matthieulel commited on
Commit
b0b7d70
·
verified ·
1 Parent(s): 395fb22

End of training

Browse files
README.md CHANGED
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [facebook/convnextv2-tiny-1k-224](https://huggingface.co/facebook/convnextv2-tiny-1k-224) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 0.3980
21
- - Accuracy: 0.8647
22
 
23
  ## Model description
24
 
 
17
 
18
  This model is a fine-tuned version of [facebook/convnextv2-tiny-1k-224](https://huggingface.co/facebook/convnextv2-tiny-1k-224) on an unknown dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 0.4169
21
+ - Accuracy: 0.8673
22
 
23
  ## Model description
24
 
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.977728285077951,
3
+ "eval_accuracy": 0.8672510958046337,
4
+ "eval_loss": 0.41685062646865845,
5
+ "eval_runtime": 12.3483,
6
+ "eval_samples_per_second": 129.329,
7
+ "eval_steps_per_second": 4.049,
8
+ "total_flos": 3.6084187126879396e+18,
9
+ "train_loss": 0.7042633635657174,
10
+ "train_runtime": 2366.3213,
11
+ "train_samples_per_second": 60.706,
12
+ "train_steps_per_second": 0.473
13
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.977728285077951,
3
+ "eval_accuracy": 0.8672510958046337,
4
+ "eval_loss": 0.41685062646865845,
5
+ "eval_runtime": 12.3483,
6
+ "eval_samples_per_second": 129.329,
7
+ "eval_steps_per_second": 4.049
8
+ }
runs/Apr26_13-48-56_f5c94994a90d/events.out.tfevents.1714141729.f5c94994a90d.529.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b96467d9f6bb66fc64cb7597ee1328838e58fde1a5189918a612f48a00ea5274
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 9.977728285077951,
3
+ "total_flos": 3.6084187126879396e+18,
4
+ "train_loss": 0.7042633635657174,
5
+ "train_runtime": 2366.3213,
6
+ "train_samples_per_second": 60.706,
7
+ "train_steps_per_second": 0.473
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,904 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.8672510958046337,
3
+ "best_model_checkpoint": "convnextv2-tiny-1k-224-finetuned-galaxy10-decals/checkpoint-785",
4
+ "epoch": 9.977728285077951,
5
+ "eval_steps": 500,
6
+ "global_step": 1120,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.08908685968819599,
13
+ "grad_norm": 2.9069173336029053,
14
+ "learning_rate": 4.464285714285715e-06,
15
+ "loss": 2.3159,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.17817371937639198,
20
+ "grad_norm": 4.669096946716309,
21
+ "learning_rate": 8.92857142857143e-06,
22
+ "loss": 2.2657,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.267260579064588,
27
+ "grad_norm": 5.892533302307129,
28
+ "learning_rate": 1.3392857142857144e-05,
29
+ "loss": 2.1852,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.35634743875278396,
34
+ "grad_norm": 4.657855033874512,
35
+ "learning_rate": 1.785714285714286e-05,
36
+ "loss": 2.0843,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.44543429844098,
41
+ "grad_norm": 4.414278030395508,
42
+ "learning_rate": 2.2321428571428575e-05,
43
+ "loss": 1.9644,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.534521158129176,
48
+ "grad_norm": 6.150153636932373,
49
+ "learning_rate": 2.6785714285714288e-05,
50
+ "loss": 1.7921,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.623608017817372,
55
+ "grad_norm": 7.54302978515625,
56
+ "learning_rate": 3.125e-05,
57
+ "loss": 1.5743,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.7126948775055679,
62
+ "grad_norm": 11.624669075012207,
63
+ "learning_rate": 3.571428571428572e-05,
64
+ "loss": 1.42,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.8017817371937639,
69
+ "grad_norm": 11.175118446350098,
70
+ "learning_rate": 4.017857142857143e-05,
71
+ "loss": 1.301,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.89086859688196,
76
+ "grad_norm": 13.484392166137695,
77
+ "learning_rate": 4.464285714285715e-05,
78
+ "loss": 1.2246,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.9799554565701559,
83
+ "grad_norm": 22.607799530029297,
84
+ "learning_rate": 4.910714285714286e-05,
85
+ "loss": 1.0664,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.9977728285077951,
90
+ "eval_accuracy": 0.6725109580463369,
91
+ "eval_loss": 0.9818494915962219,
92
+ "eval_runtime": 12.2808,
93
+ "eval_samples_per_second": 130.041,
94
+ "eval_steps_per_second": 4.071,
95
+ "step": 112
96
+ },
97
+ {
98
+ "epoch": 1.069042316258352,
99
+ "grad_norm": 11.895453453063965,
100
+ "learning_rate": 4.960317460317461e-05,
101
+ "loss": 1.0415,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 1.158129175946548,
106
+ "grad_norm": 15.956673622131348,
107
+ "learning_rate": 4.910714285714286e-05,
108
+ "loss": 1.0021,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 1.247216035634744,
113
+ "grad_norm": 12.564276695251465,
114
+ "learning_rate": 4.8611111111111115e-05,
115
+ "loss": 0.9518,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 1.3363028953229399,
120
+ "grad_norm": 15.347941398620605,
121
+ "learning_rate": 4.811507936507937e-05,
122
+ "loss": 0.9316,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 1.4253897550111359,
127
+ "grad_norm": 13.298521041870117,
128
+ "learning_rate": 4.761904761904762e-05,
129
+ "loss": 0.8316,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 1.5144766146993318,
134
+ "grad_norm": 12.444368362426758,
135
+ "learning_rate": 4.7123015873015876e-05,
136
+ "loss": 0.8863,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 1.6035634743875278,
141
+ "grad_norm": 10.18455982208252,
142
+ "learning_rate": 4.662698412698413e-05,
143
+ "loss": 0.8304,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 1.692650334075724,
148
+ "grad_norm": 13.635807991027832,
149
+ "learning_rate": 4.613095238095239e-05,
150
+ "loss": 0.7748,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 1.7817371937639197,
155
+ "grad_norm": 9.778557777404785,
156
+ "learning_rate": 4.563492063492064e-05,
157
+ "loss": 0.7561,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 1.8708240534521159,
162
+ "grad_norm": 16.08139419555664,
163
+ "learning_rate": 4.5138888888888894e-05,
164
+ "loss": 0.8051,
165
+ "step": 210
166
+ },
167
+ {
168
+ "epoch": 1.9599109131403119,
169
+ "grad_norm": 12.761980056762695,
170
+ "learning_rate": 4.464285714285715e-05,
171
+ "loss": 0.8019,
172
+ "step": 220
173
+ },
174
+ {
175
+ "epoch": 1.9955456570155903,
176
+ "eval_accuracy": 0.7989981214777708,
177
+ "eval_loss": 0.6332681775093079,
178
+ "eval_runtime": 12.3035,
179
+ "eval_samples_per_second": 129.8,
180
+ "eval_steps_per_second": 4.064,
181
+ "step": 224
182
+ },
183
+ {
184
+ "epoch": 2.048997772828508,
185
+ "grad_norm": 10.136615753173828,
186
+ "learning_rate": 4.41468253968254e-05,
187
+ "loss": 0.7221,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 2.138084632516704,
192
+ "grad_norm": 13.145553588867188,
193
+ "learning_rate": 4.3650793650793655e-05,
194
+ "loss": 0.6877,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 2.2271714922048997,
199
+ "grad_norm": 14.941648483276367,
200
+ "learning_rate": 4.315476190476191e-05,
201
+ "loss": 0.7462,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 2.316258351893096,
206
+ "grad_norm": 9.842031478881836,
207
+ "learning_rate": 4.265873015873016e-05,
208
+ "loss": 0.6886,
209
+ "step": 260
210
+ },
211
+ {
212
+ "epoch": 2.4053452115812917,
213
+ "grad_norm": 11.978471755981445,
214
+ "learning_rate": 4.2162698412698416e-05,
215
+ "loss": 0.6843,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 2.494432071269488,
220
+ "grad_norm": 12.24606990814209,
221
+ "learning_rate": 4.166666666666667e-05,
222
+ "loss": 0.7038,
223
+ "step": 280
224
+ },
225
+ {
226
+ "epoch": 2.5835189309576836,
227
+ "grad_norm": 12.357331275939941,
228
+ "learning_rate": 4.117063492063492e-05,
229
+ "loss": 0.696,
230
+ "step": 290
231
+ },
232
+ {
233
+ "epoch": 2.6726057906458798,
234
+ "grad_norm": 16.765913009643555,
235
+ "learning_rate": 4.067460317460318e-05,
236
+ "loss": 0.7174,
237
+ "step": 300
238
+ },
239
+ {
240
+ "epoch": 2.7616926503340755,
241
+ "grad_norm": 13.57442569732666,
242
+ "learning_rate": 4.017857142857143e-05,
243
+ "loss": 0.7305,
244
+ "step": 310
245
+ },
246
+ {
247
+ "epoch": 2.8507795100222717,
248
+ "grad_norm": 13.48181438446045,
249
+ "learning_rate": 3.968253968253968e-05,
250
+ "loss": 0.6957,
251
+ "step": 320
252
+ },
253
+ {
254
+ "epoch": 2.939866369710468,
255
+ "grad_norm": 9.809232711791992,
256
+ "learning_rate": 3.918650793650794e-05,
257
+ "loss": 0.6524,
258
+ "step": 330
259
+ },
260
+ {
261
+ "epoch": 2.9933184855233854,
262
+ "eval_accuracy": 0.8340638697557922,
263
+ "eval_loss": 0.5247990489006042,
264
+ "eval_runtime": 12.2883,
265
+ "eval_samples_per_second": 129.961,
266
+ "eval_steps_per_second": 4.069,
267
+ "step": 336
268
+ },
269
+ {
270
+ "epoch": 3.0289532293986636,
271
+ "grad_norm": 16.003482818603516,
272
+ "learning_rate": 3.8690476190476195e-05,
273
+ "loss": 0.6928,
274
+ "step": 340
275
+ },
276
+ {
277
+ "epoch": 3.11804008908686,
278
+ "grad_norm": 23.448598861694336,
279
+ "learning_rate": 3.8194444444444444e-05,
280
+ "loss": 0.6696,
281
+ "step": 350
282
+ },
283
+ {
284
+ "epoch": 3.2071269487750556,
285
+ "grad_norm": 12.516254425048828,
286
+ "learning_rate": 3.76984126984127e-05,
287
+ "loss": 0.6665,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 3.2962138084632517,
292
+ "grad_norm": 13.503238677978516,
293
+ "learning_rate": 3.7202380952380956e-05,
294
+ "loss": 0.6189,
295
+ "step": 370
296
+ },
297
+ {
298
+ "epoch": 3.3853006681514475,
299
+ "grad_norm": 14.721902847290039,
300
+ "learning_rate": 3.6706349206349205e-05,
301
+ "loss": 0.6405,
302
+ "step": 380
303
+ },
304
+ {
305
+ "epoch": 3.4743875278396437,
306
+ "grad_norm": 11.428637504577637,
307
+ "learning_rate": 3.621031746031746e-05,
308
+ "loss": 0.6502,
309
+ "step": 390
310
+ },
311
+ {
312
+ "epoch": 3.5634743875278394,
313
+ "grad_norm": 8.628026008605957,
314
+ "learning_rate": 3.571428571428572e-05,
315
+ "loss": 0.6335,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 3.6525612472160356,
320
+ "grad_norm": 12.637211799621582,
321
+ "learning_rate": 3.521825396825397e-05,
322
+ "loss": 0.6457,
323
+ "step": 410
324
+ },
325
+ {
326
+ "epoch": 3.7416481069042318,
327
+ "grad_norm": 13.72917652130127,
328
+ "learning_rate": 3.472222222222222e-05,
329
+ "loss": 0.6338,
330
+ "step": 420
331
+ },
332
+ {
333
+ "epoch": 3.8307349665924275,
334
+ "grad_norm": 14.159635543823242,
335
+ "learning_rate": 3.422619047619048e-05,
336
+ "loss": 0.6318,
337
+ "step": 430
338
+ },
339
+ {
340
+ "epoch": 3.9198218262806237,
341
+ "grad_norm": 12.676724433898926,
342
+ "learning_rate": 3.3730158730158734e-05,
343
+ "loss": 0.6339,
344
+ "step": 440
345
+ },
346
+ {
347
+ "epoch": 4.0,
348
+ "eval_accuracy": 0.8447088290544772,
349
+ "eval_loss": 0.4730662703514099,
350
+ "eval_runtime": 12.4926,
351
+ "eval_samples_per_second": 127.835,
352
+ "eval_steps_per_second": 4.002,
353
+ "step": 449
354
+ },
355
+ {
356
+ "epoch": 4.008908685968819,
357
+ "grad_norm": 11.317361831665039,
358
+ "learning_rate": 3.3234126984126983e-05,
359
+ "loss": 0.613,
360
+ "step": 450
361
+ },
362
+ {
363
+ "epoch": 4.097995545657016,
364
+ "grad_norm": 14.402922630310059,
365
+ "learning_rate": 3.273809523809524e-05,
366
+ "loss": 0.6124,
367
+ "step": 460
368
+ },
369
+ {
370
+ "epoch": 4.187082405345212,
371
+ "grad_norm": 9.939033508300781,
372
+ "learning_rate": 3.2242063492063495e-05,
373
+ "loss": 0.5868,
374
+ "step": 470
375
+ },
376
+ {
377
+ "epoch": 4.276169265033408,
378
+ "grad_norm": 10.611005783081055,
379
+ "learning_rate": 3.1746031746031745e-05,
380
+ "loss": 0.5786,
381
+ "step": 480
382
+ },
383
+ {
384
+ "epoch": 4.365256124721603,
385
+ "grad_norm": 11.104296684265137,
386
+ "learning_rate": 3.125e-05,
387
+ "loss": 0.544,
388
+ "step": 490
389
+ },
390
+ {
391
+ "epoch": 4.4543429844097995,
392
+ "grad_norm": 14.008048057556152,
393
+ "learning_rate": 3.075396825396826e-05,
394
+ "loss": 0.6175,
395
+ "step": 500
396
+ },
397
+ {
398
+ "epoch": 4.543429844097996,
399
+ "grad_norm": 9.320144653320312,
400
+ "learning_rate": 3.0257936507936506e-05,
401
+ "loss": 0.5999,
402
+ "step": 510
403
+ },
404
+ {
405
+ "epoch": 4.632516703786192,
406
+ "grad_norm": 9.274946212768555,
407
+ "learning_rate": 2.9761904761904762e-05,
408
+ "loss": 0.5709,
409
+ "step": 520
410
+ },
411
+ {
412
+ "epoch": 4.721603563474387,
413
+ "grad_norm": 12.640064239501953,
414
+ "learning_rate": 2.9265873015873018e-05,
415
+ "loss": 0.6231,
416
+ "step": 530
417
+ },
418
+ {
419
+ "epoch": 4.810690423162583,
420
+ "grad_norm": 11.968724250793457,
421
+ "learning_rate": 2.876984126984127e-05,
422
+ "loss": 0.6206,
423
+ "step": 540
424
+ },
425
+ {
426
+ "epoch": 4.8997772828507795,
427
+ "grad_norm": 11.681157112121582,
428
+ "learning_rate": 2.8273809523809523e-05,
429
+ "loss": 0.6031,
430
+ "step": 550
431
+ },
432
+ {
433
+ "epoch": 4.988864142538976,
434
+ "grad_norm": 13.320256233215332,
435
+ "learning_rate": 2.777777777777778e-05,
436
+ "loss": 0.5178,
437
+ "step": 560
438
+ },
439
+ {
440
+ "epoch": 4.997772828507795,
441
+ "eval_accuracy": 0.8503443957420163,
442
+ "eval_loss": 0.4537416100502014,
443
+ "eval_runtime": 12.2913,
444
+ "eval_samples_per_second": 129.93,
445
+ "eval_steps_per_second": 4.068,
446
+ "step": 561
447
+ },
448
+ {
449
+ "epoch": 5.077951002227172,
450
+ "grad_norm": 15.332996368408203,
451
+ "learning_rate": 2.7281746031746032e-05,
452
+ "loss": 0.5617,
453
+ "step": 570
454
+ },
455
+ {
456
+ "epoch": 5.167037861915367,
457
+ "grad_norm": 14.994087219238281,
458
+ "learning_rate": 2.6785714285714288e-05,
459
+ "loss": 0.5797,
460
+ "step": 580
461
+ },
462
+ {
463
+ "epoch": 5.256124721603563,
464
+ "grad_norm": 13.461969375610352,
465
+ "learning_rate": 2.628968253968254e-05,
466
+ "loss": 0.5524,
467
+ "step": 590
468
+ },
469
+ {
470
+ "epoch": 5.3452115812917596,
471
+ "grad_norm": 12.29080581665039,
472
+ "learning_rate": 2.5793650793650796e-05,
473
+ "loss": 0.5824,
474
+ "step": 600
475
+ },
476
+ {
477
+ "epoch": 5.434298440979956,
478
+ "grad_norm": 11.07197380065918,
479
+ "learning_rate": 2.529761904761905e-05,
480
+ "loss": 0.554,
481
+ "step": 610
482
+ },
483
+ {
484
+ "epoch": 5.523385300668151,
485
+ "grad_norm": 9.797560691833496,
486
+ "learning_rate": 2.4801587301587305e-05,
487
+ "loss": 0.5108,
488
+ "step": 620
489
+ },
490
+ {
491
+ "epoch": 5.612472160356347,
492
+ "grad_norm": 10.469209671020508,
493
+ "learning_rate": 2.4305555555555558e-05,
494
+ "loss": 0.5586,
495
+ "step": 630
496
+ },
497
+ {
498
+ "epoch": 5.701559020044543,
499
+ "grad_norm": 13.22735595703125,
500
+ "learning_rate": 2.380952380952381e-05,
501
+ "loss": 0.5358,
502
+ "step": 640
503
+ },
504
+ {
505
+ "epoch": 5.79064587973274,
506
+ "grad_norm": 8.305063247680664,
507
+ "learning_rate": 2.3313492063492066e-05,
508
+ "loss": 0.5295,
509
+ "step": 650
510
+ },
511
+ {
512
+ "epoch": 5.879732739420936,
513
+ "grad_norm": 18.399051666259766,
514
+ "learning_rate": 2.281746031746032e-05,
515
+ "loss": 0.5442,
516
+ "step": 660
517
+ },
518
+ {
519
+ "epoch": 5.968819599109131,
520
+ "grad_norm": 8.103595733642578,
521
+ "learning_rate": 2.2321428571428575e-05,
522
+ "loss": 0.5907,
523
+ "step": 670
524
+ },
525
+ {
526
+ "epoch": 5.99554565701559,
527
+ "eval_accuracy": 0.8472135253600501,
528
+ "eval_loss": 0.4555535316467285,
529
+ "eval_runtime": 12.2927,
530
+ "eval_samples_per_second": 129.914,
531
+ "eval_steps_per_second": 4.067,
532
+ "step": 673
533
+ },
534
+ {
535
+ "epoch": 6.057906458797327,
536
+ "grad_norm": 10.681763648986816,
537
+ "learning_rate": 2.1825396825396827e-05,
538
+ "loss": 0.5332,
539
+ "step": 680
540
+ },
541
+ {
542
+ "epoch": 6.146993318485523,
543
+ "grad_norm": 10.129424095153809,
544
+ "learning_rate": 2.132936507936508e-05,
545
+ "loss": 0.4747,
546
+ "step": 690
547
+ },
548
+ {
549
+ "epoch": 6.23608017817372,
550
+ "grad_norm": 16.834814071655273,
551
+ "learning_rate": 2.0833333333333336e-05,
552
+ "loss": 0.5576,
553
+ "step": 700
554
+ },
555
+ {
556
+ "epoch": 6.325167037861915,
557
+ "grad_norm": 11.258397102355957,
558
+ "learning_rate": 2.033730158730159e-05,
559
+ "loss": 0.5063,
560
+ "step": 710
561
+ },
562
+ {
563
+ "epoch": 6.414253897550111,
564
+ "grad_norm": 15.159914016723633,
565
+ "learning_rate": 1.984126984126984e-05,
566
+ "loss": 0.5385,
567
+ "step": 720
568
+ },
569
+ {
570
+ "epoch": 6.503340757238307,
571
+ "grad_norm": 10.242027282714844,
572
+ "learning_rate": 1.9345238095238097e-05,
573
+ "loss": 0.5046,
574
+ "step": 730
575
+ },
576
+ {
577
+ "epoch": 6.5924276169265035,
578
+ "grad_norm": 10.377813339233398,
579
+ "learning_rate": 1.884920634920635e-05,
580
+ "loss": 0.5247,
581
+ "step": 740
582
+ },
583
+ {
584
+ "epoch": 6.6815144766147,
585
+ "grad_norm": 12.55459213256836,
586
+ "learning_rate": 1.8353174603174602e-05,
587
+ "loss": 0.529,
588
+ "step": 750
589
+ },
590
+ {
591
+ "epoch": 6.770601336302895,
592
+ "grad_norm": 16.02656364440918,
593
+ "learning_rate": 1.785714285714286e-05,
594
+ "loss": 0.5073,
595
+ "step": 760
596
+ },
597
+ {
598
+ "epoch": 6.859688195991091,
599
+ "grad_norm": 16.140487670898438,
600
+ "learning_rate": 1.736111111111111e-05,
601
+ "loss": 0.5414,
602
+ "step": 770
603
+ },
604
+ {
605
+ "epoch": 6.948775055679287,
606
+ "grad_norm": 17.321931838989258,
607
+ "learning_rate": 1.6865079365079367e-05,
608
+ "loss": 0.5292,
609
+ "step": 780
610
+ },
611
+ {
612
+ "epoch": 6.993318485523385,
613
+ "eval_accuracy": 0.8672510958046337,
614
+ "eval_loss": 0.41685062646865845,
615
+ "eval_runtime": 12.3673,
616
+ "eval_samples_per_second": 129.131,
617
+ "eval_steps_per_second": 4.043,
618
+ "step": 785
619
+ },
620
+ {
621
+ "epoch": 7.0378619153674835,
622
+ "grad_norm": 9.479693412780762,
623
+ "learning_rate": 1.636904761904762e-05,
624
+ "loss": 0.4586,
625
+ "step": 790
626
+ },
627
+ {
628
+ "epoch": 7.12694877505568,
629
+ "grad_norm": 11.711000442504883,
630
+ "learning_rate": 1.5873015873015872e-05,
631
+ "loss": 0.5188,
632
+ "step": 800
633
+ },
634
+ {
635
+ "epoch": 7.216035634743875,
636
+ "grad_norm": 11.616864204406738,
637
+ "learning_rate": 1.537698412698413e-05,
638
+ "loss": 0.5024,
639
+ "step": 810
640
+ },
641
+ {
642
+ "epoch": 7.305122494432071,
643
+ "grad_norm": 10.370725631713867,
644
+ "learning_rate": 1.4880952380952381e-05,
645
+ "loss": 0.4902,
646
+ "step": 820
647
+ },
648
+ {
649
+ "epoch": 7.394209354120267,
650
+ "grad_norm": 14.04218864440918,
651
+ "learning_rate": 1.4384920634920635e-05,
652
+ "loss": 0.5149,
653
+ "step": 830
654
+ },
655
+ {
656
+ "epoch": 7.4832962138084635,
657
+ "grad_norm": 13.194646835327148,
658
+ "learning_rate": 1.388888888888889e-05,
659
+ "loss": 0.5562,
660
+ "step": 840
661
+ },
662
+ {
663
+ "epoch": 7.57238307349666,
664
+ "grad_norm": 9.960190773010254,
665
+ "learning_rate": 1.3392857142857144e-05,
666
+ "loss": 0.4921,
667
+ "step": 850
668
+ },
669
+ {
670
+ "epoch": 7.661469933184855,
671
+ "grad_norm": 15.14493465423584,
672
+ "learning_rate": 1.2896825396825398e-05,
673
+ "loss": 0.471,
674
+ "step": 860
675
+ },
676
+ {
677
+ "epoch": 7.750556792873051,
678
+ "grad_norm": 11.185235977172852,
679
+ "learning_rate": 1.2400793650793652e-05,
680
+ "loss": 0.4963,
681
+ "step": 870
682
+ },
683
+ {
684
+ "epoch": 7.839643652561247,
685
+ "grad_norm": 12.782095909118652,
686
+ "learning_rate": 1.1904761904761905e-05,
687
+ "loss": 0.4915,
688
+ "step": 880
689
+ },
690
+ {
691
+ "epoch": 7.928730512249444,
692
+ "grad_norm": 11.89919376373291,
693
+ "learning_rate": 1.140873015873016e-05,
694
+ "loss": 0.5017,
695
+ "step": 890
696
+ },
697
+ {
698
+ "epoch": 8.0,
699
+ "eval_accuracy": 0.8597370068879149,
700
+ "eval_loss": 0.4106651544570923,
701
+ "eval_runtime": 12.3902,
702
+ "eval_samples_per_second": 128.892,
703
+ "eval_steps_per_second": 4.035,
704
+ "step": 898
705
+ },
706
+ {
707
+ "epoch": 8.017817371937639,
708
+ "grad_norm": 12.601805686950684,
709
+ "learning_rate": 1.0912698412698414e-05,
710
+ "loss": 0.5064,
711
+ "step": 900
712
+ },
713
+ {
714
+ "epoch": 8.106904231625835,
715
+ "grad_norm": 8.723831176757812,
716
+ "learning_rate": 1.0416666666666668e-05,
717
+ "loss": 0.4181,
718
+ "step": 910
719
+ },
720
+ {
721
+ "epoch": 8.195991091314031,
722
+ "grad_norm": 12.781538963317871,
723
+ "learning_rate": 9.92063492063492e-06,
724
+ "loss": 0.4427,
725
+ "step": 920
726
+ },
727
+ {
728
+ "epoch": 8.285077951002227,
729
+ "grad_norm": 12.263012886047363,
730
+ "learning_rate": 9.424603174603175e-06,
731
+ "loss": 0.5087,
732
+ "step": 930
733
+ },
734
+ {
735
+ "epoch": 8.374164810690424,
736
+ "grad_norm": 17.41984748840332,
737
+ "learning_rate": 8.92857142857143e-06,
738
+ "loss": 0.5301,
739
+ "step": 940
740
+ },
741
+ {
742
+ "epoch": 8.46325167037862,
743
+ "grad_norm": 10.731024742126465,
744
+ "learning_rate": 8.432539682539684e-06,
745
+ "loss": 0.4987,
746
+ "step": 950
747
+ },
748
+ {
749
+ "epoch": 8.552338530066816,
750
+ "grad_norm": 15.722013473510742,
751
+ "learning_rate": 7.936507936507936e-06,
752
+ "loss": 0.4613,
753
+ "step": 960
754
+ },
755
+ {
756
+ "epoch": 8.64142538975501,
757
+ "grad_norm": 11.301126480102539,
758
+ "learning_rate": 7.4404761904761905e-06,
759
+ "loss": 0.5136,
760
+ "step": 970
761
+ },
762
+ {
763
+ "epoch": 8.730512249443207,
764
+ "grad_norm": 21.23493194580078,
765
+ "learning_rate": 6.944444444444445e-06,
766
+ "loss": 0.494,
767
+ "step": 980
768
+ },
769
+ {
770
+ "epoch": 8.819599109131403,
771
+ "grad_norm": 10.211363792419434,
772
+ "learning_rate": 6.448412698412699e-06,
773
+ "loss": 0.4619,
774
+ "step": 990
775
+ },
776
+ {
777
+ "epoch": 8.908685968819599,
778
+ "grad_norm": 12.277856826782227,
779
+ "learning_rate": 5.9523809523809525e-06,
780
+ "loss": 0.4361,
781
+ "step": 1000
782
+ },
783
+ {
784
+ "epoch": 8.997772828507795,
785
+ "grad_norm": 12.28085708618164,
786
+ "learning_rate": 5.456349206349207e-06,
787
+ "loss": 0.4605,
788
+ "step": 1010
789
+ },
790
+ {
791
+ "epoch": 8.997772828507795,
792
+ "eval_accuracy": 0.8634940513462742,
793
+ "eval_loss": 0.40621063113212585,
794
+ "eval_runtime": 12.3051,
795
+ "eval_samples_per_second": 129.783,
796
+ "eval_steps_per_second": 4.063,
797
+ "step": 1010
798
+ },
799
+ {
800
+ "epoch": 9.086859688195991,
801
+ "grad_norm": 8.358485221862793,
802
+ "learning_rate": 4.96031746031746e-06,
803
+ "loss": 0.4403,
804
+ "step": 1020
805
+ },
806
+ {
807
+ "epoch": 9.175946547884188,
808
+ "grad_norm": 8.35409164428711,
809
+ "learning_rate": 4.464285714285715e-06,
810
+ "loss": 0.4514,
811
+ "step": 1030
812
+ },
813
+ {
814
+ "epoch": 9.265033407572384,
815
+ "grad_norm": 10.057600021362305,
816
+ "learning_rate": 3.968253968253968e-06,
817
+ "loss": 0.427,
818
+ "step": 1040
819
+ },
820
+ {
821
+ "epoch": 9.35412026726058,
822
+ "grad_norm": 7.57137393951416,
823
+ "learning_rate": 3.4722222222222224e-06,
824
+ "loss": 0.4619,
825
+ "step": 1050
826
+ },
827
+ {
828
+ "epoch": 9.443207126948774,
829
+ "grad_norm": 9.249728202819824,
830
+ "learning_rate": 2.9761904761904763e-06,
831
+ "loss": 0.4606,
832
+ "step": 1060
833
+ },
834
+ {
835
+ "epoch": 9.53229398663697,
836
+ "grad_norm": 10.303194046020508,
837
+ "learning_rate": 2.48015873015873e-06,
838
+ "loss": 0.508,
839
+ "step": 1070
840
+ },
841
+ {
842
+ "epoch": 9.621380846325167,
843
+ "grad_norm": 11.307740211486816,
844
+ "learning_rate": 1.984126984126984e-06,
845
+ "loss": 0.4667,
846
+ "step": 1080
847
+ },
848
+ {
849
+ "epoch": 9.710467706013363,
850
+ "grad_norm": 11.34073543548584,
851
+ "learning_rate": 1.4880952380952381e-06,
852
+ "loss": 0.4803,
853
+ "step": 1090
854
+ },
855
+ {
856
+ "epoch": 9.799554565701559,
857
+ "grad_norm": 12.684574127197266,
858
+ "learning_rate": 9.92063492063492e-07,
859
+ "loss": 0.4613,
860
+ "step": 1100
861
+ },
862
+ {
863
+ "epoch": 9.888641425389755,
864
+ "grad_norm": 12.430156707763672,
865
+ "learning_rate": 4.96031746031746e-07,
866
+ "loss": 0.434,
867
+ "step": 1110
868
+ },
869
+ {
870
+ "epoch": 9.977728285077951,
871
+ "grad_norm": 19.223121643066406,
872
+ "learning_rate": 0.0,
873
+ "loss": 0.4765,
874
+ "step": 1120
875
+ },
876
+ {
877
+ "epoch": 9.977728285077951,
878
+ "eval_accuracy": 0.8647463994990607,
879
+ "eval_loss": 0.3980247676372528,
880
+ "eval_runtime": 12.312,
881
+ "eval_samples_per_second": 129.711,
882
+ "eval_steps_per_second": 4.061,
883
+ "step": 1120
884
+ },
885
+ {
886
+ "epoch": 9.977728285077951,
887
+ "step": 1120,
888
+ "total_flos": 3.6084187126879396e+18,
889
+ "train_loss": 0.7042633635657174,
890
+ "train_runtime": 2366.3213,
891
+ "train_samples_per_second": 60.706,
892
+ "train_steps_per_second": 0.473
893
+ }
894
+ ],
895
+ "logging_steps": 10,
896
+ "max_steps": 1120,
897
+ "num_input_tokens_seen": 0,
898
+ "num_train_epochs": 10,
899
+ "save_steps": 500,
900
+ "total_flos": 3.6084187126879396e+18,
901
+ "train_batch_size": 32,
902
+ "trial_name": null,
903
+ "trial_params": null
904
+ }