
+ "learning_rate": 0.00017115384615384616,
376
+ "loss": 0.5067,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 1.18,
381
+ "learning_rate": 0.00017051282051282053,
382
+ "loss": 0.4973,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 1.21,
387
+ "learning_rate": 0.00016987179487179486,
388
+ "loss": 0.5721,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 1.23,
393
+ "learning_rate": 0.00016923076923076923,
394
+ "loss": 0.5467,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 1.23,
399
+ "eval_accuracy": 0.7244668911335578,
400
+ "eval_loss": 0.7471520900726318,
401
+ "eval_runtime": 32.6752,
402
+ "eval_samples_per_second": 109.196,
403
+ "eval_steps_per_second": 13.65,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 1.26,
408
+ "learning_rate": 0.0001685897435897436,
409
+ "loss": 0.5494,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 1.28,
414
+ "learning_rate": 0.00016794871794871796,
415
+ "loss": 0.6158,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 1.31,
420
+ "learning_rate": 0.00016730769230769232,
421
+ "loss": 0.5412,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 1.33,
426
+ "learning_rate": 0.0001666666666666667,
427
+ "loss": 0.388,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 1.33,
432
+ "eval_accuracy": 0.7716049382716049,
433
+ "eval_loss": 0.6508304476737976,
434
+ "eval_runtime": 36.0929,
435
+ "eval_samples_per_second": 98.856,
436
+ "eval_steps_per_second": 12.357,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 1.36,
441
+ "learning_rate": 0.00016602564102564105,
442
+ "loss": 0.4511,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 1.38,
447
+ "learning_rate": 0.0001653846153846154,
448
+ "loss": 0.566,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 1.41,
453
+ "learning_rate": 0.00016474358974358976,
454
+ "loss": 0.4476,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 1.44,
459
+ "learning_rate": 0.0001641025641025641,
460
+ "loss": 0.4699,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 1.44,
465
+ "eval_accuracy": 0.7244668911335578,
466
+ "eval_loss": 0.773759663105011,
467
+ "eval_runtime": 31.7272,
468
+ "eval_samples_per_second": 112.459,
469
+ "eval_steps_per_second": 14.057,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 1.46,
474
+ "learning_rate": 0.00016346153846153846,
475
+ "loss": 0.402,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 1.49,
480
+ "learning_rate": 0.00016282051282051282,
481
+ "loss": 0.425,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 1.51,
486
+ "learning_rate": 0.0001621794871794872,
487
+ "loss": 0.5171,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 1.54,
492
+ "learning_rate": 0.00016153846153846155,
493
+ "loss": 0.5344,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 1.54,
498
+ "eval_accuracy": 0.7328843995510662,
499
+ "eval_loss": 0.7624912858009338,
500
+ "eval_runtime": 32.007,
501
+ "eval_samples_per_second": 111.476,
502
+ "eval_steps_per_second": 13.934,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 1.56,
507
+ "learning_rate": 0.00016089743589743592,
508
+ "loss": 0.4913,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 1.59,
513
+ "learning_rate": 0.00016025641025641028,
514
+ "loss": 0.5042,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 1.62,
519
+ "learning_rate": 0.00015961538461538462,
520
+ "loss": 0.6051,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 1.64,
525
+ "learning_rate": 0.00015897435897435896,
526
+ "loss": 0.5753,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 1.64,
531
+ "eval_accuracy": 0.7598204264870931,
532
+ "eval_loss": 0.6743783950805664,
533
+ "eval_runtime": 30.705,
534
+ "eval_samples_per_second": 116.203,
535
+ "eval_steps_per_second": 14.525,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 1.67,
540
+ "learning_rate": 0.00015833333333333332,
541
+ "loss": 0.524,
542
+ "step": 650
543
+ },
544
+ {
545
+ "epoch": 1.69,
546
+ "learning_rate": 0.0001576923076923077,
547
+ "loss": 0.5007,
548
+ "step": 660
549
+ },
550
+ {
551
+ "epoch": 1.72,
552
+ "learning_rate": 0.00015705128205128205,
553
+ "loss": 0.57,
554
+ "step": 670
555
+ },
556
+ {
557
+ "epoch": 1.74,
558
+ "learning_rate": 0.00015641025641025642,
559
+ "loss": 0.5533,
560
+ "step": 680
561
+ },
562
+ {
563
+ "epoch": 1.74,
564
+ "eval_accuracy": 0.734006734006734,
565
+ "eval_loss": 0.7230738997459412,
566
+ "eval_runtime": 33.3034,
567
+ "eval_samples_per_second": 107.136,
568
+ "eval_steps_per_second": 13.392,
569
+ "step": 680
570
+ },
571
+ {
572
+ "epoch": 1.77,
573
+ "learning_rate": 0.00015576923076923078,
574
+ "loss": 0.4399,
575
+ "step": 690
576
+ },
577
+ {
578
+ "epoch": 1.79,
579
+ "learning_rate": 0.00015512820512820515,
580
+ "loss": 0.453,
581
+ "step": 700
582
+ },
583
+ {
584
+ "epoch": 1.82,
585
+ "learning_rate": 0.00015448717948717951,
586
+ "loss": 0.4504,
587
+ "step": 710
588
+ },
589
+ {
590
+ "epoch": 1.85,
591
+ "learning_rate": 0.00015384615384615385,
592
+ "loss": 0.4244,
593
+ "step": 720
594
+ },
595
+ {
596
+ "epoch": 1.85,
597
+ "eval_accuracy": 0.7687991021324355,
598
+ "eval_loss": 0.6673141717910767,
599
+ "eval_runtime": 32.6948,
600
+ "eval_samples_per_second": 109.131,
601
+ "eval_steps_per_second": 13.641,
602
+ "step": 720
603
+ },
604
+ {
605
+ "epoch": 1.87,
606
+ "learning_rate": 0.00015320512820512822,
607
+ "loss": 0.46,
608
+ "step": 730
609
+ },
610
+ {
611
+ "epoch": 1.9,
612
+ "learning_rate": 0.00015256410256410255,
613
+ "loss": 0.4132,
614
+ "step": 740
615
+ },
616
+ {
617
+ "epoch": 1.92,
618
+ "learning_rate": 0.00015192307692307692,
619
+ "loss": 0.4849,
620
+ "step": 750
621
+ },
622
+ {
623
+ "epoch": 1.95,
624
+ "learning_rate": 0.00015128205128205128,
625
+ "loss": 0.5423,
626
+ "step": 760
627
+ },
628
+ {
629
+ "epoch": 1.95,
630
+ "eval_accuracy": 0.7413019079685746,
631
+ "eval_loss": 0.7185856699943542,
632
+ "eval_runtime": 35.9539,
633
+ "eval_samples_per_second": 99.238,
634
+ "eval_steps_per_second": 12.405,
635
+ "step": 760
636
+ },
637
+ {
638
+ "epoch": 1.97,
639
+ "learning_rate": 0.00015064102564102565,
640
+ "loss": 0.3997,
641
+ "step": 770
642
+ },
643
+ {
644
+ "epoch": 2.0,
645
+ "learning_rate": 0.00015000000000000001,
646
+ "loss": 0.4744,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 2.03,
651
+ "learning_rate": 0.00014935897435897438,
652
+ "loss": 0.4717,
653
+ "step": 790
654
+ },
655
+ {
656
+ "epoch": 2.05,
657
+ "learning_rate": 0.00014871794871794872,
658
+ "loss": 0.3384,
659
+ "step": 800
660
+ },
661
+ {
662
+ "epoch": 2.05,
663
+ "eval_accuracy": 0.7351290684624018,
664
+ "eval_loss": 0.7269920706748962,
665
+ "eval_runtime": 33.5897,
666
+ "eval_samples_per_second": 106.223,
667
+ "eval_steps_per_second": 13.278,
668
+ "step": 800
669
+ },
670
+ {
671
+ "epoch": 2.08,
672
+ "learning_rate": 0.00014807692307692308,
673
+ "loss": 0.2602,
674
+ "step": 810
675
+ },
676
+ {
677
+ "epoch": 2.1,
678
+ "learning_rate": 0.00014743589743589745,
679
+ "loss": 0.2345,
680
+ "step": 820
681
+ },
682
+ {
683
+ "epoch": 2.13,
684
+ "learning_rate": 0.00014679487179487178,
685
+ "loss": 0.2135,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 2.15,
690
+ "learning_rate": 0.00014615384615384615,
691
+ "loss": 0.2797,
692
+ "step": 840
693
+ },
694
+ {
695
+ "epoch": 2.15,
696
+ "eval_accuracy": 0.7497194163860831,
697
+ "eval_loss": 0.7742622494697571,
698
+ "eval_runtime": 31.2996,
699
+ "eval_samples_per_second": 113.995,
700
+ "eval_steps_per_second": 14.249,
701
+ "step": 840
702
+ },
703
+ {
704
+ "epoch": 2.18,
705
+ "learning_rate": 0.00014551282051282051,
706
+ "loss": 0.1862,
707
+ "step": 850
708
+ },
709
+ {
710
+ "epoch": 2.21,
711
+ "learning_rate": 0.00014487179487179488,
712
+ "loss": 0.2334,
713
+ "step": 860
714
+ },
715
+ {
716
+ "epoch": 2.23,
717
+ "learning_rate": 0.00014423076923076924,
718
+ "loss": 0.3361,
719
+ "step": 870
720
+ },
721
+ {
722
+ "epoch": 2.26,
723
+ "learning_rate": 0.0001435897435897436,
724
+ "loss": 0.2939,
725
+ "step": 880
726
+ },
727
+ {
728
+ "epoch": 2.26,
729
+ "eval_accuracy": 0.7732884399551067,
730
+ "eval_loss": 0.6985616087913513,
731
+ "eval_runtime": 32.9047,
732
+ "eval_samples_per_second": 108.435,
733
+ "eval_steps_per_second": 13.554,
734
+ "step": 880
735
+ },
736
+ {
737
+ "epoch": 2.28,
738
+ "learning_rate": 0.00014294871794871795,
739
+ "loss": 0.2846,
740
+ "step": 890
741
+ },
742
+ {
743
+ "epoch": 2.31,
744
+ "learning_rate": 0.0001423076923076923,
745
+ "loss": 0.2555,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 2.33,
750
+ "learning_rate": 0.00014166666666666668,
751
+ "loss": 0.2629,
752
+ "step": 910
753
+ },
754
+ {
755
+ "epoch": 2.36,
756
+ "learning_rate": 0.00014102564102564104,
757
+ "loss": 0.2288,
758
+ "step": 920
759
+ },
760
+ {
761
+ "epoch": 2.36,
762
+ "eval_accuracy": 0.7480359147025814,
763
+ "eval_loss": 0.761025607585907,
764
+ "eval_runtime": 34.1426,
765
+ "eval_samples_per_second": 104.503,
766
+ "eval_steps_per_second": 13.063,
767
+ "step": 920
768
+ },
769
+ {
770
+ "epoch": 2.38,
771
+ "learning_rate": 0.00014038461538461538,
772
+ "loss": 0.2115,
773
+ "step": 930
774
+ },
775
+ {
776
+ "epoch": 2.41,
777
+ "learning_rate": 0.00013974358974358974,
778
+ "loss": 0.2036,
779
+ "step": 940
780
+ },
781
+ {
782
+ "epoch": 2.44,
783
+ "learning_rate": 0.0001391025641025641,
784
+ "loss": 0.2913,
785
+ "step": 950
786
+ },
787
+ {
788
+ "epoch": 2.46,
789
+ "learning_rate": 0.00013846153846153847,
790
+ "loss": 0.2204,
791
+ "step": 960
792
+ },
793
+ {
794
+ "epoch": 2.46,
795
+ "eval_accuracy": 0.7570145903479237,
796
+ "eval_loss": 0.7840890884399414,
797
+ "eval_runtime": 35.7983,
798
+ "eval_samples_per_second": 99.67,
799
+ "eval_steps_per_second": 12.459,
800
+ "step": 960
801
+ },
802
+ {
803
+ "epoch": 2.49,
804
+ "learning_rate": 0.00013782051282051284,
805
+ "loss": 0.2759,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 2.51,
810
+ "learning_rate": 0.00013717948717948718,
811
+ "loss": 0.3945,
812
+ "step": 980
813
+ },
814
+ {
815
+ "epoch": 2.54,
816
+ "learning_rate": 0.00013653846153846154,
817
+ "loss": 0.3639,
818
+ "step": 990
819
+ },
820
+ {
821
+ "epoch": 2.56,
822
+ "learning_rate": 0.0001358974358974359,
823
+ "loss": 0.4397,
824
+ "step": 1000
825
+ },
826
+ {
827
+ "epoch": 2.56,
828
+ "eval_accuracy": 0.7789001122334456,
829
+ "eval_loss": 0.6565700173377991,
830
+ "eval_runtime": 31.6553,
831
+ "eval_samples_per_second": 112.714,
832
+ "eval_steps_per_second": 14.089,
833
+ "step": 1000
834
+ },
835
+ {
836
+ "epoch": 2.59,
837
+ "learning_rate": 0.00013525641025641027,
838
+ "loss": 0.2707,
839
+ "step": 1010
840
+ },
841
+ {
842
+ "epoch": 2.62,
843
+ "learning_rate": 0.00013461538461538464,
844
+ "loss": 0.3451,
845
+ "step": 1020
846
+ },
847
+ {
848
+ "epoch": 2.64,
849
+ "learning_rate": 0.00013397435897435897,
850
+ "loss": 0.3277,
851
+ "step": 1030
852
+ },
853
+ {
854
+ "epoch": 2.67,
855
+ "learning_rate": 0.00013333333333333334,
856
+ "loss": 0.2219,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 2.67,
861
+ "eval_accuracy": 0.7581369248035915,
862
+ "eval_loss": 0.6900522112846375,
863
+ "eval_runtime": 31.7845,
864
+ "eval_samples_per_second": 112.256,
865
+ "eval_steps_per_second": 14.032,
866
+ "step": 1040
867
+ },
868
+ {
869
+ "epoch": 2.69,
870
+ "learning_rate": 0.0001326923076923077,
871
+ "loss": 0.3174,
872
+ "step": 1050
873
+ },
874
+ {
875
+ "epoch": 2.72,
876
+ "learning_rate": 0.00013205128205128204,
877
+ "loss": 0.189,
878
+ "step": 1060
879
+ },
880
+ {
881
+ "epoch": 2.74,
882
+ "learning_rate": 0.0001314102564102564,
883
+ "loss": 0.2619,
884
+ "step": 1070
885
+ },
886
+ {
887
+ "epoch": 2.77,
888
+ "learning_rate": 0.00013076923076923077,
889
+ "loss": 0.2297,
890
+ "step": 1080
891
+ },
892
+ {
893
+ "epoch": 2.77,
894
+ "eval_accuracy": 0.7581369248035915,
895
+ "eval_loss": 0.7095093131065369,
896
+ "eval_runtime": 38.4561,
897
+ "eval_samples_per_second": 92.781,
898
+ "eval_steps_per_second": 11.598,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 2.79,
903
+ "learning_rate": 0.00013012820512820514,
904
+ "loss": 0.2201,
905
+ "step": 1090
906
+ },
907
+ {
908
+ "epoch": 2.82,
909
+ "learning_rate": 0.0001294871794871795,
910
+ "loss": 0.2261,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 2.85,
915
+ "learning_rate": 0.00012884615384615387,
916
+ "loss": 0.2095,
917
+ "step": 1110
918
+ },
919
+ {
920
+ "epoch": 2.87,
921
+ "learning_rate": 0.00012820512820512823,
922
+ "loss": 0.209,
923
+ "step": 1120
924
+ },
925
+ {
926
+ "epoch": 2.87,
927
+ "eval_accuracy": 0.7738496071829405,
928
+ "eval_loss": 0.7127913236618042,
929
+ "eval_runtime": 32.9997,
930
+ "eval_samples_per_second": 108.122,
931
+ "eval_steps_per_second": 13.515,
932
+ "step": 1120
933
+ },
934
+ {
935
+ "epoch": 2.9,
936
+ "learning_rate": 0.00012756410256410257,
937
+ "loss": 0.3365,
938
+ "step": 1130
939
+ },
940
+ {
941
+ "epoch": 2.92,
942
+ "learning_rate": 0.00012692307692307693,
943
+ "loss": 0.2639,
944
+ "step": 1140
945
+ },
946
+ {
947
+ "epoch": 2.95,
948
+ "learning_rate": 0.00012628205128205127,
949
+ "loss": 0.1892,
950
+ "step": 1150
951
+ },
952
+ {
953
+ "epoch": 2.97,
954
+ "learning_rate": 0.00012564102564102564,
955
+ "loss": 0.1853,
956
+ "step": 1160
957
+ },
958
+ {
959
+ "epoch": 2.97,
960
+ "eval_accuracy": 0.7710437710437711,
961
+ "eval_loss": 0.6986069083213806,
962
+ "eval_runtime": 32.2598,
963
+ "eval_samples_per_second": 110.602,
964
+ "eval_steps_per_second": 13.825,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 3.0,
969
+ "learning_rate": 0.000125,
970
+ "loss": 0.1807,
971
+ "step": 1170
972
+ },
973
+ {
974
+ "epoch": 3.03,
975
+ "learning_rate": 0.00012435897435897437,
976
+ "loss": 0.1554,
977
+ "step": 1180
978
+ },
979
+ {
980
+ "epoch": 3.05,
981
+ "learning_rate": 0.00012371794871794873,
982
+ "loss": 0.1845,
983
+ "step": 1190
984
+ },
985
+ {
986
+ "epoch": 3.08,
987
+ "learning_rate": 0.0001230769230769231,
988
+ "loss": 0.1322,
989
+ "step": 1200
990
+ },
991
+ {
992
+ "epoch": 3.08,
993
+ "eval_accuracy": 0.7962962962962963,
994
+ "eval_loss": 0.6381103992462158,
995
+ "eval_runtime": 35.1257,
996
+ "eval_samples_per_second": 101.578,
997
+ "eval_steps_per_second": 12.697,
998
+ "step": 1200
999
+ }
1000
+ ],
1001
+ "max_steps": 3120,
1002
+ "num_train_epochs": 8,
1003
+ "total_flos": 1.4858071822502093e+18,
1004
+ "trial_name": null,
1005
+ "trial_params": null
1006
+ }
checkpoint-1200/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38c25b2b280e8f8dde2511a004bf1b097d7641bf1aae99dd53711356509cd78e
3
+ size 3899
checkpoint-3120/config.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Brazil",
13
+ "1": "Canada",
14
+ "2": "Finland",
15
+ "3": "Japan",
16
+ "4": "United-Kingdom",
17
+ "5": "United_States"
18
+ },
19
+ "image_size": 224,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Brazil": "0",
24
+ "Canada": "1",
25
+ "Finland": "2",
26
+ "Japan": "3",
27
+ "United-Kingdom": "4",
28
+ "United_States": "5"
29
+ },
30
+ "layer_norm_eps": 1e-12,
31
+ "model_type": "vit",
32
+ "num_attention_heads": 12,
33
+ "num_channels": 3,
34
+ "num_hidden_layers": 12,
35
+ "patch_size": 16,
36
+ "problem_type": "single_label_classification",
37
+ "qkv_bias": true,
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.31.0"
40
+ }
checkpoint-3120/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a4a8f329c33f0b869453a941fbff52267fa7967b9aeeff87f529f98ab308f8f8
3
+ size 686537861
checkpoint-3120/preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
checkpoint-3120/pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6f7265735d9427d338f6934c4ac7bc62a2c11b9c0842576be5b8f6032cfa8508
3
+ size 343278253
checkpoint-3120/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5603187b2e7109f06291eba7b2b6fdeed2976f8fc03c3ec1a91df1a163133bf0
3
+ size 13617
checkpoint-3120/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac683e6fa1db5f0ebc121b7681a83c3a8a325b59f341f4ccfd5c5bedb8cb70f3
3
+ size 627
checkpoint-3120/trainer_state.json ADDED
@@ -0,0 +1,2590 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.6381103992462158,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-1200",
4
+ "epoch": 8.0,
5
+ "global_step": 3120,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.03,
12
+ "learning_rate": 0.00019935897435897437,
13
+ "loss": 1.7678,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 0.00019871794871794874,
19
+ "loss": 1.7014,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.08,
24
+ "learning_rate": 0.0001980769230769231,
25
+ "loss": 1.4957,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 0.00019743589743589744,
31
+ "loss": 1.5043,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.1,
36
+ "eval_accuracy": 0.44556677890011226,
37
+ "eval_loss": 1.493240237236023,
38
+ "eval_runtime": 59.3866,
39
+ "eval_samples_per_second": 60.081,
40
+ "eval_steps_per_second": 7.51,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.13,
45
+ "learning_rate": 0.00019679487179487178,
46
+ "loss": 1.3931,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.15,
51
+ "learning_rate": 0.00019615384615384615,
52
+ "loss": 1.3029,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.18,
57
+ "learning_rate": 0.0001955128205128205,
58
+ "loss": 1.1753,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.21,
63
+ "learning_rate": 0.00019487179487179487,
64
+ "loss": 1.3894,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.21,
69
+ "eval_accuracy": 0.5297418630751964,
70
+ "eval_loss": 1.2430940866470337,
71
+ "eval_runtime": 38.1635,
72
+ "eval_samples_per_second": 93.492,
73
+ "eval_steps_per_second": 11.687,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.23,
78
+ "learning_rate": 0.00019423076923076924,
79
+ "loss": 1.118,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.26,
84
+ "learning_rate": 0.0001935897435897436,
85
+ "loss": 1.1395,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.28,
90
+ "learning_rate": 0.00019294871794871797,
91
+ "loss": 1.0946,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.31,
96
+ "learning_rate": 0.00019230769230769233,
97
+ "loss": 1.208,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.31,
102
+ "eval_accuracy": 0.6464646464646465,
103
+ "eval_loss": 0.9881709218025208,
104
+ "eval_runtime": 34.6718,
105
+ "eval_samples_per_second": 102.908,
106
+ "eval_steps_per_second": 12.863,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.33,
111
+ "learning_rate": 0.00019166666666666667,
112
+ "loss": 0.9775,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.36,
117
+ "learning_rate": 0.00019102564102564104,
118
+ "loss": 1.0216,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.38,
123
+ "learning_rate": 0.00019038461538461538,
124
+ "loss": 1.1942,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.41,
129
+ "learning_rate": 0.00018974358974358974,
130
+ "loss": 0.9847,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.41,
135
+ "eval_accuracy": 0.6599326599326599,
136
+ "eval_loss": 0.950343132019043,
137
+ "eval_runtime": 40.3536,
138
+ "eval_samples_per_second": 88.418,
139
+ "eval_steps_per_second": 11.052,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.44,
144
+ "learning_rate": 0.0001891025641025641,
145
+ "loss": 0.9707,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.46,
150
+ "learning_rate": 0.00018846153846153847,
151
+ "loss": 1.1941,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.49,
156
+ "learning_rate": 0.00018782051282051283,
157
+ "loss": 1.0309,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 0.51,
162
+ "learning_rate": 0.0001871794871794872,
163
+ "loss": 1.102,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 0.51,
168
+ "eval_accuracy": 0.6374859708193041,
169
+ "eval_loss": 0.9868837594985962,
170
+ "eval_runtime": 32.1858,
171
+ "eval_samples_per_second": 110.856,
172
+ "eval_steps_per_second": 13.857,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 0.54,
177
+ "learning_rate": 0.00018653846153846154,
178
+ "loss": 0.9899,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 0.56,
183
+ "learning_rate": 0.0001858974358974359,
184
+ "loss": 0.8926,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 0.59,
189
+ "learning_rate": 0.00018525641025641027,
190
+ "loss": 0.9552,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 0.62,
195
+ "learning_rate": 0.00018461538461538463,
196
+ "loss": 0.9109,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 0.62,
201
+ "eval_accuracy": 0.6380471380471381,
202
+ "eval_loss": 0.9887320399284363,
203
+ "eval_runtime": 32.5312,
204
+ "eval_samples_per_second": 109.679,
205
+ "eval_steps_per_second": 13.71,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 0.64,
210
+ "learning_rate": 0.00018397435897435897,
211
+ "loss": 0.832,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 0.67,
216
+ "learning_rate": 0.00018333333333333334,
217
+ "loss": 0.9323,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 0.69,
222
+ "learning_rate": 0.0001826923076923077,
223
+ "loss": 1.0061,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 0.72,
228
+ "learning_rate": 0.00018205128205128207,
229
+ "loss": 0.9627,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 0.72,
234
+ "eval_accuracy": 0.6156004489337823,
235
+ "eval_loss": 1.057395100593567,
236
+ "eval_runtime": 34.1152,
237
+ "eval_samples_per_second": 104.587,
238
+ "eval_steps_per_second": 13.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 0.74,
243
+ "learning_rate": 0.00018141025641025643,
244
+ "loss": 0.9613,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 0.77,
249
+ "learning_rate": 0.00018076923076923077,
250
+ "loss": 0.8986,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 0.79,
255
+ "learning_rate": 0.00018012820512820513,
256
+ "loss": 0.9344,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 0.82,
261
+ "learning_rate": 0.0001794871794871795,
262
+ "loss": 0.7031,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 0.82,
267
+ "eval_accuracy": 0.7087542087542088,
268
+ "eval_loss": 0.813530683517456,
269
+ "eval_runtime": 39.1342,
270
+ "eval_samples_per_second": 91.173,
271
+ "eval_steps_per_second": 11.397,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 0.85,
276
+ "learning_rate": 0.00017884615384615386,
277
+ "loss": 0.7596,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 0.87,
282
+ "learning_rate": 0.00017820512820512823,
283
+ "loss": 0.7207,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 0.9,
288
+ "learning_rate": 0.00017756410256410257,
289
+ "loss": 0.8904,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 0.92,
294
+ "learning_rate": 0.00017692307692307693,
295
+ "loss": 0.7605,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 0.92,
300
+ "eval_accuracy": 0.7272727272727273,
301
+ "eval_loss": 0.7620912790298462,
302
+ "eval_runtime": 33.0054,
303
+ "eval_samples_per_second": 108.103,
304
+ "eval_steps_per_second": 13.513,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 0.95,
309
+ "learning_rate": 0.0001762820512820513,
310
+ "loss": 0.8093,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 0.97,
315
+ "learning_rate": 0.00017564102564102566,
316
+ "loss": 0.8198,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.0,
321
+ "learning_rate": 0.000175,
322
+ "loss": 0.9515,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 1.03,
327
+ "learning_rate": 0.00017435897435897436,
328
+ "loss": 0.8467,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 1.03,
333
+ "eval_accuracy": 0.7244668911335578,
334
+ "eval_loss": 0.7974632978439331,
335
+ "eval_runtime": 33.3512,
336
+ "eval_samples_per_second": 106.983,
337
+ "eval_steps_per_second": 13.373,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 1.05,
342
+ "learning_rate": 0.00017371794871794873,
343
+ "loss": 0.5829,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 1.08,
348
+ "learning_rate": 0.0001730769230769231,
349
+ "loss": 0.5576,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 1.1,
354
+ "learning_rate": 0.00017243589743589746,
355
+ "loss": 0.6035,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 1.13,
360
+ "learning_rate": 0.0001717948717948718,
361
+ "loss": 0.6653,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 1.13,
366
+ "eval_accuracy": 0.7317620650953984,
367
+ "eval_loss": 0.7574185132980347,
368
+ "eval_runtime": 32.4855,
369
+ "eval_samples_per_second": 109.834,
370
+ "eval_steps_per_second": 13.729,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 1.15,
375
+ "learning_rate": 0.00017115384615384616,
376
+ "loss": 0.5067,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 1.18,
381
+ "learning_rate": 0.00017051282051282053,
382
+ "loss": 0.4973,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 1.21,
387
+ "learning_rate": 0.00016987179487179486,
388
+ "loss": 0.5721,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 1.23,
393
+ "learning_rate": 0.00016923076923076923,
394
+ "loss": 0.5467,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 1.23,
399
+ "eval_accuracy": 0.7244668911335578,
400
+ "eval_loss": 0.7471520900726318,
401
+ "eval_runtime": 32.6752,
402
+ "eval_samples_per_second": 109.196,
403
+ "eval_steps_per_second": 13.65,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 1.26,
408
+ "learning_rate": 0.0001685897435897436,
409
+ "loss": 0.5494,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 1.28,
414
+ "learning_rate": 0.00016794871794871796,
415
+ "loss": 0.6158,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 1.31,
420
+ "learning_rate": 0.00016730769230769232,
421
+ "loss": 0.5412,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 1.33,
426
+ "learning_rate": 0.0001666666666666667,
427
+ "loss": 0.388,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 1.33,
432
+ "eval_accuracy": 0.7716049382716049,
433
+ "eval_loss": 0.6508304476737976,
434
+ "eval_runtime": 36.0929,
435
+ "eval_samples_per_second": 98.856,
436
+ "eval_steps_per_second": 12.357,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 1.36,
441
+ "learning_rate": 0.00016602564102564105,
442
+ "loss": 0.4511,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 1.38,
447
+ "learning_rate": 0.0001653846153846154,
448
+ "loss": 0.566,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 1.41,
453
+ "learning_rate": 0.00016474358974358976,
454
+ "loss": 0.4476,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 1.44,
459
+ "learning_rate": 0.0001641025641025641,
460
+ "loss": 0.4699,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 1.44,
465
+ "eval_accuracy": 0.7244668911335578,
466
+ "eval_loss": 0.773759663105011,
467
+ "eval_runtime": 31.7272,
468
+ "eval_samples_per_second": 112.459,
469
+ "eval_steps_per_second": 14.057,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 1.46,
474
+ "learning_rate": 0.00016346153846153846,
475
+ "loss": 0.402,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 1.49,
480
+ "learning_rate": 0.00016282051282051282,
481
+ "loss": 0.425,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 1.51,
486
+ "learning_rate": 0.0001621794871794872,
487
+ "loss": 0.5171,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 1.54,
492
+ "learning_rate": 0.00016153846153846155,
493
+ "loss": 0.5344,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 1.54,
498
+ "eval_accuracy": 0.7328843995510662,
499
+ "eval_loss": 0.7624912858009338,
500
+ "eval_runtime": 32.007,
501
+ "eval_samples_per_second": 111.476,
502
+ "eval_steps_per_second": 13.934,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 1.56,
507
+ "learning_rate": 0.00016089743589743592,
508
+ "loss": 0.4913,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 1.59,
513
+ "learning_rate": 0.00016025641025641028,
514
+ "loss": 0.5042,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 1.62,
519
+ "learning_rate": 0.00015961538461538462,
520
+ "loss": 0.6051,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 1.64,
525
+ "learning_rate": 0.00015897435897435896,
526
+ "loss": 0.5753,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 1.64,
531
+ "eval_accuracy": 0.7598204264870931,
532
+ "eval_loss": 0.6743783950805664,
533
+ "eval_runtime": 30.705,
534
+ "eval_samples_per_second": 116.203,
535
+ "eval_steps_per_second": 14.525,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 1.67,
540
+ "learning_rate": 0.00015833333333333332,
541
+ "loss": 0.524,
542
+ "step": 650
543
+ },
544
+ {
545
+ "epoch": 1.69,
546
+ "learning_rate": 0.0001576923076923077,
547
+ "loss": 0.5007,
548
+ "step": 660
549
+ },
550
+ {
551
+ "epoch": 1.72,
552
+ "learning_rate": 0.00015705128205128205,
553
+ "loss": 0.57,
554
+ "step": 670
555
+ },
556
+ {
557
+ "epoch": 1.74,
558
+ "learning_rate": 0.00015641025641025642,
559
+ "loss": 0.5533,
560
+ "step": 680
561
+ },
562
+ {
563
+ "epoch": 1.74,
564
+ "eval_accuracy": 0.734006734006734,
565
+ "eval_loss": 0.7230738997459412,
566
+ "eval_runtime": 33.3034,
567
+ "eval_samples_per_second": 107.136,
568
+ "eval_steps_per_second": 13.392,
569
+ "step": 680
570
+ },
571
+ {
572
+ "epoch": 1.77,
573
+ "learning_rate": 0.00015576923076923078,
574
+ "loss": 0.4399,
575
+ "step": 690
576
+ },
577
+ {
578
+ "epoch": 1.79,
579
+ "learning_rate": 0.00015512820512820515,
580
+ "loss": 0.453,
581
+ "step": 700
582
+ },
583
+ {
584
+ "epoch": 1.82,
585
+ "learning_rate": 0.00015448717948717951,
586
+ "loss": 0.4504,
587
+ "step": 710
588
+ },
589
+ {
590
+ "epoch": 1.85,
591
+ "learning_rate": 0.00015384615384615385,
592
+ "loss": 0.4244,
593
+ "step": 720
594
+ },
595
+ {
596
+ "epoch": 1.85,
597
+ "eval_accuracy": 0.7687991021324355,
598
+ "eval_loss": 0.6673141717910767,
599
+ "eval_runtime": 32.6948,
600
+ "eval_samples_per_second": 109.131,
601
+ "eval_steps_per_second": 13.641,
602
+ "step": 720
603
+ },
604
+ {
605
+ "epoch": 1.87,
606
+ "learning_rate": 0.00015320512820512822,
607
+ "loss": 0.46,
608
+ "step": 730
609
+ },
610
+ {
611
+ "epoch": 1.9,
612
+ "learning_rate": 0.00015256410256410255,
613
+ "loss": 0.4132,
614
+ "step": 740
615
+ },
616
+ {
617
+ "epoch": 1.92,
618
+ "learning_rate": 0.00015192307692307692,
619
+ "loss": 0.4849,
620
+ "step": 750
621
+ },
622
+ {
623
+ "epoch": 1.95,
624
+ "learning_rate": 0.00015128205128205128,
625
+ "loss": 0.5423,
626
+ "step": 760
627
+ },
628
+ {
629
+ "epoch": 1.95,
630
+ "eval_accuracy": 0.7413019079685746,
631
+ "eval_loss": 0.7185856699943542,
632
+ "eval_runtime": 35.9539,
633
+ "eval_samples_per_second": 99.238,
634
+ "eval_steps_per_second": 12.405,
635
+ "step": 760
636
+ },
637
+ {
638
+ "epoch": 1.97,
639
+ "learning_rate": 0.00015064102564102565,
640
+ "loss": 0.3997,
641
+ "step": 770
642
+ },
643
+ {
644
+ "epoch": 2.0,
645
+ "learning_rate": 0.00015000000000000001,
646
+ "loss": 0.4744,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 2.03,
651
+ "learning_rate": 0.00014935897435897438,
652
+ "loss": 0.4717,
653
+ "step": 790
654
+ },
655
+ {
656
+ "epoch": 2.05,
657
+ "learning_rate": 0.00014871794871794872,
658
+ "loss": 0.3384,
659
+ "step": 800
660
+ },
661
+ {
662
+ "epoch": 2.05,
663
+ "eval_accuracy": 0.7351290684624018,
664
+ "eval_loss": 0.7269920706748962,
665
+ "eval_runtime": 33.5897,
666
+ "eval_samples_per_second": 106.223,
667
+ "eval_steps_per_second": 13.278,
668
+ "step": 800
669
+ },
670
+ {
671
+ "epoch": 2.08,
672
+ "learning_rate": 0.00014807692307692308,
673
+ "loss": 0.2602,
674
+ "step": 810
675
+ },
676
+ {
677
+ "epoch": 2.1,
678
+ "learning_rate": 0.00014743589743589745,
679
+ "loss": 0.2345,
680
+ "step": 820
681
+ },
682
+ {
683
+ "epoch": 2.13,
684
+ "learning_rate": 0.00014679487179487178,
685
+ "loss": 0.2135,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 2.15,
690
+ "learning_rate": 0.00014615384615384615,
691
+ "loss": 0.2797,
692
+ "step": 840
693
+ },
694
+ {
695
+ "epoch": 2.15,
696
+ "eval_accuracy": 0.7497194163860831,
697
+ "eval_loss": 0.7742622494697571,
698
+ "eval_runtime": 31.2996,
699
+ "eval_samples_per_second": 113.995,
700
+ "eval_steps_per_second": 14.249,
701
+ "step": 840
702
+ },
703
+ {
704
+ "epoch": 2.18,
705
+ "learning_rate": 0.00014551282051282051,
706
+ "loss": 0.1862,
707
+ "step": 850
708
+ },
709
+ {
710
+ "epoch": 2.21,
711
+ "learning_rate": 0.00014487179487179488,
712
+ "loss": 0.2334,
713
+ "step": 860
714
+ },
715
+ {
716
+ "epoch": 2.23,
717
+ "learning_rate": 0.00014423076923076924,
718
+ "loss": 0.3361,
719
+ "step": 870
720
+ },
721
+ {
722
+ "epoch": 2.26,
723
+ "learning_rate": 0.0001435897435897436,
724
+ "loss": 0.2939,
725
+ "step": 880
726
+ },
727
+ {
728
+ "epoch": 2.26,
729
+ "eval_accuracy": 0.7732884399551067,
730
+ "eval_loss": 0.6985616087913513,
731
+ "eval_runtime": 32.9047,
732
+ "eval_samples_per_second": 108.435,
733
+ "eval_steps_per_second": 13.554,
734
+ "step": 880
735
+ },
736
+ {
737
+ "epoch": 2.28,
738
+ "learning_rate": 0.00014294871794871795,
739
+ "loss": 0.2846,
740
+ "step": 890
741
+ },
742
+ {
743
+ "epoch": 2.31,
744
+ "learning_rate": 0.0001423076923076923,
745
+ "loss": 0.2555,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 2.33,
750
+ "learning_rate": 0.00014166666666666668,
751
+ "loss": 0.2629,
752
+ "step": 910
753
+ },
754
+ {
755
+ "epoch": 2.36,
756
+ "learning_rate": 0.00014102564102564104,
757
+ "loss": 0.2288,
758
+ "step": 920
759
+ },
760
+ {
761
+ "epoch": 2.36,
762
+ "eval_accuracy": 0.7480359147025814,
763
+ "eval_loss": 0.761025607585907,
764
+ "eval_runtime": 34.1426,
765
+ "eval_samples_per_second": 104.503,
766
+ "eval_steps_per_second": 13.063,
767
+ "step": 920
768
+ },
769
+ {
770
+ "epoch": 2.38,
771
+ "learning_rate": 0.00014038461538461538,
772
+ "loss": 0.2115,
773
+ "step": 930
774
+ },
775
+ {
776
+ "epoch": 2.41,
777
+ "learning_rate": 0.00013974358974358974,
778
+ "loss": 0.2036,
779
+ "step": 940
780
+ },
781
+ {
782
+ "epoch": 2.44,
783
+ "learning_rate": 0.0001391025641025641,
784
+ "loss": 0.2913,
785
+ "step": 950
786
+ },
787
+ {
788
+ "epoch": 2.46,
789
+ "learning_rate": 0.00013846153846153847,
790
+ "loss": 0.2204,
791
+ "step": 960
792
+ },
793
+ {
794
+ "epoch": 2.46,
795
+ "eval_accuracy": 0.7570145903479237,
796
+ "eval_loss": 0.7840890884399414,
797
+ "eval_runtime": 35.7983,
798
+ "eval_samples_per_second": 99.67,
799
+ "eval_steps_per_second": 12.459,
800
+ "step": 960
801
+ },
802
+ {
803
+ "epoch": 2.49,
804
+ "learning_rate": 0.00013782051282051284,
805
+ "loss": 0.2759,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 2.51,
810
+ "learning_rate": 0.00013717948717948718,
811
+ "loss": 0.3945,
812
+ "step": 980
813
+ },
814
+ {
815
+ "epoch": 2.54,
816
+ "learning_rate": 0.00013653846153846154,
817
+ "loss": 0.3639,
818
+ "step": 990
819
+ },
820
+ {
821
+ "epoch": 2.56,
822
+ "learning_rate": 0.0001358974358974359,
823
+ "loss": 0.4397,
824
+ "step": 1000
825
+ },
826
+ {
827
+ "epoch": 2.56,
828
+ "eval_accuracy": 0.7789001122334456,
829
+ "eval_loss": 0.6565700173377991,
830
+ "eval_runtime": 31.6553,
831
+ "eval_samples_per_second": 112.714,
832
+ "eval_steps_per_second": 14.089,
833
+ "step": 1000
834
+ },
835
+ {
836
+ "epoch": 2.59,
837
+ "learning_rate": 0.00013525641025641027,
838
+ "loss": 0.2707,
839
+ "step": 1010
840
+ },
841
+ {
842
+ "epoch": 2.62,
843
+ "learning_rate": 0.00013461538461538464,
844
+ "loss": 0.3451,
845
+ "step": 1020
846
+ },
847
+ {
848
+ "epoch": 2.64,
849
+ "learning_rate": 0.00013397435897435897,
850
+ "loss": 0.3277,
851
+ "step": 1030
852
+ },
853
+ {
854
+ "epoch": 2.67,
855
+ "learning_rate": 0.00013333333333333334,
856
+ "loss": 0.2219,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 2.67,
861
+ "eval_accuracy": 0.7581369248035915,
862
+ "eval_loss": 0.6900522112846375,
863
+ "eval_runtime": 31.7845,
864
+ "eval_samples_per_second": 112.256,
865
+ "eval_steps_per_second": 14.032,
866
+ "step": 1040
867
+ },
868
+ {
869
+ "epoch": 2.69,
870
+ "learning_rate": 0.0001326923076923077,
871
+ "loss": 0.3174,
872
+ "step": 1050
873
+ },
874
+ {
875
+ "epoch": 2.72,
876
+ "learning_rate": 0.00013205128205128204,
877
+ "loss": 0.189,
878
+ "step": 1060
879
+ },
880
+ {
881
+ "epoch": 2.74,
882
+ "learning_rate": 0.0001314102564102564,
883
+ "loss": 0.2619,
884
+ "step": 1070
885
+ },
886
+ {
887
+ "epoch": 2.77,
888
+ "learning_rate": 0.00013076923076923077,
889
+ "loss": 0.2297,
890
+ "step": 1080
891
+ },
892
+ {
893
+ "epoch": 2.77,
894
+ "eval_accuracy": 0.7581369248035915,
895
+ "eval_loss": 0.7095093131065369,
896
+ "eval_runtime": 38.4561,
897
+ "eval_samples_per_second": 92.781,
898
+ "eval_steps_per_second": 11.598,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 2.79,
903
+ "learning_rate": 0.00013012820512820514,
904
+ "loss": 0.2201,
905
+ "step": 1090
906
+ },
907
+ {
908
+ "epoch": 2.82,
909
+ "learning_rate": 0.0001294871794871795,
910
+ "loss": 0.2261,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 2.85,
915
+ "learning_rate": 0.00012884615384615387,
916
+ "loss": 0.2095,
917
+ "step": 1110
918
+ },
919
+ {
920
+ "epoch": 2.87,
921
+ "learning_rate": 0.00012820512820512823,
922
+ "loss": 0.209,
923
+ "step": 1120
924
+ },
925
+ {
926
+ "epoch": 2.87,
927
+ "eval_accuracy": 0.7738496071829405,
928
+ "eval_loss": 0.7127913236618042,
929
+ "eval_runtime": 32.9997,
930
+ "eval_samples_per_second": 108.122,
931
+ "eval_steps_per_second": 13.515,
932
+ "step": 1120
933
+ },
934
+ {
935
+ "epoch": 2.9,
936
+ "learning_rate": 0.00012756410256410257,
937
+ "loss": 0.3365,
938
+ "step": 1130
939
+ },
940
+ {
941
+ "epoch": 2.92,
942
+ "learning_rate": 0.00012692307692307693,
943
+ "loss": 0.2639,
944
+ "step": 1140
945
+ },
946
+ {
947
+ "epoch": 2.95,
948
+ "learning_rate": 0.00012628205128205127,
949
+ "loss": 0.1892,
950
+ "step": 1150
951
+ },
952
+ {
953
+ "epoch": 2.97,
954
+ "learning_rate": 0.00012564102564102564,
955
+ "loss": 0.1853,
956
+ "step": 1160
957
+ },
958
+ {
959
+ "epoch": 2.97,
960
+ "eval_accuracy": 0.7710437710437711,
961
+ "eval_loss": 0.6986069083213806,
962
+ "eval_runtime": 32.2598,
963
+ "eval_samples_per_second": 110.602,
964
+ "eval_steps_per_second": 13.825,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 3.0,
969
+ "learning_rate": 0.000125,
970
+ "loss": 0.1807,
971
+ "step": 1170
972
+ },
973
+ {
974
+ "epoch": 3.03,
975
+ "learning_rate": 0.00012435897435897437,
976
+ "loss": 0.1554,
977
+ "step": 1180
978
+ },
979
+ {
980
+ "epoch": 3.05,
981
+ "learning_rate": 0.00012371794871794873,
982
+ "loss": 0.1845,
983
+ "step": 1190
984
+ },
985
+ {
986
+ "epoch": 3.08,
987
+ "learning_rate": 0.0001230769230769231,
988
+ "loss": 0.1322,
989
+ "step": 1200
990
+ },
991
+ {
992
+ "epoch": 3.08,
993
+ "eval_accuracy": 0.7962962962962963,
994
+ "eval_loss": 0.6381103992462158,
995
+ "eval_runtime": 35.1257,
996
+ "eval_samples_per_second": 101.578,
997
+ "eval_steps_per_second": 12.697,
998
+ "step": 1200
999
+ },
1000
+ {
1001
+ "epoch": 3.1,
1002
+ "learning_rate": 0.00012243589743589746,
1003
+ "loss": 0.0598,
1004
+ "step": 1210
1005
+ },
1006
+ {
1007
+ "epoch": 3.13,
1008
+ "learning_rate": 0.00012179487179487179,
1009
+ "loss": 0.0853,
1010
+ "step": 1220
1011
+ },
1012
+ {
1013
+ "epoch": 3.15,
1014
+ "learning_rate": 0.00012115384615384615,
1015
+ "loss": 0.1616,
1016
+ "step": 1230
1017
+ },
1018
+ {
1019
+ "epoch": 3.18,
1020
+ "learning_rate": 0.00012051282051282052,
1021
+ "loss": 0.2603,
1022
+ "step": 1240
1023
+ },
1024
+ {
1025
+ "epoch": 3.18,
1026
+ "eval_accuracy": 0.7682379349046016,
1027
+ "eval_loss": 0.7860389351844788,
1028
+ "eval_runtime": 35.648,
1029
+ "eval_samples_per_second": 100.09,
1030
+ "eval_steps_per_second": 12.511,
1031
+ "step": 1240
1032
+ },
1033
+ {
1034
+ "epoch": 3.21,
1035
+ "learning_rate": 0.00011987179487179487,
1036
+ "loss": 0.1104,
1037
+ "step": 1250
1038
+ },
1039
+ {
1040
+ "epoch": 3.23,
1041
+ "learning_rate": 0.00011923076923076923,
1042
+ "loss": 0.0634,
1043
+ "step": 1260
1044
+ },
1045
+ {
1046
+ "epoch": 3.26,
1047
+ "learning_rate": 0.0001185897435897436,
1048
+ "loss": 0.0717,
1049
+ "step": 1270
1050
+ },
1051
+ {
1052
+ "epoch": 3.28,
1053
+ "learning_rate": 0.00011794871794871796,
1054
+ "loss": 0.1031,
1055
+ "step": 1280
1056
+ },
1057
+ {
1058
+ "epoch": 3.28,
1059
+ "eval_accuracy": 0.7895622895622896,
1060
+ "eval_loss": 0.7322177886962891,
1061
+ "eval_runtime": 30.6139,
1062
+ "eval_samples_per_second": 116.549,
1063
+ "eval_steps_per_second": 14.569,
1064
+ "step": 1280
1065
+ },
1066
+ {
1067
+ "epoch": 3.31,
1068
+ "learning_rate": 0.00011730769230769231,
1069
+ "loss": 0.054,
1070
+ "step": 1290
1071
+ },
1072
+ {
1073
+ "epoch": 3.33,
1074
+ "learning_rate": 0.00011666666666666668,
1075
+ "loss": 0.0883,
1076
+ "step": 1300
1077
+ },
1078
+ {
1079
+ "epoch": 3.36,
1080
+ "learning_rate": 0.00011602564102564104,
1081
+ "loss": 0.1443,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 3.38,
1086
+ "learning_rate": 0.00011538461538461538,
1087
+ "loss": 0.0763,
1088
+ "step": 1320
1089
+ },
1090
+ {
1091
+ "epoch": 3.38,
1092
+ "eval_accuracy": 0.7839506172839507,
1093
+ "eval_loss": 0.77773118019104,
1094
+ "eval_runtime": 31.4828,
1095
+ "eval_samples_per_second": 113.332,
1096
+ "eval_steps_per_second": 14.166,
1097
+ "step": 1320
1098
+ },
1099
+ {
1100
+ "epoch": 3.41,
1101
+ "learning_rate": 0.00011474358974358975,
1102
+ "loss": 0.1328,
1103
+ "step": 1330
1104
+ },
1105
+ {
1106
+ "epoch": 3.44,
1107
+ "learning_rate": 0.0001141025641025641,
1108
+ "loss": 0.0751,
1109
+ "step": 1340
1110
+ },
1111
+ {
1112
+ "epoch": 3.46,
1113
+ "learning_rate": 0.00011346153846153846,
1114
+ "loss": 0.1193,
1115
+ "step": 1350
1116
+ },
1117
+ {
1118
+ "epoch": 3.49,
1119
+ "learning_rate": 0.00011282051282051283,
1120
+ "loss": 0.1437,
1121
+ "step": 1360
1122
+ },
1123
+ {
1124
+ "epoch": 3.49,
1125
+ "eval_accuracy": 0.7665544332210998,
1126
+ "eval_loss": 0.9293356537818909,
1127
+ "eval_runtime": 31.1681,
1128
+ "eval_samples_per_second": 114.476,
1129
+ "eval_steps_per_second": 14.31,
1130
+ "step": 1360
1131
+ },
1132
+ {
1133
+ "epoch": 3.51,
1134
+ "learning_rate": 0.00011217948717948718,
1135
+ "loss": 0.1805,
1136
+ "step": 1370
1137
+ },
1138
+ {
1139
+ "epoch": 3.54,
1140
+ "learning_rate": 0.00011153846153846154,
1141
+ "loss": 0.0931,
1142
+ "step": 1380
1143
+ },
1144
+ {
1145
+ "epoch": 3.56,
1146
+ "learning_rate": 0.00011089743589743591,
1147
+ "loss": 0.1227,
1148
+ "step": 1390
1149
+ },
1150
+ {
1151
+ "epoch": 3.59,
1152
+ "learning_rate": 0.00011025641025641027,
1153
+ "loss": 0.0818,
1154
+ "step": 1400
1155
+ },
1156
+ {
1157
+ "epoch": 3.59,
1158
+ "eval_accuracy": 0.7940516273849607,
1159
+ "eval_loss": 0.796922504901886,
1160
+ "eval_runtime": 31.9044,
1161
+ "eval_samples_per_second": 111.834,
1162
+ "eval_steps_per_second": 13.979,
1163
+ "step": 1400
1164
+ },
1165
+ {
1166
+ "epoch": 3.62,
1167
+ "learning_rate": 0.00010961538461538463,
1168
+ "loss": 0.1193,
1169
+ "step": 1410
1170
+ },
1171
+ {
1172
+ "epoch": 3.64,
1173
+ "learning_rate": 0.00010897435897435896,
1174
+ "loss": 0.1237,
1175
+ "step": 1420
1176
+ },
1177
+ {
1178
+ "epoch": 3.67,
1179
+ "learning_rate": 0.00010833333333333333,
1180
+ "loss": 0.1074,
1181
+ "step": 1430
1182
+ },
1183
+ {
1184
+ "epoch": 3.69,
1185
+ "learning_rate": 0.0001076923076923077,
1186
+ "loss": 0.1355,
1187
+ "step": 1440
1188
+ },
1189
+ {
1190
+ "epoch": 3.69,
1191
+ "eval_accuracy": 0.7716049382716049,
1192
+ "eval_loss": 0.8145824670791626,
1193
+ "eval_runtime": 37.6519,
1194
+ "eval_samples_per_second": 94.763,
1195
+ "eval_steps_per_second": 11.845,
1196
+ "step": 1440
1197
+ },
1198
+ {
1199
+ "epoch": 3.72,
1200
+ "learning_rate": 0.00010705128205128206,
1201
+ "loss": 0.0844,
1202
+ "step": 1450
1203
+ },
1204
+ {
1205
+ "epoch": 3.74,
1206
+ "learning_rate": 0.00010641025641025641,
1207
+ "loss": 0.1067,
1208
+ "step": 1460
1209
+ },
1210
+ {
1211
+ "epoch": 3.77,
1212
+ "learning_rate": 0.00010576923076923077,
1213
+ "loss": 0.0342,
1214
+ "step": 1470
1215
+ },
1216
+ {
1217
+ "epoch": 3.79,
1218
+ "learning_rate": 0.00010512820512820514,
1219
+ "loss": 0.0802,
1220
+ "step": 1480
1221
+ },
1222
+ {
1223
+ "epoch": 3.79,
1224
+ "eval_accuracy": 0.8075196408529742,
1225
+ "eval_loss": 0.6976819038391113,
1226
+ "eval_runtime": 32.2393,
1227
+ "eval_samples_per_second": 110.672,
1228
+ "eval_steps_per_second": 13.834,
1229
+ "step": 1480
1230
+ },
1231
+ {
1232
+ "epoch": 3.82,
1233
+ "learning_rate": 0.0001044871794871795,
1234
+ "loss": 0.1143,
1235
+ "step": 1490
1236
+ },
1237
+ {
1238
+ "epoch": 3.85,
1239
+ "learning_rate": 0.00010384615384615386,
1240
+ "loss": 0.1332,
1241
+ "step": 1500
1242
+ },
1243
+ {
1244
+ "epoch": 3.87,
1245
+ "learning_rate": 0.00010320512820512822,
1246
+ "loss": 0.069,
1247
+ "step": 1510
1248
+ },
1249
+ {
1250
+ "epoch": 3.9,
1251
+ "learning_rate": 0.00010256410256410256,
1252
+ "loss": 0.032,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 3.9,
1257
+ "eval_accuracy": 0.7929292929292929,
1258
+ "eval_loss": 0.8349816203117371,
1259
+ "eval_runtime": 32.9637,
1260
+ "eval_samples_per_second": 108.24,
1261
+ "eval_steps_per_second": 13.53,
1262
+ "step": 1520
1263
+ },
1264
+ {
1265
+ "epoch": 3.92,
1266
+ "learning_rate": 0.00010192307692307692,
1267
+ "loss": 0.0722,
1268
+ "step": 1530
1269
+ },
1270
+ {
1271
+ "epoch": 3.95,
1272
+ "learning_rate": 0.00010128205128205129,
1273
+ "loss": 0.1383,
1274
+ "step": 1540
1275
+ },
1276
+ {
1277
+ "epoch": 3.97,
1278
+ "learning_rate": 0.00010064102564102564,
1279
+ "loss": 0.0861,
1280
+ "step": 1550
1281
+ },
1282
+ {
1283
+ "epoch": 4.0,
1284
+ "learning_rate": 0.0001,
1285
+ "loss": 0.1836,
1286
+ "step": 1560
1287
+ },
1288
+ {
1289
+ "epoch": 4.0,
1290
+ "eval_accuracy": 0.7968574635241302,
1291
+ "eval_loss": 0.775885820388794,
1292
+ "eval_runtime": 33.0589,
1293
+ "eval_samples_per_second": 107.928,
1294
+ "eval_steps_per_second": 13.491,
1295
+ "step": 1560
1296
+ },
1297
+ {
1298
+ "epoch": 4.03,
1299
+ "learning_rate": 9.935897435897437e-05,
1300
+ "loss": 0.0486,
1301
+ "step": 1570
1302
+ },
1303
+ {
1304
+ "epoch": 4.05,
1305
+ "learning_rate": 9.871794871794872e-05,
1306
+ "loss": 0.0369,
1307
+ "step": 1580
1308
+ },
1309
+ {
1310
+ "epoch": 4.08,
1311
+ "learning_rate": 9.807692307692307e-05,
1312
+ "loss": 0.0209,
1313
+ "step": 1590
1314
+ },
1315
+ {
1316
+ "epoch": 4.1,
1317
+ "learning_rate": 9.743589743589744e-05,
1318
+ "loss": 0.0391,
1319
+ "step": 1600
1320
+ },
1321
+ {
1322
+ "epoch": 4.1,
1323
+ "eval_accuracy": 0.8013468013468014,
1324
+ "eval_loss": 0.8052350282669067,
1325
+ "eval_runtime": 32.0537,
1326
+ "eval_samples_per_second": 111.313,
1327
+ "eval_steps_per_second": 13.914,
1328
+ "step": 1600
1329
+ },
1330
+ {
1331
+ "epoch": 4.13,
1332
+ "learning_rate": 9.67948717948718e-05,
1333
+ "loss": 0.0354,
1334
+ "step": 1610
1335
+ },
1336
+ {
1337
+ "epoch": 4.15,
1338
+ "learning_rate": 9.615384615384617e-05,
1339
+ "loss": 0.0738,
1340
+ "step": 1620
1341
+ },
1342
+ {
1343
+ "epoch": 4.18,
1344
+ "learning_rate": 9.551282051282052e-05,
1345
+ "loss": 0.0275,
1346
+ "step": 1630
1347
+ },
1348
+ {
1349
+ "epoch": 4.21,
1350
+ "learning_rate": 9.487179487179487e-05,
1351
+ "loss": 0.0406,
1352
+ "step": 1640
1353
+ },
1354
+ {
1355
+ "epoch": 4.21,
1356
+ "eval_accuracy": 0.7828282828282829,
1357
+ "eval_loss": 0.9251748919487,
1358
+ "eval_runtime": 31.6082,
1359
+ "eval_samples_per_second": 112.882,
1360
+ "eval_steps_per_second": 14.11,
1361
+ "step": 1640
1362
+ },
1363
+ {
1364
+ "epoch": 4.23,
1365
+ "learning_rate": 9.423076923076924e-05,
1366
+ "loss": 0.1377,
1367
+ "step": 1650
1368
+ },
1369
+ {
1370
+ "epoch": 4.26,
1371
+ "learning_rate": 9.35897435897436e-05,
1372
+ "loss": 0.0526,
1373
+ "step": 1660
1374
+ },
1375
+ {
1376
+ "epoch": 4.28,
1377
+ "learning_rate": 9.294871794871795e-05,
1378
+ "loss": 0.152,
1379
+ "step": 1670
1380
+ },
1381
+ {
1382
+ "epoch": 4.31,
1383
+ "learning_rate": 9.230769230769232e-05,
1384
+ "loss": 0.0488,
1385
+ "step": 1680
1386
+ },
1387
+ {
1388
+ "epoch": 4.31,
1389
+ "eval_accuracy": 0.7901234567901234,
1390
+ "eval_loss": 0.8203706741333008,
1391
+ "eval_runtime": 31.8606,
1392
+ "eval_samples_per_second": 111.988,
1393
+ "eval_steps_per_second": 13.998,
1394
+ "step": 1680
1395
+ },
1396
+ {
1397
+ "epoch": 4.33,
1398
+ "learning_rate": 9.166666666666667e-05,
1399
+ "loss": 0.0372,
1400
+ "step": 1690
1401
+ },
1402
+ {
1403
+ "epoch": 4.36,
1404
+ "learning_rate": 9.102564102564103e-05,
1405
+ "loss": 0.0968,
1406
+ "step": 1700
1407
+ },
1408
+ {
1409
+ "epoch": 4.38,
1410
+ "learning_rate": 9.038461538461538e-05,
1411
+ "loss": 0.0147,
1412
+ "step": 1710
1413
+ },
1414
+ {
1415
+ "epoch": 4.41,
1416
+ "learning_rate": 8.974358974358975e-05,
1417
+ "loss": 0.016,
1418
+ "step": 1720
1419
+ },
1420
+ {
1421
+ "epoch": 4.41,
1422
+ "eval_accuracy": 0.7794612794612794,
1423
+ "eval_loss": 0.9169319868087769,
1424
+ "eval_runtime": 33.724,
1425
+ "eval_samples_per_second": 105.8,
1426
+ "eval_steps_per_second": 13.225,
1427
+ "step": 1720
1428
+ },
1429
+ {
1430
+ "epoch": 4.44,
1431
+ "learning_rate": 8.910256410256411e-05,
1432
+ "loss": 0.0393,
1433
+ "step": 1730
1434
+ },
1435
+ {
1436
+ "epoch": 4.46,
1437
+ "learning_rate": 8.846153846153847e-05,
1438
+ "loss": 0.0258,
1439
+ "step": 1740
1440
+ },
1441
+ {
1442
+ "epoch": 4.49,
1443
+ "learning_rate": 8.782051282051283e-05,
1444
+ "loss": 0.009,
1445
+ "step": 1750
1446
+ },
1447
+ {
1448
+ "epoch": 4.51,
1449
+ "learning_rate": 8.717948717948718e-05,
1450
+ "loss": 0.0994,
1451
+ "step": 1760
1452
+ },
1453
+ {
1454
+ "epoch": 4.51,
1455
+ "eval_accuracy": 0.8120089786756454,
1456
+ "eval_loss": 0.7930631637573242,
1457
+ "eval_runtime": 34.0477,
1458
+ "eval_samples_per_second": 104.794,
1459
+ "eval_steps_per_second": 13.099,
1460
+ "step": 1760
1461
+ },
1462
+ {
1463
+ "epoch": 4.54,
1464
+ "learning_rate": 8.653846153846155e-05,
1465
+ "loss": 0.0418,
1466
+ "step": 1770
1467
+ },
1468
+ {
1469
+ "epoch": 4.56,
1470
+ "learning_rate": 8.58974358974359e-05,
1471
+ "loss": 0.0793,
1472
+ "step": 1780
1473
+ },
1474
+ {
1475
+ "epoch": 4.59,
1476
+ "learning_rate": 8.525641025641026e-05,
1477
+ "loss": 0.0122,
1478
+ "step": 1790
1479
+ },
1480
+ {
1481
+ "epoch": 4.62,
1482
+ "learning_rate": 8.461538461538461e-05,
1483
+ "loss": 0.07,
1484
+ "step": 1800
1485
+ },
1486
+ {
1487
+ "epoch": 4.62,
1488
+ "eval_accuracy": 0.8103254769921436,
1489
+ "eval_loss": 0.8245030045509338,
1490
+ "eval_runtime": 32.6583,
1491
+ "eval_samples_per_second": 109.252,
1492
+ "eval_steps_per_second": 13.657,
1493
+ "step": 1800
1494
+ },
1495
+ {
1496
+ "epoch": 4.64,
1497
+ "learning_rate": 8.397435897435898e-05,
1498
+ "loss": 0.0656,
1499
+ "step": 1810
1500
+ },
1501
+ {
1502
+ "epoch": 4.67,
1503
+ "learning_rate": 8.333333333333334e-05,
1504
+ "loss": 0.0364,
1505
+ "step": 1820
1506
+ },
1507
+ {
1508
+ "epoch": 4.69,
1509
+ "learning_rate": 8.26923076923077e-05,
1510
+ "loss": 0.0639,
1511
+ "step": 1830
1512
+ },
1513
+ {
1514
+ "epoch": 4.72,
1515
+ "learning_rate": 8.205128205128205e-05,
1516
+ "loss": 0.0088,
1517
+ "step": 1840
1518
+ },
1519
+ {
1520
+ "epoch": 4.72,
1521
+ "eval_accuracy": 0.7985409652076318,
1522
+ "eval_loss": 0.907173752784729,
1523
+ "eval_runtime": 40.0812,
1524
+ "eval_samples_per_second": 89.019,
1525
+ "eval_steps_per_second": 11.127,
1526
+ "step": 1840
1527
+ },
1528
+ {
1529
+ "epoch": 4.74,
1530
+ "learning_rate": 8.141025641025641e-05,
1531
+ "loss": 0.0557,
1532
+ "step": 1850
1533
+ },
1534
+ {
1535
+ "epoch": 4.77,
1536
+ "learning_rate": 8.076923076923078e-05,
1537
+ "loss": 0.0548,
1538
+ "step": 1860
1539
+ },
1540
+ {
1541
+ "epoch": 4.79,
1542
+ "learning_rate": 8.012820512820514e-05,
1543
+ "loss": 0.0055,
1544
+ "step": 1870
1545
+ },
1546
+ {
1547
+ "epoch": 4.82,
1548
+ "learning_rate": 7.948717948717948e-05,
1549
+ "loss": 0.0085,
1550
+ "step": 1880
1551
+ },
1552
+ {
1553
+ "epoch": 4.82,
1554
+ "eval_accuracy": 0.8114478114478114,
1555
+ "eval_loss": 0.7862226963043213,
1556
+ "eval_runtime": 33.9761,
1557
+ "eval_samples_per_second": 105.015,
1558
+ "eval_steps_per_second": 13.127,
1559
+ "step": 1880
1560
+ },
1561
+ {
1562
+ "epoch": 4.85,
1563
+ "learning_rate": 7.884615384615384e-05,
1564
+ "loss": 0.0423,
1565
+ "step": 1890
1566
+ },
1567
+ {
1568
+ "epoch": 4.87,
1569
+ "learning_rate": 7.820512820512821e-05,
1570
+ "loss": 0.0581,
1571
+ "step": 1900
1572
+ },
1573
+ {
1574
+ "epoch": 4.9,
1575
+ "learning_rate": 7.756410256410257e-05,
1576
+ "loss": 0.0043,
1577
+ "step": 1910
1578
+ },
1579
+ {
1580
+ "epoch": 4.92,
1581
+ "learning_rate": 7.692307692307693e-05,
1582
+ "loss": 0.083,
1583
+ "step": 1920
1584
+ },
1585
+ {
1586
+ "epoch": 4.92,
1587
+ "eval_accuracy": 0.8198653198653199,
1588
+ "eval_loss": 0.7796981334686279,
1589
+ "eval_runtime": 32.3864,
1590
+ "eval_samples_per_second": 110.17,
1591
+ "eval_steps_per_second": 13.771,
1592
+ "step": 1920
1593
+ },
1594
+ {
1595
+ "epoch": 4.95,
1596
+ "learning_rate": 7.628205128205128e-05,
1597
+ "loss": 0.1308,
1598
+ "step": 1930
1599
+ },
1600
+ {
1601
+ "epoch": 4.97,
1602
+ "learning_rate": 7.564102564102564e-05,
1603
+ "loss": 0.0063,
1604
+ "step": 1940
1605
+ },
1606
+ {
1607
+ "epoch": 5.0,
1608
+ "learning_rate": 7.500000000000001e-05,
1609
+ "loss": 0.0132,
1610
+ "step": 1950
1611
+ },
1612
+ {
1613
+ "epoch": 5.03,
1614
+ "learning_rate": 7.435897435897436e-05,
1615
+ "loss": 0.0055,
1616
+ "step": 1960
1617
+ },
1618
+ {
1619
+ "epoch": 5.03,
1620
+ "eval_accuracy": 0.8069584736251403,
1621
+ "eval_loss": 0.8723996877670288,
1622
+ "eval_runtime": 33.6185,
1623
+ "eval_samples_per_second": 106.132,
1624
+ "eval_steps_per_second": 13.267,
1625
+ "step": 1960
1626
+ },
1627
+ {
1628
+ "epoch": 5.05,
1629
+ "learning_rate": 7.371794871794872e-05,
1630
+ "loss": 0.0366,
1631
+ "step": 1970
1632
+ },
1633
+ {
1634
+ "epoch": 5.08,
1635
+ "learning_rate": 7.307692307692307e-05,
1636
+ "loss": 0.06,
1637
+ "step": 1980
1638
+ },
1639
+ {
1640
+ "epoch": 5.1,
1641
+ "learning_rate": 7.243589743589744e-05,
1642
+ "loss": 0.0065,
1643
+ "step": 1990
1644
+ },
1645
+ {
1646
+ "epoch": 5.13,
1647
+ "learning_rate": 7.17948717948718e-05,
1648
+ "loss": 0.0223,
1649
+ "step": 2000
1650
+ },
1651
+ {
1652
+ "epoch": 5.13,
1653
+ "eval_accuracy": 0.8249158249158249,
1654
+ "eval_loss": 0.7518730163574219,
1655
+ "eval_runtime": 43.1922,
1656
+ "eval_samples_per_second": 82.607,
1657
+ "eval_steps_per_second": 10.326,
1658
+ "step": 2000
1659
+ },
1660
+ {
1661
+ "epoch": 5.15,
1662
+ "learning_rate": 7.115384615384616e-05,
1663
+ "loss": 0.0215,
1664
+ "step": 2010
1665
+ },
1666
+ {
1667
+ "epoch": 5.18,
1668
+ "learning_rate": 7.051282051282052e-05,
1669
+ "loss": 0.0494,
1670
+ "step": 2020
1671
+ },
1672
+ {
1673
+ "epoch": 5.21,
1674
+ "learning_rate": 6.987179487179487e-05,
1675
+ "loss": 0.0441,
1676
+ "step": 2030
1677
+ },
1678
+ {
1679
+ "epoch": 5.23,
1680
+ "learning_rate": 6.923076923076924e-05,
1681
+ "loss": 0.0042,
1682
+ "step": 2040
1683
+ },
1684
+ {
1685
+ "epoch": 5.23,
1686
+ "eval_accuracy": 0.8232323232323232,
1687
+ "eval_loss": 0.7583897113800049,
1688
+ "eval_runtime": 31.893,
1689
+ "eval_samples_per_second": 111.874,
1690
+ "eval_steps_per_second": 13.984,
1691
+ "step": 2040
1692
+ },
1693
+ {
1694
+ "epoch": 5.26,
1695
+ "learning_rate": 6.858974358974359e-05,
1696
+ "loss": 0.0131,
1697
+ "step": 2050
1698
+ },
1699
+ {
1700
+ "epoch": 5.28,
1701
+ "learning_rate": 6.794871794871795e-05,
1702
+ "loss": 0.0037,
1703
+ "step": 2060
1704
+ },
1705
+ {
1706
+ "epoch": 5.31,
1707
+ "learning_rate": 6.730769230769232e-05,
1708
+ "loss": 0.02,
1709
+ "step": 2070
1710
+ },
1711
+ {
1712
+ "epoch": 5.33,
1713
+ "learning_rate": 6.666666666666667e-05,
1714
+ "loss": 0.0178,
1715
+ "step": 2080
1716
+ },
1717
+ {
1718
+ "epoch": 5.33,
1719
+ "eval_accuracy": 0.8080808080808081,
1720
+ "eval_loss": 0.8523600697517395,
1721
+ "eval_runtime": 32.5102,
1722
+ "eval_samples_per_second": 109.75,
1723
+ "eval_steps_per_second": 13.719,
1724
+ "step": 2080
1725
+ },
1726
+ {
1727
+ "epoch": 5.36,
1728
+ "learning_rate": 6.602564102564102e-05,
1729
+ "loss": 0.0035,
1730
+ "step": 2090
1731
+ },
1732
+ {
1733
+ "epoch": 5.38,
1734
+ "learning_rate": 6.538461538461539e-05,
1735
+ "loss": 0.0137,
1736
+ "step": 2100
1737
+ },
1738
+ {
1739
+ "epoch": 5.41,
1740
+ "learning_rate": 6.474358974358975e-05,
1741
+ "loss": 0.0033,
1742
+ "step": 2110
1743
+ },
1744
+ {
1745
+ "epoch": 5.44,
1746
+ "learning_rate": 6.410256410256412e-05,
1747
+ "loss": 0.0172,
1748
+ "step": 2120
1749
+ },
1750
+ {
1751
+ "epoch": 5.44,
1752
+ "eval_accuracy": 0.8215488215488216,
1753
+ "eval_loss": 0.7729219794273376,
1754
+ "eval_runtime": 31.7305,
1755
+ "eval_samples_per_second": 112.447,
1756
+ "eval_steps_per_second": 14.056,
1757
+ "step": 2120
1758
+ },
1759
+ {
1760
+ "epoch": 5.46,
1761
+ "learning_rate": 6.346153846153847e-05,
1762
+ "loss": 0.0252,
1763
+ "step": 2130
1764
+ },
1765
+ {
1766
+ "epoch": 5.49,
1767
+ "learning_rate": 6.282051282051282e-05,
1768
+ "loss": 0.0034,
1769
+ "step": 2140
1770
+ },
1771
+ {
1772
+ "epoch": 5.51,
1773
+ "learning_rate": 6.217948717948718e-05,
1774
+ "loss": 0.0045,
1775
+ "step": 2150
1776
+ },
1777
+ {
1778
+ "epoch": 5.54,
1779
+ "learning_rate": 6.153846153846155e-05,
1780
+ "loss": 0.0044,
1781
+ "step": 2160
1782
+ },
1783
+ {
1784
+ "epoch": 5.54,
1785
+ "eval_accuracy": 0.808641975308642,
1786
+ "eval_loss": 0.8701183795928955,
1787
+ "eval_runtime": 31.0299,
1788
+ "eval_samples_per_second": 114.986,
1789
+ "eval_steps_per_second": 14.373,
1790
+ "step": 2160
1791
+ },
1792
+ {
1793
+ "epoch": 5.56,
1794
+ "learning_rate": 6.089743589743589e-05,
1795
+ "loss": 0.0373,
1796
+ "step": 2170
1797
+ },
1798
+ {
1799
+ "epoch": 5.59,
1800
+ "learning_rate": 6.025641025641026e-05,
1801
+ "loss": 0.0029,
1802
+ "step": 2180
1803
+ },
1804
+ {
1805
+ "epoch": 5.62,
1806
+ "learning_rate": 5.9615384615384616e-05,
1807
+ "loss": 0.0032,
1808
+ "step": 2190
1809
+ },
1810
+ {
1811
+ "epoch": 5.64,
1812
+ "learning_rate": 5.897435897435898e-05,
1813
+ "loss": 0.003,
1814
+ "step": 2200
1815
+ },
1816
+ {
1817
+ "epoch": 5.64,
1818
+ "eval_accuracy": 0.8237934904601572,
1819
+ "eval_loss": 0.8060529828071594,
1820
+ "eval_runtime": 32.4421,
1821
+ "eval_samples_per_second": 109.98,
1822
+ "eval_steps_per_second": 13.748,
1823
+ "step": 2200
1824
+ },
1825
+ {
1826
+ "epoch": 5.67,
1827
+ "learning_rate": 5.833333333333334e-05,
1828
+ "loss": 0.0034,
1829
+ "step": 2210
1830
+ },
1831
+ {
1832
+ "epoch": 5.69,
1833
+ "learning_rate": 5.769230769230769e-05,
1834
+ "loss": 0.0025,
1835
+ "step": 2220
1836
+ },
1837
+ {
1838
+ "epoch": 5.72,
1839
+ "learning_rate": 5.705128205128205e-05,
1840
+ "loss": 0.0029,
1841
+ "step": 2230
1842
+ },
1843
+ {
1844
+ "epoch": 5.74,
1845
+ "learning_rate": 5.6410256410256414e-05,
1846
+ "loss": 0.0033,
1847
+ "step": 2240
1848
+ },
1849
+ {
1850
+ "epoch": 5.74,
1851
+ "eval_accuracy": 0.8226711560044894,
1852
+ "eval_loss": 0.813855767250061,
1853
+ "eval_runtime": 32.378,
1854
+ "eval_samples_per_second": 110.198,
1855
+ "eval_steps_per_second": 13.775,
1856
+ "step": 2240
1857
+ },
1858
+ {
1859
+ "epoch": 5.77,
1860
+ "learning_rate": 5.576923076923077e-05,
1861
+ "loss": 0.0129,
1862
+ "step": 2250
1863
+ },
1864
+ {
1865
+ "epoch": 5.79,
1866
+ "learning_rate": 5.512820512820514e-05,
1867
+ "loss": 0.0079,
1868
+ "step": 2260
1869
+ },
1870
+ {
1871
+ "epoch": 5.82,
1872
+ "learning_rate": 5.448717948717948e-05,
1873
+ "loss": 0.0117,
1874
+ "step": 2270
1875
+ },
1876
+ {
1877
+ "epoch": 5.85,
1878
+ "learning_rate": 5.384615384615385e-05,
1879
+ "loss": 0.0023,
1880
+ "step": 2280
1881
+ },
1882
+ {
1883
+ "epoch": 5.85,
1884
+ "eval_accuracy": 0.8164983164983165,
1885
+ "eval_loss": 0.8478395342826843,
1886
+ "eval_runtime": 35.321,
1887
+ "eval_samples_per_second": 101.016,
1888
+ "eval_steps_per_second": 12.627,
1889
+ "step": 2280
1890
+ },
1891
+ {
1892
+ "epoch": 5.87,
1893
+ "learning_rate": 5.3205128205128205e-05,
1894
+ "loss": 0.0075,
1895
+ "step": 2290
1896
+ },
1897
+ {
1898
+ "epoch": 5.9,
1899
+ "learning_rate": 5.256410256410257e-05,
1900
+ "loss": 0.0023,
1901
+ "step": 2300
1902
+ },
1903
+ {
1904
+ "epoch": 5.92,
1905
+ "learning_rate": 5.192307692307693e-05,
1906
+ "loss": 0.0058,
1907
+ "step": 2310
1908
+ },
1909
+ {
1910
+ "epoch": 5.95,
1911
+ "learning_rate": 5.128205128205128e-05,
1912
+ "loss": 0.003,
1913
+ "step": 2320
1914
+ },
1915
+ {
1916
+ "epoch": 5.95,
1917
+ "eval_accuracy": 0.8120089786756454,
1918
+ "eval_loss": 0.8443180918693542,
1919
+ "eval_runtime": 32.2001,
1920
+ "eval_samples_per_second": 110.807,
1921
+ "eval_steps_per_second": 13.851,
1922
+ "step": 2320
1923
+ },
1924
+ {
1925
+ "epoch": 5.97,
1926
+ "learning_rate": 5.0641025641025644e-05,
1927
+ "loss": 0.0512,
1928
+ "step": 2330
1929
+ },
1930
+ {
1931
+ "epoch": 6.0,
1932
+ "learning_rate": 5e-05,
1933
+ "loss": 0.0024,
1934
+ "step": 2340
1935
+ },
1936
+ {
1937
+ "epoch": 6.03,
1938
+ "learning_rate": 4.935897435897436e-05,
1939
+ "loss": 0.0022,
1940
+ "step": 2350
1941
+ },
1942
+ {
1943
+ "epoch": 6.05,
1944
+ "learning_rate": 4.871794871794872e-05,
1945
+ "loss": 0.031,
1946
+ "step": 2360
1947
+ },
1948
+ {
1949
+ "epoch": 6.05,
1950
+ "eval_accuracy": 0.8047138047138047,
1951
+ "eval_loss": 0.9272196888923645,
1952
+ "eval_runtime": 33.2145,
1953
+ "eval_samples_per_second": 107.423,
1954
+ "eval_steps_per_second": 13.428,
1955
+ "step": 2360
1956
+ },
1957
+ {
1958
+ "epoch": 6.08,
1959
+ "learning_rate": 4.8076923076923084e-05,
1960
+ "loss": 0.0456,
1961
+ "step": 2370
1962
+ },
1963
+ {
1964
+ "epoch": 6.1,
1965
+ "learning_rate": 4.7435897435897435e-05,
1966
+ "loss": 0.0024,
1967
+ "step": 2380
1968
+ },
1969
+ {
1970
+ "epoch": 6.13,
1971
+ "learning_rate": 4.67948717948718e-05,
1972
+ "loss": 0.0023,
1973
+ "step": 2390
1974
+ },
1975
+ {
1976
+ "epoch": 6.15,
1977
+ "learning_rate": 4.615384615384616e-05,
1978
+ "loss": 0.0021,
1979
+ "step": 2400
1980
+ },
1981
+ {
1982
+ "epoch": 6.15,
1983
+ "eval_accuracy": 0.8204264870931538,
1984
+ "eval_loss": 0.8369048237800598,
1985
+ "eval_runtime": 36.1395,
1986
+ "eval_samples_per_second": 98.728,
1987
+ "eval_steps_per_second": 12.341,
1988
+ "step": 2400
1989
+ },
1990
+ {
1991
+ "epoch": 6.18,
1992
+ "learning_rate": 4.5512820512820516e-05,
1993
+ "loss": 0.0021,
1994
+ "step": 2410
1995
+ },
1996
+ {
1997
+ "epoch": 6.21,
1998
+ "learning_rate": 4.4871794871794874e-05,
1999
+ "loss": 0.0026,
2000
+ "step": 2420
2001
+ },
2002
+ {
2003
+ "epoch": 6.23,
2004
+ "learning_rate": 4.423076923076923e-05,
2005
+ "loss": 0.0021,
2006
+ "step": 2430
2007
+ },
2008
+ {
2009
+ "epoch": 6.26,
2010
+ "learning_rate": 4.358974358974359e-05,
2011
+ "loss": 0.0019,
2012
+ "step": 2440
2013
+ },
2014
+ {
2015
+ "epoch": 6.26,
2016
+ "eval_accuracy": 0.8209876543209876,
2017
+ "eval_loss": 0.828059196472168,
2018
+ "eval_runtime": 32.1149,
2019
+ "eval_samples_per_second": 111.101,
2020
+ "eval_steps_per_second": 13.888,
2021
+ "step": 2440
2022
+ },
2023
+ {
2024
+ "epoch": 6.28,
2025
+ "learning_rate": 4.294871794871795e-05,
2026
+ "loss": 0.0024,
2027
+ "step": 2450
2028
+ },
2029
+ {
2030
+ "epoch": 6.31,
2031
+ "learning_rate": 4.230769230769231e-05,
2032
+ "loss": 0.0019,
2033
+ "step": 2460
2034
+ },
2035
+ {
2036
+ "epoch": 6.33,
2037
+ "learning_rate": 4.166666666666667e-05,
2038
+ "loss": 0.002,
2039
+ "step": 2470
2040
+ },
2041
+ {
2042
+ "epoch": 6.36,
2043
+ "learning_rate": 4.1025641025641023e-05,
2044
+ "loss": 0.0019,
2045
+ "step": 2480
2046
+ },
2047
+ {
2048
+ "epoch": 6.36,
2049
+ "eval_accuracy": 0.8198653198653199,
2050
+ "eval_loss": 0.834019124507904,
2051
+ "eval_runtime": 35.8439,
2052
+ "eval_samples_per_second": 99.543,
2053
+ "eval_steps_per_second": 12.443,
2054
+ "step": 2480
2055
+ },
2056
+ {
2057
+ "epoch": 6.38,
2058
+ "learning_rate": 4.038461538461539e-05,
2059
+ "loss": 0.002,
2060
+ "step": 2490
2061
+ },
2062
+ {
2063
+ "epoch": 6.41,
2064
+ "learning_rate": 3.974358974358974e-05,
2065
+ "loss": 0.0019,
2066
+ "step": 2500
2067
+ },
2068
+ {
2069
+ "epoch": 6.44,
2070
+ "learning_rate": 3.9102564102564105e-05,
2071
+ "loss": 0.0018,
2072
+ "step": 2510
2073
+ },
2074
+ {
2075
+ "epoch": 6.46,
2076
+ "learning_rate": 3.846153846153846e-05,
2077
+ "loss": 0.0018,
2078
+ "step": 2520
2079
+ },
2080
+ {
2081
+ "epoch": 6.46,
2082
+ "eval_accuracy": 0.8204264870931538,
2083
+ "eval_loss": 0.8339148759841919,
2084
+ "eval_runtime": 34.0288,
2085
+ "eval_samples_per_second": 104.853,
2086
+ "eval_steps_per_second": 13.107,
2087
+ "step": 2520
2088
+ },
2089
+ {
2090
+ "epoch": 6.49,
2091
+ "learning_rate": 3.782051282051282e-05,
2092
+ "loss": 0.0018,
2093
+ "step": 2530
2094
+ },
2095
+ {
2096
+ "epoch": 6.51,
2097
+ "learning_rate": 3.717948717948718e-05,
2098
+ "loss": 0.0018,
2099
+ "step": 2540
2100
+ },
2101
+ {
2102
+ "epoch": 6.54,
2103
+ "learning_rate": 3.653846153846154e-05,
2104
+ "loss": 0.0018,
2105
+ "step": 2550
2106
+ },
2107
+ {
2108
+ "epoch": 6.56,
2109
+ "learning_rate": 3.58974358974359e-05,
2110
+ "loss": 0.002,
2111
+ "step": 2560
2112
+ },
2113
+ {
2114
+ "epoch": 6.56,
2115
+ "eval_accuracy": 0.8209876543209876,
2116
+ "eval_loss": 0.8293877840042114,
2117
+ "eval_runtime": 33.9846,
2118
+ "eval_samples_per_second": 104.989,
2119
+ "eval_steps_per_second": 13.124,
2120
+ "step": 2560
2121
+ },
2122
+ {
2123
+ "epoch": 6.59,
2124
+ "learning_rate": 3.525641025641026e-05,
2125
+ "loss": 0.0017,
2126
+ "step": 2570
2127
+ },
2128
+ {
2129
+ "epoch": 6.62,
2130
+ "learning_rate": 3.461538461538462e-05,
2131
+ "loss": 0.0017,
2132
+ "step": 2580
2133
+ },
2134
+ {
2135
+ "epoch": 6.64,
2136
+ "learning_rate": 3.397435897435898e-05,
2137
+ "loss": 0.0018,
2138
+ "step": 2590
2139
+ },
2140
+ {
2141
+ "epoch": 6.67,
2142
+ "learning_rate": 3.3333333333333335e-05,
2143
+ "loss": 0.0017,
2144
+ "step": 2600
2145
+ },
2146
+ {
2147
+ "epoch": 6.67,
2148
+ "eval_accuracy": 0.8209876543209876,
2149
+ "eval_loss": 0.8266403079032898,
2150
+ "eval_runtime": 36.0162,
2151
+ "eval_samples_per_second": 99.067,
2152
+ "eval_steps_per_second": 12.383,
2153
+ "step": 2600
2154
+ },
2155
+ {
2156
+ "epoch": 6.69,
2157
+ "learning_rate": 3.269230769230769e-05,
2158
+ "loss": 0.002,
2159
+ "step": 2610
2160
+ },
2161
+ {
2162
+ "epoch": 6.72,
2163
+ "learning_rate": 3.205128205128206e-05,
2164
+ "loss": 0.0017,
2165
+ "step": 2620
2166
+ },
2167
+ {
2168
+ "epoch": 6.74,
2169
+ "learning_rate": 3.141025641025641e-05,
2170
+ "loss": 0.0017,
2171
+ "step": 2630
2172
+ },
2173
+ {
2174
+ "epoch": 6.77,
2175
+ "learning_rate": 3.0769230769230774e-05,
2176
+ "loss": 0.0017,
2177
+ "step": 2640
2178
+ },
2179
+ {
2180
+ "epoch": 6.77,
2181
+ "eval_accuracy": 0.8226711560044894,
2182
+ "eval_loss": 0.8177886009216309,
2183
+ "eval_runtime": 36.3329,
2184
+ "eval_samples_per_second": 98.203,
2185
+ "eval_steps_per_second": 12.275,
2186
+ "step": 2640
2187
+ },
2188
+ {
2189
+ "epoch": 6.79,
2190
+ "learning_rate": 3.012820512820513e-05,
2191
+ "loss": 0.0016,
2192
+ "step": 2650
2193
+ },
2194
+ {
2195
+ "epoch": 6.82,
2196
+ "learning_rate": 2.948717948717949e-05,
2197
+ "loss": 0.0016,
2198
+ "step": 2660
2199
+ },
2200
+ {
2201
+ "epoch": 6.85,
2202
+ "learning_rate": 2.8846153846153845e-05,
2203
+ "loss": 0.0016,
2204
+ "step": 2670
2205
+ },
2206
+ {
2207
+ "epoch": 6.87,
2208
+ "learning_rate": 2.8205128205128207e-05,
2209
+ "loss": 0.0017,
2210
+ "step": 2680
2211
+ },
2212
+ {
2213
+ "epoch": 6.87,
2214
+ "eval_accuracy": 0.824354657687991,
2215
+ "eval_loss": 0.8174560070037842,
2216
+ "eval_runtime": 32.3063,
2217
+ "eval_samples_per_second": 110.443,
2218
+ "eval_steps_per_second": 13.805,
2219
+ "step": 2680
2220
+ },
2221
+ {
2222
+ "epoch": 6.9,
2223
+ "learning_rate": 2.756410256410257e-05,
2224
+ "loss": 0.0017,
2225
+ "step": 2690
2226
+ },
2227
+ {
2228
+ "epoch": 6.92,
2229
+ "learning_rate": 2.6923076923076923e-05,
2230
+ "loss": 0.0017,
2231
+ "step": 2700
2232
+ },
2233
+ {
2234
+ "epoch": 6.95,
2235
+ "learning_rate": 2.6282051282051285e-05,
2236
+ "loss": 0.0017,
2237
+ "step": 2710
2238
+ },
2239
+ {
2240
+ "epoch": 6.97,
2241
+ "learning_rate": 2.564102564102564e-05,
2242
+ "loss": 0.0017,
2243
+ "step": 2720
2244
+ },
2245
+ {
2246
+ "epoch": 6.97,
2247
+ "eval_accuracy": 0.8237934904601572,
2248
+ "eval_loss": 0.8166114687919617,
2249
+ "eval_runtime": 33.2936,
2250
+ "eval_samples_per_second": 107.168,
2251
+ "eval_steps_per_second": 13.396,
2252
+ "step": 2720
2253
+ },
2254
+ {
2255
+ "epoch": 7.0,
2256
+ "learning_rate": 2.5e-05,
2257
+ "loss": 0.0015,
2258
+ "step": 2730
2259
+ },
2260
+ {
2261
+ "epoch": 7.03,
2262
+ "learning_rate": 2.435897435897436e-05,
2263
+ "loss": 0.0016,
2264
+ "step": 2740
2265
+ },
2266
+ {
2267
+ "epoch": 7.05,
2268
+ "learning_rate": 2.3717948717948718e-05,
2269
+ "loss": 0.0016,
2270
+ "step": 2750
2271
+ },
2272
+ {
2273
+ "epoch": 7.08,
2274
+ "learning_rate": 2.307692307692308e-05,
2275
+ "loss": 0.0015,
2276
+ "step": 2760
2277
+ },
2278
+ {
2279
+ "epoch": 7.08,
2280
+ "eval_accuracy": 0.824354657687991,
2281
+ "eval_loss": 0.8176218271255493,
2282
+ "eval_runtime": 32.3011,
2283
+ "eval_samples_per_second": 110.46,
2284
+ "eval_steps_per_second": 13.808,
2285
+ "step": 2760
2286
+ },
2287
+ {
2288
+ "epoch": 7.1,
2289
+ "learning_rate": 2.2435897435897437e-05,
2290
+ "loss": 0.0015,
2291
+ "step": 2770
2292
+ },
2293
+ {
2294
+ "epoch": 7.13,
2295
+ "learning_rate": 2.1794871794871795e-05,
2296
+ "loss": 0.0015,
2297
+ "step": 2780
2298
+ },
2299
+ {
2300
+ "epoch": 7.15,
2301
+ "learning_rate": 2.1153846153846154e-05,
2302
+ "loss": 0.0015,
2303
+ "step": 2790
2304
+ },
2305
+ {
2306
+ "epoch": 7.18,
2307
+ "learning_rate": 2.0512820512820512e-05,
2308
+ "loss": 0.0015,
2309
+ "step": 2800
2310
+ },
2311
+ {
2312
+ "epoch": 7.18,
2313
+ "eval_accuracy": 0.824354657687991,
2314
+ "eval_loss": 0.8185549378395081,
2315
+ "eval_runtime": 32.7126,
2316
+ "eval_samples_per_second": 109.071,
2317
+ "eval_steps_per_second": 13.634,
2318
+ "step": 2800
2319
+ },
2320
+ {
2321
+ "epoch": 7.21,
2322
+ "learning_rate": 1.987179487179487e-05,
2323
+ "loss": 0.0015,
2324
+ "step": 2810
2325
+ },
2326
+ {
2327
+ "epoch": 7.23,
2328
+ "learning_rate": 1.923076923076923e-05,
2329
+ "loss": 0.0015,
2330
+ "step": 2820
2331
+ },
2332
+ {
2333
+ "epoch": 7.26,
2334
+ "learning_rate": 1.858974358974359e-05,
2335
+ "loss": 0.0015,
2336
+ "step": 2830
2337
+ },
2338
+ {
2339
+ "epoch": 7.28,
2340
+ "learning_rate": 1.794871794871795e-05,
2341
+ "loss": 0.0015,
2342
+ "step": 2840
2343
+ },
2344
+ {
2345
+ "epoch": 7.28,
2346
+ "eval_accuracy": 0.8249158249158249,
2347
+ "eval_loss": 0.819684624671936,
2348
+ "eval_runtime": 35.8925,
2349
+ "eval_samples_per_second": 99.408,
2350
+ "eval_steps_per_second": 12.426,
2351
+ "step": 2840
2352
+ },
2353
+ {
2354
+ "epoch": 7.31,
2355
+ "learning_rate": 1.730769230769231e-05,
2356
+ "loss": 0.0016,
2357
+ "step": 2850
2358
+ },
2359
+ {
2360
+ "epoch": 7.33,
2361
+ "learning_rate": 1.6666666666666667e-05,
2362
+ "loss": 0.0015,
2363
+ "step": 2860
2364
+ },
2365
+ {
2366
+ "epoch": 7.36,
2367
+ "learning_rate": 1.602564102564103e-05,
2368
+ "loss": 0.0015,
2369
+ "step": 2870
2370
+ },
2371
+ {
2372
+ "epoch": 7.38,
2373
+ "learning_rate": 1.5384615384615387e-05,
2374
+ "loss": 0.0015,
2375
+ "step": 2880
2376
+ },
2377
+ {
2378
+ "epoch": 7.38,
2379
+ "eval_accuracy": 0.8249158249158249,
2380
+ "eval_loss": 0.8204275965690613,
2381
+ "eval_runtime": 31.8471,
2382
+ "eval_samples_per_second": 112.035,
2383
+ "eval_steps_per_second": 14.004,
2384
+ "step": 2880
2385
+ },
2386
+ {
2387
+ "epoch": 7.41,
2388
+ "learning_rate": 1.4743589743589745e-05,
2389
+ "loss": 0.0015,
2390
+ "step": 2890
2391
+ },
2392
+ {
2393
+ "epoch": 7.44,
2394
+ "learning_rate": 1.4102564102564104e-05,
2395
+ "loss": 0.0015,
2396
+ "step": 2900
2397
+ },
2398
+ {
2399
+ "epoch": 7.46,
2400
+ "learning_rate": 1.3461538461538462e-05,
2401
+ "loss": 0.0015,
2402
+ "step": 2910
2403
+ },
2404
+ {
2405
+ "epoch": 7.49,
2406
+ "learning_rate": 1.282051282051282e-05,
2407
+ "loss": 0.0015,
2408
+ "step": 2920
2409
+ },
2410
+ {
2411
+ "epoch": 7.49,
2412
+ "eval_accuracy": 0.8249158249158249,
2413
+ "eval_loss": 0.8213893175125122,
2414
+ "eval_runtime": 30.8496,
2415
+ "eval_samples_per_second": 115.658,
2416
+ "eval_steps_per_second": 14.457,
2417
+ "step": 2920
2418
+ },
2419
+ {
2420
+ "epoch": 7.51,
2421
+ "learning_rate": 1.217948717948718e-05,
2422
+ "loss": 0.0015,
2423
+ "step": 2930
2424
+ },
2425
+ {
2426
+ "epoch": 7.54,
2427
+ "learning_rate": 1.153846153846154e-05,
2428
+ "loss": 0.0015,
2429
+ "step": 2940
2430
+ },
2431
+ {
2432
+ "epoch": 7.56,
2433
+ "learning_rate": 1.0897435897435898e-05,
2434
+ "loss": 0.0015,
2435
+ "step": 2950
2436
+ },
2437
+ {
2438
+ "epoch": 7.59,
2439
+ "learning_rate": 1.0256410256410256e-05,
2440
+ "loss": 0.0015,
2441
+ "step": 2960
2442
+ },
2443
+ {
2444
+ "epoch": 7.59,
2445
+ "eval_accuracy": 0.8249158249158249,
2446
+ "eval_loss": 0.8220140933990479,
2447
+ "eval_runtime": 31.6074,
2448
+ "eval_samples_per_second": 112.885,
2449
+ "eval_steps_per_second": 14.111,
2450
+ "step": 2960
2451
+ },
2452
+ {
2453
+ "epoch": 7.62,
2454
+ "learning_rate": 9.615384615384616e-06,
2455
+ "loss": 0.0016,
2456
+ "step": 2970
2457
+ },
2458
+ {
2459
+ "epoch": 7.64,
2460
+ "learning_rate": 8.974358974358976e-06,
2461
+ "loss": 0.0014,
2462
+ "step": 2980
2463
+ },
2464
+ {
2465
+ "epoch": 7.67,
2466
+ "learning_rate": 8.333333333333334e-06,
2467
+ "loss": 0.0015,
2468
+ "step": 2990
2469
+ },
2470
+ {
2471
+ "epoch": 7.69,
2472
+ "learning_rate": 7.692307692307694e-06,
2473
+ "loss": 0.0015,
2474
+ "step": 3000
2475
+ },
2476
+ {
2477
+ "epoch": 7.69,
2478
+ "eval_accuracy": 0.8254769921436588,
2479
+ "eval_loss": 0.8220102787017822,
2480
+ "eval_runtime": 30.2325,
2481
+ "eval_samples_per_second": 118.019,
2482
+ "eval_steps_per_second": 14.752,
2483
+ "step": 3000
2484
+ },
2485
+ {
2486
+ "epoch": 7.72,
2487
+ "learning_rate": 7.051282051282052e-06,
2488
+ "loss": 0.0014,
2489
+ "step": 3010
2490
+ },
2491
+ {
2492
+ "epoch": 7.74,
2493
+ "learning_rate": 6.41025641025641e-06,
2494
+ "loss": 0.0014,
2495
+ "step": 3020
2496
+ },
2497
+ {
2498
+ "epoch": 7.77,
2499
+ "learning_rate": 5.76923076923077e-06,
2500
+ "loss": 0.0015,
2501
+ "step": 3030
2502
+ },
2503
+ {
2504
+ "epoch": 7.79,
2505
+ "learning_rate": 5.128205128205128e-06,
2506
+ "loss": 0.0015,
2507
+ "step": 3040
2508
+ },
2509
+ {
2510
+ "epoch": 7.79,
2511
+ "eval_accuracy": 0.8254769921436588,
2512
+ "eval_loss": 0.8217305541038513,
2513
+ "eval_runtime": 38.9983,
2514
+ "eval_samples_per_second": 91.491,
2515
+ "eval_steps_per_second": 11.436,
2516
+ "step": 3040
2517
+ },
2518
+ {
2519
+ "epoch": 7.82,
2520
+ "learning_rate": 4.487179487179488e-06,
2521
+ "loss": 0.0014,
2522
+ "step": 3050
2523
+ },
2524
+ {
2525
+ "epoch": 7.85,
2526
+ "learning_rate": 3.846153846153847e-06,
2527
+ "loss": 0.0015,
2528
+ "step": 3060
2529
+ },
2530
+ {
2531
+ "epoch": 7.87,
2532
+ "learning_rate": 3.205128205128205e-06,
2533
+ "loss": 0.0014,
2534
+ "step": 3070
2535
+ },
2536
+ {
2537
+ "epoch": 7.9,
2538
+ "learning_rate": 2.564102564102564e-06,
2539
+ "loss": 0.0014,
2540
+ "step": 3080
2541
+ },
2542
+ {
2543
+ "epoch": 7.9,
2544
+ "eval_accuracy": 0.8254769921436588,
2545
+ "eval_loss": 0.8224795460700989,
2546
+ "eval_runtime": 31.9073,
2547
+ "eval_samples_per_second": 111.824,
2548
+ "eval_steps_per_second": 13.978,
2549
+ "step": 3080
2550
+ },
2551
+ {
2552
+ "epoch": 7.92,
2553
+ "learning_rate": 1.9230769230769234e-06,
2554
+ "loss": 0.0014,
2555
+ "step": 3090
2556
+ },
2557
+ {
2558
+ "epoch": 7.95,
2559
+ "learning_rate": 1.282051282051282e-06,
2560
+ "loss": 0.0015,
2561
+ "step": 3100
2562
+ },
2563
+ {
2564
+ "epoch": 7.97,
2565
+ "learning_rate": 6.41025641025641e-07,
2566
+ "loss": 0.0014,
2567
+ "step": 3110
2568
+ },
2569
+ {
2570
+ "epoch": 8.0,
2571
+ "learning_rate": 0.0,
2572
+ "loss": 0.0014,
2573
+ "step": 3120
2574
+ },
2575
+ {
2576
+ "epoch": 8.0,
2577
+ "eval_accuracy": 0.8260381593714927,
2578
+ "eval_loss": 0.8224756717681885,
2579
+ "eval_runtime": 30.7985,
2580
+ "eval_samples_per_second": 115.85,
2581
+ "eval_steps_per_second": 14.481,
2582
+ "step": 3120
2583
+ }
2584
+ ],
2585
+ "max_steps": 3120,
2586
+ "num_train_epochs": 8,
2587
+ "total_flos": 3.8629591832685773e+18,
2588
+ "trial_name": null,
2589
+ "trial_params": null
2590
+ }
checkpoint-3120/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38c25b2b280e8f8dde2511a004bf1b097d7641bf1aae99dd53711356509cd78e
3
+ size 3899
config.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": "Brazil",
13
+ "1": "Canada",
14
+ "2": "Finland",
15
+ "3": "Japan",
16
+ "4": "United-Kingdom",
17
+ "5": "United_States"
18
+ },
19
+ "image_size": 224,
20
+ "initializer_range": 0.02,
21
+ "intermediate_size": 3072,
22
+ "label2id": {
23
+ "Brazil": "0",
24
+ "Canada": "1",
25
+ "Finland": "2",
26
+ "Japan": "3",
27
+ "United-Kingdom": "4",
28
+ "United_States": "5"
29
+ },
30
+ "layer_norm_eps": 1e-12,
31
+ "model_type": "vit",
32
+ "num_attention_heads": 12,
33
+ "num_channels": 3,
34
+ "num_hidden_layers": 12,
35
+ "patch_size": 16,
36
+ "problem_type": "single_label_classification",
37
+ "qkv_bias": true,
38
+ "torch_dtype": "float32",
39
+ "transformers_version": "4.31.0"
40
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.5,
7
+ 0.5,
8
+ 0.5
9
+ ],
10
+ "image_processor_type": "ViTFeatureExtractor",
11
+ "image_std": [
12
+ 0.5,
13
+ 0.5,
14
+ 0.5
15
+ ],
16
+ "resample": 2,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 224,
20
+ "width": 224
21
+ }
22
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e4cf7738c400c63d307d24c20ed9c77e1a7e56df0099b73936659a2e13d9670
3
+ size 343278253
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 8.0,
3
+ "total_flos": 3.8629591832685773e+18,
4
+ "train_loss": 0.2533179052472592,
5
+ "train_runtime": 4957.8804,
6
+ "train_samples_per_second": 10.069,
7
+ "train_steps_per_second": 0.629
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2599 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.6381103992462158,
3
+ "best_model_checkpoint": "./vit-base-beans/checkpoint-1200",
4
+ "epoch": 8.0,
5
+ "global_step": 3120,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.03,
12
+ "learning_rate": 0.00019935897435897437,
13
+ "loss": 1.7678,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.05,
18
+ "learning_rate": 0.00019871794871794874,
19
+ "loss": 1.7014,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.08,
24
+ "learning_rate": 0.0001980769230769231,
25
+ "loss": 1.4957,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.1,
30
+ "learning_rate": 0.00019743589743589744,
31
+ "loss": 1.5043,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.1,
36
+ "eval_accuracy": 0.44556677890011226,
37
+ "eval_loss": 1.493240237236023,
38
+ "eval_runtime": 59.3866,
39
+ "eval_samples_per_second": 60.081,
40
+ "eval_steps_per_second": 7.51,
41
+ "step": 40
42
+ },
43
+ {
44
+ "epoch": 0.13,
45
+ "learning_rate": 0.00019679487179487178,
46
+ "loss": 1.3931,
47
+ "step": 50
48
+ },
49
+ {
50
+ "epoch": 0.15,
51
+ "learning_rate": 0.00019615384615384615,
52
+ "loss": 1.3029,
53
+ "step": 60
54
+ },
55
+ {
56
+ "epoch": 0.18,
57
+ "learning_rate": 0.0001955128205128205,
58
+ "loss": 1.1753,
59
+ "step": 70
60
+ },
61
+ {
62
+ "epoch": 0.21,
63
+ "learning_rate": 0.00019487179487179487,
64
+ "loss": 1.3894,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.21,
69
+ "eval_accuracy": 0.5297418630751964,
70
+ "eval_loss": 1.2430940866470337,
71
+ "eval_runtime": 38.1635,
72
+ "eval_samples_per_second": 93.492,
73
+ "eval_steps_per_second": 11.687,
74
+ "step": 80
75
+ },
76
+ {
77
+ "epoch": 0.23,
78
+ "learning_rate": 0.00019423076923076924,
79
+ "loss": 1.118,
80
+ "step": 90
81
+ },
82
+ {
83
+ "epoch": 0.26,
84
+ "learning_rate": 0.0001935897435897436,
85
+ "loss": 1.1395,
86
+ "step": 100
87
+ },
88
+ {
89
+ "epoch": 0.28,
90
+ "learning_rate": 0.00019294871794871797,
91
+ "loss": 1.0946,
92
+ "step": 110
93
+ },
94
+ {
95
+ "epoch": 0.31,
96
+ "learning_rate": 0.00019230769230769233,
97
+ "loss": 1.208,
98
+ "step": 120
99
+ },
100
+ {
101
+ "epoch": 0.31,
102
+ "eval_accuracy": 0.6464646464646465,
103
+ "eval_loss": 0.9881709218025208,
104
+ "eval_runtime": 34.6718,
105
+ "eval_samples_per_second": 102.908,
106
+ "eval_steps_per_second": 12.863,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 0.33,
111
+ "learning_rate": 0.00019166666666666667,
112
+ "loss": 0.9775,
113
+ "step": 130
114
+ },
115
+ {
116
+ "epoch": 0.36,
117
+ "learning_rate": 0.00019102564102564104,
118
+ "loss": 1.0216,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 0.38,
123
+ "learning_rate": 0.00019038461538461538,
124
+ "loss": 1.1942,
125
+ "step": 150
126
+ },
127
+ {
128
+ "epoch": 0.41,
129
+ "learning_rate": 0.00018974358974358974,
130
+ "loss": 0.9847,
131
+ "step": 160
132
+ },
133
+ {
134
+ "epoch": 0.41,
135
+ "eval_accuracy": 0.6599326599326599,
136
+ "eval_loss": 0.950343132019043,
137
+ "eval_runtime": 40.3536,
138
+ "eval_samples_per_second": 88.418,
139
+ "eval_steps_per_second": 11.052,
140
+ "step": 160
141
+ },
142
+ {
143
+ "epoch": 0.44,
144
+ "learning_rate": 0.0001891025641025641,
145
+ "loss": 0.9707,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 0.46,
150
+ "learning_rate": 0.00018846153846153847,
151
+ "loss": 1.1941,
152
+ "step": 180
153
+ },
154
+ {
155
+ "epoch": 0.49,
156
+ "learning_rate": 0.00018782051282051283,
157
+ "loss": 1.0309,
158
+ "step": 190
159
+ },
160
+ {
161
+ "epoch": 0.51,
162
+ "learning_rate": 0.0001871794871794872,
163
+ "loss": 1.102,
164
+ "step": 200
165
+ },
166
+ {
167
+ "epoch": 0.51,
168
+ "eval_accuracy": 0.6374859708193041,
169
+ "eval_loss": 0.9868837594985962,
170
+ "eval_runtime": 32.1858,
171
+ "eval_samples_per_second": 110.856,
172
+ "eval_steps_per_second": 13.857,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 0.54,
177
+ "learning_rate": 0.00018653846153846154,
178
+ "loss": 0.9899,
179
+ "step": 210
180
+ },
181
+ {
182
+ "epoch": 0.56,
183
+ "learning_rate": 0.0001858974358974359,
184
+ "loss": 0.8926,
185
+ "step": 220
186
+ },
187
+ {
188
+ "epoch": 0.59,
189
+ "learning_rate": 0.00018525641025641027,
190
+ "loss": 0.9552,
191
+ "step": 230
192
+ },
193
+ {
194
+ "epoch": 0.62,
195
+ "learning_rate": 0.00018461538461538463,
196
+ "loss": 0.9109,
197
+ "step": 240
198
+ },
199
+ {
200
+ "epoch": 0.62,
201
+ "eval_accuracy": 0.6380471380471381,
202
+ "eval_loss": 0.9887320399284363,
203
+ "eval_runtime": 32.5312,
204
+ "eval_samples_per_second": 109.679,
205
+ "eval_steps_per_second": 13.71,
206
+ "step": 240
207
+ },
208
+ {
209
+ "epoch": 0.64,
210
+ "learning_rate": 0.00018397435897435897,
211
+ "loss": 0.832,
212
+ "step": 250
213
+ },
214
+ {
215
+ "epoch": 0.67,
216
+ "learning_rate": 0.00018333333333333334,
217
+ "loss": 0.9323,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 0.69,
222
+ "learning_rate": 0.0001826923076923077,
223
+ "loss": 1.0061,
224
+ "step": 270
225
+ },
226
+ {
227
+ "epoch": 0.72,
228
+ "learning_rate": 0.00018205128205128207,
229
+ "loss": 0.9627,
230
+ "step": 280
231
+ },
232
+ {
233
+ "epoch": 0.72,
234
+ "eval_accuracy": 0.6156004489337823,
235
+ "eval_loss": 1.057395100593567,
236
+ "eval_runtime": 34.1152,
237
+ "eval_samples_per_second": 104.587,
238
+ "eval_steps_per_second": 13.073,
239
+ "step": 280
240
+ },
241
+ {
242
+ "epoch": 0.74,
243
+ "learning_rate": 0.00018141025641025643,
244
+ "loss": 0.9613,
245
+ "step": 290
246
+ },
247
+ {
248
+ "epoch": 0.77,
249
+ "learning_rate": 0.00018076923076923077,
250
+ "loss": 0.8986,
251
+ "step": 300
252
+ },
253
+ {
254
+ "epoch": 0.79,
255
+ "learning_rate": 0.00018012820512820513,
256
+ "loss": 0.9344,
257
+ "step": 310
258
+ },
259
+ {
260
+ "epoch": 0.82,
261
+ "learning_rate": 0.0001794871794871795,
262
+ "loss": 0.7031,
263
+ "step": 320
264
+ },
265
+ {
266
+ "epoch": 0.82,
267
+ "eval_accuracy": 0.7087542087542088,
268
+ "eval_loss": 0.813530683517456,
269
+ "eval_runtime": 39.1342,
270
+ "eval_samples_per_second": 91.173,
271
+ "eval_steps_per_second": 11.397,
272
+ "step": 320
273
+ },
274
+ {
275
+ "epoch": 0.85,
276
+ "learning_rate": 0.00017884615384615386,
277
+ "loss": 0.7596,
278
+ "step": 330
279
+ },
280
+ {
281
+ "epoch": 0.87,
282
+ "learning_rate": 0.00017820512820512823,
283
+ "loss": 0.7207,
284
+ "step": 340
285
+ },
286
+ {
287
+ "epoch": 0.9,
288
+ "learning_rate": 0.00017756410256410257,
289
+ "loss": 0.8904,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 0.92,
294
+ "learning_rate": 0.00017692307692307693,
295
+ "loss": 0.7605,
296
+ "step": 360
297
+ },
298
+ {
299
+ "epoch": 0.92,
300
+ "eval_accuracy": 0.7272727272727273,
301
+ "eval_loss": 0.7620912790298462,
302
+ "eval_runtime": 33.0054,
303
+ "eval_samples_per_second": 108.103,
304
+ "eval_steps_per_second": 13.513,
305
+ "step": 360
306
+ },
307
+ {
308
+ "epoch": 0.95,
309
+ "learning_rate": 0.0001762820512820513,
310
+ "loss": 0.8093,
311
+ "step": 370
312
+ },
313
+ {
314
+ "epoch": 0.97,
315
+ "learning_rate": 0.00017564102564102566,
316
+ "loss": 0.8198,
317
+ "step": 380
318
+ },
319
+ {
320
+ "epoch": 1.0,
321
+ "learning_rate": 0.000175,
322
+ "loss": 0.9515,
323
+ "step": 390
324
+ },
325
+ {
326
+ "epoch": 1.03,
327
+ "learning_rate": 0.00017435897435897436,
328
+ "loss": 0.8467,
329
+ "step": 400
330
+ },
331
+ {
332
+ "epoch": 1.03,
333
+ "eval_accuracy": 0.7244668911335578,
334
+ "eval_loss": 0.7974632978439331,
335
+ "eval_runtime": 33.3512,
336
+ "eval_samples_per_second": 106.983,
337
+ "eval_steps_per_second": 13.373,
338
+ "step": 400
339
+ },
340
+ {
341
+ "epoch": 1.05,
342
+ "learning_rate": 0.00017371794871794873,
343
+ "loss": 0.5829,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 1.08,
348
+ "learning_rate": 0.0001730769230769231,
349
+ "loss": 0.5576,
350
+ "step": 420
351
+ },
352
+ {
353
+ "epoch": 1.1,
354
+ "learning_rate": 0.00017243589743589746,
355
+ "loss": 0.6035,
356
+ "step": 430
357
+ },
358
+ {
359
+ "epoch": 1.13,
360
+ "learning_rate": 0.0001717948717948718,
361
+ "loss": 0.6653,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 1.13,
366
+ "eval_accuracy": 0.7317620650953984,
367
+ "eval_loss": 0.7574185132980347,
368
+ "eval_runtime": 32.4855,
369
+ "eval_samples_per_second": 109.834,
370
+ "eval_steps_per_second": 13.729,
371
+ "step": 440
372
+ },
373
+ {
374
+ "epoch": 1.15,
375
+ "learning_rate": 0.00017115384615384616,
376
+ "loss": 0.5067,
377
+ "step": 450
378
+ },
379
+ {
380
+ "epoch": 1.18,
381
+ "learning_rate": 0.00017051282051282053,
382
+ "loss": 0.4973,
383
+ "step": 460
384
+ },
385
+ {
386
+ "epoch": 1.21,
387
+ "learning_rate": 0.00016987179487179486,
388
+ "loss": 0.5721,
389
+ "step": 470
390
+ },
391
+ {
392
+ "epoch": 1.23,
393
+ "learning_rate": 0.00016923076923076923,
394
+ "loss": 0.5467,
395
+ "step": 480
396
+ },
397
+ {
398
+ "epoch": 1.23,
399
+ "eval_accuracy": 0.7244668911335578,
400
+ "eval_loss": 0.7471520900726318,
401
+ "eval_runtime": 32.6752,
402
+ "eval_samples_per_second": 109.196,
403
+ "eval_steps_per_second": 13.65,
404
+ "step": 480
405
+ },
406
+ {
407
+ "epoch": 1.26,
408
+ "learning_rate": 0.0001685897435897436,
409
+ "loss": 0.5494,
410
+ "step": 490
411
+ },
412
+ {
413
+ "epoch": 1.28,
414
+ "learning_rate": 0.00016794871794871796,
415
+ "loss": 0.6158,
416
+ "step": 500
417
+ },
418
+ {
419
+ "epoch": 1.31,
420
+ "learning_rate": 0.00016730769230769232,
421
+ "loss": 0.5412,
422
+ "step": 510
423
+ },
424
+ {
425
+ "epoch": 1.33,
426
+ "learning_rate": 0.0001666666666666667,
427
+ "loss": 0.388,
428
+ "step": 520
429
+ },
430
+ {
431
+ "epoch": 1.33,
432
+ "eval_accuracy": 0.7716049382716049,
433
+ "eval_loss": 0.6508304476737976,
434
+ "eval_runtime": 36.0929,
435
+ "eval_samples_per_second": 98.856,
436
+ "eval_steps_per_second": 12.357,
437
+ "step": 520
438
+ },
439
+ {
440
+ "epoch": 1.36,
441
+ "learning_rate": 0.00016602564102564105,
442
+ "loss": 0.4511,
443
+ "step": 530
444
+ },
445
+ {
446
+ "epoch": 1.38,
447
+ "learning_rate": 0.0001653846153846154,
448
+ "loss": 0.566,
449
+ "step": 540
450
+ },
451
+ {
452
+ "epoch": 1.41,
453
+ "learning_rate": 0.00016474358974358976,
454
+ "loss": 0.4476,
455
+ "step": 550
456
+ },
457
+ {
458
+ "epoch": 1.44,
459
+ "learning_rate": 0.0001641025641025641,
460
+ "loss": 0.4699,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 1.44,
465
+ "eval_accuracy": 0.7244668911335578,
466
+ "eval_loss": 0.773759663105011,
467
+ "eval_runtime": 31.7272,
468
+ "eval_samples_per_second": 112.459,
469
+ "eval_steps_per_second": 14.057,
470
+ "step": 560
471
+ },
472
+ {
473
+ "epoch": 1.46,
474
+ "learning_rate": 0.00016346153846153846,
475
+ "loss": 0.402,
476
+ "step": 570
477
+ },
478
+ {
479
+ "epoch": 1.49,
480
+ "learning_rate": 0.00016282051282051282,
481
+ "loss": 0.425,
482
+ "step": 580
483
+ },
484
+ {
485
+ "epoch": 1.51,
486
+ "learning_rate": 0.0001621794871794872,
487
+ "loss": 0.5171,
488
+ "step": 590
489
+ },
490
+ {
491
+ "epoch": 1.54,
492
+ "learning_rate": 0.00016153846153846155,
493
+ "loss": 0.5344,
494
+ "step": 600
495
+ },
496
+ {
497
+ "epoch": 1.54,
498
+ "eval_accuracy": 0.7328843995510662,
499
+ "eval_loss": 0.7624912858009338,
500
+ "eval_runtime": 32.007,
501
+ "eval_samples_per_second": 111.476,
502
+ "eval_steps_per_second": 13.934,
503
+ "step": 600
504
+ },
505
+ {
506
+ "epoch": 1.56,
507
+ "learning_rate": 0.00016089743589743592,
508
+ "loss": 0.4913,
509
+ "step": 610
510
+ },
511
+ {
512
+ "epoch": 1.59,
513
+ "learning_rate": 0.00016025641025641028,
514
+ "loss": 0.5042,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 1.62,
519
+ "learning_rate": 0.00015961538461538462,
520
+ "loss": 0.6051,
521
+ "step": 630
522
+ },
523
+ {
524
+ "epoch": 1.64,
525
+ "learning_rate": 0.00015897435897435896,
526
+ "loss": 0.5753,
527
+ "step": 640
528
+ },
529
+ {
530
+ "epoch": 1.64,
531
+ "eval_accuracy": 0.7598204264870931,
532
+ "eval_loss": 0.6743783950805664,
533
+ "eval_runtime": 30.705,
534
+ "eval_samples_per_second": 116.203,
535
+ "eval_steps_per_second": 14.525,
536
+ "step": 640
537
+ },
538
+ {
539
+ "epoch": 1.67,
540
+ "learning_rate": 0.00015833333333333332,
541
+ "loss": 0.524,
542
+ "step": 650
543
+ },
544
+ {
545
+ "epoch": 1.69,
546
+ "learning_rate": 0.0001576923076923077,
547
+ "loss": 0.5007,
548
+ "step": 660
549
+ },
550
+ {
551
+ "epoch": 1.72,
552
+ "learning_rate": 0.00015705128205128205,
553
+ "loss": 0.57,
554
+ "step": 670
555
+ },
556
+ {
557
+ "epoch": 1.74,
558
+ "learning_rate": 0.00015641025641025642,
559
+ "loss": 0.5533,
560
+ "step": 680
561
+ },
562
+ {
563
+ "epoch": 1.74,
564
+ "eval_accuracy": 0.734006734006734,
565
+ "eval_loss": 0.7230738997459412,
566
+ "eval_runtime": 33.3034,
567
+ "eval_samples_per_second": 107.136,
568
+ "eval_steps_per_second": 13.392,
569
+ "step": 680
570
+ },
571
+ {
572
+ "epoch": 1.77,
573
+ "learning_rate": 0.00015576923076923078,
574
+ "loss": 0.4399,
575
+ "step": 690
576
+ },
577
+ {
578
+ "epoch": 1.79,
579
+ "learning_rate": 0.00015512820512820515,
580
+ "loss": 0.453,
581
+ "step": 700
582
+ },
583
+ {
584
+ "epoch": 1.82,
585
+ "learning_rate": 0.00015448717948717951,
586
+ "loss": 0.4504,
587
+ "step": 710
588
+ },
589
+ {
590
+ "epoch": 1.85,
591
+ "learning_rate": 0.00015384615384615385,
592
+ "loss": 0.4244,
593
+ "step": 720
594
+ },
595
+ {
596
+ "epoch": 1.85,
597
+ "eval_accuracy": 0.7687991021324355,
598
+ "eval_loss": 0.6673141717910767,
599
+ "eval_runtime": 32.6948,
600
+ "eval_samples_per_second": 109.131,
601
+ "eval_steps_per_second": 13.641,
602
+ "step": 720
603
+ },
604
+ {
605
+ "epoch": 1.87,
606
+ "learning_rate": 0.00015320512820512822,
607
+ "loss": 0.46,
608
+ "step": 730
609
+ },
610
+ {
611
+ "epoch": 1.9,
612
+ "learning_rate": 0.00015256410256410255,
613
+ "loss": 0.4132,
614
+ "step": 740
615
+ },
616
+ {
617
+ "epoch": 1.92,
618
+ "learning_rate": 0.00015192307692307692,
619
+ "loss": 0.4849,
620
+ "step": 750
621
+ },
622
+ {
623
+ "epoch": 1.95,
624
+ "learning_rate": 0.00015128205128205128,
625
+ "loss": 0.5423,
626
+ "step": 760
627
+ },
628
+ {
629
+ "epoch": 1.95,
630
+ "eval_accuracy": 0.7413019079685746,
631
+ "eval_loss": 0.7185856699943542,
632
+ "eval_runtime": 35.9539,
633
+ "eval_samples_per_second": 99.238,
634
+ "eval_steps_per_second": 12.405,
635
+ "step": 760
636
+ },
637
+ {
638
+ "epoch": 1.97,
639
+ "learning_rate": 0.00015064102564102565,
640
+ "loss": 0.3997,
641
+ "step": 770
642
+ },
643
+ {
644
+ "epoch": 2.0,
645
+ "learning_rate": 0.00015000000000000001,
646
+ "loss": 0.4744,
647
+ "step": 780
648
+ },
649
+ {
650
+ "epoch": 2.03,
651
+ "learning_rate": 0.00014935897435897438,
652
+ "loss": 0.4717,
653
+ "step": 790
654
+ },
655
+ {
656
+ "epoch": 2.05,
657
+ "learning_rate": 0.00014871794871794872,
658
+ "loss": 0.3384,
659
+ "step": 800
660
+ },
661
+ {
662
+ "epoch": 2.05,
663
+ "eval_accuracy": 0.7351290684624018,
664
+ "eval_loss": 0.7269920706748962,
665
+ "eval_runtime": 33.5897,
666
+ "eval_samples_per_second": 106.223,
667
+ "eval_steps_per_second": 13.278,
668
+ "step": 800
669
+ },
670
+ {
671
+ "epoch": 2.08,
672
+ "learning_rate": 0.00014807692307692308,
673
+ "loss": 0.2602,
674
+ "step": 810
675
+ },
676
+ {
677
+ "epoch": 2.1,
678
+ "learning_rate": 0.00014743589743589745,
679
+ "loss": 0.2345,
680
+ "step": 820
681
+ },
682
+ {
683
+ "epoch": 2.13,
684
+ "learning_rate": 0.00014679487179487178,
685
+ "loss": 0.2135,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 2.15,
690
+ "learning_rate": 0.00014615384615384615,
691
+ "loss": 0.2797,
692
+ "step": 840
693
+ },
694
+ {
695
+ "epoch": 2.15,
696
+ "eval_accuracy": 0.7497194163860831,
697
+ "eval_loss": 0.7742622494697571,
698
+ "eval_runtime": 31.2996,
699
+ "eval_samples_per_second": 113.995,
700
+ "eval_steps_per_second": 14.249,
701
+ "step": 840
702
+ },
703
+ {
704
+ "epoch": 2.18,
705
+ "learning_rate": 0.00014551282051282051,
706
+ "loss": 0.1862,
707
+ "step": 850
708
+ },
709
+ {
710
+ "epoch": 2.21,
711
+ "learning_rate": 0.00014487179487179488,
712
+ "loss": 0.2334,
713
+ "step": 860
714
+ },
715
+ {
716
+ "epoch": 2.23,
717
+ "learning_rate": 0.00014423076923076924,
718
+ "loss": 0.3361,
719
+ "step": 870
720
+ },
721
+ {
722
+ "epoch": 2.26,
723
+ "learning_rate": 0.0001435897435897436,
724
+ "loss": 0.2939,
725
+ "step": 880
726
+ },
727
+ {
728
+ "epoch": 2.26,
729
+ "eval_accuracy": 0.7732884399551067,
730
+ "eval_loss": 0.6985616087913513,
731
+ "eval_runtime": 32.9047,
732
+ "eval_samples_per_second": 108.435,
733
+ "eval_steps_per_second": 13.554,
734
+ "step": 880
735
+ },
736
+ {
737
+ "epoch": 2.28,
738
+ "learning_rate": 0.00014294871794871795,
739
+ "loss": 0.2846,
740
+ "step": 890
741
+ },
742
+ {
743
+ "epoch": 2.31,
744
+ "learning_rate": 0.0001423076923076923,
745
+ "loss": 0.2555,
746
+ "step": 900
747
+ },
748
+ {
749
+ "epoch": 2.33,
750
+ "learning_rate": 0.00014166666666666668,
751
+ "loss": 0.2629,
752
+ "step": 910
753
+ },
754
+ {
755
+ "epoch": 2.36,
756
+ "learning_rate": 0.00014102564102564104,
757
+ "loss": 0.2288,
758
+ "step": 920
759
+ },
760
+ {
761
+ "epoch": 2.36,
762
+ "eval_accuracy": 0.7480359147025814,
763
+ "eval_loss": 0.761025607585907,
764
+ "eval_runtime": 34.1426,
765
+ "eval_samples_per_second": 104.503,
766
+ "eval_steps_per_second": 13.063,
767
+ "step": 920
768
+ },
769
+ {
770
+ "epoch": 2.38,
771
+ "learning_rate": 0.00014038461538461538,
772
+ "loss": 0.2115,
773
+ "step": 930
774
+ },
775
+ {
776
+ "epoch": 2.41,
777
+ "learning_rate": 0.00013974358974358974,
778
+ "loss": 0.2036,
779
+ "step": 940
780
+ },
781
+ {
782
+ "epoch": 2.44,
783
+ "learning_rate": 0.0001391025641025641,
784
+ "loss": 0.2913,
785
+ "step": 950
786
+ },
787
+ {
788
+ "epoch": 2.46,
789
+ "learning_rate": 0.00013846153846153847,
790
+ "loss": 0.2204,
791
+ "step": 960
792
+ },
793
+ {
794
+ "epoch": 2.46,
795
+ "eval_accuracy": 0.7570145903479237,
796
+ "eval_loss": 0.7840890884399414,
797
+ "eval_runtime": 35.7983,
798
+ "eval_samples_per_second": 99.67,
799
+ "eval_steps_per_second": 12.459,
800
+ "step": 960
801
+ },
802
+ {
803
+ "epoch": 2.49,
804
+ "learning_rate": 0.00013782051282051284,
805
+ "loss": 0.2759,
806
+ "step": 970
807
+ },
808
+ {
809
+ "epoch": 2.51,
810
+ "learning_rate": 0.00013717948717948718,
811
+ "loss": 0.3945,
812
+ "step": 980
813
+ },
814
+ {
815
+ "epoch": 2.54,
816
+ "learning_rate": 0.00013653846153846154,
817
+ "loss": 0.3639,
818
+ "step": 990
819
+ },
820
+ {
821
+ "epoch": 2.56,
822
+ "learning_rate": 0.0001358974358974359,
823
+ "loss": 0.4397,
824
+ "step": 1000
825
+ },
826
+ {
827
+ "epoch": 2.56,
828
+ "eval_accuracy": 0.7789001122334456,
829
+ "eval_loss": 0.6565700173377991,
830
+ "eval_runtime": 31.6553,
831
+ "eval_samples_per_second": 112.714,
832
+ "eval_steps_per_second": 14.089,
833
+ "step": 1000
834
+ },
835
+ {
836
+ "epoch": 2.59,
837
+ "learning_rate": 0.00013525641025641027,
838
+ "loss": 0.2707,
839
+ "step": 1010
840
+ },
841
+ {
842
+ "epoch": 2.62,
843
+ "learning_rate": 0.00013461538461538464,
844
+ "loss": 0.3451,
845
+ "step": 1020
846
+ },
847
+ {
848
+ "epoch": 2.64,
849
+ "learning_rate": 0.00013397435897435897,
850
+ "loss": 0.3277,
851
+ "step": 1030
852
+ },
853
+ {
854
+ "epoch": 2.67,
855
+ "learning_rate": 0.00013333333333333334,
856
+ "loss": 0.2219,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 2.67,
861
+ "eval_accuracy": 0.7581369248035915,
862
+ "eval_loss": 0.6900522112846375,
863
+ "eval_runtime": 31.7845,
864
+ "eval_samples_per_second": 112.256,
865
+ "eval_steps_per_second": 14.032,
866
+ "step": 1040
867
+ },
868
+ {
869
+ "epoch": 2.69,
870
+ "learning_rate": 0.0001326923076923077,
871
+ "loss": 0.3174,
872
+ "step": 1050
873
+ },
874
+ {
875
+ "epoch": 2.72,
876
+ "learning_rate": 0.00013205128205128204,
877
+ "loss": 0.189,
878
+ "step": 1060
879
+ },
880
+ {
881
+ "epoch": 2.74,
882
+ "learning_rate": 0.0001314102564102564,
883
+ "loss": 0.2619,
884
+ "step": 1070
885
+ },
886
+ {
887
+ "epoch": 2.77,
888
+ "learning_rate": 0.00013076923076923077,
889
+ "loss": 0.2297,
890
+ "step": 1080
891
+ },
892
+ {
893
+ "epoch": 2.77,
894
+ "eval_accuracy": 0.7581369248035915,
895
+ "eval_loss": 0.7095093131065369,
896
+ "eval_runtime": 38.4561,
897
+ "eval_samples_per_second": 92.781,
898
+ "eval_steps_per_second": 11.598,
899
+ "step": 1080
900
+ },
901
+ {
902
+ "epoch": 2.79,
903
+ "learning_rate": 0.00013012820512820514,
904
+ "loss": 0.2201,
905
+ "step": 1090
906
+ },
907
+ {
908
+ "epoch": 2.82,
909
+ "learning_rate": 0.0001294871794871795,
910
+ "loss": 0.2261,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 2.85,
915
+ "learning_rate": 0.00012884615384615387,
916
+ "loss": 0.2095,
917
+ "step": 1110
918
+ },
919
+ {
920
+ "epoch": 2.87,
921
+ "learning_rate": 0.00012820512820512823,
922
+ "loss": 0.209,
923
+ "step": 1120
924
+ },
925
+ {
926
+ "epoch": 2.87,
927
+ "eval_accuracy": 0.7738496071829405,
928
+ "eval_loss": 0.7127913236618042,
929
+ "eval_runtime": 32.9997,
930
+ "eval_samples_per_second": 108.122,
931
+ "eval_steps_per_second": 13.515,
932
+ "step": 1120
933
+ },
934
+ {
935
+ "epoch": 2.9,
936
+ "learning_rate": 0.00012756410256410257,
937
+ "loss": 0.3365,
938
+ "step": 1130
939
+ },
940
+ {
941
+ "epoch": 2.92,
942
+ "learning_rate": 0.00012692307692307693,
943
+ "loss": 0.2639,
944
+ "step": 1140
945
+ },
946
+ {
947
+ "epoch": 2.95,
948
+ "learning_rate": 0.00012628205128205127,
949
+ "loss": 0.1892,
950
+ "step": 1150
951
+ },
952
+ {
953
+ "epoch": 2.97,
954
+ "learning_rate": 0.00012564102564102564,
955
+ "loss": 0.1853,
956
+ "step": 1160
957
+ },
958
+ {
959
+ "epoch": 2.97,
960
+ "eval_accuracy": 0.7710437710437711,
961
+ "eval_loss": 0.6986069083213806,
962
+ "eval_runtime": 32.2598,
963
+ "eval_samples_per_second": 110.602,
964
+ "eval_steps_per_second": 13.825,
965
+ "step": 1160
966
+ },
967
+ {
968
+ "epoch": 3.0,
969
+ "learning_rate": 0.000125,
970
+ "loss": 0.1807,
971
+ "step": 1170
972
+ },
973
+ {
974
+ "epoch": 3.03,
975
+ "learning_rate": 0.00012435897435897437,
976
+ "loss": 0.1554,
977
+ "step": 1180
978
+ },
979
+ {
980
+ "epoch": 3.05,
981
+ "learning_rate": 0.00012371794871794873,
982
+ "loss": 0.1845,
983
+ "step": 1190
984
+ },
985
+ {
986
+ "epoch": 3.08,
987
+ "learning_rate": 0.0001230769230769231,
988
+ "loss": 0.1322,
989
+ "step": 1200
990
+ },
991
+ {
992
+ "epoch": 3.08,
993
+ "eval_accuracy": 0.7962962962962963,
994
+ "eval_loss": 0.6381103992462158,
995
+ "eval_runtime": 35.1257,
996
+ "eval_samples_per_second": 101.578,
997
+ "eval_steps_per_second": 12.697,
998
+ "step": 1200
999
+ },
1000
+ {
1001
+ "epoch": 3.1,
1002
+ "learning_rate": 0.00012243589743589746,
1003
+ "loss": 0.0598,
1004
+ "step": 1210
1005
+ },
1006
+ {
1007
+ "epoch": 3.13,
1008
+ "learning_rate": 0.00012179487179487179,
1009
+ "loss": 0.0853,
1010
+ "step": 1220
1011
+ },
1012
+ {
1013
+ "epoch": 3.15,
1014
+ "learning_rate": 0.00012115384615384615,
1015
+ "loss": 0.1616,
1016
+ "step": 1230
1017
+ },
1018
+ {
1019
+ "epoch": 3.18,
1020
+ "learning_rate": 0.00012051282051282052,
1021
+ "loss": 0.2603,
1022
+ "step": 1240
1023
+ },
1024
+ {
1025
+ "epoch": 3.18,
1026
+ "eval_accuracy": 0.7682379349046016,
1027
+ "eval_loss": 0.7860389351844788,
1028
+ "eval_runtime": 35.648,
1029
+ "eval_samples_per_second": 100.09,
1030
+ "eval_steps_per_second": 12.511,
1031
+ "step": 1240
1032
+ },
1033
+ {
1034
+ "epoch": 3.21,
1035
+ "learning_rate": 0.00011987179487179487,
1036
+ "loss": 0.1104,
1037
+ "step": 1250
1038
+ },
1039
+ {
1040
+ "epoch": 3.23,
1041
+ "learning_rate": 0.00011923076923076923,
1042
+ "loss": 0.0634,
1043
+ "step": 1260
1044
+ },
1045
+ {
1046
+ "epoch": 3.26,
1047
+ "learning_rate": 0.0001185897435897436,
1048
+ "loss": 0.0717,
1049
+ "step": 1270
1050
+ },
1051
+ {
1052
+ "epoch": 3.28,
1053
+ "learning_rate": 0.00011794871794871796,
1054
+ "loss": 0.1031,
1055
+ "step": 1280
1056
+ },
1057
+ {
1058
+ "epoch": 3.28,
1059
+ "eval_accuracy": 0.7895622895622896,
1060
+ "eval_loss": 0.7322177886962891,
1061
+ "eval_runtime": 30.6139,
1062
+ "eval_samples_per_second": 116.549,
1063
+ "eval_steps_per_second": 14.569,
1064
+ "step": 1280
1065
+ },
1066
+ {
1067
+ "epoch": 3.31,
1068
+ "learning_rate": 0.00011730769230769231,
1069
+ "loss": 0.054,
1070
+ "step": 1290
1071
+ },
1072
+ {
1073
+ "epoch": 3.33,
1074
+ "learning_rate": 0.00011666666666666668,
1075
+ "loss": 0.0883,
1076
+ "step": 1300
1077
+ },
1078
+ {
1079
+ "epoch": 3.36,
1080
+ "learning_rate": 0.00011602564102564104,
1081
+ "loss": 0.1443,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 3.38,
1086
+ "learning_rate": 0.00011538461538461538,
1087
+ "loss": 0.0763,
1088
+ "step": 1320
1089
+ },
1090
+ {
1091
+ "epoch": 3.38,
1092
+ "eval_accuracy": 0.7839506172839507,
1093
+ "eval_loss": 0.77773118019104,
1094
+ "eval_runtime": 31.4828,
1095
+ "eval_samples_per_second": 113.332,
1096
+ "eval_steps_per_second": 14.166,
1097
+ "step": 1320
1098
+ },
1099
+ {
1100
+ "epoch": 3.41,
1101
+ "learning_rate": 0.00011474358974358975,
1102
+ "loss": 0.1328,
1103
+ "step": 1330
1104
+ },
1105
+ {
1106
+ "epoch": 3.44,
1107
+ "learning_rate": 0.0001141025641025641,
1108
+ "loss": 0.0751,
1109
+ "step": 1340
1110
+ },
1111
+ {
1112
+ "epoch": 3.46,
1113
+ "learning_rate": 0.00011346153846153846,
1114
+ "loss": 0.1193,
1115
+ "step": 1350
1116
+ },
1117
+ {
1118
+ "epoch": 3.49,
1119
+ "learning_rate": 0.00011282051282051283,
1120
+ "loss": 0.1437,
1121
+ "step": 1360
1122
+ },
1123
+ {
1124
+ "epoch": 3.49,
1125
+ "eval_accuracy": 0.7665544332210998,
1126
+ "eval_loss": 0.9293356537818909,
1127
+ "eval_runtime": 31.1681,
1128
+ "eval_samples_per_second": 114.476,
1129
+ "eval_steps_per_second": 14.31,
1130
+ "step": 1360
1131
+ },
1132
+ {
1133
+ "epoch": 3.51,
1134
+ "learning_rate": 0.00011217948717948718,
1135
+ "loss": 0.1805,
1136
+ "step": 1370
1137
+ },
1138
+ {
1139
+ "epoch": 3.54,
1140
+ "learning_rate": 0.00011153846153846154,
1141
+ "loss": 0.0931,
1142
+ "step": 1380
1143
+ },
1144
+ {
1145
+ "epoch": 3.56,
1146
+ "learning_rate": 0.00011089743589743591,
1147
+ "loss": 0.1227,
1148
+ "step": 1390
1149
+ },
1150
+ {
1151
+ "epoch": 3.59,
1152
+ "learning_rate": 0.00011025641025641027,
1153
+ "loss": 0.0818,
1154
+ "step": 1400
1155
+ },
1156
+ {
1157
+ "epoch": 3.59,
1158
+ "eval_accuracy": 0.7940516273849607,
1159
+ "eval_loss": 0.796922504901886,
1160
+ "eval_runtime": 31.9044,
1161
+ "eval_samples_per_second": 111.834,
1162
+ "eval_steps_per_second": 13.979,
1163
+ "step": 1400
1164
+ },
1165
+ {
1166
+ "epoch": 3.62,
1167
+ "learning_rate": 0.00010961538461538463,
1168
+ "loss": 0.1193,
1169
+ "step": 1410
1170
+ },
1171
+ {
1172
+ "epoch": 3.64,
1173
+ "learning_rate": 0.00010897435897435896,
1174
+ "loss": 0.1237,
1175
+ "step": 1420
1176
+ },
1177
+ {
1178
+ "epoch": 3.67,
1179
+ "learning_rate": 0.00010833333333333333,
1180
+ "loss": 0.1074,
1181
+ "step": 1430
1182
+ },
1183
+ {
1184
+ "epoch": 3.69,
1185
+ "learning_rate": 0.0001076923076923077,
1186
+ "loss": 0.1355,
1187
+ "step": 1440
1188
+ },
1189
+ {
1190
+ "epoch": 3.69,
1191
+ "eval_accuracy": 0.7716049382716049,
1192
+ "eval_loss": 0.8145824670791626,
1193
+ "eval_runtime": 37.6519,
1194
+ "eval_samples_per_second": 94.763,
1195
+ "eval_steps_per_second": 11.845,
1196
+ "step": 1440
1197
+ },
1198
+ {
1199
+ "epoch": 3.72,
1200
+ "learning_rate": 0.00010705128205128206,
1201
+ "loss": 0.0844,
1202
+ "step": 1450
1203
+ },
1204
+ {
1205
+ "epoch": 3.74,
1206
+ "learning_rate": 0.00010641025641025641,
1207
+ "loss": 0.1067,
1208
+ "step": 1460
1209
+ },
1210
+ {
1211
+ "epoch": 3.77,
1212
+ "learning_rate": 0.00010576923076923077,
1213
+ "loss": 0.0342,
1214
+ "step": 1470
1215
+ },
1216
+ {
1217
+ "epoch": 3.79,
1218
+ "learning_rate": 0.00010512820512820514,
1219
+ "loss": 0.0802,
1220
+ "step": 1480
1221
+ },
1222
+ {
1223
+ "epoch": 3.79,
1224
+ "eval_accuracy": 0.8075196408529742,
1225
+ "eval_loss": 0.6976819038391113,
1226
+ "eval_runtime": 32.2393,
1227
+ "eval_samples_per_second": 110.672,
1228
+ "eval_steps_per_second": 13.834,
1229
+ "step": 1480
1230
+ },
1231
+ {
1232
+ "epoch": 3.82,
1233
+ "learning_rate": 0.0001044871794871795,
1234
+ "loss": 0.1143,
1235
+ "step": 1490
1236
+ },
1237
+ {
1238
+ "epoch": 3.85,
1239
+ "learning_rate": 0.00010384615384615386,
1240
+ "loss": 0.1332,
1241
+ "step": 1500
1242
+ },
1243
+ {
1244
+ "epoch": 3.87,
1245
+ "learning_rate": 0.00010320512820512822,
1246
+ "loss": 0.069,
1247
+ "step": 1510
1248
+ },
1249
+ {
1250
+ "epoch": 3.9,
1251
+ "learning_rate": 0.00010256410256410256,
1252
+ "loss": 0.032,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 3.9,
1257
+ "eval_accuracy": 0.7929292929292929,
1258
+ "eval_loss": 0.8349816203117371,
1259
+ "eval_runtime": 32.9637,
1260
+ "eval_samples_per_second": 108.24,
1261
+ "eval_steps_per_second": 13.53,
1262
+ "step": 1520
1263
+ },
1264
+ {
1265
+ "epoch": 3.92,
1266
+ "learning_rate": 0.00010192307692307692,
1267
+ "loss": 0.0722,
1268
+ "step": 1530
1269
+ },
1270
+ {
1271
+ "epoch": 3.95,
1272
+ "learning_rate": 0.00010128205128205129,
1273
+ "loss": 0.1383,
1274
+ "step": 1540
1275
+ },
1276
+ {
1277
+ "epoch": 3.97,
1278
+ "learning_rate": 0.00010064102564102564,
1279
+ "loss": 0.0861,
1280
+ "step": 1550
1281
+ },
1282
+ {
1283
+ "epoch": 4.0,
1284
+ "learning_rate": 0.0001,
1285
+ "loss": 0.1836,
1286
+ "step": 1560
1287
+ },
1288
+ {
1289
+ "epoch": 4.0,
1290
+ "eval_accuracy": 0.7968574635241302,
1291
+ "eval_loss": 0.775885820388794,
1292
+ "eval_runtime": 33.0589,
1293
+ "eval_samples_per_second": 107.928,
1294
+ "eval_steps_per_second": 13.491,
1295
+ "step": 1560
1296
+ },
1297
+ {
1298
+ "epoch": 4.03,
1299
+ "learning_rate": 9.935897435897437e-05,
1300
+ "loss": 0.0486,
1301
+ "step": 1570
1302
+ },
1303
+ {
1304
+ "epoch": 4.05,
1305
+ "learning_rate": 9.871794871794872e-05,
1306
+ "loss": 0.0369,
1307
+ "step": 1580
1308
+ },
1309
+ {
1310
+ "epoch": 4.08,
1311
+ "learning_rate": 9.807692307692307e-05,
1312
+ "loss": 0.0209,
1313
+ "step": 1590
1314
+ },
1315
+ {
1316
+ "epoch": 4.1,
1317
+ "learning_rate": 9.743589743589744e-05,
1318
+ "loss": 0.0391,
1319
+ "step": 1600
1320
+ },
1321
+ {
1322
+ "epoch": 4.1,
1323
+ "eval_accuracy": 0.8013468013468014,
1324
+ "eval_loss": 0.8052350282669067,
1325
+ "eval_runtime": 32.0537,
1326
+ "eval_samples_per_second": 111.313,
1327
+ "eval_steps_per_second": 13.914,
1328
+ "step": 1600
1329
+ },
1330
+ {
1331
+ "epoch": 4.13,
1332
+ "learning_rate": 9.67948717948718e-05,
1333
+ "loss": 0.0354,
1334
+ "step": 1610
1335
+ },
1336
+ {
1337
+ "epoch": 4.15,
1338
+ "learning_rate": 9.615384615384617e-05,
1339
+ "loss": 0.0738,
1340
+ "step": 1620
1341
+ },
1342
+ {
1343
+ "epoch": 4.18,
1344
+ "learning_rate": 9.551282051282052e-05,
1345
+ "loss": 0.0275,
1346
+ "step": 1630
1347
+ },
1348
+ {
1349
+ "epoch": 4.21,
1350
+ "learning_rate": 9.487179487179487e-05,
1351
+ "loss": 0.0406,
1352
+ "step": 1640
1353
+ },
1354
+ {
1355
+ "epoch": 4.21,
1356
+ "eval_accuracy": 0.7828282828282829,
1357
+ "eval_loss": 0.9251748919487,
1358
+ "eval_runtime": 31.6082,
1359
+ "eval_samples_per_second": 112.882,
1360
+ "eval_steps_per_second": 14.11,
1361
+ "step": 1640
1362
+ },
1363
+ {
1364
+ "epoch": 4.23,
1365
+ "learning_rate": 9.423076923076924e-05,
1366
+ "loss": 0.1377,
1367
+ "step": 1650
1368
+ },
1369
+ {
1370
+ "epoch": 4.26,
1371
+ "learning_rate": 9.35897435897436e-05,
1372
+ "loss": 0.0526,
1373
+ "step": 1660
1374
+ },
1375
+ {
1376
+ "epoch": 4.28,
1377
+ "learning_rate": 9.294871794871795e-05,
1378
+ "loss": 0.152,
1379
+ "step": 1670
1380
+ },
1381
+ {
1382
+ "epoch": 4.31,
1383
+ "learning_rate": 9.230769230769232e-05,
1384
+ "loss": 0.0488,
1385
+ "step": 1680
1386
+ },
1387
+ {
1388
+ "epoch": 4.31,
1389
+ "eval_accuracy": 0.7901234567901234,
1390
+ "eval_loss": 0.8203706741333008,
1391
+ "eval_runtime": 31.8606,
1392
+ "eval_samples_per_second": 111.988,
1393
+ "eval_steps_per_second": 13.998,
1394
+ "step": 1680
1395
+ },
1396
+ {
1397
+ "epoch": 4.33,
1398
+ "learning_rate": 9.166666666666667e-05,
1399
+ "loss": 0.0372,
1400
+ "step": 1690
1401
+ },
1402
+ {
1403
+ "epoch": 4.36,
1404
+ "learning_rate": 9.102564102564103e-05,
1405
+ "loss": 0.0968,
1406
+ "step": 1700
1407
+ },
1408
+ {
1409
+ "epoch": 4.38,
1410
+ "learning_rate": 9.038461538461538e-05,
1411
+ "loss": 0.0147,
1412
+ "step": 1710
1413
+ },
1414
+ {
1415
+ "epoch": 4.41,
1416
+ "learning_rate": 8.974358974358975e-05,
1417
+ "loss": 0.016,
1418
+ "step": 1720
1419
+ },
1420
+ {
1421
+ "epoch": 4.41,
1422
+ "eval_accuracy": 0.7794612794612794,
1423
+ "eval_loss": 0.9169319868087769,
1424
+ "eval_runtime": 33.724,
1425
+ "eval_samples_per_second": 105.8,
1426
+ "eval_steps_per_second": 13.225,
1427
+ "step": 1720
1428
+ },
1429
+ {
1430
+ "epoch": 4.44,
1431
+ "learning_rate": 8.910256410256411e-05,
1432
+ "loss": 0.0393,
1433
+ "step": 1730
1434
+ },
1435
+ {
1436
+ "epoch": 4.46,
1437
+ "learning_rate": 8.846153846153847e-05,
1438
+ "loss": 0.0258,
1439
+ "step": 1740
1440
+ },
1441
+ {
1442
+ "epoch": 4.49,
1443
+ "learning_rate": 8.782051282051283e-05,
1444
+ "loss": 0.009,
1445
+ "step": 1750
1446
+ },
1447
+ {
1448
+ "epoch": 4.51,
1449
+ "learning_rate": 8.717948717948718e-05,
1450
+ "loss": 0.0994,
1451
+ "step": 1760
1452
+ },
1453
+ {
1454
+ "epoch": 4.51,
1455
+ "eval_accuracy": 0.8120089786756454,
1456
+ "eval_loss": 0.7930631637573242,
1457
+ "eval_runtime": 34.0477,
1458
+ "eval_samples_per_second": 104.794,
1459
+ "eval_steps_per_second": 13.099,
1460
+ "step": 1760
1461
+ },
1462
+ {
1463
+ "epoch": 4.54,
1464
+ "learning_rate": 8.653846153846155e-05,
1465
+ "loss": 0.0418,
1466
+ "step": 1770
1467
+ },
1468
+ {
1469
+ "epoch": 4.56,
1470
+ "learning_rate": 8.58974358974359e-05,
1471
+ "loss": 0.0793,
1472
+ "step": 1780
1473
+ },
1474
+ {
1475
+ "epoch": 4.59,
1476
+ "learning_rate": 8.525641025641026e-05,
1477
+ "loss": 0.0122,
1478
+ "step": 1790
1479
+ },
1480
+ {
1481
+ "epoch": 4.62,
1482
+ "learning_rate": 8.461538461538461e-05,
1483
+ "loss": 0.07,
1484
+ "step": 1800
1485
+ },
1486
+ {
1487
+ "epoch": 4.62,
1488
+ "eval_accuracy": 0.8103254769921436,
1489
+ "eval_loss": 0.8245030045509338,
1490
+ "eval_runtime": 32.6583,
1491
+ "eval_samples_per_second": 109.252,
1492
+ "eval_steps_per_second": 13.657,
1493
+ "step": 1800
1494
+ },
1495
+ {
1496
+ "epoch": 4.64,
1497
+ "learning_rate": 8.397435897435898e-05,
1498
+ "loss": 0.0656,
1499
+ "step": 1810
1500
+ },
1501
+ {
1502
+ "epoch": 4.67,
1503
+ "learning_rate": 8.333333333333334e-05,
1504
+ "loss": 0.0364,
1505
+ "step": 1820
1506
+ },
1507
+ {
1508
+ "epoch": 4.69,
1509
+ "learning_rate": 8.26923076923077e-05,
1510
+ "loss": 0.0639,
1511
+ "step": 1830
1512
+ },
1513
+ {
1514
+ "epoch": 4.72,
1515
+ "learning_rate": 8.205128205128205e-05,
1516
+ "loss": 0.0088,
1517
+ "step": 1840
1518
+ },
1519
+ {
1520
+ "epoch": 4.72,
1521
+ "eval_accuracy": 0.7985409652076318,
1522
+ "eval_loss": 0.907173752784729,
1523
+ "eval_runtime": 40.0812,
1524
+ "eval_samples_per_second": 89.019,
1525
+ "eval_steps_per_second": 11.127,
1526
+ "step": 1840
1527
+ },
1528
+ {
1529
+ "epoch": 4.74,
1530
+ "learning_rate": 8.141025641025641e-05,
1531
+ "loss": 0.0557,
1532
+ "step": 1850
1533
+ },
1534
+ {
1535
+ "epoch": 4.77,
1536
+ "learning_rate": 8.076923076923078e-05,
1537
+ "loss": 0.0548,
1538
+ "step": 1860
1539
+ },
1540
+ {
1541
+ "epoch": 4.79,
1542
+ "learning_rate": 8.012820512820514e-05,
1543
+ "loss": 0.0055,
1544
+ "step": 1870
1545
+ },
1546
+ {
1547
+ "epoch": 4.82,
1548
+ "learning_rate": 7.948717948717948e-05,
1549
+ "loss": 0.0085,
1550
+ "step": 1880
1551
+ },
1552
+ {
1553
+ "epoch": 4.82,
1554
+ "eval_accuracy": 0.8114478114478114,
1555
+ "eval_loss": 0.7862226963043213,
1556
+ "eval_runtime": 33.9761,
1557
+ "eval_samples_per_second": 105.015,
1558
+ "eval_steps_per_second": 13.127,
1559
+ "step": 1880
1560
+ },
1561
+ {
1562
+ "epoch": 4.85,
1563
+ "learning_rate": 7.884615384615384e-05,
1564
+ "loss": 0.0423,
1565
+ "step": 1890
1566
+ },
1567
+ {
1568
+ "epoch": 4.87,
1569
+ "learning_rate": 7.820512820512821e-05,
1570
+ "loss": 0.0581,
1571
+ "step": 1900
1572
+ },
1573
+ {
1574
+ "epoch": 4.9,
1575
+ "learning_rate": 7.756410256410257e-05,
1576
+ "loss": 0.0043,
1577
+ "step": 1910
1578
+ },
1579
+ {
1580
+ "epoch": 4.92,
1581
+ "learning_rate": 7.692307692307693e-05,
1582
+ "loss": 0.083,
1583
+ "step": 1920
1584
+ },
1585
+ {
1586
+ "epoch": 4.92,
1587
+ "eval_accuracy": 0.8198653198653199,
1588
+ "eval_loss": 0.7796981334686279,
1589
+ "eval_runtime": 32.3864,
1590
+ "eval_samples_per_second": 110.17,
1591
+ "eval_steps_per_second": 13.771,
1592
+ "step": 1920
1593
+ },
1594
+ {
1595
+ "epoch": 4.95,
1596
+ "learning_rate": 7.628205128205128e-05,
1597
+ "loss": 0.1308,
1598
+ "step": 1930
1599
+ },
1600
+ {
1601
+ "epoch": 4.97,
1602
+ "learning_rate": 7.564102564102564e-05,
1603
+ "loss": 0.0063,
1604
+ "step": 1940
1605
+ },
1606
+ {
1607
+ "epoch": 5.0,
1608
+ "learning_rate": 7.500000000000001e-05,
1609
+ "loss": 0.0132,
1610
+ "step": 1950
1611
+ },
1612
+ {
1613
+ "epoch": 5.03,
1614
+ "learning_rate": 7.435897435897436e-05,
1615
+ "loss": 0.0055,
1616
+ "step": 1960
1617
+ },
1618
+ {
1619
+ "epoch": 5.03,
1620
+ "eval_accuracy": 0.8069584736251403,
1621
+ "eval_loss": 0.8723996877670288,
1622
+ "eval_runtime": 33.6185,
1623
+ "eval_samples_per_second": 106.132,
1624
+ "eval_steps_per_second": 13.267,
1625
+ "step": 1960
1626
+ },
1627
+ {
1628
+ "epoch": 5.05,
1629
+ "learning_rate": 7.371794871794872e-05,
1630
+ "loss": 0.0366,
1631
+ "step": 1970
1632
+ },
1633
+ {
1634
+ "epoch": 5.08,
1635
+ "learning_rate": 7.307692307692307e-05,
1636
+ "loss": 0.06,
1637
+ "step": 1980
1638
+ },
1639
+ {
1640
+ "epoch": 5.1,
1641
+ "learning_rate": 7.243589743589744e-05,
1642
+ "loss": 0.0065,
1643
+ "step": 1990
1644
+ },
1645
+ {
1646
+ "epoch": 5.13,
1647
+ "learning_rate": 7.17948717948718e-05,
1648
+ "loss": 0.0223,
1649
+ "step": 2000
1650
+ },
1651
+ {
1652
+ "epoch": 5.13,
1653
+ "eval_accuracy": 0.8249158249158249,
1654
+ "eval_loss": 0.7518730163574219,
1655
+ "eval_runtime": 43.1922,
1656
+ "eval_samples_per_second": 82.607,
1657
+ "eval_steps_per_second": 10.326,
1658
+ "step": 2000
1659
+ },
1660
+ {
1661
+ "epoch": 5.15,
1662
+ "learning_rate": 7.115384615384616e-05,
1663
+ "loss": 0.0215,
1664
+ "step": 2010
1665
+ },
1666
+ {
1667
+ "epoch": 5.18,
1668
+ "learning_rate": 7.051282051282052e-05,
1669
+ "loss": 0.0494,
1670
+ "step": 2020
1671
+ },
1672
+ {
1673
+ "epoch": 5.21,
1674
+ "learning_rate": 6.987179487179487e-05,
1675
+ "loss": 0.0441,
1676
+ "step": 2030
1677
+ },
1678
+ {
1679
+ "epoch": 5.23,
1680
+ "learning_rate": 6.923076923076924e-05,
1681
+ "loss": 0.0042,
1682
+ "step": 2040
1683
+ },
1684
+ {
1685
+ "epoch": 5.23,
1686
+ "eval_accuracy": 0.8232323232323232,
1687
+ "eval_loss": 0.7583897113800049,
1688
+ "eval_runtime": 31.893,
1689
+ "eval_samples_per_second": 111.874,
1690
+ "eval_steps_per_second": 13.984,
1691
+ "step": 2040
1692
+ },
1693
+ {
1694
+ "epoch": 5.26,
1695
+ "learning_rate": 6.858974358974359e-05,
1696
+ "loss": 0.0131,
1697
+ "step": 2050
1698
+ },
1699
+ {
1700
+ "epoch": 5.28,
1701
+ "learning_rate": 6.794871794871795e-05,
1702
+ "loss": 0.0037,
1703
+ "step": 2060
1704
+ },
1705
+ {
1706
+ "epoch": 5.31,
1707
+ "learning_rate": 6.730769230769232e-05,
1708
+ "loss": 0.02,
1709
+ "step": 2070
1710
+ },
1711
+ {
1712
+ "epoch": 5.33,
1713
+ "learning_rate": 6.666666666666667e-05,
1714
+ "loss": 0.0178,
1715
+ "step": 2080
1716
+ },
1717
+ {
1718
+ "epoch": 5.33,
1719
+ "eval_accuracy": 0.8080808080808081,
1720
+ "eval_loss": 0.8523600697517395,
1721
+ "eval_runtime": 32.5102,
1722
+ "eval_samples_per_second": 109.75,
1723
+ "eval_steps_per_second": 13.719,
1724
+ "step": 2080
1725
+ },
1726
+ {
1727
+ "epoch": 5.36,
1728
+ "learning_rate": 6.602564102564102e-05,
1729
+ "loss": 0.0035,
1730
+ "step": 2090
1731
+ },
1732
+ {
1733
+ "epoch": 5.38,
1734
+ "learning_rate": 6.538461538461539e-05,
1735
+ "loss": 0.0137,
1736
+ "step": 2100
1737
+ },
1738
+ {
1739
+ "epoch": 5.41,
1740
+ "learning_rate": 6.474358974358975e-05,
1741
+ "loss": 0.0033,
1742
+ "step": 2110
1743
+ },
1744
+ {
1745
+ "epoch": 5.44,
1746
+ "learning_rate": 6.410256410256412e-05,
1747
+ "loss": 0.0172,
1748
+ "step": 2120
1749
+ },
1750
+ {
1751
+ "epoch": 5.44,
1752
+ "eval_accuracy": 0.8215488215488216,
1753
+ "eval_loss": 0.7729219794273376,
1754
+ "eval_runtime": 31.7305,
1755
+ "eval_samples_per_second": 112.447,
1756
+ "eval_steps_per_second": 14.056,
1757
+ "step": 2120
1758
+ },
1759
+ {
1760
+ "epoch": 5.46,
1761
+ "learning_rate": 6.346153846153847e-05,
1762
+ "loss": 0.0252,
1763
+ "step": 2130
1764
+ },
1765
+ {
1766
+ "epoch": 5.49,
1767
+ "learning_rate": 6.282051282051282e-05,
1768
+ "loss": 0.0034,
1769
+ "step": 2140
1770
+ },
1771
+ {
1772
+ "epoch": 5.51,
1773
+ "learning_rate": 6.217948717948718e-05,
1774
+ "loss": 0.0045,
1775
+ "step": 2150
1776
+ },
1777
+ {
1778
+ "epoch": 5.54,
1779
+ "learning_rate": 6.153846153846155e-05,
1780
+ "loss": 0.0044,
1781
+ "step": 2160
1782
+ },
1783
+ {
1784
+ "epoch": 5.54,
1785
+ "eval_accuracy": 0.808641975308642,
1786
+ "eval_loss": 0.8701183795928955,
1787
+ "eval_runtime": 31.0299,
1788
+ "eval_samples_per_second": 114.986,
1789
+ "eval_steps_per_second": 14.373,
1790
+ "step": 2160
1791
+ },
1792
+ {
1793
+ "epoch": 5.56,
1794
+ "learning_rate": 6.089743589743589e-05,
1795
+ "loss": 0.0373,
1796
+ "step": 2170
1797
+ },
1798
+ {
1799
+ "epoch": 5.59,
1800
+ "learning_rate": 6.025641025641026e-05,
1801
+ "loss": 0.0029,
1802
+ "step": 2180
1803
+ },
1804
+ {
1805
+ "epoch": 5.62,
1806
+ "learning_rate": 5.9615384615384616e-05,
1807
+ "loss": 0.0032,
1808
+ "step": 2190
1809
+ },
1810
+ {
1811
+ "epoch": 5.64,
1812
+ "learning_rate": 5.897435897435898e-05,
1813
+ "loss": 0.003,
1814
+ "step": 2200
1815
+ },
1816
+ {
1817
+ "epoch": 5.64,
1818
+ "eval_accuracy": 0.8237934904601572,
1819
+ "eval_loss": 0.8060529828071594,
1820
+ "eval_runtime": 32.4421,
1821
+ "eval_samples_per_second": 109.98,
1822
+ "eval_steps_per_second": 13.748,
1823
+ "step": 2200
1824
+ },
1825
+ {
1826
+ "epoch": 5.67,
1827
+ "learning_rate": 5.833333333333334e-05,
1828
+ "loss": 0.0034,
1829
+ "step": 2210
1830
+ },
1831
+ {
1832
+ "epoch": 5.69,
1833
+ "learning_rate": 5.769230769230769e-05,
1834
+ "loss": 0.0025,
1835
+ "step": 2220
1836
+ },
1837
+ {
1838
+ "epoch": 5.72,
1839
+ "learning_rate": 5.705128205128205e-05,
1840
+ "loss": 0.0029,
1841
+ "step": 2230
1842
+ },
1843
+ {
1844
+ "epoch": 5.74,
1845
+ "learning_rate": 5.6410256410256414e-05,
1846
+ "loss": 0.0033,
1847
+ "step": 2240
1848
+ },
1849
+ {
1850
+ "epoch": 5.74,
1851
+ "eval_accuracy": 0.8226711560044894,
1852
+ "eval_loss": 0.813855767250061,
1853
+ "eval_runtime": 32.378,
1854
+ "eval_samples_per_second": 110.198,
1855
+ "eval_steps_per_second": 13.775,
1856
+ "step": 2240
1857
+ },
1858
+ {
1859
+ "epoch": 5.77,
1860
+ "learning_rate": 5.576923076923077e-05,
1861
+ "loss": 0.0129,
1862
+ "step": 2250
1863
+ },
1864
+ {
1865
+ "epoch": 5.79,
1866
+ "learning_rate": 5.512820512820514e-05,
1867
+ "loss": 0.0079,
1868
+ "step": 2260
1869
+ },
1870
+ {
1871
+ "epoch": 5.82,
1872
+ "learning_rate": 5.448717948717948e-05,
1873
+ "loss": 0.0117,
1874
+ "step": 2270
1875
+ },
1876
+ {
1877
+ "epoch": 5.85,
1878
+ "learning_rate": 5.384615384615385e-05,
1879
+ "loss": 0.0023,
1880
+ "step": 2280
1881
+ },
1882
+ {
1883
+ "epoch": 5.85,
1884
+ "eval_accuracy": 0.8164983164983165,
1885
+ "eval_loss": 0.8478395342826843,
1886
+ "eval_runtime": 35.321,
1887
+ "eval_samples_per_second": 101.016,
1888
+ "eval_steps_per_second": 12.627,
1889
+ "step": 2280
1890
+ },
1891
+ {
1892
+ "epoch": 5.87,
1893
+ "learning_rate": 5.3205128205128205e-05,
1894
+ "loss": 0.0075,
1895
+ "step": 2290
1896
+ },
1897
+ {
1898
+ "epoch": 5.9,
1899
+ "learning_rate": 5.256410256410257e-05,
1900
+ "loss": 0.0023,
1901
+ "step": 2300
1902
+ },
1903
+ {
1904
+ "epoch": 5.92,
1905
+ "learning_rate": 5.192307692307693e-05,
1906
+ "loss": 0.0058,
1907
+ "step": 2310
1908
+ },
1909
+ {
1910
+ "epoch": 5.95,
1911
+ "learning_rate": 5.128205128205128e-05,
1912
+ "loss": 0.003,
1913
+ "step": 2320
1914
+ },
1915
+ {
1916
+ "epoch": 5.95,
1917
+ "eval_accuracy": 0.8120089786756454,
1918
+ "eval_loss": 0.8443180918693542,
1919
+ "eval_runtime": 32.2001,
1920
+ "eval_samples_per_second": 110.807,
1921
+ "eval_steps_per_second": 13.851,
1922
+ "step": 2320
1923
+ },
1924
+ {
1925
+ "epoch": 5.97,
1926
+ "learning_rate": 5.0641025641025644e-05,
1927
+ "loss": 0.0512,
1928
+ "step": 2330
1929
+ },
1930
+ {
1931
+ "epoch": 6.0,
1932
+ "learning_rate": 5e-05,
1933
+ "loss": 0.0024,
1934
+ "step": 2340
1935
+ },
1936
+ {
1937
+ "epoch": 6.03,
1938
+ "learning_rate": 4.935897435897436e-05,
1939
+ "loss": 0.0022,
1940
+ "step": 2350
1941
+ },
1942
+ {
1943
+ "epoch": 6.05,
1944
+ "learning_rate": 4.871794871794872e-05,
1945
+ "loss": 0.031,
1946
+ "step": 2360
1947
+ },
1948
+ {
1949
+ "epoch": 6.05,
1950
+ "eval_accuracy": 0.8047138047138047,
1951
+ "eval_loss": 0.9272196888923645,
1952
+ "eval_runtime": 33.2145,
1953
+ "eval_samples_per_second": 107.423,
1954
+ "eval_steps_per_second": 13.428,
1955
+ "step": 2360
1956
+ },
1957
+ {
1958
+ "epoch": 6.08,
1959
+ "learning_rate": 4.8076923076923084e-05,
1960
+ "loss": 0.0456,
1961
+ "step": 2370
1962
+ },
1963
+ {
1964
+ "epoch": 6.1,
1965
+ "learning_rate": 4.7435897435897435e-05,
1966
+ "loss": 0.0024,
1967
+ "step": 2380
1968
+ },
1969
+ {
1970
+ "epoch": 6.13,
1971
+ "learning_rate": 4.67948717948718e-05,
1972
+ "loss": 0.0023,
1973
+ "step": 2390
1974
+ },
1975
+ {
1976
+ "epoch": 6.15,
1977
+ "learning_rate": 4.615384615384616e-05,
1978
+ "loss": 0.0021,
1979
+ "step": 2400
1980
+ },
1981
+ {
1982
+ "epoch": 6.15,
1983
+ "eval_accuracy": 0.8204264870931538,
1984
+ "eval_loss": 0.8369048237800598,
1985
+ "eval_runtime": 36.1395,
1986
+ "eval_samples_per_second": 98.728,
1987
+ "eval_steps_per_second": 12.341,
1988
+ "step": 2400
1989
+ },
1990
+ {
1991
+ "epoch": 6.18,
1992
+ "learning_rate": 4.5512820512820516e-05,
1993
+ "loss": 0.0021,
1994
+ "step": 2410
1995
+ },
1996
+ {
1997
+ "epoch": 6.21,
1998
+ "learning_rate": 4.4871794871794874e-05,
1999
+ "loss": 0.0026,
2000
+ "step": 2420
2001
+ },
2002
+ {
2003
+ "epoch": 6.23,
2004
+ "learning_rate": 4.423076923076923e-05,
2005
+ "loss": 0.0021,
2006
+ "step": 2430
2007
+ },
2008
+ {
2009
+ "epoch": 6.26,
2010
+ "learning_rate": 4.358974358974359e-05,
2011
+ "loss": 0.0019,
2012
+ "step": 2440
2013
+ },
2014
+ {
2015
+ "epoch": 6.26,
2016
+ "eval_accuracy": 0.8209876543209876,
2017
+ "eval_loss": 0.828059196472168,
2018
+ "eval_runtime": 32.1149,
2019
+ "eval_samples_per_second": 111.101,
2020
+ "eval_steps_per_second": 13.888,
2021
+ "step": 2440
2022
+ },
2023
+ {
2024
+ "epoch": 6.28,
2025
+ "learning_rate": 4.294871794871795e-05,
2026
+ "loss": 0.0024,
2027
+ "step": 2450
2028
+ },
2029
+ {
2030
+ "epoch": 6.31,
2031
+ "learning_rate": 4.230769230769231e-05,
2032
+ "loss": 0.0019,
2033
+ "step": 2460
2034
+ },
2035
+ {
2036
+ "epoch": 6.33,
2037
+ "learning_rate": 4.166666666666667e-05,
2038
+ "loss": 0.002,
2039
+ "step": 2470
2040
+ },
2041
+ {
2042
+ "epoch": 6.36,
2043
+ "learning_rate": 4.1025641025641023e-05,
2044
+ "loss": 0.0019,
2045
+ "step": 2480
2046
+ },
2047
+ {
2048
+ "epoch": 6.36,
2049
+ "eval_accuracy": 0.8198653198653199,
2050
+ "eval_loss": 0.834019124507904,
2051
+ "eval_runtime": 35.8439,
2052
+ "eval_samples_per_second": 99.543,
2053
+ "eval_steps_per_second": 12.443,
2054
+ "step": 2480
2055
+ },
2056
+ {
2057
+ "epoch": 6.38,
2058
+ "learning_rate": 4.038461538461539e-05,
2059
+ "loss": 0.002,
2060
+ "step": 2490
2061
+ },
2062
+ {
2063
+ "epoch": 6.41,
2064
+ "learning_rate": 3.974358974358974e-05,
2065
+ "loss": 0.0019,
2066
+ "step": 2500
2067
+ },
2068
+ {
2069
+ "epoch": 6.44,
2070
+ "learning_rate": 3.9102564102564105e-05,
2071
+ "loss": 0.0018,
2072
+ "step": 2510
2073
+ },
2074
+ {
2075
+ "epoch": 6.46,
2076
+ "learning_rate": 3.846153846153846e-05,
2077
+ "loss": 0.0018,
2078
+ "step": 2520
2079
+ },
2080
+ {
2081
+ "epoch": 6.46,
2082
+ "eval_accuracy": 0.8204264870931538,
2083
+ "eval_loss": 0.8339148759841919,
2084
+ "eval_runtime": 34.0288,
2085
+ "eval_samples_per_second": 104.853,
2086
+ "eval_steps_per_second": 13.107,
2087
+ "step": 2520
2088
+ },
2089
+ {
2090
+ "epoch": 6.49,
2091
+ "learning_rate": 3.782051282051282e-05,
2092
+ "loss": 0.0018,
2093
+ "step": 2530
2094
+ },
2095
+ {
2096
+ "epoch": 6.51,
2097
+ "learning_rate": 3.717948717948718e-05,
2098
+ "loss": 0.0018,
2099
+ "step": 2540
2100
+ },
2101
+ {
2102
+ "epoch": 6.54,
2103
+ "learning_rate": 3.653846153846154e-05,
2104
+ "loss": 0.0018,
2105
+ "step": 2550
2106
+ },
2107
+ {
2108
+ "epoch": 6.56,
2109
+ "learning_rate": 3.58974358974359e-05,
2110
+ "loss": 0.002,
2111
+ "step": 2560
2112
+ },
2113
+ {
2114
+ "epoch": 6.56,
2115
+ "eval_accuracy": 0.8209876543209876,
2116
+ "eval_loss": 0.8293877840042114,
2117
+ "eval_runtime": 33.9846,
2118
+ "eval_samples_per_second": 104.989,
2119
+ "eval_steps_per_second": 13.124,
2120
+ "step": 2560
2121
+ },
2122
+ {
2123
+ "epoch": 6.59,
2124
+ "learning_rate": 3.525641025641026e-05,
2125
+ "loss": 0.0017,
2126
+ "step": 2570
2127
+ },
2128
+ {
2129
+ "epoch": 6.62,
2130
+ "learning_rate": 3.461538461538462e-05,
2131
+ "loss": 0.0017,
2132
+ "step": 2580
2133
+ },
2134
+ {
2135
+ "epoch": 6.64,
2136
+ "learning_rate": 3.397435897435898e-05,
2137
+ "loss": 0.0018,
2138
+ "step": 2590
2139
+ },
2140
+ {
2141
+ "epoch": 6.67,
2142
+ "learning_rate": 3.3333333333333335e-05,
2143
+ "loss": 0.0017,
2144
+ "step": 2600
2145
+ },
2146
+ {
2147
+ "epoch": 6.67,
2148
+ "eval_accuracy": 0.8209876543209876,
2149
+ "eval_loss": 0.8266403079032898,
2150
+ "eval_runtime": 36.0162,
2151
+ "eval_samples_per_second": 99.067,
2152
+ "eval_steps_per_second": 12.383,
2153
+ "step": 2600
2154
+ },
2155
+ {
2156
+ "epoch": 6.69,
2157
+ "learning_rate": 3.269230769230769e-05,
2158
+ "loss": 0.002,
2159
+ "step": 2610
2160
+ },
2161
+ {
2162
+ "epoch": 6.72,
2163
+ "learning_rate": 3.205128205128206e-05,
2164
+ "loss": 0.0017,
2165
+ "step": 2620
2166
+ },
2167
+ {
2168
+ "epoch": 6.74,
2169
+ "learning_rate": 3.141025641025641e-05,
2170
+ "loss": 0.0017,
2171
+ "step": 2630
2172
+ },
2173
+ {
2174
+ "epoch": 6.77,
2175
+ "learning_rate": 3.0769230769230774e-05,
2176
+ "loss": 0.0017,
2177
+ "step": 2640
2178
+ },
2179
+ {
2180
+ "epoch": 6.77,
2181
+ "eval_accuracy": 0.8226711560044894,
2182
+ "eval_loss": 0.8177886009216309,
2183
+ "eval_runtime": 36.3329,
2184
+ "eval_samples_per_second": 98.203,
2185
+ "eval_steps_per_second": 12.275,
2186
+ "step": 2640
2187
+ },
2188
+ {
2189
+ "epoch": 6.79,
2190
+ "learning_rate": 3.012820512820513e-05,
2191
+ "loss": 0.0016,
2192
+ "step": 2650
2193
+ },
2194
+ {
2195
+ "epoch": 6.82,
2196
+ "learning_rate": 2.948717948717949e-05,
2197
+ "loss": 0.0016,
2198
+ "step": 2660
2199
+ },
2200
+ {
2201
+ "epoch": 6.85,
2202
+ "learning_rate": 2.8846153846153845e-05,
2203
+ "loss": 0.0016,
2204
+ "step": 2670
2205
+ },
2206
+ {
2207
+ "epoch": 6.87,
2208
+ "learning_rate": 2.8205128205128207e-05,
2209
+ "loss": 0.0017,
2210
+ "step": 2680
2211
+ },
2212
+ {
2213
+ "epoch": 6.87,
2214
+ "eval_accuracy": 0.824354657687991,
2215
+ "eval_loss": 0.8174560070037842,
2216
+ "eval_runtime": 32.3063,
2217
+ "eval_samples_per_second": 110.443,
2218
+ "eval_steps_per_second": 13.805,
2219
+ "step": 2680
2220
+ },
2221
+ {
2222
+ "epoch": 6.9,
2223
+ "learning_rate": 2.756410256410257e-05,
2224
+ "loss": 0.0017,
2225
+ "step": 2690
2226
+ },
2227
+ {
2228
+ "epoch": 6.92,
2229
+ "learning_rate": 2.6923076923076923e-05,
2230
+ "loss": 0.0017,
2231
+ "step": 2700
2232
+ },
2233
+ {
2234
+ "epoch": 6.95,
2235
+ "learning_rate": 2.6282051282051285e-05,
2236
+ "loss": 0.0017,
2237
+ "step": 2710
2238
+ },
2239
+ {
2240
+ "epoch": 6.97,
2241
+ "learning_rate": 2.564102564102564e-05,
2242
+ "loss": 0.0017,
2243
+ "step": 2720
2244
+ },
2245
+ {
2246
+ "epoch": 6.97,
2247
+ "eval_accuracy": 0.8237934904601572,
2248
+ "eval_loss": 0.8166114687919617,
2249
+ "eval_runtime": 33.2936,
2250
+ "eval_samples_per_second": 107.168,
2251
+ "eval_steps_per_second": 13.396,
2252
+ "step": 2720
2253
+ },
2254
+ {
2255
+ "epoch": 7.0,
2256
+ "learning_rate": 2.5e-05,
2257
+ "loss": 0.0015,
2258
+ "step": 2730
2259
+ },
2260
+ {
2261
+ "epoch": 7.03,
2262
+ "learning_rate": 2.435897435897436e-05,
2263
+ "loss": 0.0016,
2264
+ "step": 2740
2265
+ },
2266
+ {
2267
+ "epoch": 7.05,
2268
+ "learning_rate": 2.3717948717948718e-05,
2269
+ "loss": 0.0016,
2270
+ "step": 2750
2271
+ },
2272
+ {
2273
+ "epoch": 7.08,
2274
+ "learning_rate": 2.307692307692308e-05,
2275
+ "loss": 0.0015,
2276
+ "step": 2760
2277
+ },
2278
+ {
2279
+ "epoch": 7.08,
2280
+ "eval_accuracy": 0.824354657687991,
2281
+ "eval_loss": 0.8176218271255493,
2282
+ "eval_runtime": 32.3011,
2283
+ "eval_samples_per_second": 110.46,
2284
+ "eval_steps_per_second": 13.808,
2285
+ "step": 2760
2286
+ },
2287
+ {
2288
+ "epoch": 7.1,
2289
+ "learning_rate": 2.2435897435897437e-05,
2290
+ "loss": 0.0015,
2291
+ "step": 2770
2292
+ },
2293
+ {
2294
+ "epoch": 7.13,
2295
+ "learning_rate": 2.1794871794871795e-05,
2296
+ "loss": 0.0015,
2297
+ "step": 2780
2298
+ },
2299
+ {
2300
+ "epoch": 7.15,
2301
+ "learning_rate": 2.1153846153846154e-05,
2302
+ "loss": 0.0015,
2303
+ "step": 2790
2304
+ },
2305
+ {
2306
+ "epoch": 7.18,
2307
+ "learning_rate": 2.0512820512820512e-05,
2308
+ "loss": 0.0015,
2309
+ "step": 2800
2310
+ },
2311
+ {
2312
+ "epoch": 7.18,
2313
+ "eval_accuracy": 0.824354657687991,
2314
+ "eval_loss": 0.8185549378395081,
2315
+ "eval_runtime": 32.7126,
2316
+ "eval_samples_per_second": 109.071,
2317
+ "eval_steps_per_second": 13.634,
2318
+ "step": 2800
2319
+ },
2320
+ {
2321
+ "epoch": 7.21,
2322
+ "learning_rate": 1.987179487179487e-05,
2323
+ "loss": 0.0015,
2324
+ "step": 2810
2325
+ },
2326
+ {
2327
+ "epoch": 7.23,
2328
+ "learning_rate": 1.923076923076923e-05,
2329
+ "loss": 0.0015,
2330
+ "step": 2820
2331
+ },
2332
+ {
2333
+ "epoch": 7.26,
2334
+ "learning_rate": 1.858974358974359e-05,
2335
+ "loss": 0.0015,
2336
+ "step": 2830
2337
+ },
2338
+ {
2339
+ "epoch": 7.28,
2340
+ "learning_rate": 1.794871794871795e-05,
2341
+ "loss": 0.0015,
2342
+ "step": 2840
2343
+ },
2344
+ {
2345
+ "epoch": 7.28,
2346
+ "eval_accuracy": 0.8249158249158249,
2347
+ "eval_loss": 0.819684624671936,
2348
+ "eval_runtime": 35.8925,
2349
+ "eval_samples_per_second": 99.408,
2350
+ "eval_steps_per_second": 12.426,
2351
+ "step": 2840
2352
+ },
2353
+ {
2354
+ "epoch": 7.31,
2355
+ "learning_rate": 1.730769230769231e-05,
2356
+ "loss": 0.0016,
2357
+ "step": 2850
2358
+ },
2359
+ {
2360
+ "epoch": 7.33,
2361
+ "learning_rate": 1.6666666666666667e-05,
2362
+ "loss": 0.0015,
2363
+ "step": 2860
2364
+ },
2365
+ {
2366
+ "epoch": 7.36,
2367
+ "learning_rate": 1.602564102564103e-05,
2368
+ "loss": 0.0015,
2369
+ "step": 2870
2370
+ },
2371
+ {
2372
+ "epoch": 7.38,
2373
+ "learning_rate": 1.5384615384615387e-05,
2374
+ "loss": 0.0015,
2375
+ "step": 2880
2376
+ },
2377
+ {
2378
+ "epoch": 7.38,
2379
+ "eval_accuracy": 0.8249158249158249,
2380
+ "eval_loss": 0.8204275965690613,
2381
+ "eval_runtime": 31.8471,
2382
+ "eval_samples_per_second": 112.035,
2383
+ "eval_steps_per_second": 14.004,
2384
+ "step": 2880
2385
+ },
2386
+ {
2387
+ "epoch": 7.41,
2388
+ "learning_rate": 1.4743589743589745e-05,
2389
+ "loss": 0.0015,
2390
+ "step": 2890
2391
+ },
2392
+ {
2393
+ "epoch": 7.44,
2394
+ "learning_rate": 1.4102564102564104e-05,
2395
+ "loss": 0.0015,
2396
+ "step": 2900
2397
+ },
2398
+ {
2399
+ "epoch": 7.46,
2400
+ "learning_rate": 1.3461538461538462e-05,
2401
+ "loss": 0.0015,
2402
+ "step": 2910
2403
+ },
2404
+ {
2405
+ "epoch": 7.49,
2406
+ "learning_rate": 1.282051282051282e-05,
2407
+ "loss": 0.0015,
2408
+ "step": 2920
2409
+ },
2410
+ {
2411
+ "epoch": 7.49,
2412
+ "eval_accuracy": 0.8249158249158249,
2413
+ "eval_loss": 0.8213893175125122,
2414
+ "eval_runtime": 30.8496,
2415
+ "eval_samples_per_second": 115.658,
2416
+ "eval_steps_per_second": 14.457,
2417
+ "step": 2920
2418
+ },
2419
+ {
2420
+ "epoch": 7.51,
2421
+ "learning_rate": 1.217948717948718e-05,
2422
+ "loss": 0.0015,
2423
+ "step": 2930
2424
+ },
2425
+ {
2426
+ "epoch": 7.54,
2427
+ "learning_rate": 1.153846153846154e-05,
2428
+ "loss": 0.0015,
2429
+ "step": 2940
2430
+ },
2431
+ {
2432
+ "epoch": 7.56,
2433
+ "learning_rate": 1.0897435897435898e-05,
2434
+ "loss": 0.0015,
2435
+ "step": 2950
2436
+ },
2437
+ {
2438
+ "epoch": 7.59,
2439
+ "learning_rate": 1.0256410256410256e-05,
2440
+ "loss": 0.0015,
2441
+ "step": 2960
2442
+ },
2443
+ {
2444
+ "epoch": 7.59,
2445
+ "eval_accuracy": 0.8249158249158249,
2446
+ "eval_loss": 0.8220140933990479,
2447
+ "eval_runtime": 31.6074,
2448
+ "eval_samples_per_second": 112.885,
2449
+ "eval_steps_per_second": 14.111,
2450
+ "step": 2960
2451
+ },
2452
+ {
2453
+ "epoch": 7.62,
2454
+ "learning_rate": 9.615384615384616e-06,
2455
+ "loss": 0.0016,
2456
+ "step": 2970
2457
+ },
2458
+ {
2459
+ "epoch": 7.64,
2460
+ "learning_rate": 8.974358974358976e-06,
2461
+ "loss": 0.0014,
2462
+ "step": 2980
2463
+ },
2464
+ {
2465
+ "epoch": 7.67,
2466
+ "learning_rate": 8.333333333333334e-06,
2467
+ "loss": 0.0015,
2468
+ "step": 2990
2469
+ },
2470
+ {
2471
+ "epoch": 7.69,
2472
+ "learning_rate": 7.692307692307694e-06,
2473
+ "loss": 0.0015,
2474
+ "step": 3000
2475
+ },
2476
+ {
2477
+ "epoch": 7.69,
2478
+ "eval_accuracy": 0.8254769921436588,
2479
+ "eval_loss": 0.8220102787017822,
2480
+ "eval_runtime": 30.2325,
2481
+ "eval_samples_per_second": 118.019,
2482
+ "eval_steps_per_second": 14.752,
2483
+ "step": 3000
2484
+ },
2485
+ {
2486
+ "epoch": 7.72,
2487
+ "learning_rate": 7.051282051282052e-06,
2488
+ "loss": 0.0014,
2489
+ "step": 3010
2490
+ },
2491
+ {
2492
+ "epoch": 7.74,
2493
+ "learning_rate": 6.41025641025641e-06,
2494
+ "loss": 0.0014,
2495
+ "step": 3020
2496
+ },
2497
+ {
2498
+ "epoch": 7.77,
2499
+ "learning_rate": 5.76923076923077e-06,
2500
+ "loss": 0.0015,
2501
+ "step": 3030
2502
+ },
2503
+ {
2504
+ "epoch": 7.79,
2505
+ "learning_rate": 5.128205128205128e-06,
2506
+ "loss": 0.0015,
2507
+ "step": 3040
2508
+ },
2509
+ {
2510
+ "epoch": 7.79,
2511
+ "eval_accuracy": 0.8254769921436588,
2512
+ "eval_loss": 0.8217305541038513,
2513
+ "eval_runtime": 38.9983,
2514
+ "eval_samples_per_second": 91.491,
2515
+ "eval_steps_per_second": 11.436,
2516
+ "step": 3040
2517
+ },
2518
+ {
2519
+ "epoch": 7.82,
2520
+ "learning_rate": 4.487179487179488e-06,
2521
+ "loss": 0.0014,
2522
+ "step": 3050
2523
+ },
2524
+ {
2525
+ "epoch": 7.85,
2526
+ "learning_rate": 3.846153846153847e-06,
2527
+ "loss": 0.0015,
2528
+ "step": 3060
2529
+ },
2530
+ {
2531
+ "epoch": 7.87,
2532
+ "learning_rate": 3.205128205128205e-06,
2533
+ "loss": 0.0014,
2534
+ "step": 3070
2535
+ },
2536
+ {
2537
+ "epoch": 7.9,
2538
+ "learning_rate": 2.564102564102564e-06,
2539
+ "loss": 0.0014,
2540
+ "step": 3080
2541
+ },
2542
+ {
2543
+ "epoch": 7.9,
2544
+ "eval_accuracy": 0.8254769921436588,
2545
+ "eval_loss": 0.8224795460700989,
2546
+ "eval_runtime": 31.9073,
2547
+ "eval_samples_per_second": 111.824,
2548
+ "eval_steps_per_second": 13.978,
2549
+ "step": 3080
2550
+ },
2551
+ {
2552
+ "epoch": 7.92,
2553
+ "learning_rate": 1.9230769230769234e-06,
2554
+ "loss": 0.0014,
2555
+ "step": 3090
2556
+ },
2557
+ {
2558
+ "epoch": 7.95,
2559
+ "learning_rate": 1.282051282051282e-06,
2560
+ "loss": 0.0015,
2561
+ "step": 3100
2562
+ },
2563
+ {
2564
+ "epoch": 7.97,
2565
+ "learning_rate": 6.41025641025641e-07,
2566
+ "loss": 0.0014,
2567
+ "step": 3110
2568
+ },
2569
+ {
2570
+ "epoch": 8.0,
2571
+ "learning_rate": 0.0,
2572
+ "loss": 0.0014,
2573
+ "step": 3120
2574
+ },
2575
+ {
2576
+ "epoch": 8.0,
2577
+ "eval_accuracy": 0.8260381593714927,
2578
+ "eval_loss": 0.8224756717681885,
2579
+ "eval_runtime": 30.7985,
2580
+ "eval_samples_per_second": 115.85,
2581
+ "eval_steps_per_second": 14.481,
2582
+ "step": 3120
2583
+ },
2584
+ {
2585
+ "epoch": 8.0,
2586
+ "step": 3120,
2587
+ "total_flos": 3.8629591832685773e+18,
2588
+ "train_loss": 0.2533179052472592,
2589
+ "train_runtime": 4957.8804,
2590
+ "train_samples_per_second": 10.069,
2591
+ "train_steps_per_second": 0.629
2592
+ }
2593
+ ],
2594
+ "max_steps": 3120,
2595
+ "num_train_epochs": 8,
2596
+ "total_flos": 3.8629591832685773e+18,
2597
+ "trial_name": null,
2598
+ "trial_params": null
2599
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38c25b2b280e8f8dde2511a004bf1b097d7641bf1aae99dd53711356509cd78e
3
+ size 3899