amey6056 commited on
Commit
37db72f
·
1 Parent(s): b2f5418

amey6056/PlantC

Browse files
Files changed (4) hide show
  1. all_results.json +13 -0
  2. eval_results.json +8 -0
  3. train_results.json +8 -0
  4. trainer_state.json +1717 -0
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.9648304120191213,
4
+ "eval_loss": 0.13630270957946777,
5
+ "eval_runtime": 102.7768,
6
+ "eval_samples_per_second": 170.972,
7
+ "eval_steps_per_second": 5.351,
8
+ "total_flos": 7.483912529986142e+18,
9
+ "train_loss": 1.0410587748977438,
10
+ "train_runtime": 5056.7157,
11
+ "train_samples_per_second": 69.507,
12
+ "train_steps_per_second": 0.543
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "eval_accuracy": 0.9648304120191213,
4
+ "eval_loss": 0.13630270957946777,
5
+ "eval_runtime": 102.7768,
6
+ "eval_samples_per_second": 170.972,
7
+ "eval_steps_per_second": 5.351
8
+ }
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 5.0,
3
+ "total_flos": 7.483912529986142e+18,
4
+ "train_loss": 1.0410587748977438,
5
+ "train_runtime": 5056.7157,
6
+ "train_samples_per_second": 69.507,
7
+ "train_steps_per_second": 0.543
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1717 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9648304120191213,
3
+ "best_model_checkpoint": "resnet-50-finetuned-eurosat/checkpoint-2745",
4
+ "epoch": 4.997724169321803,
5
+ "eval_steps": 500,
6
+ "global_step": 2745,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.02,
13
+ "learning_rate": 1.818181818181818e-06,
14
+ "loss": 3.6402,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.04,
19
+ "learning_rate": 3.636363636363636e-06,
20
+ "loss": 3.6399,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.05,
25
+ "learning_rate": 5.4545454545454545e-06,
26
+ "loss": 3.6392,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.07,
31
+ "learning_rate": 7.272727272727272e-06,
32
+ "loss": 3.637,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.09,
37
+ "learning_rate": 9.090909090909091e-06,
38
+ "loss": 3.6383,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.11,
43
+ "learning_rate": 1.0909090909090909e-05,
44
+ "loss": 3.6362,
45
+ "step": 60
46
+ },
47
+ {
48
+ "epoch": 0.13,
49
+ "learning_rate": 1.2727272727272727e-05,
50
+ "loss": 3.6312,
51
+ "step": 70
52
+ },
53
+ {
54
+ "epoch": 0.15,
55
+ "learning_rate": 1.4545454545454545e-05,
56
+ "loss": 3.6275,
57
+ "step": 80
58
+ },
59
+ {
60
+ "epoch": 0.16,
61
+ "learning_rate": 1.6363636363636366e-05,
62
+ "loss": 3.6237,
63
+ "step": 90
64
+ },
65
+ {
66
+ "epoch": 0.18,
67
+ "learning_rate": 1.8181818181818182e-05,
68
+ "loss": 3.6269,
69
+ "step": 100
70
+ },
71
+ {
72
+ "epoch": 0.2,
73
+ "learning_rate": 2e-05,
74
+ "loss": 3.6201,
75
+ "step": 110
76
+ },
77
+ {
78
+ "epoch": 0.22,
79
+ "learning_rate": 2.1818181818181818e-05,
80
+ "loss": 3.6147,
81
+ "step": 120
82
+ },
83
+ {
84
+ "epoch": 0.24,
85
+ "learning_rate": 2.3636363636363637e-05,
86
+ "loss": 3.6125,
87
+ "step": 130
88
+ },
89
+ {
90
+ "epoch": 0.25,
91
+ "learning_rate": 2.5454545454545454e-05,
92
+ "loss": 3.6079,
93
+ "step": 140
94
+ },
95
+ {
96
+ "epoch": 0.27,
97
+ "learning_rate": 2.7272727272727273e-05,
98
+ "loss": 3.6001,
99
+ "step": 150
100
+ },
101
+ {
102
+ "epoch": 0.29,
103
+ "learning_rate": 2.909090909090909e-05,
104
+ "loss": 3.5919,
105
+ "step": 160
106
+ },
107
+ {
108
+ "epoch": 0.31,
109
+ "learning_rate": 3.090909090909091e-05,
110
+ "loss": 3.5876,
111
+ "step": 170
112
+ },
113
+ {
114
+ "epoch": 0.33,
115
+ "learning_rate": 3.272727272727273e-05,
116
+ "loss": 3.5793,
117
+ "step": 180
118
+ },
119
+ {
120
+ "epoch": 0.35,
121
+ "learning_rate": 3.454545454545455e-05,
122
+ "loss": 3.5733,
123
+ "step": 190
124
+ },
125
+ {
126
+ "epoch": 0.36,
127
+ "learning_rate": 3.6363636363636364e-05,
128
+ "loss": 3.5642,
129
+ "step": 200
130
+ },
131
+ {
132
+ "epoch": 0.38,
133
+ "learning_rate": 3.818181818181819e-05,
134
+ "loss": 3.556,
135
+ "step": 210
136
+ },
137
+ {
138
+ "epoch": 0.4,
139
+ "learning_rate": 4e-05,
140
+ "loss": 3.5382,
141
+ "step": 220
142
+ },
143
+ {
144
+ "epoch": 0.42,
145
+ "learning_rate": 4.181818181818182e-05,
146
+ "loss": 3.5308,
147
+ "step": 230
148
+ },
149
+ {
150
+ "epoch": 0.44,
151
+ "learning_rate": 4.3636363636363636e-05,
152
+ "loss": 3.5158,
153
+ "step": 240
154
+ },
155
+ {
156
+ "epoch": 0.46,
157
+ "learning_rate": 4.545454545454546e-05,
158
+ "loss": 3.4985,
159
+ "step": 250
160
+ },
161
+ {
162
+ "epoch": 0.47,
163
+ "learning_rate": 4.7272727272727275e-05,
164
+ "loss": 3.4826,
165
+ "step": 260
166
+ },
167
+ {
168
+ "epoch": 0.49,
169
+ "learning_rate": 4.909090909090909e-05,
170
+ "loss": 3.4614,
171
+ "step": 270
172
+ },
173
+ {
174
+ "epoch": 0.51,
175
+ "learning_rate": 4.989878542510122e-05,
176
+ "loss": 3.4349,
177
+ "step": 280
178
+ },
179
+ {
180
+ "epoch": 0.53,
181
+ "learning_rate": 4.9696356275303645e-05,
182
+ "loss": 3.4079,
183
+ "step": 290
184
+ },
185
+ {
186
+ "epoch": 0.55,
187
+ "learning_rate": 4.9493927125506076e-05,
188
+ "loss": 3.3765,
189
+ "step": 300
190
+ },
191
+ {
192
+ "epoch": 0.56,
193
+ "learning_rate": 4.9291497975708506e-05,
194
+ "loss": 3.3423,
195
+ "step": 310
196
+ },
197
+ {
198
+ "epoch": 0.58,
199
+ "learning_rate": 4.908906882591093e-05,
200
+ "loss": 3.2997,
201
+ "step": 320
202
+ },
203
+ {
204
+ "epoch": 0.6,
205
+ "learning_rate": 4.888663967611336e-05,
206
+ "loss": 3.2531,
207
+ "step": 330
208
+ },
209
+ {
210
+ "epoch": 0.62,
211
+ "learning_rate": 4.868421052631579e-05,
212
+ "loss": 3.2099,
213
+ "step": 340
214
+ },
215
+ {
216
+ "epoch": 0.64,
217
+ "learning_rate": 4.848178137651822e-05,
218
+ "loss": 3.1511,
219
+ "step": 350
220
+ },
221
+ {
222
+ "epoch": 0.66,
223
+ "learning_rate": 4.8279352226720646e-05,
224
+ "loss": 3.1053,
225
+ "step": 360
226
+ },
227
+ {
228
+ "epoch": 0.67,
229
+ "learning_rate": 4.8076923076923084e-05,
230
+ "loss": 3.03,
231
+ "step": 370
232
+ },
233
+ {
234
+ "epoch": 0.69,
235
+ "learning_rate": 4.787449392712551e-05,
236
+ "loss": 2.9668,
237
+ "step": 380
238
+ },
239
+ {
240
+ "epoch": 0.71,
241
+ "learning_rate": 4.767206477732794e-05,
242
+ "loss": 2.8766,
243
+ "step": 390
244
+ },
245
+ {
246
+ "epoch": 0.73,
247
+ "learning_rate": 4.746963562753037e-05,
248
+ "loss": 2.7964,
249
+ "step": 400
250
+ },
251
+ {
252
+ "epoch": 0.75,
253
+ "learning_rate": 4.726720647773279e-05,
254
+ "loss": 2.7181,
255
+ "step": 410
256
+ },
257
+ {
258
+ "epoch": 0.76,
259
+ "learning_rate": 4.7064777327935223e-05,
260
+ "loss": 2.6812,
261
+ "step": 420
262
+ },
263
+ {
264
+ "epoch": 0.78,
265
+ "learning_rate": 4.6862348178137654e-05,
266
+ "loss": 2.5486,
267
+ "step": 430
268
+ },
269
+ {
270
+ "epoch": 0.8,
271
+ "learning_rate": 4.6659919028340085e-05,
272
+ "loss": 2.4858,
273
+ "step": 440
274
+ },
275
+ {
276
+ "epoch": 0.82,
277
+ "learning_rate": 4.6457489878542516e-05,
278
+ "loss": 2.3997,
279
+ "step": 450
280
+ },
281
+ {
282
+ "epoch": 0.84,
283
+ "learning_rate": 4.625506072874494e-05,
284
+ "loss": 2.3232,
285
+ "step": 460
286
+ },
287
+ {
288
+ "epoch": 0.86,
289
+ "learning_rate": 4.605263157894737e-05,
290
+ "loss": 2.2581,
291
+ "step": 470
292
+ },
293
+ {
294
+ "epoch": 0.87,
295
+ "learning_rate": 4.58502024291498e-05,
296
+ "loss": 2.1588,
297
+ "step": 480
298
+ },
299
+ {
300
+ "epoch": 0.89,
301
+ "learning_rate": 4.564777327935223e-05,
302
+ "loss": 2.1098,
303
+ "step": 490
304
+ },
305
+ {
306
+ "epoch": 0.91,
307
+ "learning_rate": 4.5445344129554655e-05,
308
+ "loss": 2.0018,
309
+ "step": 500
310
+ },
311
+ {
312
+ "epoch": 0.93,
313
+ "learning_rate": 4.524291497975709e-05,
314
+ "loss": 1.9727,
315
+ "step": 510
316
+ },
317
+ {
318
+ "epoch": 0.95,
319
+ "learning_rate": 4.504048582995952e-05,
320
+ "loss": 1.8964,
321
+ "step": 520
322
+ },
323
+ {
324
+ "epoch": 0.96,
325
+ "learning_rate": 4.483805668016194e-05,
326
+ "loss": 1.7941,
327
+ "step": 530
328
+ },
329
+ {
330
+ "epoch": 0.98,
331
+ "learning_rate": 4.463562753036438e-05,
332
+ "loss": 1.7502,
333
+ "step": 540
334
+ },
335
+ {
336
+ "epoch": 1.0,
337
+ "eval_accuracy": 0.795356248577282,
338
+ "eval_loss": 1.328504204750061,
339
+ "eval_runtime": 106.9422,
340
+ "eval_samples_per_second": 164.313,
341
+ "eval_steps_per_second": 5.143,
342
+ "step": 549
343
+ },
344
+ {
345
+ "epoch": 1.0,
346
+ "learning_rate": 4.44331983805668e-05,
347
+ "loss": 1.7101,
348
+ "step": 550
349
+ },
350
+ {
351
+ "epoch": 1.02,
352
+ "learning_rate": 4.423076923076923e-05,
353
+ "loss": 1.6815,
354
+ "step": 560
355
+ },
356
+ {
357
+ "epoch": 1.04,
358
+ "learning_rate": 4.402834008097166e-05,
359
+ "loss": 1.5733,
360
+ "step": 570
361
+ },
362
+ {
363
+ "epoch": 1.06,
364
+ "learning_rate": 4.3825910931174094e-05,
365
+ "loss": 1.5261,
366
+ "step": 580
367
+ },
368
+ {
369
+ "epoch": 1.07,
370
+ "learning_rate": 4.362348178137652e-05,
371
+ "loss": 1.479,
372
+ "step": 590
373
+ },
374
+ {
375
+ "epoch": 1.09,
376
+ "learning_rate": 4.342105263157895e-05,
377
+ "loss": 1.4515,
378
+ "step": 600
379
+ },
380
+ {
381
+ "epoch": 1.11,
382
+ "learning_rate": 4.321862348178138e-05,
383
+ "loss": 1.3921,
384
+ "step": 610
385
+ },
386
+ {
387
+ "epoch": 1.13,
388
+ "learning_rate": 4.30161943319838e-05,
389
+ "loss": 1.3221,
390
+ "step": 620
391
+ },
392
+ {
393
+ "epoch": 1.15,
394
+ "learning_rate": 4.281376518218624e-05,
395
+ "loss": 1.332,
396
+ "step": 630
397
+ },
398
+ {
399
+ "epoch": 1.17,
400
+ "learning_rate": 4.2611336032388664e-05,
401
+ "loss": 1.2475,
402
+ "step": 640
403
+ },
404
+ {
405
+ "epoch": 1.18,
406
+ "learning_rate": 4.2408906882591095e-05,
407
+ "loss": 1.2658,
408
+ "step": 650
409
+ },
410
+ {
411
+ "epoch": 1.2,
412
+ "learning_rate": 4.2206477732793526e-05,
413
+ "loss": 1.2328,
414
+ "step": 660
415
+ },
416
+ {
417
+ "epoch": 1.22,
418
+ "learning_rate": 4.200404858299595e-05,
419
+ "loss": 1.1655,
420
+ "step": 670
421
+ },
422
+ {
423
+ "epoch": 1.24,
424
+ "learning_rate": 4.180161943319838e-05,
425
+ "loss": 1.1272,
426
+ "step": 680
427
+ },
428
+ {
429
+ "epoch": 1.26,
430
+ "learning_rate": 4.159919028340081e-05,
431
+ "loss": 1.0837,
432
+ "step": 690
433
+ },
434
+ {
435
+ "epoch": 1.27,
436
+ "learning_rate": 4.139676113360324e-05,
437
+ "loss": 1.1115,
438
+ "step": 700
439
+ },
440
+ {
441
+ "epoch": 1.29,
442
+ "learning_rate": 4.1194331983805666e-05,
443
+ "loss": 1.078,
444
+ "step": 710
445
+ },
446
+ {
447
+ "epoch": 1.31,
448
+ "learning_rate": 4.09919028340081e-05,
449
+ "loss": 1.0328,
450
+ "step": 720
451
+ },
452
+ {
453
+ "epoch": 1.33,
454
+ "learning_rate": 4.078947368421053e-05,
455
+ "loss": 1.0057,
456
+ "step": 730
457
+ },
458
+ {
459
+ "epoch": 1.35,
460
+ "learning_rate": 4.058704453441296e-05,
461
+ "loss": 0.9995,
462
+ "step": 740
463
+ },
464
+ {
465
+ "epoch": 1.37,
466
+ "learning_rate": 4.038461538461539e-05,
467
+ "loss": 0.975,
468
+ "step": 750
469
+ },
470
+ {
471
+ "epoch": 1.38,
472
+ "learning_rate": 4.018218623481781e-05,
473
+ "loss": 0.9552,
474
+ "step": 760
475
+ },
476
+ {
477
+ "epoch": 1.4,
478
+ "learning_rate": 3.997975708502025e-05,
479
+ "loss": 0.9118,
480
+ "step": 770
481
+ },
482
+ {
483
+ "epoch": 1.42,
484
+ "learning_rate": 3.9777327935222674e-05,
485
+ "loss": 0.9089,
486
+ "step": 780
487
+ },
488
+ {
489
+ "epoch": 1.44,
490
+ "learning_rate": 3.9574898785425104e-05,
491
+ "loss": 0.8996,
492
+ "step": 790
493
+ },
494
+ {
495
+ "epoch": 1.46,
496
+ "learning_rate": 3.9372469635627535e-05,
497
+ "loss": 0.8773,
498
+ "step": 800
499
+ },
500
+ {
501
+ "epoch": 1.47,
502
+ "learning_rate": 3.917004048582996e-05,
503
+ "loss": 0.8736,
504
+ "step": 810
505
+ },
506
+ {
507
+ "epoch": 1.49,
508
+ "learning_rate": 3.896761133603239e-05,
509
+ "loss": 0.8889,
510
+ "step": 820
511
+ },
512
+ {
513
+ "epoch": 1.51,
514
+ "learning_rate": 3.876518218623482e-05,
515
+ "loss": 0.8332,
516
+ "step": 830
517
+ },
518
+ {
519
+ "epoch": 1.53,
520
+ "learning_rate": 3.856275303643725e-05,
521
+ "loss": 0.7772,
522
+ "step": 840
523
+ },
524
+ {
525
+ "epoch": 1.55,
526
+ "learning_rate": 3.8360323886639675e-05,
527
+ "loss": 0.7924,
528
+ "step": 850
529
+ },
530
+ {
531
+ "epoch": 1.57,
532
+ "learning_rate": 3.815789473684211e-05,
533
+ "loss": 0.8023,
534
+ "step": 860
535
+ },
536
+ {
537
+ "epoch": 1.58,
538
+ "learning_rate": 3.7955465587044536e-05,
539
+ "loss": 0.7885,
540
+ "step": 870
541
+ },
542
+ {
543
+ "epoch": 1.6,
544
+ "learning_rate": 3.775303643724697e-05,
545
+ "loss": 0.753,
546
+ "step": 880
547
+ },
548
+ {
549
+ "epoch": 1.62,
550
+ "learning_rate": 3.75506072874494e-05,
551
+ "loss": 0.7889,
552
+ "step": 890
553
+ },
554
+ {
555
+ "epoch": 1.64,
556
+ "learning_rate": 3.734817813765182e-05,
557
+ "loss": 0.7486,
558
+ "step": 900
559
+ },
560
+ {
561
+ "epoch": 1.66,
562
+ "learning_rate": 3.714574898785425e-05,
563
+ "loss": 0.7106,
564
+ "step": 910
565
+ },
566
+ {
567
+ "epoch": 1.68,
568
+ "learning_rate": 3.694331983805668e-05,
569
+ "loss": 0.7279,
570
+ "step": 920
571
+ },
572
+ {
573
+ "epoch": 1.69,
574
+ "learning_rate": 3.6740890688259113e-05,
575
+ "loss": 0.727,
576
+ "step": 930
577
+ },
578
+ {
579
+ "epoch": 1.71,
580
+ "learning_rate": 3.653846153846154e-05,
581
+ "loss": 0.7469,
582
+ "step": 940
583
+ },
584
+ {
585
+ "epoch": 1.73,
586
+ "learning_rate": 3.633603238866397e-05,
587
+ "loss": 0.6788,
588
+ "step": 950
589
+ },
590
+ {
591
+ "epoch": 1.75,
592
+ "learning_rate": 3.61336032388664e-05,
593
+ "loss": 0.6962,
594
+ "step": 960
595
+ },
596
+ {
597
+ "epoch": 1.77,
598
+ "learning_rate": 3.593117408906882e-05,
599
+ "loss": 0.6872,
600
+ "step": 970
601
+ },
602
+ {
603
+ "epoch": 1.78,
604
+ "learning_rate": 3.572874493927126e-05,
605
+ "loss": 0.6464,
606
+ "step": 980
607
+ },
608
+ {
609
+ "epoch": 1.8,
610
+ "learning_rate": 3.5526315789473684e-05,
611
+ "loss": 0.6389,
612
+ "step": 990
613
+ },
614
+ {
615
+ "epoch": 1.82,
616
+ "learning_rate": 3.5323886639676115e-05,
617
+ "loss": 0.6092,
618
+ "step": 1000
619
+ },
620
+ {
621
+ "epoch": 1.84,
622
+ "learning_rate": 3.5121457489878545e-05,
623
+ "loss": 0.6159,
624
+ "step": 1010
625
+ },
626
+ {
627
+ "epoch": 1.86,
628
+ "learning_rate": 3.491902834008097e-05,
629
+ "loss": 0.6301,
630
+ "step": 1020
631
+ },
632
+ {
633
+ "epoch": 1.88,
634
+ "learning_rate": 3.47165991902834e-05,
635
+ "loss": 0.6138,
636
+ "step": 1030
637
+ },
638
+ {
639
+ "epoch": 1.89,
640
+ "learning_rate": 3.451417004048583e-05,
641
+ "loss": 0.6002,
642
+ "step": 1040
643
+ },
644
+ {
645
+ "epoch": 1.91,
646
+ "learning_rate": 3.431174089068826e-05,
647
+ "loss": 0.6109,
648
+ "step": 1050
649
+ },
650
+ {
651
+ "epoch": 1.93,
652
+ "learning_rate": 3.410931174089069e-05,
653
+ "loss": 0.6136,
654
+ "step": 1060
655
+ },
656
+ {
657
+ "epoch": 1.95,
658
+ "learning_rate": 3.390688259109312e-05,
659
+ "loss": 0.5631,
660
+ "step": 1070
661
+ },
662
+ {
663
+ "epoch": 1.97,
664
+ "learning_rate": 3.3704453441295546e-05,
665
+ "loss": 0.6395,
666
+ "step": 1080
667
+ },
668
+ {
669
+ "epoch": 1.98,
670
+ "learning_rate": 3.350202429149798e-05,
671
+ "loss": 0.6126,
672
+ "step": 1090
673
+ },
674
+ {
675
+ "epoch": 2.0,
676
+ "eval_accuracy": 0.9301160937855679,
677
+ "eval_loss": 0.32050490379333496,
678
+ "eval_runtime": 103.9225,
679
+ "eval_samples_per_second": 169.088,
680
+ "eval_steps_per_second": 5.292,
681
+ "step": 1098
682
+ },
683
+ {
684
+ "epoch": 2.0,
685
+ "learning_rate": 3.329959514170041e-05,
686
+ "loss": 0.5663,
687
+ "step": 1100
688
+ },
689
+ {
690
+ "epoch": 2.02,
691
+ "learning_rate": 3.309716599190283e-05,
692
+ "loss": 0.5179,
693
+ "step": 1110
694
+ },
695
+ {
696
+ "epoch": 2.04,
697
+ "learning_rate": 3.289473684210527e-05,
698
+ "loss": 0.5657,
699
+ "step": 1120
700
+ },
701
+ {
702
+ "epoch": 2.06,
703
+ "learning_rate": 3.269230769230769e-05,
704
+ "loss": 0.5623,
705
+ "step": 1130
706
+ },
707
+ {
708
+ "epoch": 2.08,
709
+ "learning_rate": 3.2489878542510124e-05,
710
+ "loss": 0.581,
711
+ "step": 1140
712
+ },
713
+ {
714
+ "epoch": 2.09,
715
+ "learning_rate": 3.2287449392712554e-05,
716
+ "loss": 0.5664,
717
+ "step": 1150
718
+ },
719
+ {
720
+ "epoch": 2.11,
721
+ "learning_rate": 3.208502024291498e-05,
722
+ "loss": 0.5485,
723
+ "step": 1160
724
+ },
725
+ {
726
+ "epoch": 2.13,
727
+ "learning_rate": 3.188259109311741e-05,
728
+ "loss": 0.5518,
729
+ "step": 1170
730
+ },
731
+ {
732
+ "epoch": 2.15,
733
+ "learning_rate": 3.168016194331984e-05,
734
+ "loss": 0.5452,
735
+ "step": 1180
736
+ },
737
+ {
738
+ "epoch": 2.17,
739
+ "learning_rate": 3.147773279352227e-05,
740
+ "loss": 0.5394,
741
+ "step": 1190
742
+ },
743
+ {
744
+ "epoch": 2.18,
745
+ "learning_rate": 3.1275303643724694e-05,
746
+ "loss": 0.5483,
747
+ "step": 1200
748
+ },
749
+ {
750
+ "epoch": 2.2,
751
+ "learning_rate": 3.107287449392713e-05,
752
+ "loss": 0.5503,
753
+ "step": 1210
754
+ },
755
+ {
756
+ "epoch": 2.22,
757
+ "learning_rate": 3.0870445344129556e-05,
758
+ "loss": 0.5164,
759
+ "step": 1220
760
+ },
761
+ {
762
+ "epoch": 2.24,
763
+ "learning_rate": 3.0668016194331986e-05,
764
+ "loss": 0.4868,
765
+ "step": 1230
766
+ },
767
+ {
768
+ "epoch": 2.26,
769
+ "learning_rate": 3.0465587044534417e-05,
770
+ "loss": 0.4483,
771
+ "step": 1240
772
+ },
773
+ {
774
+ "epoch": 2.28,
775
+ "learning_rate": 3.0263157894736844e-05,
776
+ "loss": 0.5105,
777
+ "step": 1250
778
+ },
779
+ {
780
+ "epoch": 2.29,
781
+ "learning_rate": 3.006072874493927e-05,
782
+ "loss": 0.4933,
783
+ "step": 1260
784
+ },
785
+ {
786
+ "epoch": 2.31,
787
+ "learning_rate": 2.9858299595141702e-05,
788
+ "loss": 0.4973,
789
+ "step": 1270
790
+ },
791
+ {
792
+ "epoch": 2.33,
793
+ "learning_rate": 2.965587044534413e-05,
794
+ "loss": 0.4855,
795
+ "step": 1280
796
+ },
797
+ {
798
+ "epoch": 2.35,
799
+ "learning_rate": 2.9453441295546557e-05,
800
+ "loss": 0.4828,
801
+ "step": 1290
802
+ },
803
+ {
804
+ "epoch": 2.37,
805
+ "learning_rate": 2.925101214574899e-05,
806
+ "loss": 0.4806,
807
+ "step": 1300
808
+ },
809
+ {
810
+ "epoch": 2.39,
811
+ "learning_rate": 2.9048582995951418e-05,
812
+ "loss": 0.5003,
813
+ "step": 1310
814
+ },
815
+ {
816
+ "epoch": 2.4,
817
+ "learning_rate": 2.8846153846153845e-05,
818
+ "loss": 0.5143,
819
+ "step": 1320
820
+ },
821
+ {
822
+ "epoch": 2.42,
823
+ "learning_rate": 2.8643724696356276e-05,
824
+ "loss": 0.4758,
825
+ "step": 1330
826
+ },
827
+ {
828
+ "epoch": 2.44,
829
+ "learning_rate": 2.8441295546558703e-05,
830
+ "loss": 0.4782,
831
+ "step": 1340
832
+ },
833
+ {
834
+ "epoch": 2.46,
835
+ "learning_rate": 2.823886639676113e-05,
836
+ "loss": 0.4825,
837
+ "step": 1350
838
+ },
839
+ {
840
+ "epoch": 2.48,
841
+ "learning_rate": 2.8036437246963565e-05,
842
+ "loss": 0.4637,
843
+ "step": 1360
844
+ },
845
+ {
846
+ "epoch": 2.49,
847
+ "learning_rate": 2.7834008097165992e-05,
848
+ "loss": 0.4574,
849
+ "step": 1370
850
+ },
851
+ {
852
+ "epoch": 2.51,
853
+ "learning_rate": 2.7631578947368426e-05,
854
+ "loss": 0.443,
855
+ "step": 1380
856
+ },
857
+ {
858
+ "epoch": 2.53,
859
+ "learning_rate": 2.7429149797570853e-05,
860
+ "loss": 0.4742,
861
+ "step": 1390
862
+ },
863
+ {
864
+ "epoch": 2.55,
865
+ "learning_rate": 2.722672064777328e-05,
866
+ "loss": 0.4719,
867
+ "step": 1400
868
+ },
869
+ {
870
+ "epoch": 2.57,
871
+ "learning_rate": 2.702429149797571e-05,
872
+ "loss": 0.4539,
873
+ "step": 1410
874
+ },
875
+ {
876
+ "epoch": 2.59,
877
+ "learning_rate": 2.682186234817814e-05,
878
+ "loss": 0.4393,
879
+ "step": 1420
880
+ },
881
+ {
882
+ "epoch": 2.6,
883
+ "learning_rate": 2.6619433198380566e-05,
884
+ "loss": 0.4418,
885
+ "step": 1430
886
+ },
887
+ {
888
+ "epoch": 2.62,
889
+ "learning_rate": 2.6417004048583e-05,
890
+ "loss": 0.4801,
891
+ "step": 1440
892
+ },
893
+ {
894
+ "epoch": 2.64,
895
+ "learning_rate": 2.6214574898785427e-05,
896
+ "loss": 0.4524,
897
+ "step": 1450
898
+ },
899
+ {
900
+ "epoch": 2.66,
901
+ "learning_rate": 2.6012145748987855e-05,
902
+ "loss": 0.4154,
903
+ "step": 1460
904
+ },
905
+ {
906
+ "epoch": 2.68,
907
+ "learning_rate": 2.5809716599190285e-05,
908
+ "loss": 0.4374,
909
+ "step": 1470
910
+ },
911
+ {
912
+ "epoch": 2.69,
913
+ "learning_rate": 2.5607287449392713e-05,
914
+ "loss": 0.3937,
915
+ "step": 1480
916
+ },
917
+ {
918
+ "epoch": 2.71,
919
+ "learning_rate": 2.540485829959514e-05,
920
+ "loss": 0.4008,
921
+ "step": 1490
922
+ },
923
+ {
924
+ "epoch": 2.73,
925
+ "learning_rate": 2.5202429149797574e-05,
926
+ "loss": 0.4617,
927
+ "step": 1500
928
+ },
929
+ {
930
+ "epoch": 2.75,
931
+ "learning_rate": 2.5e-05,
932
+ "loss": 0.4357,
933
+ "step": 1510
934
+ },
935
+ {
936
+ "epoch": 2.77,
937
+ "learning_rate": 2.4797570850202432e-05,
938
+ "loss": 0.4234,
939
+ "step": 1520
940
+ },
941
+ {
942
+ "epoch": 2.79,
943
+ "learning_rate": 2.459514170040486e-05,
944
+ "loss": 0.4479,
945
+ "step": 1530
946
+ },
947
+ {
948
+ "epoch": 2.8,
949
+ "learning_rate": 2.439271255060729e-05,
950
+ "loss": 0.3867,
951
+ "step": 1540
952
+ },
953
+ {
954
+ "epoch": 2.82,
955
+ "learning_rate": 2.4190283400809717e-05,
956
+ "loss": 0.4096,
957
+ "step": 1550
958
+ },
959
+ {
960
+ "epoch": 2.84,
961
+ "learning_rate": 2.3987854251012144e-05,
962
+ "loss": 0.4076,
963
+ "step": 1560
964
+ },
965
+ {
966
+ "epoch": 2.86,
967
+ "learning_rate": 2.3785425101214575e-05,
968
+ "loss": 0.4417,
969
+ "step": 1570
970
+ },
971
+ {
972
+ "epoch": 2.88,
973
+ "learning_rate": 2.3582995951417006e-05,
974
+ "loss": 0.4439,
975
+ "step": 1580
976
+ },
977
+ {
978
+ "epoch": 2.89,
979
+ "learning_rate": 2.3380566801619436e-05,
980
+ "loss": 0.4154,
981
+ "step": 1590
982
+ },
983
+ {
984
+ "epoch": 2.91,
985
+ "learning_rate": 2.3178137651821864e-05,
986
+ "loss": 0.4498,
987
+ "step": 1600
988
+ },
989
+ {
990
+ "epoch": 2.93,
991
+ "learning_rate": 2.2975708502024294e-05,
992
+ "loss": 0.4068,
993
+ "step": 1610
994
+ },
995
+ {
996
+ "epoch": 2.95,
997
+ "learning_rate": 2.2773279352226722e-05,
998
+ "loss": 0.3884,
999
+ "step": 1620
1000
+ },
1001
+ {
1002
+ "epoch": 2.97,
1003
+ "learning_rate": 2.257085020242915e-05,
1004
+ "loss": 0.3952,
1005
+ "step": 1630
1006
+ },
1007
+ {
1008
+ "epoch": 2.99,
1009
+ "learning_rate": 2.236842105263158e-05,
1010
+ "loss": 0.4063,
1011
+ "step": 1640
1012
+ },
1013
+ {
1014
+ "epoch": 3.0,
1015
+ "eval_accuracy": 0.9550990211700433,
1016
+ "eval_loss": 0.1893182247877121,
1017
+ "eval_runtime": 104.3895,
1018
+ "eval_samples_per_second": 168.331,
1019
+ "eval_steps_per_second": 5.269,
1020
+ "step": 1647
1021
+ },
1022
+ {
1023
+ "epoch": 3.0,
1024
+ "learning_rate": 2.216599190283401e-05,
1025
+ "loss": 0.4239,
1026
+ "step": 1650
1027
+ },
1028
+ {
1029
+ "epoch": 3.02,
1030
+ "learning_rate": 2.1963562753036438e-05,
1031
+ "loss": 0.4083,
1032
+ "step": 1660
1033
+ },
1034
+ {
1035
+ "epoch": 3.04,
1036
+ "learning_rate": 2.1761133603238868e-05,
1037
+ "loss": 0.4241,
1038
+ "step": 1670
1039
+ },
1040
+ {
1041
+ "epoch": 3.06,
1042
+ "learning_rate": 2.15587044534413e-05,
1043
+ "loss": 0.407,
1044
+ "step": 1680
1045
+ },
1046
+ {
1047
+ "epoch": 3.08,
1048
+ "learning_rate": 2.1356275303643726e-05,
1049
+ "loss": 0.3937,
1050
+ "step": 1690
1051
+ },
1052
+ {
1053
+ "epoch": 3.1,
1054
+ "learning_rate": 2.1153846153846154e-05,
1055
+ "loss": 0.3753,
1056
+ "step": 1700
1057
+ },
1058
+ {
1059
+ "epoch": 3.11,
1060
+ "learning_rate": 2.0951417004048584e-05,
1061
+ "loss": 0.3942,
1062
+ "step": 1710
1063
+ },
1064
+ {
1065
+ "epoch": 3.13,
1066
+ "learning_rate": 2.074898785425101e-05,
1067
+ "loss": 0.4204,
1068
+ "step": 1720
1069
+ },
1070
+ {
1071
+ "epoch": 3.15,
1072
+ "learning_rate": 2.0546558704453442e-05,
1073
+ "loss": 0.4329,
1074
+ "step": 1730
1075
+ },
1076
+ {
1077
+ "epoch": 3.17,
1078
+ "learning_rate": 2.0344129554655873e-05,
1079
+ "loss": 0.3618,
1080
+ "step": 1740
1081
+ },
1082
+ {
1083
+ "epoch": 3.19,
1084
+ "learning_rate": 2.0141700404858304e-05,
1085
+ "loss": 0.3692,
1086
+ "step": 1750
1087
+ },
1088
+ {
1089
+ "epoch": 3.2,
1090
+ "learning_rate": 1.9939271255060727e-05,
1091
+ "loss": 0.3712,
1092
+ "step": 1760
1093
+ },
1094
+ {
1095
+ "epoch": 3.22,
1096
+ "learning_rate": 1.9736842105263158e-05,
1097
+ "loss": 0.419,
1098
+ "step": 1770
1099
+ },
1100
+ {
1101
+ "epoch": 3.24,
1102
+ "learning_rate": 1.953441295546559e-05,
1103
+ "loss": 0.3825,
1104
+ "step": 1780
1105
+ },
1106
+ {
1107
+ "epoch": 3.26,
1108
+ "learning_rate": 1.9331983805668016e-05,
1109
+ "loss": 0.3896,
1110
+ "step": 1790
1111
+ },
1112
+ {
1113
+ "epoch": 3.28,
1114
+ "learning_rate": 1.9129554655870447e-05,
1115
+ "loss": 0.3531,
1116
+ "step": 1800
1117
+ },
1118
+ {
1119
+ "epoch": 3.3,
1120
+ "learning_rate": 1.8927125506072877e-05,
1121
+ "loss": 0.3676,
1122
+ "step": 1810
1123
+ },
1124
+ {
1125
+ "epoch": 3.31,
1126
+ "learning_rate": 1.8724696356275305e-05,
1127
+ "loss": 0.3713,
1128
+ "step": 1820
1129
+ },
1130
+ {
1131
+ "epoch": 3.33,
1132
+ "learning_rate": 1.8522267206477732e-05,
1133
+ "loss": 0.3826,
1134
+ "step": 1830
1135
+ },
1136
+ {
1137
+ "epoch": 3.35,
1138
+ "learning_rate": 1.8319838056680163e-05,
1139
+ "loss": 0.3865,
1140
+ "step": 1840
1141
+ },
1142
+ {
1143
+ "epoch": 3.37,
1144
+ "learning_rate": 1.811740890688259e-05,
1145
+ "loss": 0.362,
1146
+ "step": 1850
1147
+ },
1148
+ {
1149
+ "epoch": 3.39,
1150
+ "learning_rate": 1.791497975708502e-05,
1151
+ "loss": 0.3906,
1152
+ "step": 1860
1153
+ },
1154
+ {
1155
+ "epoch": 3.4,
1156
+ "learning_rate": 1.771255060728745e-05,
1157
+ "loss": 0.3623,
1158
+ "step": 1870
1159
+ },
1160
+ {
1161
+ "epoch": 3.42,
1162
+ "learning_rate": 1.751012145748988e-05,
1163
+ "loss": 0.3954,
1164
+ "step": 1880
1165
+ },
1166
+ {
1167
+ "epoch": 3.44,
1168
+ "learning_rate": 1.730769230769231e-05,
1169
+ "loss": 0.373,
1170
+ "step": 1890
1171
+ },
1172
+ {
1173
+ "epoch": 3.46,
1174
+ "learning_rate": 1.7105263157894737e-05,
1175
+ "loss": 0.3897,
1176
+ "step": 1900
1177
+ },
1178
+ {
1179
+ "epoch": 3.48,
1180
+ "learning_rate": 1.6902834008097167e-05,
1181
+ "loss": 0.3627,
1182
+ "step": 1910
1183
+ },
1184
+ {
1185
+ "epoch": 3.5,
1186
+ "learning_rate": 1.6700404858299595e-05,
1187
+ "loss": 0.3843,
1188
+ "step": 1920
1189
+ },
1190
+ {
1191
+ "epoch": 3.51,
1192
+ "learning_rate": 1.6497975708502025e-05,
1193
+ "loss": 0.3501,
1194
+ "step": 1930
1195
+ },
1196
+ {
1197
+ "epoch": 3.53,
1198
+ "learning_rate": 1.6295546558704456e-05,
1199
+ "loss": 0.3681,
1200
+ "step": 1940
1201
+ },
1202
+ {
1203
+ "epoch": 3.55,
1204
+ "learning_rate": 1.6093117408906883e-05,
1205
+ "loss": 0.359,
1206
+ "step": 1950
1207
+ },
1208
+ {
1209
+ "epoch": 3.57,
1210
+ "learning_rate": 1.5890688259109314e-05,
1211
+ "loss": 0.3595,
1212
+ "step": 1960
1213
+ },
1214
+ {
1215
+ "epoch": 3.59,
1216
+ "learning_rate": 1.568825910931174e-05,
1217
+ "loss": 0.3684,
1218
+ "step": 1970
1219
+ },
1220
+ {
1221
+ "epoch": 3.6,
1222
+ "learning_rate": 1.548582995951417e-05,
1223
+ "loss": 0.377,
1224
+ "step": 1980
1225
+ },
1226
+ {
1227
+ "epoch": 3.62,
1228
+ "learning_rate": 1.52834008097166e-05,
1229
+ "loss": 0.3604,
1230
+ "step": 1990
1231
+ },
1232
+ {
1233
+ "epoch": 3.64,
1234
+ "learning_rate": 1.508097165991903e-05,
1235
+ "loss": 0.3432,
1236
+ "step": 2000
1237
+ },
1238
+ {
1239
+ "epoch": 3.66,
1240
+ "learning_rate": 1.4878542510121457e-05,
1241
+ "loss": 0.3641,
1242
+ "step": 2010
1243
+ },
1244
+ {
1245
+ "epoch": 3.68,
1246
+ "learning_rate": 1.4676113360323888e-05,
1247
+ "loss": 0.3588,
1248
+ "step": 2020
1249
+ },
1250
+ {
1251
+ "epoch": 3.7,
1252
+ "learning_rate": 1.4473684210526317e-05,
1253
+ "loss": 0.3579,
1254
+ "step": 2030
1255
+ },
1256
+ {
1257
+ "epoch": 3.71,
1258
+ "learning_rate": 1.4271255060728744e-05,
1259
+ "loss": 0.326,
1260
+ "step": 2040
1261
+ },
1262
+ {
1263
+ "epoch": 3.73,
1264
+ "learning_rate": 1.4068825910931175e-05,
1265
+ "loss": 0.3464,
1266
+ "step": 2050
1267
+ },
1268
+ {
1269
+ "epoch": 3.75,
1270
+ "learning_rate": 1.3866396761133604e-05,
1271
+ "loss": 0.3513,
1272
+ "step": 2060
1273
+ },
1274
+ {
1275
+ "epoch": 3.77,
1276
+ "learning_rate": 1.3663967611336034e-05,
1277
+ "loss": 0.4124,
1278
+ "step": 2070
1279
+ },
1280
+ {
1281
+ "epoch": 3.79,
1282
+ "learning_rate": 1.3461538461538462e-05,
1283
+ "loss": 0.3532,
1284
+ "step": 2080
1285
+ },
1286
+ {
1287
+ "epoch": 3.81,
1288
+ "learning_rate": 1.3259109311740892e-05,
1289
+ "loss": 0.3482,
1290
+ "step": 2090
1291
+ },
1292
+ {
1293
+ "epoch": 3.82,
1294
+ "learning_rate": 1.3056680161943321e-05,
1295
+ "loss": 0.3489,
1296
+ "step": 2100
1297
+ },
1298
+ {
1299
+ "epoch": 3.84,
1300
+ "learning_rate": 1.2854251012145749e-05,
1301
+ "loss": 0.3168,
1302
+ "step": 2110
1303
+ },
1304
+ {
1305
+ "epoch": 3.86,
1306
+ "learning_rate": 1.265182186234818e-05,
1307
+ "loss": 0.3575,
1308
+ "step": 2120
1309
+ },
1310
+ {
1311
+ "epoch": 3.88,
1312
+ "learning_rate": 1.2449392712550607e-05,
1313
+ "loss": 0.3506,
1314
+ "step": 2130
1315
+ },
1316
+ {
1317
+ "epoch": 3.9,
1318
+ "learning_rate": 1.2246963562753037e-05,
1319
+ "loss": 0.3338,
1320
+ "step": 2140
1321
+ },
1322
+ {
1323
+ "epoch": 3.91,
1324
+ "learning_rate": 1.2044534412955466e-05,
1325
+ "loss": 0.3388,
1326
+ "step": 2150
1327
+ },
1328
+ {
1329
+ "epoch": 3.93,
1330
+ "learning_rate": 1.1842105263157895e-05,
1331
+ "loss": 0.3355,
1332
+ "step": 2160
1333
+ },
1334
+ {
1335
+ "epoch": 3.95,
1336
+ "learning_rate": 1.1639676113360324e-05,
1337
+ "loss": 0.3238,
1338
+ "step": 2170
1339
+ },
1340
+ {
1341
+ "epoch": 3.97,
1342
+ "learning_rate": 1.1437246963562753e-05,
1343
+ "loss": 0.3328,
1344
+ "step": 2180
1345
+ },
1346
+ {
1347
+ "epoch": 3.99,
1348
+ "learning_rate": 1.1234817813765184e-05,
1349
+ "loss": 0.3333,
1350
+ "step": 2190
1351
+ },
1352
+ {
1353
+ "epoch": 4.0,
1354
+ "eval_accuracy": 0.9624402458456636,
1355
+ "eval_loss": 0.15150564908981323,
1356
+ "eval_runtime": 102.6282,
1357
+ "eval_samples_per_second": 171.22,
1358
+ "eval_steps_per_second": 5.359,
1359
+ "step": 2197
1360
+ },
1361
+ {
1362
+ "epoch": 4.01,
1363
+ "learning_rate": 1.1032388663967611e-05,
1364
+ "loss": 0.3703,
1365
+ "step": 2200
1366
+ },
1367
+ {
1368
+ "epoch": 4.02,
1369
+ "learning_rate": 1.082995951417004e-05,
1370
+ "loss": 0.3558,
1371
+ "step": 2210
1372
+ },
1373
+ {
1374
+ "epoch": 4.04,
1375
+ "learning_rate": 1.062753036437247e-05,
1376
+ "loss": 0.3665,
1377
+ "step": 2220
1378
+ },
1379
+ {
1380
+ "epoch": 4.06,
1381
+ "learning_rate": 1.04251012145749e-05,
1382
+ "loss": 0.3178,
1383
+ "step": 2230
1384
+ },
1385
+ {
1386
+ "epoch": 4.08,
1387
+ "learning_rate": 1.0222672064777327e-05,
1388
+ "loss": 0.3195,
1389
+ "step": 2240
1390
+ },
1391
+ {
1392
+ "epoch": 4.1,
1393
+ "learning_rate": 1.0020242914979758e-05,
1394
+ "loss": 0.3246,
1395
+ "step": 2250
1396
+ },
1397
+ {
1398
+ "epoch": 4.11,
1399
+ "learning_rate": 9.817813765182187e-06,
1400
+ "loss": 0.3535,
1401
+ "step": 2260
1402
+ },
1403
+ {
1404
+ "epoch": 4.13,
1405
+ "learning_rate": 9.615384615384616e-06,
1406
+ "loss": 0.3413,
1407
+ "step": 2270
1408
+ },
1409
+ {
1410
+ "epoch": 4.15,
1411
+ "learning_rate": 9.412955465587045e-06,
1412
+ "loss": 0.3725,
1413
+ "step": 2280
1414
+ },
1415
+ {
1416
+ "epoch": 4.17,
1417
+ "learning_rate": 9.210526315789474e-06,
1418
+ "loss": 0.347,
1419
+ "step": 2290
1420
+ },
1421
+ {
1422
+ "epoch": 4.19,
1423
+ "learning_rate": 9.008097165991904e-06,
1424
+ "loss": 0.3258,
1425
+ "step": 2300
1426
+ },
1427
+ {
1428
+ "epoch": 4.21,
1429
+ "learning_rate": 8.805668016194332e-06,
1430
+ "loss": 0.3556,
1431
+ "step": 2310
1432
+ },
1433
+ {
1434
+ "epoch": 4.22,
1435
+ "learning_rate": 8.60323886639676e-06,
1436
+ "loss": 0.3161,
1437
+ "step": 2320
1438
+ },
1439
+ {
1440
+ "epoch": 4.24,
1441
+ "learning_rate": 8.400809716599191e-06,
1442
+ "loss": 0.3404,
1443
+ "step": 2330
1444
+ },
1445
+ {
1446
+ "epoch": 4.26,
1447
+ "learning_rate": 8.19838056680162e-06,
1448
+ "loss": 0.2928,
1449
+ "step": 2340
1450
+ },
1451
+ {
1452
+ "epoch": 4.28,
1453
+ "learning_rate": 7.99595141700405e-06,
1454
+ "loss": 0.3363,
1455
+ "step": 2350
1456
+ },
1457
+ {
1458
+ "epoch": 4.3,
1459
+ "learning_rate": 7.793522267206478e-06,
1460
+ "loss": 0.2995,
1461
+ "step": 2360
1462
+ },
1463
+ {
1464
+ "epoch": 4.31,
1465
+ "learning_rate": 7.591093117408906e-06,
1466
+ "loss": 0.3457,
1467
+ "step": 2370
1468
+ },
1469
+ {
1470
+ "epoch": 4.33,
1471
+ "learning_rate": 7.388663967611337e-06,
1472
+ "loss": 0.3192,
1473
+ "step": 2380
1474
+ },
1475
+ {
1476
+ "epoch": 4.35,
1477
+ "learning_rate": 7.186234817813765e-06,
1478
+ "loss": 0.3464,
1479
+ "step": 2390
1480
+ },
1481
+ {
1482
+ "epoch": 4.37,
1483
+ "learning_rate": 6.983805668016195e-06,
1484
+ "loss": 0.3351,
1485
+ "step": 2400
1486
+ },
1487
+ {
1488
+ "epoch": 4.39,
1489
+ "learning_rate": 6.781376518218624e-06,
1490
+ "loss": 0.3412,
1491
+ "step": 2410
1492
+ },
1493
+ {
1494
+ "epoch": 4.41,
1495
+ "learning_rate": 6.578947368421053e-06,
1496
+ "loss": 0.3266,
1497
+ "step": 2420
1498
+ },
1499
+ {
1500
+ "epoch": 4.42,
1501
+ "learning_rate": 6.376518218623483e-06,
1502
+ "loss": 0.349,
1503
+ "step": 2430
1504
+ },
1505
+ {
1506
+ "epoch": 4.44,
1507
+ "learning_rate": 6.174089068825911e-06,
1508
+ "loss": 0.3293,
1509
+ "step": 2440
1510
+ },
1511
+ {
1512
+ "epoch": 4.46,
1513
+ "learning_rate": 5.971659919028341e-06,
1514
+ "loss": 0.34,
1515
+ "step": 2450
1516
+ },
1517
+ {
1518
+ "epoch": 4.48,
1519
+ "learning_rate": 5.76923076923077e-06,
1520
+ "loss": 0.3109,
1521
+ "step": 2460
1522
+ },
1523
+ {
1524
+ "epoch": 4.5,
1525
+ "learning_rate": 5.566801619433199e-06,
1526
+ "loss": 0.3153,
1527
+ "step": 2470
1528
+ },
1529
+ {
1530
+ "epoch": 4.52,
1531
+ "learning_rate": 5.364372469635628e-06,
1532
+ "loss": 0.3318,
1533
+ "step": 2480
1534
+ },
1535
+ {
1536
+ "epoch": 4.53,
1537
+ "learning_rate": 5.161943319838057e-06,
1538
+ "loss": 0.3565,
1539
+ "step": 2490
1540
+ },
1541
+ {
1542
+ "epoch": 4.55,
1543
+ "learning_rate": 4.9595141700404865e-06,
1544
+ "loss": 0.3399,
1545
+ "step": 2500
1546
+ },
1547
+ {
1548
+ "epoch": 4.57,
1549
+ "learning_rate": 4.757085020242915e-06,
1550
+ "loss": 0.3014,
1551
+ "step": 2510
1552
+ },
1553
+ {
1554
+ "epoch": 4.59,
1555
+ "learning_rate": 4.5546558704453445e-06,
1556
+ "loss": 0.3368,
1557
+ "step": 2520
1558
+ },
1559
+ {
1560
+ "epoch": 4.61,
1561
+ "learning_rate": 4.3522267206477735e-06,
1562
+ "loss": 0.3252,
1563
+ "step": 2530
1564
+ },
1565
+ {
1566
+ "epoch": 4.62,
1567
+ "learning_rate": 4.149797570850203e-06,
1568
+ "loss": 0.313,
1569
+ "step": 2540
1570
+ },
1571
+ {
1572
+ "epoch": 4.64,
1573
+ "learning_rate": 3.9473684210526315e-06,
1574
+ "loss": 0.3495,
1575
+ "step": 2550
1576
+ },
1577
+ {
1578
+ "epoch": 4.66,
1579
+ "learning_rate": 3.744939271255061e-06,
1580
+ "loss": 0.3318,
1581
+ "step": 2560
1582
+ },
1583
+ {
1584
+ "epoch": 4.68,
1585
+ "learning_rate": 3.5425101214574903e-06,
1586
+ "loss": 0.3463,
1587
+ "step": 2570
1588
+ },
1589
+ {
1590
+ "epoch": 4.7,
1591
+ "learning_rate": 3.3400809716599197e-06,
1592
+ "loss": 0.3212,
1593
+ "step": 2580
1594
+ },
1595
+ {
1596
+ "epoch": 4.72,
1597
+ "learning_rate": 3.1376518218623482e-06,
1598
+ "loss": 0.2976,
1599
+ "step": 2590
1600
+ },
1601
+ {
1602
+ "epoch": 4.73,
1603
+ "learning_rate": 2.9352226720647772e-06,
1604
+ "loss": 0.3084,
1605
+ "step": 2600
1606
+ },
1607
+ {
1608
+ "epoch": 4.75,
1609
+ "learning_rate": 2.7327935222672066e-06,
1610
+ "loss": 0.3233,
1611
+ "step": 2610
1612
+ },
1613
+ {
1614
+ "epoch": 4.77,
1615
+ "learning_rate": 2.5303643724696356e-06,
1616
+ "loss": 0.3123,
1617
+ "step": 2620
1618
+ },
1619
+ {
1620
+ "epoch": 4.79,
1621
+ "learning_rate": 2.327935222672065e-06,
1622
+ "loss": 0.3201,
1623
+ "step": 2630
1624
+ },
1625
+ {
1626
+ "epoch": 4.81,
1627
+ "learning_rate": 2.125506072874494e-06,
1628
+ "loss": 0.378,
1629
+ "step": 2640
1630
+ },
1631
+ {
1632
+ "epoch": 4.82,
1633
+ "learning_rate": 1.9230769230769234e-06,
1634
+ "loss": 0.3351,
1635
+ "step": 2650
1636
+ },
1637
+ {
1638
+ "epoch": 4.84,
1639
+ "learning_rate": 1.7206477732793522e-06,
1640
+ "loss": 0.3003,
1641
+ "step": 2660
1642
+ },
1643
+ {
1644
+ "epoch": 4.86,
1645
+ "learning_rate": 1.5182186234817814e-06,
1646
+ "loss": 0.3341,
1647
+ "step": 2670
1648
+ },
1649
+ {
1650
+ "epoch": 4.88,
1651
+ "learning_rate": 1.3157894736842106e-06,
1652
+ "loss": 0.3253,
1653
+ "step": 2680
1654
+ },
1655
+ {
1656
+ "epoch": 4.9,
1657
+ "learning_rate": 1.1133603238866398e-06,
1658
+ "loss": 0.306,
1659
+ "step": 2690
1660
+ },
1661
+ {
1662
+ "epoch": 4.92,
1663
+ "learning_rate": 9.109311740890688e-07,
1664
+ "loss": 0.2901,
1665
+ "step": 2700
1666
+ },
1667
+ {
1668
+ "epoch": 4.93,
1669
+ "learning_rate": 7.08502024291498e-07,
1670
+ "loss": 0.3485,
1671
+ "step": 2710
1672
+ },
1673
+ {
1674
+ "epoch": 4.95,
1675
+ "learning_rate": 5.060728744939271e-07,
1676
+ "loss": 0.328,
1677
+ "step": 2720
1678
+ },
1679
+ {
1680
+ "epoch": 4.97,
1681
+ "learning_rate": 3.036437246963563e-07,
1682
+ "loss": 0.3459,
1683
+ "step": 2730
1684
+ },
1685
+ {
1686
+ "epoch": 4.99,
1687
+ "learning_rate": 1.0121457489878543e-07,
1688
+ "loss": 0.3365,
1689
+ "step": 2740
1690
+ },
1691
+ {
1692
+ "epoch": 5.0,
1693
+ "eval_accuracy": 0.9648304120191213,
1694
+ "eval_loss": 0.13630270957946777,
1695
+ "eval_runtime": 102.6364,
1696
+ "eval_samples_per_second": 171.206,
1697
+ "eval_steps_per_second": 5.359,
1698
+ "step": 2745
1699
+ },
1700
+ {
1701
+ "epoch": 5.0,
1702
+ "step": 2745,
1703
+ "total_flos": 7.483912529986142e+18,
1704
+ "train_loss": 1.0410587748977438,
1705
+ "train_runtime": 5056.7157,
1706
+ "train_samples_per_second": 69.507,
1707
+ "train_steps_per_second": 0.543
1708
+ }
1709
+ ],
1710
+ "logging_steps": 10,
1711
+ "max_steps": 2745,
1712
+ "num_train_epochs": 5,
1713
+ "save_steps": 500,
1714
+ "total_flos": 7.483912529986142e+18,
1715
+ "trial_name": null,
1716
+ "trial_params": null
1717
+ }