vuongnhathien commited on
Commit
b1aa8c6
1 Parent(s): 6d9297f

End of training

Browse files
README.md CHANGED
@@ -22,7 +22,7 @@ model-index:
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
- value: 0.9431411530815109
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,8 +32,8 @@ should probably proofread and complete it, then remove this comment. -->
32
 
33
  This model is a fine-tuned version of [facebook/convnextv2-base-22k-384](https://huggingface.co/facebook/convnextv2-base-22k-384) on the imagefolder dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 0.3192
36
- - Accuracy: 0.9431
37
 
38
  ## Model description
39
 
 
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
+ value: 0.942063492063492
26
  ---
27
 
28
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
32
 
33
  This model is a fine-tuned version of [facebook/convnextv2-base-22k-384](https://huggingface.co/facebook/convnextv2-base-22k-384) on the imagefolder dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 0.3165
36
+ - Accuracy: 0.9421
37
 
38
  ## Model description
39
 
all_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 6.140249030814106e+19,
4
+ "train_loss": 0.2272892904238228,
5
+ "train_runtime": 26538.9058,
6
+ "train_samples_per_second": 9.937,
7
+ "train_steps_per_second": 0.621
8
+ }
runs/May25_17-44-28_ab94a2ea8078/events.out.tfevents.1716685921.ab94a2ea8078.24.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a17a44e2431fee34471ad1a90a90c2668204132fc337afac6330e93e00a6b791
3
+ size 417
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 15.0,
3
+ "total_flos": 6.140249030814106e+19,
4
+ "train_loss": 0.2272892904238228,
5
+ "train_runtime": 26538.9058,
6
+ "train_samples_per_second": 9.937,
7
+ "train_steps_per_second": 0.621
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1313 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.31903526186943054,
3
+ "best_model_checkpoint": "./convnext-base-15ep/checkpoint-15386",
4
+ "epoch": 15.0,
5
+ "eval_steps": 500,
6
+ "global_step": 16485,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.09,
13
+ "grad_norm": 12.268571853637695,
14
+ "learning_rate": 9.999092077649917e-05,
15
+ "loss": 1.8816,
16
+ "step": 100
17
+ },
18
+ {
19
+ "epoch": 0.18,
20
+ "grad_norm": 16.745285034179688,
21
+ "learning_rate": 9.996368640328861e-05,
22
+ "loss": 0.9784,
23
+ "step": 200
24
+ },
25
+ {
26
+ "epoch": 0.27,
27
+ "grad_norm": 15.278342247009277,
28
+ "learning_rate": 9.991830677104683e-05,
29
+ "loss": 0.8074,
30
+ "step": 300
31
+ },
32
+ {
33
+ "epoch": 0.36,
34
+ "grad_norm": 12.999685287475586,
35
+ "learning_rate": 9.985479836024671e-05,
36
+ "loss": 0.6924,
37
+ "step": 400
38
+ },
39
+ {
40
+ "epoch": 0.45,
41
+ "grad_norm": 15.637765884399414,
42
+ "learning_rate": 9.977318423517052e-05,
43
+ "loss": 0.6877,
44
+ "step": 500
45
+ },
46
+ {
47
+ "epoch": 0.55,
48
+ "grad_norm": 8.964929580688477,
49
+ "learning_rate": 9.967349403553353e-05,
50
+ "loss": 0.6085,
51
+ "step": 600
52
+ },
53
+ {
54
+ "epoch": 0.64,
55
+ "grad_norm": 20.98185157775879,
56
+ "learning_rate": 9.95557639657199e-05,
57
+ "loss": 0.6267,
58
+ "step": 700
59
+ },
60
+ {
61
+ "epoch": 0.73,
62
+ "grad_norm": 17.35112953186035,
63
+ "learning_rate": 9.942003678163429e-05,
64
+ "loss": 0.6133,
65
+ "step": 800
66
+ },
67
+ {
68
+ "epoch": 0.82,
69
+ "grad_norm": 4.6055192947387695,
70
+ "learning_rate": 9.926636177517427e-05,
71
+ "loss": 0.5813,
72
+ "step": 900
73
+ },
74
+ {
75
+ "epoch": 0.91,
76
+ "grad_norm": 10.87624454498291,
77
+ "learning_rate": 9.909479475632904e-05,
78
+ "loss": 0.5717,
79
+ "step": 1000
80
+ },
81
+ {
82
+ "epoch": 1.0,
83
+ "eval_accuracy": 0.8572564612326043,
84
+ "eval_loss": 0.4616268575191498,
85
+ "eval_runtime": 111.9566,
86
+ "eval_samples_per_second": 22.464,
87
+ "eval_steps_per_second": 1.411,
88
+ "step": 1099
89
+ },
90
+ {
91
+ "epoch": 1.0,
92
+ "grad_norm": 12.744146347045898,
93
+ "learning_rate": 9.8905398032911e-05,
94
+ "loss": 0.5917,
95
+ "step": 1100
96
+ },
97
+ {
98
+ "epoch": 1.09,
99
+ "grad_norm": 7.046997547149658,
100
+ "learning_rate": 9.869824038792741e-05,
101
+ "loss": 0.4628,
102
+ "step": 1200
103
+ },
104
+ {
105
+ "epoch": 1.18,
106
+ "grad_norm": 14.269908905029297,
107
+ "learning_rate": 9.847339705460064e-05,
108
+ "loss": 0.4921,
109
+ "step": 1300
110
+ },
111
+ {
112
+ "epoch": 1.27,
113
+ "grad_norm": 13.306988716125488,
114
+ "learning_rate": 9.823094968904572e-05,
115
+ "loss": 0.474,
116
+ "step": 1400
117
+ },
118
+ {
119
+ "epoch": 1.36,
120
+ "grad_norm": 11.327611923217773,
121
+ "learning_rate": 9.797098634061542e-05,
122
+ "loss": 0.4723,
123
+ "step": 1500
124
+ },
125
+ {
126
+ "epoch": 1.46,
127
+ "grad_norm": 10.74616813659668,
128
+ "learning_rate": 9.769360141992343e-05,
129
+ "loss": 0.4207,
130
+ "step": 1600
131
+ },
132
+ {
133
+ "epoch": 1.55,
134
+ "grad_norm": 12.389139175415039,
135
+ "learning_rate": 9.739889566455738e-05,
136
+ "loss": 0.492,
137
+ "step": 1700
138
+ },
139
+ {
140
+ "epoch": 1.64,
141
+ "grad_norm": 13.068366050720215,
142
+ "learning_rate": 9.708697610249406e-05,
143
+ "loss": 0.4583,
144
+ "step": 1800
145
+ },
146
+ {
147
+ "epoch": 1.73,
148
+ "grad_norm": 6.97837495803833,
149
+ "learning_rate": 9.675795601323023e-05,
150
+ "loss": 0.4619,
151
+ "step": 1900
152
+ },
153
+ {
154
+ "epoch": 1.82,
155
+ "grad_norm": 11.903929710388184,
156
+ "learning_rate": 9.641195488664292e-05,
157
+ "loss": 0.4033,
158
+ "step": 2000
159
+ },
160
+ {
161
+ "epoch": 1.91,
162
+ "grad_norm": 8.415858268737793,
163
+ "learning_rate": 9.604909837959455e-05,
164
+ "loss": 0.4653,
165
+ "step": 2100
166
+ },
167
+ {
168
+ "epoch": 2.0,
169
+ "eval_accuracy": 0.8970178926441352,
170
+ "eval_loss": 0.36072757840156555,
171
+ "eval_runtime": 111.9718,
172
+ "eval_samples_per_second": 22.461,
173
+ "eval_steps_per_second": 1.411,
174
+ "step": 2198
175
+ },
176
+ {
177
+ "epoch": 2.0,
178
+ "grad_norm": 19.435712814331055,
179
+ "learning_rate": 9.566951827029816e-05,
180
+ "loss": 0.442,
181
+ "step": 2200
182
+ },
183
+ {
184
+ "epoch": 2.09,
185
+ "grad_norm": 6.7451982498168945,
186
+ "learning_rate": 9.52733524104597e-05,
187
+ "loss": 0.3708,
188
+ "step": 2300
189
+ },
190
+ {
191
+ "epoch": 2.18,
192
+ "grad_norm": 15.736494064331055,
193
+ "learning_rate": 9.486074467521456e-05,
194
+ "loss": 0.3583,
195
+ "step": 2400
196
+ },
197
+ {
198
+ "epoch": 2.27,
199
+ "grad_norm": 10.238139152526855,
200
+ "learning_rate": 9.44318449108766e-05,
201
+ "loss": 0.3513,
202
+ "step": 2500
203
+ },
204
+ {
205
+ "epoch": 2.37,
206
+ "grad_norm": 5.870657444000244,
207
+ "learning_rate": 9.398680888051863e-05,
208
+ "loss": 0.4243,
209
+ "step": 2600
210
+ },
211
+ {
212
+ "epoch": 2.46,
213
+ "grad_norm": 10.592901229858398,
214
+ "learning_rate": 9.352579820740405e-05,
215
+ "loss": 0.3681,
216
+ "step": 2700
217
+ },
218
+ {
219
+ "epoch": 2.55,
220
+ "grad_norm": 2.5265157222747803,
221
+ "learning_rate": 9.304898031629036e-05,
222
+ "loss": 0.3971,
223
+ "step": 2800
224
+ },
225
+ {
226
+ "epoch": 2.64,
227
+ "grad_norm": 20.8736572265625,
228
+ "learning_rate": 9.25565283726257e-05,
229
+ "loss": 0.4244,
230
+ "step": 2900
231
+ },
232
+ {
233
+ "epoch": 2.73,
234
+ "grad_norm": 2.61696457862854,
235
+ "learning_rate": 9.204862121966044e-05,
236
+ "loss": 0.3762,
237
+ "step": 3000
238
+ },
239
+ {
240
+ "epoch": 2.82,
241
+ "grad_norm": 11.449930191040039,
242
+ "learning_rate": 9.152544331349694e-05,
243
+ "loss": 0.3584,
244
+ "step": 3100
245
+ },
246
+ {
247
+ "epoch": 2.91,
248
+ "grad_norm": 9.69878101348877,
249
+ "learning_rate": 9.098718465610088e-05,
250
+ "loss": 0.3449,
251
+ "step": 3200
252
+ },
253
+ {
254
+ "epoch": 3.0,
255
+ "eval_accuracy": 0.8950298210735587,
256
+ "eval_loss": 0.41042593121528625,
257
+ "eval_runtime": 111.0632,
258
+ "eval_samples_per_second": 22.645,
259
+ "eval_steps_per_second": 1.423,
260
+ "step": 3297
261
+ },
262
+ {
263
+ "epoch": 3.0,
264
+ "grad_norm": 2.2799901962280273,
265
+ "learning_rate": 9.043404072629829e-05,
266
+ "loss": 0.394,
267
+ "step": 3300
268
+ },
269
+ {
270
+ "epoch": 3.09,
271
+ "grad_norm": 11.792191505432129,
272
+ "learning_rate": 8.986621240878385e-05,
273
+ "loss": 0.2869,
274
+ "step": 3400
275
+ },
276
+ {
277
+ "epoch": 3.18,
278
+ "grad_norm": 1.5572706460952759,
279
+ "learning_rate": 8.928390592116575e-05,
280
+ "loss": 0.2879,
281
+ "step": 3500
282
+ },
283
+ {
284
+ "epoch": 3.28,
285
+ "grad_norm": 7.704630374908447,
286
+ "learning_rate": 8.86873327390739e-05,
287
+ "loss": 0.3059,
288
+ "step": 3600
289
+ },
290
+ {
291
+ "epoch": 3.37,
292
+ "grad_norm": 10.151177406311035,
293
+ "learning_rate": 8.807670951935846e-05,
294
+ "loss": 0.3329,
295
+ "step": 3700
296
+ },
297
+ {
298
+ "epoch": 3.46,
299
+ "grad_norm": 9.523168563842773,
300
+ "learning_rate": 8.745225802140691e-05,
301
+ "loss": 0.287,
302
+ "step": 3800
303
+ },
304
+ {
305
+ "epoch": 3.55,
306
+ "grad_norm": 9.30478572845459,
307
+ "learning_rate": 8.681420502660786e-05,
308
+ "loss": 0.2992,
309
+ "step": 3900
310
+ },
311
+ {
312
+ "epoch": 3.64,
313
+ "grad_norm": 7.661346912384033,
314
+ "learning_rate": 8.616278225599111e-05,
315
+ "loss": 0.3259,
316
+ "step": 4000
317
+ },
318
+ {
319
+ "epoch": 3.73,
320
+ "grad_norm": 5.623369216918945,
321
+ "learning_rate": 8.54982262860738e-05,
322
+ "loss": 0.334,
323
+ "step": 4100
324
+ },
325
+ {
326
+ "epoch": 3.82,
327
+ "grad_norm": 0.08087614178657532,
328
+ "learning_rate": 8.482077846294308e-05,
329
+ "loss": 0.3385,
330
+ "step": 4200
331
+ },
332
+ {
333
+ "epoch": 3.91,
334
+ "grad_norm": 12.912845611572266,
335
+ "learning_rate": 8.413068481460686e-05,
336
+ "loss": 0.3522,
337
+ "step": 4300
338
+ },
339
+ {
340
+ "epoch": 4.0,
341
+ "eval_accuracy": 0.9025844930417495,
342
+ "eval_loss": 0.37546443939208984,
343
+ "eval_runtime": 111.8065,
344
+ "eval_samples_per_second": 22.494,
345
+ "eval_steps_per_second": 1.413,
346
+ "step": 4396
347
+ },
348
+ {
349
+ "epoch": 4.0,
350
+ "grad_norm": 2.0130882263183594,
351
+ "learning_rate": 8.342819596164387e-05,
352
+ "loss": 0.2963,
353
+ "step": 4400
354
+ },
355
+ {
356
+ "epoch": 4.09,
357
+ "grad_norm": 9.422698020935059,
358
+ "learning_rate": 8.271356702618626e-05,
359
+ "loss": 0.263,
360
+ "step": 4500
361
+ },
362
+ {
363
+ "epoch": 4.19,
364
+ "grad_norm": 4.466526985168457,
365
+ "learning_rate": 8.198705753926704e-05,
366
+ "loss": 0.2767,
367
+ "step": 4600
368
+ },
369
+ {
370
+ "epoch": 4.28,
371
+ "grad_norm": 8.486282348632812,
372
+ "learning_rate": 8.12489313465665e-05,
373
+ "loss": 0.2691,
374
+ "step": 4700
375
+ },
376
+ {
377
+ "epoch": 4.37,
378
+ "grad_norm": 1.178945541381836,
379
+ "learning_rate": 8.049945651259163e-05,
380
+ "loss": 0.2454,
381
+ "step": 4800
382
+ },
383
+ {
384
+ "epoch": 4.46,
385
+ "grad_norm": 8.335172653198242,
386
+ "learning_rate": 7.973890522332348e-05,
387
+ "loss": 0.2906,
388
+ "step": 4900
389
+ },
390
+ {
391
+ "epoch": 4.55,
392
+ "grad_norm": 3.436420440673828,
393
+ "learning_rate": 7.89675536873676e-05,
394
+ "loss": 0.2316,
395
+ "step": 5000
396
+ },
397
+ {
398
+ "epoch": 4.64,
399
+ "grad_norm": 7.467193603515625,
400
+ "learning_rate": 7.818568203564374e-05,
401
+ "loss": 0.2738,
402
+ "step": 5100
403
+ },
404
+ {
405
+ "epoch": 4.73,
406
+ "grad_norm": 10.7122163772583,
407
+ "learning_rate": 7.739357421965086e-05,
408
+ "loss": 0.281,
409
+ "step": 5200
410
+ },
411
+ {
412
+ "epoch": 4.82,
413
+ "grad_norm": 14.049674034118652,
414
+ "learning_rate": 7.65915179083449e-05,
415
+ "loss": 0.2782,
416
+ "step": 5300
417
+ },
418
+ {
419
+ "epoch": 4.91,
420
+ "grad_norm": 16.04802703857422,
421
+ "learning_rate": 7.577980438366628e-05,
422
+ "loss": 0.28,
423
+ "step": 5400
424
+ },
425
+ {
426
+ "epoch": 5.0,
427
+ "eval_accuracy": 0.9065606361829026,
428
+ "eval_loss": 0.3756468892097473,
429
+ "eval_runtime": 111.8227,
430
+ "eval_samples_per_second": 22.491,
431
+ "eval_steps_per_second": 1.413,
432
+ "step": 5495
433
+ },
434
+ {
435
+ "epoch": 5.0,
436
+ "grad_norm": 2.5667672157287598,
437
+ "learning_rate": 7.495872843475536e-05,
438
+ "loss": 0.2711,
439
+ "step": 5500
440
+ },
441
+ {
442
+ "epoch": 5.1,
443
+ "grad_norm": 5.182483673095703,
444
+ "learning_rate": 7.412858825089422e-05,
445
+ "loss": 0.1939,
446
+ "step": 5600
447
+ },
448
+ {
449
+ "epoch": 5.19,
450
+ "grad_norm": 13.568757057189941,
451
+ "learning_rate": 7.32896853132135e-05,
452
+ "loss": 0.2178,
453
+ "step": 5700
454
+ },
455
+ {
456
+ "epoch": 5.28,
457
+ "grad_norm": 10.223276138305664,
458
+ "learning_rate": 7.244232428520383e-05,
459
+ "loss": 0.2185,
460
+ "step": 5800
461
+ },
462
+ {
463
+ "epoch": 5.37,
464
+ "grad_norm": 0.07271099835634232,
465
+ "learning_rate": 7.158681290207163e-05,
466
+ "loss": 0.2218,
467
+ "step": 5900
468
+ },
469
+ {
470
+ "epoch": 5.46,
471
+ "grad_norm": 2.859640598297119,
472
+ "learning_rate": 7.07234618589791e-05,
473
+ "loss": 0.1974,
474
+ "step": 6000
475
+ },
476
+ {
477
+ "epoch": 5.55,
478
+ "grad_norm": 4.893543243408203,
479
+ "learning_rate": 6.985258469820939e-05,
480
+ "loss": 0.2201,
481
+ "step": 6100
482
+ },
483
+ {
484
+ "epoch": 5.64,
485
+ "grad_norm": 9.013633728027344,
486
+ "learning_rate": 6.897449769529792e-05,
487
+ "loss": 0.2031,
488
+ "step": 6200
489
+ },
490
+ {
491
+ "epoch": 5.73,
492
+ "grad_norm": 8.197441101074219,
493
+ "learning_rate": 6.808951974417078e-05,
494
+ "loss": 0.2532,
495
+ "step": 6300
496
+ },
497
+ {
498
+ "epoch": 5.82,
499
+ "grad_norm": 0.8769248127937317,
500
+ "learning_rate": 6.719797224133242e-05,
501
+ "loss": 0.2399,
502
+ "step": 6400
503
+ },
504
+ {
505
+ "epoch": 5.91,
506
+ "grad_norm": 9.442893981933594,
507
+ "learning_rate": 6.630017896914446e-05,
508
+ "loss": 0.2456,
509
+ "step": 6500
510
+ },
511
+ {
512
+ "epoch": 6.0,
513
+ "eval_accuracy": 0.9172962226640159,
514
+ "eval_loss": 0.349565327167511,
515
+ "eval_runtime": 110.6537,
516
+ "eval_samples_per_second": 22.729,
517
+ "eval_steps_per_second": 1.428,
518
+ "step": 6594
519
+ },
520
+ {
521
+ "epoch": 6.01,
522
+ "grad_norm": 9.589459419250488,
523
+ "learning_rate": 6.539646597823791e-05,
524
+ "loss": 0.2373,
525
+ "step": 6600
526
+ },
527
+ {
528
+ "epoch": 6.1,
529
+ "grad_norm": 11.073722839355469,
530
+ "learning_rate": 6.44871614691018e-05,
531
+ "loss": 0.1756,
532
+ "step": 6700
533
+ },
534
+ {
535
+ "epoch": 6.19,
536
+ "grad_norm": 15.033493995666504,
537
+ "learning_rate": 6.357259567289082e-05,
538
+ "loss": 0.1883,
539
+ "step": 6800
540
+ },
541
+ {
542
+ "epoch": 6.28,
543
+ "grad_norm": 2.1004364490509033,
544
+ "learning_rate": 6.265310073149584e-05,
545
+ "loss": 0.1599,
546
+ "step": 6900
547
+ },
548
+ {
549
+ "epoch": 6.37,
550
+ "grad_norm": 2.42246675491333,
551
+ "learning_rate": 6.172901057692007e-05,
552
+ "loss": 0.2066,
553
+ "step": 7000
554
+ },
555
+ {
556
+ "epoch": 6.46,
557
+ "grad_norm": 5.9083147048950195,
558
+ "learning_rate": 6.0800660810005416e-05,
559
+ "loss": 0.1966,
560
+ "step": 7100
561
+ },
562
+ {
563
+ "epoch": 6.55,
564
+ "grad_norm": 2.3564085960388184,
565
+ "learning_rate": 5.9868388578552734e-05,
566
+ "loss": 0.1641,
567
+ "step": 7200
568
+ },
569
+ {
570
+ "epoch": 6.64,
571
+ "grad_norm": 0.16256462037563324,
572
+ "learning_rate": 5.893253245488015e-05,
573
+ "loss": 0.1643,
574
+ "step": 7300
575
+ },
576
+ {
577
+ "epoch": 6.73,
578
+ "grad_norm": 6.537099838256836,
579
+ "learning_rate": 5.79934323128641e-05,
580
+ "loss": 0.1932,
581
+ "step": 7400
582
+ },
583
+ {
584
+ "epoch": 6.82,
585
+ "grad_norm": 9.104645729064941,
586
+ "learning_rate": 5.705142920450777e-05,
587
+ "loss": 0.1993,
588
+ "step": 7500
589
+ },
590
+ {
591
+ "epoch": 6.92,
592
+ "grad_norm": 13.837006568908691,
593
+ "learning_rate": 5.610686523608151e-05,
594
+ "loss": 0.2141,
595
+ "step": 7600
596
+ },
597
+ {
598
+ "epoch": 7.0,
599
+ "eval_accuracy": 0.920079522862823,
600
+ "eval_loss": 0.3611968457698822,
601
+ "eval_runtime": 111.0708,
602
+ "eval_samples_per_second": 22.643,
603
+ "eval_steps_per_second": 1.423,
604
+ "step": 7693
605
+ },
606
+ {
607
+ "epoch": 7.01,
608
+ "grad_norm": 13.015904426574707,
609
+ "learning_rate": 5.516008344388053e-05,
610
+ "loss": 0.1581,
611
+ "step": 7700
612
+ },
613
+ {
614
+ "epoch": 7.1,
615
+ "grad_norm": 5.245505332946777,
616
+ "learning_rate": 5.421142766964474e-05,
617
+ "loss": 0.1366,
618
+ "step": 7800
619
+ },
620
+ {
621
+ "epoch": 7.19,
622
+ "grad_norm": 1.4385249614715576,
623
+ "learning_rate": 5.326124243568617e-05,
624
+ "loss": 0.1867,
625
+ "step": 7900
626
+ },
627
+ {
628
+ "epoch": 7.28,
629
+ "grad_norm": 1.0775200128555298,
630
+ "learning_rate": 5.230987281976901e-05,
631
+ "loss": 0.1415,
632
+ "step": 8000
633
+ },
634
+ {
635
+ "epoch": 7.37,
636
+ "grad_norm": 1.4409250020980835,
637
+ "learning_rate": 5.135766432978829e-05,
638
+ "loss": 0.2004,
639
+ "step": 8100
640
+ },
641
+ {
642
+ "epoch": 7.46,
643
+ "grad_norm": 0.05134790390729904,
644
+ "learning_rate": 5.0404962778292e-05,
645
+ "loss": 0.159,
646
+ "step": 8200
647
+ },
648
+ {
649
+ "epoch": 7.55,
650
+ "grad_norm": 7.336070537567139,
651
+ "learning_rate": 4.945211415689278e-05,
652
+ "loss": 0.1578,
653
+ "step": 8300
654
+ },
655
+ {
656
+ "epoch": 7.64,
657
+ "grad_norm": 0.08769059181213379,
658
+ "learning_rate": 4.849946451061443e-05,
659
+ "loss": 0.1883,
660
+ "step": 8400
661
+ },
662
+ {
663
+ "epoch": 7.73,
664
+ "grad_norm": 12.924542427062988,
665
+ "learning_rate": 4.754735981221927e-05,
666
+ "loss": 0.1717,
667
+ "step": 8500
668
+ },
669
+ {
670
+ "epoch": 7.83,
671
+ "grad_norm": 6.11922025680542,
672
+ "learning_rate": 4.659614583656138e-05,
673
+ "loss": 0.1557,
674
+ "step": 8600
675
+ },
676
+ {
677
+ "epoch": 7.92,
678
+ "grad_norm": 20.91149139404297,
679
+ "learning_rate": 4.564616803501205e-05,
680
+ "loss": 0.1458,
681
+ "step": 8700
682
+ },
683
+ {
684
+ "epoch": 8.0,
685
+ "eval_accuracy": 0.9304174950298211,
686
+ "eval_loss": 0.33909356594085693,
687
+ "eval_runtime": 111.9164,
688
+ "eval_samples_per_second": 22.472,
689
+ "eval_steps_per_second": 1.412,
690
+ "step": 8792
691
+ },
692
+ {
693
+ "epoch": 8.01,
694
+ "grad_norm": 2.7522988319396973,
695
+ "learning_rate": 4.469777141000255e-05,
696
+ "loss": 0.1309,
697
+ "step": 8800
698
+ },
699
+ {
700
+ "epoch": 8.1,
701
+ "grad_norm": 8.19546890258789,
702
+ "learning_rate": 4.375130038972988e-05,
703
+ "loss": 0.1502,
704
+ "step": 8900
705
+ },
706
+ {
707
+ "epoch": 8.19,
708
+ "grad_norm": 3.5378799438476562,
709
+ "learning_rate": 4.2807098703071255e-05,
710
+ "loss": 0.1528,
711
+ "step": 9000
712
+ },
713
+ {
714
+ "epoch": 8.28,
715
+ "grad_norm": 8.663042068481445,
716
+ "learning_rate": 4.18655092547524e-05,
717
+ "loss": 0.0897,
718
+ "step": 9100
719
+ },
720
+ {
721
+ "epoch": 8.37,
722
+ "grad_norm": 0.019041290506720543,
723
+ "learning_rate": 4.092687400081522e-05,
724
+ "loss": 0.16,
725
+ "step": 9200
726
+ },
727
+ {
728
+ "epoch": 8.46,
729
+ "grad_norm": 0.053332068026065826,
730
+ "learning_rate": 3.999153382442995e-05,
731
+ "loss": 0.1378,
732
+ "step": 9300
733
+ },
734
+ {
735
+ "epoch": 8.55,
736
+ "grad_norm": 5.864047050476074,
737
+ "learning_rate": 3.9059828412097024e-05,
738
+ "loss": 0.1176,
739
+ "step": 9400
740
+ },
741
+ {
742
+ "epoch": 8.64,
743
+ "grad_norm": 11.58228588104248,
744
+ "learning_rate": 3.8132096130283455e-05,
745
+ "loss": 0.1336,
746
+ "step": 9500
747
+ },
748
+ {
749
+ "epoch": 8.74,
750
+ "grad_norm": 0.3950587511062622,
751
+ "learning_rate": 3.7208673902538706e-05,
752
+ "loss": 0.1525,
753
+ "step": 9600
754
+ },
755
+ {
756
+ "epoch": 8.83,
757
+ "grad_norm": 9.62333869934082,
758
+ "learning_rate": 3.628989708713436e-05,
759
+ "loss": 0.1565,
760
+ "step": 9700
761
+ },
762
+ {
763
+ "epoch": 8.92,
764
+ "grad_norm": 12.212615966796875,
765
+ "learning_rate": 3.537609935527264e-05,
766
+ "loss": 0.1842,
767
+ "step": 9800
768
+ },
769
+ {
770
+ "epoch": 9.0,
771
+ "eval_accuracy": 0.932803180914513,
772
+ "eval_loss": 0.33526965975761414,
773
+ "eval_runtime": 111.0268,
774
+ "eval_samples_per_second": 22.652,
775
+ "eval_steps_per_second": 1.423,
776
+ "step": 9891
777
+ },
778
+ {
779
+ "epoch": 9.01,
780
+ "grad_norm": 2.527280807495117,
781
+ "learning_rate": 3.446761256990723e-05,
782
+ "loss": 0.1074,
783
+ "step": 9900
784
+ },
785
+ {
786
+ "epoch": 9.1,
787
+ "grad_norm": 0.02450253628194332,
788
+ "learning_rate": 3.356476666522099e-05,
789
+ "loss": 0.1001,
790
+ "step": 10000
791
+ },
792
+ {
793
+ "epoch": 9.19,
794
+ "grad_norm": 0.32657700777053833,
795
+ "learning_rate": 3.266788952680414e-05,
796
+ "loss": 0.1094,
797
+ "step": 10100
798
+ },
799
+ {
800
+ "epoch": 9.28,
801
+ "grad_norm": 2.115306854248047,
802
+ "learning_rate": 3.177730687257639e-05,
803
+ "loss": 0.1069,
804
+ "step": 10200
805
+ },
806
+ {
807
+ "epoch": 9.37,
808
+ "grad_norm": 0.252822607755661,
809
+ "learning_rate": 3.0893342134496295e-05,
810
+ "loss": 0.1074,
811
+ "step": 10300
812
+ },
813
+ {
814
+ "epoch": 9.46,
815
+ "grad_norm": 10.717082023620605,
816
+ "learning_rate": 3.0016316341100808e-05,
817
+ "loss": 0.1075,
818
+ "step": 10400
819
+ },
820
+ {
821
+ "epoch": 9.55,
822
+ "grad_norm": 0.030081748962402344,
823
+ "learning_rate": 2.914654800091768e-05,
824
+ "loss": 0.1344,
825
+ "step": 10500
826
+ },
827
+ {
828
+ "epoch": 9.65,
829
+ "grad_norm": 13.859551429748535,
830
+ "learning_rate": 2.8284352986793094e-05,
831
+ "loss": 0.1205,
832
+ "step": 10600
833
+ },
834
+ {
835
+ "epoch": 9.74,
836
+ "grad_norm": 1.8139934539794922,
837
+ "learning_rate": 2.7430044421176447e-05,
838
+ "loss": 0.1253,
839
+ "step": 10700
840
+ },
841
+ {
842
+ "epoch": 9.83,
843
+ "grad_norm": 1.271942377090454,
844
+ "learning_rate": 2.6583932562403957e-05,
845
+ "loss": 0.1226,
846
+ "step": 10800
847
+ },
848
+ {
849
+ "epoch": 9.92,
850
+ "grad_norm": 0.060803137719631195,
851
+ "learning_rate": 2.5746324692022527e-05,
852
+ "loss": 0.1037,
853
+ "step": 10900
854
+ },
855
+ {
856
+ "epoch": 10.0,
857
+ "eval_accuracy": 0.9355864811133201,
858
+ "eval_loss": 0.33833950757980347,
859
+ "eval_runtime": 111.2929,
860
+ "eval_samples_per_second": 22.598,
861
+ "eval_steps_per_second": 1.42,
862
+ "step": 10990
863
+ },
864
+ {
865
+ "epoch": 10.01,
866
+ "grad_norm": 0.0546095035970211,
867
+ "learning_rate": 2.4917525003194624e-05,
868
+ "loss": 0.1162,
869
+ "step": 11000
870
+ },
871
+ {
872
+ "epoch": 10.1,
873
+ "grad_norm": 6.095231533050537,
874
+ "learning_rate": 2.409783449022475e-05,
875
+ "loss": 0.0934,
876
+ "step": 11100
877
+ },
878
+ {
879
+ "epoch": 10.19,
880
+ "grad_norm": 0.6238592267036438,
881
+ "learning_rate": 2.3287550839247624e-05,
882
+ "loss": 0.087,
883
+ "step": 11200
884
+ },
885
+ {
886
+ "epoch": 10.28,
887
+ "grad_norm": 24.5279598236084,
888
+ "learning_rate": 2.2486968320117907e-05,
889
+ "loss": 0.0873,
890
+ "step": 11300
891
+ },
892
+ {
893
+ "epoch": 10.37,
894
+ "grad_norm": 2.309201240539551,
895
+ "learning_rate": 2.169637767954048e-05,
896
+ "loss": 0.0853,
897
+ "step": 11400
898
+ },
899
+ {
900
+ "epoch": 10.46,
901
+ "grad_norm": 0.10698918998241425,
902
+ "learning_rate": 2.091606603548029e-05,
903
+ "loss": 0.0994,
904
+ "step": 11500
905
+ },
906
+ {
907
+ "epoch": 10.56,
908
+ "grad_norm": 0.09092956781387329,
909
+ "learning_rate": 2.0146316772889983e-05,
910
+ "loss": 0.0704,
911
+ "step": 11600
912
+ },
913
+ {
914
+ "epoch": 10.65,
915
+ "grad_norm": 10.446697235107422,
916
+ "learning_rate": 1.9387409440793386e-05,
917
+ "loss": 0.0918,
918
+ "step": 11700
919
+ },
920
+ {
921
+ "epoch": 10.74,
922
+ "grad_norm": 8.398408889770508,
923
+ "learning_rate": 1.863961965076186e-05,
924
+ "loss": 0.096,
925
+ "step": 11800
926
+ },
927
+ {
928
+ "epoch": 10.83,
929
+ "grad_norm": 0.9171711802482605,
930
+ "learning_rate": 1.790321897682083e-05,
931
+ "loss": 0.1025,
932
+ "step": 11900
933
+ },
934
+ {
935
+ "epoch": 10.92,
936
+ "grad_norm": 0.2961582839488983,
937
+ "learning_rate": 1.7178474856822456e-05,
938
+ "loss": 0.0747,
939
+ "step": 12000
940
+ },
941
+ {
942
+ "epoch": 11.0,
943
+ "eval_accuracy": 0.936779324055666,
944
+ "eval_loss": 0.33453264832496643,
945
+ "eval_runtime": 109.702,
946
+ "eval_samples_per_second": 22.926,
947
+ "eval_steps_per_second": 1.44,
948
+ "step": 12089
949
+ },
950
+ {
951
+ "epoch": 11.01,
952
+ "grad_norm": 0.05641782283782959,
953
+ "learning_rate": 1.646565049532063e-05,
954
+ "loss": 0.0575,
955
+ "step": 12100
956
+ },
957
+ {
958
+ "epoch": 11.1,
959
+ "grad_norm": 0.008614973165094852,
960
+ "learning_rate": 1.576500476798311e-05,
961
+ "loss": 0.0942,
962
+ "step": 12200
963
+ },
964
+ {
965
+ "epoch": 11.19,
966
+ "grad_norm": 0.0614984966814518,
967
+ "learning_rate": 1.5076792127576073e-05,
968
+ "loss": 0.0737,
969
+ "step": 12300
970
+ },
971
+ {
972
+ "epoch": 11.28,
973
+ "grad_norm": 0.6600818037986755,
974
+ "learning_rate": 1.4401262511554642e-05,
975
+ "loss": 0.0601,
976
+ "step": 12400
977
+ },
978
+ {
979
+ "epoch": 11.37,
980
+ "grad_norm": 7.176374435424805,
981
+ "learning_rate": 1.3738661251293423e-05,
982
+ "loss": 0.0739,
983
+ "step": 12500
984
+ },
985
+ {
986
+ "epoch": 11.46,
987
+ "grad_norm": 0.5290641784667969,
988
+ "learning_rate": 1.308922898298977e-05,
989
+ "loss": 0.069,
990
+ "step": 12600
991
+ },
992
+ {
993
+ "epoch": 11.56,
994
+ "grad_norm": 0.010565654374659061,
995
+ "learning_rate": 1.2453201560272204e-05,
996
+ "loss": 0.0818,
997
+ "step": 12700
998
+ },
999
+ {
1000
+ "epoch": 11.65,
1001
+ "grad_norm": 0.48881271481513977,
1002
+ "learning_rate": 1.183080996854562e-05,
1003
+ "loss": 0.0751,
1004
+ "step": 12800
1005
+ },
1006
+ {
1007
+ "epoch": 11.74,
1008
+ "grad_norm": 0.5463552474975586,
1009
+ "learning_rate": 1.1222280241104716e-05,
1010
+ "loss": 0.0684,
1011
+ "step": 12900
1012
+ },
1013
+ {
1014
+ "epoch": 11.83,
1015
+ "grad_norm": 5.191521167755127,
1016
+ "learning_rate": 1.062783337704557e-05,
1017
+ "loss": 0.1047,
1018
+ "step": 13000
1019
+ },
1020
+ {
1021
+ "epoch": 11.92,
1022
+ "grad_norm": 0.0762176662683487,
1023
+ "learning_rate": 1.0047685261005707e-05,
1024
+ "loss": 0.0912,
1025
+ "step": 13100
1026
+ },
1027
+ {
1028
+ "epoch": 12.0,
1029
+ "eval_accuracy": 0.9391650099403579,
1030
+ "eval_loss": 0.3244304656982422,
1031
+ "eval_runtime": 112.1314,
1032
+ "eval_samples_per_second": 22.429,
1033
+ "eval_steps_per_second": 1.409,
1034
+ "step": 13188
1035
+ },
1036
+ {
1037
+ "epoch": 12.01,
1038
+ "grad_norm": 0.017082059755921364,
1039
+ "learning_rate": 9.482046584761495e-06,
1040
+ "loss": 0.0909,
1041
+ "step": 13200
1042
+ },
1043
+ {
1044
+ "epoch": 12.1,
1045
+ "grad_norm": 16.901845932006836,
1046
+ "learning_rate": 8.931122770711425e-06,
1047
+ "loss": 0.0827,
1048
+ "step": 13300
1049
+ },
1050
+ {
1051
+ "epoch": 12.19,
1052
+ "grad_norm": 0.1826079934835434,
1053
+ "learning_rate": 8.395113897273105e-06,
1054
+ "loss": 0.086,
1055
+ "step": 13400
1056
+ },
1057
+ {
1058
+ "epoch": 12.28,
1059
+ "grad_norm": 0.14454180002212524,
1060
+ "learning_rate": 7.874214626220899e-06,
1061
+ "loss": 0.0624,
1062
+ "step": 13500
1063
+ },
1064
+ {
1065
+ "epoch": 12.37,
1066
+ "grad_norm": 3.5157103538513184,
1067
+ "learning_rate": 7.368614131990986e-06,
1068
+ "loss": 0.1045,
1069
+ "step": 13600
1070
+ },
1071
+ {
1072
+ "epoch": 12.47,
1073
+ "grad_norm": 0.0011518648825585842,
1074
+ "learning_rate": 6.8784960329789264e-06,
1075
+ "loss": 0.0802,
1076
+ "step": 13700
1077
+ },
1078
+ {
1079
+ "epoch": 12.56,
1080
+ "grad_norm": 1.2086189985275269,
1081
+ "learning_rate": 6.404038324855222e-06,
1082
+ "loss": 0.0515,
1083
+ "step": 13800
1084
+ },
1085
+ {
1086
+ "epoch": 12.65,
1087
+ "grad_norm": 5.540249824523926,
1088
+ "learning_rate": 5.945413315922826e-06,
1089
+ "loss": 0.0687,
1090
+ "step": 13900
1091
+ },
1092
+ {
1093
+ "epoch": 12.74,
1094
+ "grad_norm": 9.239886283874512,
1095
+ "learning_rate": 5.5027875645401015e-06,
1096
+ "loss": 0.0625,
1097
+ "step": 14000
1098
+ },
1099
+ {
1100
+ "epoch": 12.83,
1101
+ "grad_norm": 0.017092958092689514,
1102
+ "learning_rate": 5.076321818632018e-06,
1103
+ "loss": 0.1087,
1104
+ "step": 14100
1105
+ },
1106
+ {
1107
+ "epoch": 12.92,
1108
+ "grad_norm": 9.08121109008789,
1109
+ "learning_rate": 4.666170957311472e-06,
1110
+ "loss": 0.0733,
1111
+ "step": 14200
1112
+ },
1113
+ {
1114
+ "epoch": 13.0,
1115
+ "eval_accuracy": 0.9407554671968191,
1116
+ "eval_loss": 0.32192325592041016,
1117
+ "eval_runtime": 110.2447,
1118
+ "eval_samples_per_second": 22.813,
1119
+ "eval_steps_per_second": 1.433,
1120
+ "step": 14287
1121
+ },
1122
+ {
1123
+ "epoch": 13.01,
1124
+ "grad_norm": 0.021881932392716408,
1125
+ "learning_rate": 4.272483934632021e-06,
1126
+ "loss": 0.0548,
1127
+ "step": 14300
1128
+ },
1129
+ {
1130
+ "epoch": 13.1,
1131
+ "grad_norm": 0.6576229929924011,
1132
+ "learning_rate": 3.895403725492402e-06,
1133
+ "loss": 0.059,
1134
+ "step": 14400
1135
+ },
1136
+ {
1137
+ "epoch": 13.19,
1138
+ "grad_norm": 5.590795516967773,
1139
+ "learning_rate": 3.5350672737124725e-06,
1140
+ "loss": 0.092,
1141
+ "step": 14500
1142
+ },
1143
+ {
1144
+ "epoch": 13.28,
1145
+ "grad_norm": 0.2879085838794708,
1146
+ "learning_rate": 3.1916054422994834e-06,
1147
+ "loss": 0.0642,
1148
+ "step": 14600
1149
+ },
1150
+ {
1151
+ "epoch": 13.38,
1152
+ "grad_norm": 0.1755078285932541,
1153
+ "learning_rate": 2.86514296592269e-06,
1154
+ "loss": 0.069,
1155
+ "step": 14700
1156
+ },
1157
+ {
1158
+ "epoch": 13.47,
1159
+ "grad_norm": 11.491580963134766,
1160
+ "learning_rate": 2.5557984056135964e-06,
1161
+ "loss": 0.0692,
1162
+ "step": 14800
1163
+ },
1164
+ {
1165
+ "epoch": 13.56,
1166
+ "grad_norm": 0.22245600819587708,
1167
+ "learning_rate": 2.263684105708275e-06,
1168
+ "loss": 0.0843,
1169
+ "step": 14900
1170
+ },
1171
+ {
1172
+ "epoch": 13.65,
1173
+ "grad_norm": 19.67642593383789,
1174
+ "learning_rate": 1.9889061530473986e-06,
1175
+ "loss": 0.058,
1176
+ "step": 15000
1177
+ },
1178
+ {
1179
+ "epoch": 13.74,
1180
+ "grad_norm": 6.375917434692383,
1181
+ "learning_rate": 1.7315643384487713e-06,
1182
+ "loss": 0.0828,
1183
+ "step": 15100
1184
+ },
1185
+ {
1186
+ "epoch": 13.83,
1187
+ "grad_norm": 0.5529243350028992,
1188
+ "learning_rate": 1.4917521204664331e-06,
1189
+ "loss": 0.0743,
1190
+ "step": 15200
1191
+ },
1192
+ {
1193
+ "epoch": 13.92,
1194
+ "grad_norm": 6.179434299468994,
1195
+ "learning_rate": 1.269556591449389e-06,
1196
+ "loss": 0.0667,
1197
+ "step": 15300
1198
+ },
1199
+ {
1200
+ "epoch": 14.0,
1201
+ "eval_accuracy": 0.9435387673956263,
1202
+ "eval_loss": 0.31903526186943054,
1203
+ "eval_runtime": 110.0813,
1204
+ "eval_samples_per_second": 22.847,
1205
+ "eval_steps_per_second": 1.435,
1206
+ "step": 15386
1207
+ },
1208
+ {
1209
+ "epoch": 14.01,
1210
+ "grad_norm": 2.8753578662872314,
1211
+ "learning_rate": 1.065058445912398e-06,
1212
+ "loss": 0.0588,
1213
+ "step": 15400
1214
+ },
1215
+ {
1216
+ "epoch": 14.1,
1217
+ "grad_norm": 0.04811515659093857,
1218
+ "learning_rate": 8.783319512302102e-07,
1219
+ "loss": 0.0608,
1220
+ "step": 15500
1221
+ },
1222
+ {
1223
+ "epoch": 14.19,
1224
+ "grad_norm": 7.1183762550354,
1225
+ "learning_rate": 7.094449206659748e-07,
1226
+ "loss": 0.0768,
1227
+ "step": 15600
1228
+ },
1229
+ {
1230
+ "epoch": 14.29,
1231
+ "grad_norm": 0.26764577627182007,
1232
+ "learning_rate": 5.584586887435739e-07,
1233
+ "loss": 0.0713,
1234
+ "step": 15700
1235
+ },
1236
+ {
1237
+ "epoch": 14.38,
1238
+ "grad_norm": 5.16556978225708,
1239
+ "learning_rate": 4.254280889728068e-07,
1240
+ "loss": 0.0687,
1241
+ "step": 15800
1242
+ },
1243
+ {
1244
+ "epoch": 14.47,
1245
+ "grad_norm": 0.2050618976354599,
1246
+ "learning_rate": 3.104014339355921e-07,
1247
+ "loss": 0.0591,
1248
+ "step": 15900
1249
+ },
1250
+ {
1251
+ "epoch": 14.56,
1252
+ "grad_norm": 5.02229642868042,
1253
+ "learning_rate": 2.1342049774030758e-07,
1254
+ "loss": 0.0687,
1255
+ "step": 16000
1256
+ },
1257
+ {
1258
+ "epoch": 14.65,
1259
+ "grad_norm": 0.027897778898477554,
1260
+ "learning_rate": 1.3452050085075442e-07,
1261
+ "loss": 0.0534,
1262
+ "step": 16100
1263
+ },
1264
+ {
1265
+ "epoch": 14.74,
1266
+ "grad_norm": 14.302966117858887,
1267
+ "learning_rate": 7.37300972951771e-08,
1268
+ "loss": 0.0811,
1269
+ "step": 16200
1270
+ },
1271
+ {
1272
+ "epoch": 14.83,
1273
+ "grad_norm": 0.017526021227240562,
1274
+ "learning_rate": 3.107136425999912e-08,
1275
+ "loss": 0.0723,
1276
+ "step": 16300
1277
+ },
1278
+ {
1279
+ "epoch": 14.92,
1280
+ "grad_norm": 4.576681137084961,
1281
+ "learning_rate": 6.559794072080738e-09,
1282
+ "loss": 0.0694,
1283
+ "step": 16400
1284
+ },
1285
+ {
1286
+ "epoch": 15.0,
1287
+ "eval_accuracy": 0.9431411530815109,
1288
+ "eval_loss": 0.31916290521621704,
1289
+ "eval_runtime": 106.1086,
1290
+ "eval_samples_per_second": 23.702,
1291
+ "eval_steps_per_second": 1.489,
1292
+ "step": 16485
1293
+ },
1294
+ {
1295
+ "epoch": 15.0,
1296
+ "step": 16485,
1297
+ "total_flos": 6.140249030814106e+19,
1298
+ "train_loss": 0.2272892904238228,
1299
+ "train_runtime": 26538.9058,
1300
+ "train_samples_per_second": 9.937,
1301
+ "train_steps_per_second": 0.621
1302
+ }
1303
+ ],
1304
+ "logging_steps": 100,
1305
+ "max_steps": 16485,
1306
+ "num_input_tokens_seen": 0,
1307
+ "num_train_epochs": 15,
1308
+ "save_steps": 500,
1309
+ "total_flos": 6.140249030814106e+19,
1310
+ "train_batch_size": 16,
1311
+ "trial_name": null,
1312
+ "trial_params": null
1313
+ }