Payoto commited on
Commit
802d980
1 Parent(s): f588b0e

End of training

Browse files
Files changed (4) hide show
  1. all_results.json +12 -0
  2. eval_results.json +8 -0
  3. train_results.json +7 -0
  4. trainer_state.json +1415 -0
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_accuracy": 0.9844444444444445,
4
+ "eval_loss": 0.07232666015625,
5
+ "eval_runtime": 9.7892,
6
+ "eval_samples_per_second": 275.814,
7
+ "eval_steps_per_second": 68.953,
8
+ "train_loss": 0.3087928044884991,
9
+ "train_runtime": 984.02,
10
+ "train_samples_per_second": 74.047,
11
+ "train_steps_per_second": 2.314
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "eval_accuracy": 0.9844444444444445,
4
+ "eval_loss": 0.07232666015625,
5
+ "eval_runtime": 9.7892,
6
+ "eval_samples_per_second": 275.814,
7
+ "eval_steps_per_second": 68.953
8
+ }
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 3.0,
3
+ "train_loss": 0.3087928044884991,
4
+ "train_runtime": 984.02,
5
+ "train_samples_per_second": 74.047,
6
+ "train_steps_per_second": 2.314
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1415 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9844444444444445,
3
+ "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-eurosat/checkpoint-2277",
4
+ "epoch": 3.0,
5
+ "global_step": 2277,
6
+ "is_hyper_param_search": false,
7
+ "is_local_process_zero": true,
8
+ "is_world_process_zero": true,
9
+ "log_history": [
10
+ {
11
+ "epoch": 0.01,
12
+ "learning_rate": 2.1929824561403507e-06,
13
+ "loss": 2.3182,
14
+ "step": 10
15
+ },
16
+ {
17
+ "epoch": 0.03,
18
+ "learning_rate": 4.3859649122807014e-06,
19
+ "loss": 2.3088,
20
+ "step": 20
21
+ },
22
+ {
23
+ "epoch": 0.04,
24
+ "learning_rate": 6.578947368421053e-06,
25
+ "loss": 2.2658,
26
+ "step": 30
27
+ },
28
+ {
29
+ "epoch": 0.05,
30
+ "learning_rate": 8.771929824561403e-06,
31
+ "loss": 2.1543,
32
+ "step": 40
33
+ },
34
+ {
35
+ "epoch": 0.07,
36
+ "learning_rate": 1.0964912280701754e-05,
37
+ "loss": 2.2097,
38
+ "step": 50
39
+ },
40
+ {
41
+ "epoch": 0.08,
42
+ "learning_rate": 1.3157894736842106e-05,
43
+ "loss": 2.1236,
44
+ "step": 60
45
+ },
46
+ {
47
+ "epoch": 0.09,
48
+ "learning_rate": 1.5350877192982457e-05,
49
+ "loss": 1.9738,
50
+ "step": 70
51
+ },
52
+ {
53
+ "epoch": 0.11,
54
+ "learning_rate": 1.7543859649122806e-05,
55
+ "loss": 1.8933,
56
+ "step": 80
57
+ },
58
+ {
59
+ "epoch": 0.12,
60
+ "learning_rate": 1.9736842105263158e-05,
61
+ "loss": 1.7186,
62
+ "step": 90
63
+ },
64
+ {
65
+ "epoch": 0.13,
66
+ "learning_rate": 2.1929824561403507e-05,
67
+ "loss": 1.7101,
68
+ "step": 100
69
+ },
70
+ {
71
+ "epoch": 0.14,
72
+ "learning_rate": 2.412280701754386e-05,
73
+ "loss": 1.7188,
74
+ "step": 110
75
+ },
76
+ {
77
+ "epoch": 0.16,
78
+ "learning_rate": 2.6315789473684212e-05,
79
+ "loss": 1.0406,
80
+ "step": 120
81
+ },
82
+ {
83
+ "epoch": 0.17,
84
+ "learning_rate": 2.850877192982456e-05,
85
+ "loss": 1.1604,
86
+ "step": 130
87
+ },
88
+ {
89
+ "epoch": 0.18,
90
+ "learning_rate": 3.0701754385964913e-05,
91
+ "loss": 1.1226,
92
+ "step": 140
93
+ },
94
+ {
95
+ "epoch": 0.2,
96
+ "learning_rate": 3.289473684210527e-05,
97
+ "loss": 1.0185,
98
+ "step": 150
99
+ },
100
+ {
101
+ "epoch": 0.21,
102
+ "learning_rate": 3.508771929824561e-05,
103
+ "loss": 1.0203,
104
+ "step": 160
105
+ },
106
+ {
107
+ "epoch": 0.22,
108
+ "learning_rate": 3.728070175438597e-05,
109
+ "loss": 0.5058,
110
+ "step": 170
111
+ },
112
+ {
113
+ "epoch": 0.24,
114
+ "learning_rate": 3.9473684210526316e-05,
115
+ "loss": 0.7132,
116
+ "step": 180
117
+ },
118
+ {
119
+ "epoch": 0.25,
120
+ "learning_rate": 4.166666666666667e-05,
121
+ "loss": 0.7988,
122
+ "step": 190
123
+ },
124
+ {
125
+ "epoch": 0.26,
126
+ "learning_rate": 4.3859649122807014e-05,
127
+ "loss": 0.5992,
128
+ "step": 200
129
+ },
130
+ {
131
+ "epoch": 0.28,
132
+ "learning_rate": 4.605263157894737e-05,
133
+ "loss": 0.7898,
134
+ "step": 210
135
+ },
136
+ {
137
+ "epoch": 0.29,
138
+ "learning_rate": 4.824561403508772e-05,
139
+ "loss": 0.6248,
140
+ "step": 220
141
+ },
142
+ {
143
+ "epoch": 0.3,
144
+ "learning_rate": 4.9951195705222064e-05,
145
+ "loss": 0.6488,
146
+ "step": 230
147
+ },
148
+ {
149
+ "epoch": 0.32,
150
+ "learning_rate": 4.970717423133236e-05,
151
+ "loss": 0.6448,
152
+ "step": 240
153
+ },
154
+ {
155
+ "epoch": 0.33,
156
+ "learning_rate": 4.946315275744266e-05,
157
+ "loss": 0.6157,
158
+ "step": 250
159
+ },
160
+ {
161
+ "epoch": 0.34,
162
+ "learning_rate": 4.9219131283552955e-05,
163
+ "loss": 0.636,
164
+ "step": 260
165
+ },
166
+ {
167
+ "epoch": 0.36,
168
+ "learning_rate": 4.897510980966325e-05,
169
+ "loss": 0.6592,
170
+ "step": 270
171
+ },
172
+ {
173
+ "epoch": 0.37,
174
+ "learning_rate": 4.873108833577355e-05,
175
+ "loss": 0.6987,
176
+ "step": 280
177
+ },
178
+ {
179
+ "epoch": 0.38,
180
+ "learning_rate": 4.848706686188385e-05,
181
+ "loss": 0.5183,
182
+ "step": 290
183
+ },
184
+ {
185
+ "epoch": 0.4,
186
+ "learning_rate": 4.824304538799414e-05,
187
+ "loss": 0.2226,
188
+ "step": 300
189
+ },
190
+ {
191
+ "epoch": 0.41,
192
+ "learning_rate": 4.799902391410445e-05,
193
+ "loss": 1.0035,
194
+ "step": 310
195
+ },
196
+ {
197
+ "epoch": 0.42,
198
+ "learning_rate": 4.7755002440214744e-05,
199
+ "loss": 0.2608,
200
+ "step": 320
201
+ },
202
+ {
203
+ "epoch": 0.43,
204
+ "learning_rate": 4.751098096632504e-05,
205
+ "loss": 0.2487,
206
+ "step": 330
207
+ },
208
+ {
209
+ "epoch": 0.45,
210
+ "learning_rate": 4.726695949243534e-05,
211
+ "loss": 0.5035,
212
+ "step": 340
213
+ },
214
+ {
215
+ "epoch": 0.46,
216
+ "learning_rate": 4.7022938018545635e-05,
217
+ "loss": 0.6245,
218
+ "step": 350
219
+ },
220
+ {
221
+ "epoch": 0.47,
222
+ "learning_rate": 4.677891654465593e-05,
223
+ "loss": 0.2621,
224
+ "step": 360
225
+ },
226
+ {
227
+ "epoch": 0.49,
228
+ "learning_rate": 4.6534895070766235e-05,
229
+ "loss": 0.2333,
230
+ "step": 370
231
+ },
232
+ {
233
+ "epoch": 0.5,
234
+ "learning_rate": 4.6290873596876525e-05,
235
+ "loss": 0.1616,
236
+ "step": 380
237
+ },
238
+ {
239
+ "epoch": 0.51,
240
+ "learning_rate": 4.604685212298683e-05,
241
+ "loss": 0.2174,
242
+ "step": 390
243
+ },
244
+ {
245
+ "epoch": 0.53,
246
+ "learning_rate": 4.580283064909712e-05,
247
+ "loss": 0.5526,
248
+ "step": 400
249
+ },
250
+ {
251
+ "epoch": 0.54,
252
+ "learning_rate": 4.555880917520742e-05,
253
+ "loss": 0.2229,
254
+ "step": 410
255
+ },
256
+ {
257
+ "epoch": 0.55,
258
+ "learning_rate": 4.531478770131772e-05,
259
+ "loss": 0.1677,
260
+ "step": 420
261
+ },
262
+ {
263
+ "epoch": 0.57,
264
+ "learning_rate": 4.507076622742802e-05,
265
+ "loss": 0.14,
266
+ "step": 430
267
+ },
268
+ {
269
+ "epoch": 0.58,
270
+ "learning_rate": 4.4826744753538314e-05,
271
+ "loss": 0.1421,
272
+ "step": 440
273
+ },
274
+ {
275
+ "epoch": 0.59,
276
+ "learning_rate": 4.458272327964861e-05,
277
+ "loss": 0.1422,
278
+ "step": 450
279
+ },
280
+ {
281
+ "epoch": 0.61,
282
+ "learning_rate": 4.433870180575891e-05,
283
+ "loss": 0.1693,
284
+ "step": 460
285
+ },
286
+ {
287
+ "epoch": 0.62,
288
+ "learning_rate": 4.409468033186921e-05,
289
+ "loss": 0.1312,
290
+ "step": 470
291
+ },
292
+ {
293
+ "epoch": 0.63,
294
+ "learning_rate": 4.38506588579795e-05,
295
+ "loss": 0.5159,
296
+ "step": 480
297
+ },
298
+ {
299
+ "epoch": 0.65,
300
+ "learning_rate": 4.3606637384089805e-05,
301
+ "loss": 0.1083,
302
+ "step": 490
303
+ },
304
+ {
305
+ "epoch": 0.66,
306
+ "learning_rate": 4.33626159102001e-05,
307
+ "loss": 0.5269,
308
+ "step": 500
309
+ },
310
+ {
311
+ "epoch": 0.67,
312
+ "learning_rate": 4.31185944363104e-05,
313
+ "loss": 0.1268,
314
+ "step": 510
315
+ },
316
+ {
317
+ "epoch": 0.69,
318
+ "learning_rate": 4.2874572962420696e-05,
319
+ "loss": 0.0972,
320
+ "step": 520
321
+ },
322
+ {
323
+ "epoch": 0.7,
324
+ "learning_rate": 4.263055148853099e-05,
325
+ "loss": 0.104,
326
+ "step": 530
327
+ },
328
+ {
329
+ "epoch": 0.71,
330
+ "learning_rate": 4.238653001464129e-05,
331
+ "loss": 0.3534,
332
+ "step": 540
333
+ },
334
+ {
335
+ "epoch": 0.72,
336
+ "learning_rate": 4.2142508540751594e-05,
337
+ "loss": 0.1813,
338
+ "step": 550
339
+ },
340
+ {
341
+ "epoch": 0.74,
342
+ "learning_rate": 4.1898487066861884e-05,
343
+ "loss": 0.2723,
344
+ "step": 560
345
+ },
346
+ {
347
+ "epoch": 0.75,
348
+ "learning_rate": 4.165446559297219e-05,
349
+ "loss": 0.1537,
350
+ "step": 570
351
+ },
352
+ {
353
+ "epoch": 0.76,
354
+ "learning_rate": 4.141044411908248e-05,
355
+ "loss": 0.5187,
356
+ "step": 580
357
+ },
358
+ {
359
+ "epoch": 0.78,
360
+ "learning_rate": 4.116642264519278e-05,
361
+ "loss": 0.1365,
362
+ "step": 590
363
+ },
364
+ {
365
+ "epoch": 0.79,
366
+ "learning_rate": 4.092240117130308e-05,
367
+ "loss": 0.2843,
368
+ "step": 600
369
+ },
370
+ {
371
+ "epoch": 0.8,
372
+ "learning_rate": 4.0678379697413375e-05,
373
+ "loss": 0.385,
374
+ "step": 610
375
+ },
376
+ {
377
+ "epoch": 0.82,
378
+ "learning_rate": 4.043435822352367e-05,
379
+ "loss": 0.3066,
380
+ "step": 620
381
+ },
382
+ {
383
+ "epoch": 0.83,
384
+ "learning_rate": 4.019033674963397e-05,
385
+ "loss": 0.1013,
386
+ "step": 630
387
+ },
388
+ {
389
+ "epoch": 0.84,
390
+ "learning_rate": 3.9946315275744266e-05,
391
+ "loss": 0.5619,
392
+ "step": 640
393
+ },
394
+ {
395
+ "epoch": 0.86,
396
+ "learning_rate": 3.970229380185457e-05,
397
+ "loss": 0.1353,
398
+ "step": 650
399
+ },
400
+ {
401
+ "epoch": 0.87,
402
+ "learning_rate": 3.945827232796486e-05,
403
+ "loss": 0.2441,
404
+ "step": 660
405
+ },
406
+ {
407
+ "epoch": 0.88,
408
+ "learning_rate": 3.9214250854075164e-05,
409
+ "loss": 0.091,
410
+ "step": 670
411
+ },
412
+ {
413
+ "epoch": 0.9,
414
+ "learning_rate": 3.897022938018546e-05,
415
+ "loss": 0.1073,
416
+ "step": 680
417
+ },
418
+ {
419
+ "epoch": 0.91,
420
+ "learning_rate": 3.872620790629576e-05,
421
+ "loss": 0.1118,
422
+ "step": 690
423
+ },
424
+ {
425
+ "epoch": 0.92,
426
+ "learning_rate": 3.8482186432406055e-05,
427
+ "loss": 0.0803,
428
+ "step": 700
429
+ },
430
+ {
431
+ "epoch": 0.94,
432
+ "learning_rate": 3.823816495851635e-05,
433
+ "loss": 0.127,
434
+ "step": 710
435
+ },
436
+ {
437
+ "epoch": 0.95,
438
+ "learning_rate": 3.799414348462665e-05,
439
+ "loss": 0.109,
440
+ "step": 720
441
+ },
442
+ {
443
+ "epoch": 0.96,
444
+ "learning_rate": 3.775012201073695e-05,
445
+ "loss": 0.0735,
446
+ "step": 730
447
+ },
448
+ {
449
+ "epoch": 0.97,
450
+ "learning_rate": 3.750610053684724e-05,
451
+ "loss": 0.1187,
452
+ "step": 740
453
+ },
454
+ {
455
+ "epoch": 0.99,
456
+ "learning_rate": 3.7262079062957546e-05,
457
+ "loss": 0.2969,
458
+ "step": 750
459
+ },
460
+ {
461
+ "epoch": 1.0,
462
+ "eval_accuracy": 0.9785185185185186,
463
+ "eval_loss": 0.1263427734375,
464
+ "eval_runtime": 10.5602,
465
+ "eval_samples_per_second": 255.677,
466
+ "eval_steps_per_second": 63.919,
467
+ "step": 759
468
+ },
469
+ {
470
+ "epoch": 1.0,
471
+ "learning_rate": 3.7018057589067836e-05,
472
+ "loss": 0.2753,
473
+ "step": 760
474
+ },
475
+ {
476
+ "epoch": 1.01,
477
+ "learning_rate": 3.677403611517814e-05,
478
+ "loss": 0.403,
479
+ "step": 770
480
+ },
481
+ {
482
+ "epoch": 1.03,
483
+ "learning_rate": 3.653001464128844e-05,
484
+ "loss": 0.2884,
485
+ "step": 780
486
+ },
487
+ {
488
+ "epoch": 1.04,
489
+ "learning_rate": 3.6285993167398734e-05,
490
+ "loss": 0.4866,
491
+ "step": 790
492
+ },
493
+ {
494
+ "epoch": 1.05,
495
+ "learning_rate": 3.604197169350903e-05,
496
+ "loss": 0.4529,
497
+ "step": 800
498
+ },
499
+ {
500
+ "epoch": 1.07,
501
+ "learning_rate": 3.579795021961933e-05,
502
+ "loss": 0.063,
503
+ "step": 810
504
+ },
505
+ {
506
+ "epoch": 1.08,
507
+ "learning_rate": 3.5553928745729625e-05,
508
+ "loss": 0.0952,
509
+ "step": 820
510
+ },
511
+ {
512
+ "epoch": 1.09,
513
+ "learning_rate": 3.530990727183993e-05,
514
+ "loss": 0.0886,
515
+ "step": 830
516
+ },
517
+ {
518
+ "epoch": 1.11,
519
+ "learning_rate": 3.506588579795022e-05,
520
+ "loss": 0.2751,
521
+ "step": 840
522
+ },
523
+ {
524
+ "epoch": 1.12,
525
+ "learning_rate": 3.482186432406052e-05,
526
+ "loss": 0.09,
527
+ "step": 850
528
+ },
529
+ {
530
+ "epoch": 1.13,
531
+ "learning_rate": 3.457784285017082e-05,
532
+ "loss": 0.8028,
533
+ "step": 860
534
+ },
535
+ {
536
+ "epoch": 1.15,
537
+ "learning_rate": 3.4333821376281116e-05,
538
+ "loss": 0.0623,
539
+ "step": 870
540
+ },
541
+ {
542
+ "epoch": 1.16,
543
+ "learning_rate": 3.408979990239141e-05,
544
+ "loss": 0.0583,
545
+ "step": 880
546
+ },
547
+ {
548
+ "epoch": 1.17,
549
+ "learning_rate": 3.384577842850171e-05,
550
+ "loss": 0.5004,
551
+ "step": 890
552
+ },
553
+ {
554
+ "epoch": 1.19,
555
+ "learning_rate": 3.360175695461201e-05,
556
+ "loss": 0.0587,
557
+ "step": 900
558
+ },
559
+ {
560
+ "epoch": 1.2,
561
+ "learning_rate": 3.335773548072231e-05,
562
+ "loss": 0.066,
563
+ "step": 910
564
+ },
565
+ {
566
+ "epoch": 1.21,
567
+ "learning_rate": 3.31137140068326e-05,
568
+ "loss": 0.1028,
569
+ "step": 920
570
+ },
571
+ {
572
+ "epoch": 1.23,
573
+ "learning_rate": 3.2869692532942905e-05,
574
+ "loss": 0.069,
575
+ "step": 930
576
+ },
577
+ {
578
+ "epoch": 1.24,
579
+ "learning_rate": 3.2625671059053195e-05,
580
+ "loss": 0.1084,
581
+ "step": 940
582
+ },
583
+ {
584
+ "epoch": 1.25,
585
+ "learning_rate": 3.23816495851635e-05,
586
+ "loss": 0.8451,
587
+ "step": 950
588
+ },
589
+ {
590
+ "epoch": 1.26,
591
+ "learning_rate": 3.2137628111273795e-05,
592
+ "loss": 0.0993,
593
+ "step": 960
594
+ },
595
+ {
596
+ "epoch": 1.28,
597
+ "learning_rate": 3.189360663738409e-05,
598
+ "loss": 0.0488,
599
+ "step": 970
600
+ },
601
+ {
602
+ "epoch": 1.29,
603
+ "learning_rate": 3.164958516349439e-05,
604
+ "loss": 0.0645,
605
+ "step": 980
606
+ },
607
+ {
608
+ "epoch": 1.3,
609
+ "learning_rate": 3.1405563689604686e-05,
610
+ "loss": 0.0586,
611
+ "step": 990
612
+ },
613
+ {
614
+ "epoch": 1.32,
615
+ "learning_rate": 3.116154221571498e-05,
616
+ "loss": 0.4479,
617
+ "step": 1000
618
+ },
619
+ {
620
+ "epoch": 1.33,
621
+ "learning_rate": 3.091752074182529e-05,
622
+ "loss": 0.4036,
623
+ "step": 1010
624
+ },
625
+ {
626
+ "epoch": 1.34,
627
+ "learning_rate": 3.067349926793558e-05,
628
+ "loss": 0.0975,
629
+ "step": 1020
630
+ },
631
+ {
632
+ "epoch": 1.36,
633
+ "learning_rate": 3.042947779404588e-05,
634
+ "loss": 0.7869,
635
+ "step": 1030
636
+ },
637
+ {
638
+ "epoch": 1.37,
639
+ "learning_rate": 3.0185456320156174e-05,
640
+ "loss": 0.0552,
641
+ "step": 1040
642
+ },
643
+ {
644
+ "epoch": 1.38,
645
+ "learning_rate": 2.9941434846266475e-05,
646
+ "loss": 0.0515,
647
+ "step": 1050
648
+ },
649
+ {
650
+ "epoch": 1.4,
651
+ "learning_rate": 2.969741337237677e-05,
652
+ "loss": 0.6272,
653
+ "step": 1060
654
+ },
655
+ {
656
+ "epoch": 1.41,
657
+ "learning_rate": 2.9453391898487072e-05,
658
+ "loss": 0.0522,
659
+ "step": 1070
660
+ },
661
+ {
662
+ "epoch": 1.42,
663
+ "learning_rate": 2.9209370424597365e-05,
664
+ "loss": 0.0493,
665
+ "step": 1080
666
+ },
667
+ {
668
+ "epoch": 1.44,
669
+ "learning_rate": 2.8965348950707666e-05,
670
+ "loss": 0.4377,
671
+ "step": 1090
672
+ },
673
+ {
674
+ "epoch": 1.45,
675
+ "learning_rate": 2.872132747681796e-05,
676
+ "loss": 0.0417,
677
+ "step": 1100
678
+ },
679
+ {
680
+ "epoch": 1.46,
681
+ "learning_rate": 2.847730600292826e-05,
682
+ "loss": 0.0684,
683
+ "step": 1110
684
+ },
685
+ {
686
+ "epoch": 1.48,
687
+ "learning_rate": 2.8233284529038557e-05,
688
+ "loss": 0.0735,
689
+ "step": 1120
690
+ },
691
+ {
692
+ "epoch": 1.49,
693
+ "learning_rate": 2.7989263055148857e-05,
694
+ "loss": 0.0614,
695
+ "step": 1130
696
+ },
697
+ {
698
+ "epoch": 1.5,
699
+ "learning_rate": 2.774524158125915e-05,
700
+ "loss": 0.3421,
701
+ "step": 1140
702
+ },
703
+ {
704
+ "epoch": 1.52,
705
+ "learning_rate": 2.750122010736945e-05,
706
+ "loss": 0.104,
707
+ "step": 1150
708
+ },
709
+ {
710
+ "epoch": 1.53,
711
+ "learning_rate": 2.7257198633479748e-05,
712
+ "loss": 0.6174,
713
+ "step": 1160
714
+ },
715
+ {
716
+ "epoch": 1.54,
717
+ "learning_rate": 2.7013177159590048e-05,
718
+ "loss": 0.1763,
719
+ "step": 1170
720
+ },
721
+ {
722
+ "epoch": 1.55,
723
+ "learning_rate": 2.676915568570034e-05,
724
+ "loss": 0.2104,
725
+ "step": 1180
726
+ },
727
+ {
728
+ "epoch": 1.57,
729
+ "learning_rate": 2.6525134211810642e-05,
730
+ "loss": 0.0592,
731
+ "step": 1190
732
+ },
733
+ {
734
+ "epoch": 1.58,
735
+ "learning_rate": 2.628111273792094e-05,
736
+ "loss": 0.1282,
737
+ "step": 1200
738
+ },
739
+ {
740
+ "epoch": 1.59,
741
+ "learning_rate": 2.603709126403124e-05,
742
+ "loss": 0.0435,
743
+ "step": 1210
744
+ },
745
+ {
746
+ "epoch": 1.61,
747
+ "learning_rate": 2.5793069790141533e-05,
748
+ "loss": 0.2258,
749
+ "step": 1220
750
+ },
751
+ {
752
+ "epoch": 1.62,
753
+ "learning_rate": 2.5549048316251833e-05,
754
+ "loss": 0.2802,
755
+ "step": 1230
756
+ },
757
+ {
758
+ "epoch": 1.63,
759
+ "learning_rate": 2.530502684236213e-05,
760
+ "loss": 0.0463,
761
+ "step": 1240
762
+ },
763
+ {
764
+ "epoch": 1.65,
765
+ "learning_rate": 2.506100536847243e-05,
766
+ "loss": 0.1848,
767
+ "step": 1250
768
+ },
769
+ {
770
+ "epoch": 1.66,
771
+ "learning_rate": 2.4816983894582724e-05,
772
+ "loss": 0.0794,
773
+ "step": 1260
774
+ },
775
+ {
776
+ "epoch": 1.67,
777
+ "learning_rate": 2.457296242069302e-05,
778
+ "loss": 0.0435,
779
+ "step": 1270
780
+ },
781
+ {
782
+ "epoch": 1.69,
783
+ "learning_rate": 2.4328940946803318e-05,
784
+ "loss": 0.0921,
785
+ "step": 1280
786
+ },
787
+ {
788
+ "epoch": 1.7,
789
+ "learning_rate": 2.4084919472913618e-05,
790
+ "loss": 0.1854,
791
+ "step": 1290
792
+ },
793
+ {
794
+ "epoch": 1.71,
795
+ "learning_rate": 2.3840897999023915e-05,
796
+ "loss": 0.0384,
797
+ "step": 1300
798
+ },
799
+ {
800
+ "epoch": 1.73,
801
+ "learning_rate": 2.3596876525134212e-05,
802
+ "loss": 0.0462,
803
+ "step": 1310
804
+ },
805
+ {
806
+ "epoch": 1.74,
807
+ "learning_rate": 2.335285505124451e-05,
808
+ "loss": 0.0802,
809
+ "step": 1320
810
+ },
811
+ {
812
+ "epoch": 1.75,
813
+ "learning_rate": 2.310883357735481e-05,
814
+ "loss": 0.046,
815
+ "step": 1330
816
+ },
817
+ {
818
+ "epoch": 1.77,
819
+ "learning_rate": 2.2864812103465106e-05,
820
+ "loss": 0.0775,
821
+ "step": 1340
822
+ },
823
+ {
824
+ "epoch": 1.78,
825
+ "learning_rate": 2.2620790629575403e-05,
826
+ "loss": 0.0532,
827
+ "step": 1350
828
+ },
829
+ {
830
+ "epoch": 1.79,
831
+ "learning_rate": 2.23767691556857e-05,
832
+ "loss": 0.0615,
833
+ "step": 1360
834
+ },
835
+ {
836
+ "epoch": 1.81,
837
+ "learning_rate": 2.2132747681795997e-05,
838
+ "loss": 0.0435,
839
+ "step": 1370
840
+ },
841
+ {
842
+ "epoch": 1.82,
843
+ "learning_rate": 2.1888726207906297e-05,
844
+ "loss": 0.0372,
845
+ "step": 1380
846
+ },
847
+ {
848
+ "epoch": 1.83,
849
+ "learning_rate": 2.1644704734016594e-05,
850
+ "loss": 0.0351,
851
+ "step": 1390
852
+ },
853
+ {
854
+ "epoch": 1.84,
855
+ "learning_rate": 2.140068326012689e-05,
856
+ "loss": 0.0523,
857
+ "step": 1400
858
+ },
859
+ {
860
+ "epoch": 1.86,
861
+ "learning_rate": 2.1156661786237188e-05,
862
+ "loss": 0.0345,
863
+ "step": 1410
864
+ },
865
+ {
866
+ "epoch": 1.87,
867
+ "learning_rate": 2.0912640312347485e-05,
868
+ "loss": 0.0359,
869
+ "step": 1420
870
+ },
871
+ {
872
+ "epoch": 1.88,
873
+ "learning_rate": 2.0668618838457785e-05,
874
+ "loss": 0.5072,
875
+ "step": 1430
876
+ },
877
+ {
878
+ "epoch": 1.9,
879
+ "learning_rate": 2.0424597364568082e-05,
880
+ "loss": 0.0281,
881
+ "step": 1440
882
+ },
883
+ {
884
+ "epoch": 1.91,
885
+ "learning_rate": 2.018057589067838e-05,
886
+ "loss": 0.0366,
887
+ "step": 1450
888
+ },
889
+ {
890
+ "epoch": 1.92,
891
+ "learning_rate": 1.9936554416788676e-05,
892
+ "loss": 0.2221,
893
+ "step": 1460
894
+ },
895
+ {
896
+ "epoch": 1.94,
897
+ "learning_rate": 1.9692532942898977e-05,
898
+ "loss": 0.0413,
899
+ "step": 1470
900
+ },
901
+ {
902
+ "epoch": 1.95,
903
+ "learning_rate": 1.9448511469009274e-05,
904
+ "loss": 0.3138,
905
+ "step": 1480
906
+ },
907
+ {
908
+ "epoch": 1.96,
909
+ "learning_rate": 1.920448999511957e-05,
910
+ "loss": 0.113,
911
+ "step": 1490
912
+ },
913
+ {
914
+ "epoch": 1.98,
915
+ "learning_rate": 1.8960468521229867e-05,
916
+ "loss": 0.037,
917
+ "step": 1500
918
+ },
919
+ {
920
+ "epoch": 1.99,
921
+ "learning_rate": 1.8716447047340164e-05,
922
+ "loss": 0.0375,
923
+ "step": 1510
924
+ },
925
+ {
926
+ "epoch": 2.0,
927
+ "eval_accuracy": 0.9792592592592593,
928
+ "eval_loss": 0.09527587890625,
929
+ "eval_runtime": 9.7577,
930
+ "eval_samples_per_second": 276.704,
931
+ "eval_steps_per_second": 69.176,
932
+ "step": 1518
933
+ },
934
+ {
935
+ "epoch": 2.0,
936
+ "learning_rate": 1.8472425573450465e-05,
937
+ "loss": 0.0722,
938
+ "step": 1520
939
+ },
940
+ {
941
+ "epoch": 2.02,
942
+ "learning_rate": 1.822840409956076e-05,
943
+ "loss": 0.0311,
944
+ "step": 1530
945
+ },
946
+ {
947
+ "epoch": 2.03,
948
+ "learning_rate": 1.798438262567106e-05,
949
+ "loss": 0.4362,
950
+ "step": 1540
951
+ },
952
+ {
953
+ "epoch": 2.04,
954
+ "learning_rate": 1.7740361151781356e-05,
955
+ "loss": 0.1499,
956
+ "step": 1550
957
+ },
958
+ {
959
+ "epoch": 2.06,
960
+ "learning_rate": 1.7496339677891656e-05,
961
+ "loss": 0.2203,
962
+ "step": 1560
963
+ },
964
+ {
965
+ "epoch": 2.07,
966
+ "learning_rate": 1.7252318204001953e-05,
967
+ "loss": 0.0499,
968
+ "step": 1570
969
+ },
970
+ {
971
+ "epoch": 2.08,
972
+ "learning_rate": 1.700829673011225e-05,
973
+ "loss": 0.1047,
974
+ "step": 1580
975
+ },
976
+ {
977
+ "epoch": 2.09,
978
+ "learning_rate": 1.6764275256222547e-05,
979
+ "loss": 0.0313,
980
+ "step": 1590
981
+ },
982
+ {
983
+ "epoch": 2.11,
984
+ "learning_rate": 1.6520253782332844e-05,
985
+ "loss": 0.2069,
986
+ "step": 1600
987
+ },
988
+ {
989
+ "epoch": 2.12,
990
+ "learning_rate": 1.6276232308443144e-05,
991
+ "loss": 0.0337,
992
+ "step": 1610
993
+ },
994
+ {
995
+ "epoch": 2.13,
996
+ "learning_rate": 1.603221083455344e-05,
997
+ "loss": 0.1789,
998
+ "step": 1620
999
+ },
1000
+ {
1001
+ "epoch": 2.15,
1002
+ "learning_rate": 1.5788189360663738e-05,
1003
+ "loss": 0.0437,
1004
+ "step": 1630
1005
+ },
1006
+ {
1007
+ "epoch": 2.16,
1008
+ "learning_rate": 1.5544167886774035e-05,
1009
+ "loss": 0.0346,
1010
+ "step": 1640
1011
+ },
1012
+ {
1013
+ "epoch": 2.17,
1014
+ "learning_rate": 1.5300146412884335e-05,
1015
+ "loss": 0.0351,
1016
+ "step": 1650
1017
+ },
1018
+ {
1019
+ "epoch": 2.19,
1020
+ "learning_rate": 1.5056124938994632e-05,
1021
+ "loss": 0.2496,
1022
+ "step": 1660
1023
+ },
1024
+ {
1025
+ "epoch": 2.2,
1026
+ "learning_rate": 1.4812103465104929e-05,
1027
+ "loss": 0.1546,
1028
+ "step": 1670
1029
+ },
1030
+ {
1031
+ "epoch": 2.21,
1032
+ "learning_rate": 1.4568081991215226e-05,
1033
+ "loss": 0.6364,
1034
+ "step": 1680
1035
+ },
1036
+ {
1037
+ "epoch": 2.23,
1038
+ "learning_rate": 1.4324060517325525e-05,
1039
+ "loss": 0.0405,
1040
+ "step": 1690
1041
+ },
1042
+ {
1043
+ "epoch": 2.24,
1044
+ "learning_rate": 1.4080039043435821e-05,
1045
+ "loss": 0.083,
1046
+ "step": 1700
1047
+ },
1048
+ {
1049
+ "epoch": 2.25,
1050
+ "learning_rate": 1.383601756954612e-05,
1051
+ "loss": 0.0354,
1052
+ "step": 1710
1053
+ },
1054
+ {
1055
+ "epoch": 2.27,
1056
+ "learning_rate": 1.3591996095656417e-05,
1057
+ "loss": 0.0658,
1058
+ "step": 1720
1059
+ },
1060
+ {
1061
+ "epoch": 2.28,
1062
+ "learning_rate": 1.3347974621766716e-05,
1063
+ "loss": 0.09,
1064
+ "step": 1730
1065
+ },
1066
+ {
1067
+ "epoch": 2.29,
1068
+ "learning_rate": 1.3103953147877013e-05,
1069
+ "loss": 0.0259,
1070
+ "step": 1740
1071
+ },
1072
+ {
1073
+ "epoch": 2.31,
1074
+ "learning_rate": 1.2859931673987311e-05,
1075
+ "loss": 0.041,
1076
+ "step": 1750
1077
+ },
1078
+ {
1079
+ "epoch": 2.32,
1080
+ "learning_rate": 1.2615910200097608e-05,
1081
+ "loss": 0.1706,
1082
+ "step": 1760
1083
+ },
1084
+ {
1085
+ "epoch": 2.33,
1086
+ "learning_rate": 1.2371888726207907e-05,
1087
+ "loss": 0.0301,
1088
+ "step": 1770
1089
+ },
1090
+ {
1091
+ "epoch": 2.35,
1092
+ "learning_rate": 1.2127867252318205e-05,
1093
+ "loss": 0.2378,
1094
+ "step": 1780
1095
+ },
1096
+ {
1097
+ "epoch": 2.36,
1098
+ "learning_rate": 1.1883845778428502e-05,
1099
+ "loss": 0.0277,
1100
+ "step": 1790
1101
+ },
1102
+ {
1103
+ "epoch": 2.37,
1104
+ "learning_rate": 1.1639824304538801e-05,
1105
+ "loss": 0.0265,
1106
+ "step": 1800
1107
+ },
1108
+ {
1109
+ "epoch": 2.38,
1110
+ "learning_rate": 1.1395802830649098e-05,
1111
+ "loss": 0.036,
1112
+ "step": 1810
1113
+ },
1114
+ {
1115
+ "epoch": 2.4,
1116
+ "learning_rate": 1.1151781356759395e-05,
1117
+ "loss": 0.0587,
1118
+ "step": 1820
1119
+ },
1120
+ {
1121
+ "epoch": 2.41,
1122
+ "learning_rate": 1.0907759882869694e-05,
1123
+ "loss": 0.0388,
1124
+ "step": 1830
1125
+ },
1126
+ {
1127
+ "epoch": 2.42,
1128
+ "learning_rate": 1.066373840897999e-05,
1129
+ "loss": 0.0492,
1130
+ "step": 1840
1131
+ },
1132
+ {
1133
+ "epoch": 2.44,
1134
+ "learning_rate": 1.0419716935090289e-05,
1135
+ "loss": 0.0276,
1136
+ "step": 1850
1137
+ },
1138
+ {
1139
+ "epoch": 2.45,
1140
+ "learning_rate": 1.0175695461200586e-05,
1141
+ "loss": 0.0286,
1142
+ "step": 1860
1143
+ },
1144
+ {
1145
+ "epoch": 2.46,
1146
+ "learning_rate": 9.931673987310885e-06,
1147
+ "loss": 0.0285,
1148
+ "step": 1870
1149
+ },
1150
+ {
1151
+ "epoch": 2.48,
1152
+ "learning_rate": 9.687652513421182e-06,
1153
+ "loss": 0.0447,
1154
+ "step": 1880
1155
+ },
1156
+ {
1157
+ "epoch": 2.49,
1158
+ "learning_rate": 9.44363103953148e-06,
1159
+ "loss": 0.0258,
1160
+ "step": 1890
1161
+ },
1162
+ {
1163
+ "epoch": 2.5,
1164
+ "learning_rate": 9.199609565641777e-06,
1165
+ "loss": 0.0248,
1166
+ "step": 1900
1167
+ },
1168
+ {
1169
+ "epoch": 2.52,
1170
+ "learning_rate": 8.955588091752074e-06,
1171
+ "loss": 0.0379,
1172
+ "step": 1910
1173
+ },
1174
+ {
1175
+ "epoch": 2.53,
1176
+ "learning_rate": 8.711566617862373e-06,
1177
+ "loss": 0.0386,
1178
+ "step": 1920
1179
+ },
1180
+ {
1181
+ "epoch": 2.54,
1182
+ "learning_rate": 8.46754514397267e-06,
1183
+ "loss": 0.0938,
1184
+ "step": 1930
1185
+ },
1186
+ {
1187
+ "epoch": 2.56,
1188
+ "learning_rate": 8.223523670082968e-06,
1189
+ "loss": 0.0381,
1190
+ "step": 1940
1191
+ },
1192
+ {
1193
+ "epoch": 2.57,
1194
+ "learning_rate": 7.979502196193265e-06,
1195
+ "loss": 0.0589,
1196
+ "step": 1950
1197
+ },
1198
+ {
1199
+ "epoch": 2.58,
1200
+ "learning_rate": 7.735480722303564e-06,
1201
+ "loss": 0.3975,
1202
+ "step": 1960
1203
+ },
1204
+ {
1205
+ "epoch": 2.6,
1206
+ "learning_rate": 7.491459248413861e-06,
1207
+ "loss": 0.0307,
1208
+ "step": 1970
1209
+ },
1210
+ {
1211
+ "epoch": 2.61,
1212
+ "learning_rate": 7.247437774524159e-06,
1213
+ "loss": 0.046,
1214
+ "step": 1980
1215
+ },
1216
+ {
1217
+ "epoch": 2.62,
1218
+ "learning_rate": 7.0034163006344565e-06,
1219
+ "loss": 0.026,
1220
+ "step": 1990
1221
+ },
1222
+ {
1223
+ "epoch": 2.64,
1224
+ "learning_rate": 6.759394826744754e-06,
1225
+ "loss": 0.0749,
1226
+ "step": 2000
1227
+ },
1228
+ {
1229
+ "epoch": 2.65,
1230
+ "learning_rate": 6.515373352855052e-06,
1231
+ "loss": 0.0306,
1232
+ "step": 2010
1233
+ },
1234
+ {
1235
+ "epoch": 2.66,
1236
+ "learning_rate": 6.27135187896535e-06,
1237
+ "loss": 0.0536,
1238
+ "step": 2020
1239
+ },
1240
+ {
1241
+ "epoch": 2.67,
1242
+ "learning_rate": 6.027330405075647e-06,
1243
+ "loss": 0.0265,
1244
+ "step": 2030
1245
+ },
1246
+ {
1247
+ "epoch": 2.69,
1248
+ "learning_rate": 5.7833089311859446e-06,
1249
+ "loss": 0.0329,
1250
+ "step": 2040
1251
+ },
1252
+ {
1253
+ "epoch": 2.7,
1254
+ "learning_rate": 5.539287457296242e-06,
1255
+ "loss": 0.5737,
1256
+ "step": 2050
1257
+ },
1258
+ {
1259
+ "epoch": 2.71,
1260
+ "learning_rate": 5.29526598340654e-06,
1261
+ "loss": 0.8366,
1262
+ "step": 2060
1263
+ },
1264
+ {
1265
+ "epoch": 2.73,
1266
+ "learning_rate": 5.051244509516838e-06,
1267
+ "loss": 0.0282,
1268
+ "step": 2070
1269
+ },
1270
+ {
1271
+ "epoch": 2.74,
1272
+ "learning_rate": 4.807223035627136e-06,
1273
+ "loss": 0.0262,
1274
+ "step": 2080
1275
+ },
1276
+ {
1277
+ "epoch": 2.75,
1278
+ "learning_rate": 4.5632015617374335e-06,
1279
+ "loss": 0.1776,
1280
+ "step": 2090
1281
+ },
1282
+ {
1283
+ "epoch": 2.77,
1284
+ "learning_rate": 4.319180087847731e-06,
1285
+ "loss": 0.024,
1286
+ "step": 2100
1287
+ },
1288
+ {
1289
+ "epoch": 2.78,
1290
+ "learning_rate": 4.075158613958028e-06,
1291
+ "loss": 0.1574,
1292
+ "step": 2110
1293
+ },
1294
+ {
1295
+ "epoch": 2.79,
1296
+ "learning_rate": 3.831137140068326e-06,
1297
+ "loss": 0.1553,
1298
+ "step": 2120
1299
+ },
1300
+ {
1301
+ "epoch": 2.81,
1302
+ "learning_rate": 3.587115666178624e-06,
1303
+ "loss": 0.1287,
1304
+ "step": 2130
1305
+ },
1306
+ {
1307
+ "epoch": 2.82,
1308
+ "learning_rate": 3.3430941922889216e-06,
1309
+ "loss": 0.2801,
1310
+ "step": 2140
1311
+ },
1312
+ {
1313
+ "epoch": 2.83,
1314
+ "learning_rate": 3.0990727183992194e-06,
1315
+ "loss": 0.0326,
1316
+ "step": 2150
1317
+ },
1318
+ {
1319
+ "epoch": 2.85,
1320
+ "learning_rate": 2.855051244509517e-06,
1321
+ "loss": 0.2817,
1322
+ "step": 2160
1323
+ },
1324
+ {
1325
+ "epoch": 2.86,
1326
+ "learning_rate": 2.611029770619815e-06,
1327
+ "loss": 0.0579,
1328
+ "step": 2170
1329
+ },
1330
+ {
1331
+ "epoch": 2.87,
1332
+ "learning_rate": 2.3670082967301123e-06,
1333
+ "loss": 0.0243,
1334
+ "step": 2180
1335
+ },
1336
+ {
1337
+ "epoch": 2.89,
1338
+ "learning_rate": 2.12298682284041e-06,
1339
+ "loss": 0.0426,
1340
+ "step": 2190
1341
+ },
1342
+ {
1343
+ "epoch": 2.9,
1344
+ "learning_rate": 1.8789653489507079e-06,
1345
+ "loss": 0.1296,
1346
+ "step": 2200
1347
+ },
1348
+ {
1349
+ "epoch": 2.91,
1350
+ "learning_rate": 1.6349438750610055e-06,
1351
+ "loss": 0.5111,
1352
+ "step": 2210
1353
+ },
1354
+ {
1355
+ "epoch": 2.92,
1356
+ "learning_rate": 1.3909224011713032e-06,
1357
+ "loss": 0.1463,
1358
+ "step": 2220
1359
+ },
1360
+ {
1361
+ "epoch": 2.94,
1362
+ "learning_rate": 1.1469009272816008e-06,
1363
+ "loss": 0.05,
1364
+ "step": 2230
1365
+ },
1366
+ {
1367
+ "epoch": 2.95,
1368
+ "learning_rate": 9.028794533918985e-07,
1369
+ "loss": 0.0225,
1370
+ "step": 2240
1371
+ },
1372
+ {
1373
+ "epoch": 2.96,
1374
+ "learning_rate": 6.588579795021963e-07,
1375
+ "loss": 0.0311,
1376
+ "step": 2250
1377
+ },
1378
+ {
1379
+ "epoch": 2.98,
1380
+ "learning_rate": 4.148365056124939e-07,
1381
+ "loss": 0.2809,
1382
+ "step": 2260
1383
+ },
1384
+ {
1385
+ "epoch": 2.99,
1386
+ "learning_rate": 1.7081503172279163e-07,
1387
+ "loss": 0.0345,
1388
+ "step": 2270
1389
+ },
1390
+ {
1391
+ "epoch": 3.0,
1392
+ "eval_accuracy": 0.9844444444444445,
1393
+ "eval_loss": 0.07232666015625,
1394
+ "eval_runtime": 9.9115,
1395
+ "eval_samples_per_second": 272.41,
1396
+ "eval_steps_per_second": 68.102,
1397
+ "step": 2277
1398
+ },
1399
+ {
1400
+ "epoch": 3.0,
1401
+ "step": 2277,
1402
+ "total_flos": 1.8069699791226824e+20,
1403
+ "train_loss": 0.3087928044884991,
1404
+ "train_runtime": 984.02,
1405
+ "train_samples_per_second": 74.047,
1406
+ "train_steps_per_second": 2.314
1407
+ }
1408
+ ],
1409
+ "max_steps": 2277,
1410
+ "num_train_epochs": 3,
1411
+ "start_time": 1668807038.3244002,
1412
+ "total_flos": 1.8069699791226824e+20,
1413
+ "trial_name": null,
1414
+ "trial_params": null
1415
+ }