Augusto777 commited on
Commit
183dbd0
1 Parent(s): f226664

End of training

Browse files
README.md CHANGED
@@ -22,7 +22,7 @@ model-index:
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
- value: 0.7608695652173914
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 [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the imagefolder dataset.
34
  It achieves the following results on the evaluation set:
35
- - Loss: 1.1777
36
- - Accuracy: 0.7609
37
 
38
  ## Model description
39
 
 
22
  metrics:
23
  - name: Accuracy
24
  type: accuracy
25
+ value: 0.9347826086956522
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 [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the imagefolder dataset.
34
  It achieves the following results on the evaluation set:
35
+ - Loss: 0.3534
36
+ - Accuracy: 0.9348
37
 
38
  ## Model description
39
 
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.61,
3
+ "eval_accuracy": 0.9347826086956522,
4
+ "eval_loss": 0.35338208079338074,
5
+ "eval_runtime": 0.9809,
6
+ "eval_samples_per_second": 46.895,
7
+ "eval_steps_per_second": 6.117,
8
+ "train_loss": 0.19644491698172892,
9
+ "train_runtime": 666.0117,
10
+ "train_samples_per_second": 49.188,
11
+ "train_steps_per_second": 3.063
12
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.61,
3
+ "eval_accuracy": 0.9347826086956522,
4
+ "eval_loss": 0.35338208079338074,
5
+ "eval_runtime": 0.9809,
6
+ "eval_samples_per_second": 46.895,
7
+ "eval_steps_per_second": 6.117
8
+ }
runs/Jun23_14-03-26_DESKTOP-SKBE9FB/events.out.tfevents.1719173699.DESKTOP-SKBE9FB.3876.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a5379a3889a5a6de9265f1fc1979ab1f64d14fd02d41b1931641b2a00d0d8054
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 39.61,
3
+ "train_loss": 0.19644491698172892,
4
+ "train_runtime": 666.0117,
5
+ "train_samples_per_second": 49.188,
6
+ "train_steps_per_second": 3.063
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1614 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.9347826086956522,
3
+ "best_model_checkpoint": "vit-base-patch16-224-ve-U13b-R\\checkpoint-360",
4
+ "epoch": 39.61165048543689,
5
+ "eval_steps": 500,
6
+ "global_step": 2040,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.19,
13
+ "learning_rate": 5.392156862745098e-06,
14
+ "loss": 1.3862,
15
+ "step": 10
16
+ },
17
+ {
18
+ "epoch": 0.39,
19
+ "learning_rate": 1.0784313725490196e-05,
20
+ "loss": 1.3853,
21
+ "step": 20
22
+ },
23
+ {
24
+ "epoch": 0.58,
25
+ "learning_rate": 1.6176470588235296e-05,
26
+ "loss": 1.3787,
27
+ "step": 30
28
+ },
29
+ {
30
+ "epoch": 0.78,
31
+ "learning_rate": 2.156862745098039e-05,
32
+ "loss": 1.3618,
33
+ "step": 40
34
+ },
35
+ {
36
+ "epoch": 0.97,
37
+ "learning_rate": 2.696078431372549e-05,
38
+ "loss": 1.3157,
39
+ "step": 50
40
+ },
41
+ {
42
+ "epoch": 0.99,
43
+ "eval_accuracy": 0.34782608695652173,
44
+ "eval_loss": 1.296690583229065,
45
+ "eval_runtime": 0.977,
46
+ "eval_samples_per_second": 47.083,
47
+ "eval_steps_per_second": 6.141,
48
+ "step": 51
49
+ },
50
+ {
51
+ "epoch": 1.17,
52
+ "learning_rate": 3.235294117647059e-05,
53
+ "loss": 1.2679,
54
+ "step": 60
55
+ },
56
+ {
57
+ "epoch": 1.36,
58
+ "learning_rate": 3.774509803921569e-05,
59
+ "loss": 1.1855,
60
+ "step": 70
61
+ },
62
+ {
63
+ "epoch": 1.55,
64
+ "learning_rate": 4.313725490196078e-05,
65
+ "loss": 1.1615,
66
+ "step": 80
67
+ },
68
+ {
69
+ "epoch": 1.75,
70
+ "learning_rate": 4.8529411764705885e-05,
71
+ "loss": 1.0505,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 1.94,
76
+ "learning_rate": 5.392156862745098e-05,
77
+ "loss": 0.9801,
78
+ "step": 100
79
+ },
80
+ {
81
+ "epoch": 2.0,
82
+ "eval_accuracy": 0.5869565217391305,
83
+ "eval_loss": 0.9965764880180359,
84
+ "eval_runtime": 0.7837,
85
+ "eval_samples_per_second": 58.697,
86
+ "eval_steps_per_second": 7.656,
87
+ "step": 103
88
+ },
89
+ {
90
+ "epoch": 2.14,
91
+ "learning_rate": 5.477296181630547e-05,
92
+ "loss": 0.95,
93
+ "step": 110
94
+ },
95
+ {
96
+ "epoch": 2.33,
97
+ "learning_rate": 5.448916408668731e-05,
98
+ "loss": 0.9522,
99
+ "step": 120
100
+ },
101
+ {
102
+ "epoch": 2.52,
103
+ "learning_rate": 5.4205366357069146e-05,
104
+ "loss": 0.8341,
105
+ "step": 130
106
+ },
107
+ {
108
+ "epoch": 2.72,
109
+ "learning_rate": 5.392156862745098e-05,
110
+ "loss": 0.7555,
111
+ "step": 140
112
+ },
113
+ {
114
+ "epoch": 2.91,
115
+ "learning_rate": 5.363777089783282e-05,
116
+ "loss": 0.7385,
117
+ "step": 150
118
+ },
119
+ {
120
+ "epoch": 2.99,
121
+ "eval_accuracy": 0.717391304347826,
122
+ "eval_loss": 0.7600191235542297,
123
+ "eval_runtime": 0.7802,
124
+ "eval_samples_per_second": 58.961,
125
+ "eval_steps_per_second": 7.691,
126
+ "step": 154
127
+ },
128
+ {
129
+ "epoch": 3.11,
130
+ "learning_rate": 5.3353973168214655e-05,
131
+ "loss": 0.7079,
132
+ "step": 160
133
+ },
134
+ {
135
+ "epoch": 3.3,
136
+ "learning_rate": 5.3070175438596496e-05,
137
+ "loss": 0.5291,
138
+ "step": 170
139
+ },
140
+ {
141
+ "epoch": 3.5,
142
+ "learning_rate": 5.278637770897833e-05,
143
+ "loss": 0.5315,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 3.69,
148
+ "learning_rate": 5.2502579979360165e-05,
149
+ "loss": 0.5752,
150
+ "step": 190
151
+ },
152
+ {
153
+ "epoch": 3.88,
154
+ "learning_rate": 5.2218782249742006e-05,
155
+ "loss": 0.572,
156
+ "step": 200
157
+ },
158
+ {
159
+ "epoch": 4.0,
160
+ "eval_accuracy": 0.782608695652174,
161
+ "eval_loss": 0.6425113677978516,
162
+ "eval_runtime": 0.7782,
163
+ "eval_samples_per_second": 59.112,
164
+ "eval_steps_per_second": 7.71,
165
+ "step": 206
166
+ },
167
+ {
168
+ "epoch": 4.08,
169
+ "learning_rate": 5.193498452012384e-05,
170
+ "loss": 0.4306,
171
+ "step": 210
172
+ },
173
+ {
174
+ "epoch": 4.27,
175
+ "learning_rate": 5.165118679050568e-05,
176
+ "loss": 0.4744,
177
+ "step": 220
178
+ },
179
+ {
180
+ "epoch": 4.47,
181
+ "learning_rate": 5.1367389060887515e-05,
182
+ "loss": 0.4993,
183
+ "step": 230
184
+ },
185
+ {
186
+ "epoch": 4.66,
187
+ "learning_rate": 5.108359133126935e-05,
188
+ "loss": 0.3998,
189
+ "step": 240
190
+ },
191
+ {
192
+ "epoch": 4.85,
193
+ "learning_rate": 5.079979360165119e-05,
194
+ "loss": 0.3646,
195
+ "step": 250
196
+ },
197
+ {
198
+ "epoch": 4.99,
199
+ "eval_accuracy": 0.6956521739130435,
200
+ "eval_loss": 0.768746554851532,
201
+ "eval_runtime": 0.8062,
202
+ "eval_samples_per_second": 57.057,
203
+ "eval_steps_per_second": 7.442,
204
+ "step": 257
205
+ },
206
+ {
207
+ "epoch": 5.05,
208
+ "learning_rate": 5.0515995872033025e-05,
209
+ "loss": 0.327,
210
+ "step": 260
211
+ },
212
+ {
213
+ "epoch": 5.24,
214
+ "learning_rate": 5.023219814241486e-05,
215
+ "loss": 0.3172,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 5.44,
220
+ "learning_rate": 4.99484004127967e-05,
221
+ "loss": 0.3136,
222
+ "step": 280
223
+ },
224
+ {
225
+ "epoch": 5.63,
226
+ "learning_rate": 4.9664602683178534e-05,
227
+ "loss": 0.4175,
228
+ "step": 290
229
+ },
230
+ {
231
+ "epoch": 5.83,
232
+ "learning_rate": 4.9380804953560375e-05,
233
+ "loss": 0.3033,
234
+ "step": 300
235
+ },
236
+ {
237
+ "epoch": 6.0,
238
+ "eval_accuracy": 0.7391304347826086,
239
+ "eval_loss": 0.6335754990577698,
240
+ "eval_runtime": 0.7852,
241
+ "eval_samples_per_second": 58.586,
242
+ "eval_steps_per_second": 7.642,
243
+ "step": 309
244
+ },
245
+ {
246
+ "epoch": 6.02,
247
+ "learning_rate": 4.909700722394221e-05,
248
+ "loss": 0.2792,
249
+ "step": 310
250
+ },
251
+ {
252
+ "epoch": 6.21,
253
+ "learning_rate": 4.8813209494324044e-05,
254
+ "loss": 0.2537,
255
+ "step": 320
256
+ },
257
+ {
258
+ "epoch": 6.41,
259
+ "learning_rate": 4.8529411764705885e-05,
260
+ "loss": 0.3938,
261
+ "step": 330
262
+ },
263
+ {
264
+ "epoch": 6.6,
265
+ "learning_rate": 4.824561403508772e-05,
266
+ "loss": 0.3292,
267
+ "step": 340
268
+ },
269
+ {
270
+ "epoch": 6.8,
271
+ "learning_rate": 4.796181630546956e-05,
272
+ "loss": 0.2956,
273
+ "step": 350
274
+ },
275
+ {
276
+ "epoch": 6.99,
277
+ "learning_rate": 4.7678018575851394e-05,
278
+ "loss": 0.3073,
279
+ "step": 360
280
+ },
281
+ {
282
+ "epoch": 6.99,
283
+ "eval_accuracy": 0.9347826086956522,
284
+ "eval_loss": 0.35338208079338074,
285
+ "eval_runtime": 0.7677,
286
+ "eval_samples_per_second": 59.921,
287
+ "eval_steps_per_second": 7.816,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 7.18,
292
+ "learning_rate": 4.739422084623323e-05,
293
+ "loss": 0.2049,
294
+ "step": 370
295
+ },
296
+ {
297
+ "epoch": 7.38,
298
+ "learning_rate": 4.711042311661507e-05,
299
+ "loss": 0.2429,
300
+ "step": 380
301
+ },
302
+ {
303
+ "epoch": 7.57,
304
+ "learning_rate": 4.6826625386996904e-05,
305
+ "loss": 0.2084,
306
+ "step": 390
307
+ },
308
+ {
309
+ "epoch": 7.77,
310
+ "learning_rate": 4.6542827657378745e-05,
311
+ "loss": 0.3038,
312
+ "step": 400
313
+ },
314
+ {
315
+ "epoch": 7.96,
316
+ "learning_rate": 4.625902992776058e-05,
317
+ "loss": 0.1623,
318
+ "step": 410
319
+ },
320
+ {
321
+ "epoch": 8.0,
322
+ "eval_accuracy": 0.6739130434782609,
323
+ "eval_loss": 0.8559263348579407,
324
+ "eval_runtime": 0.8107,
325
+ "eval_samples_per_second": 56.742,
326
+ "eval_steps_per_second": 7.401,
327
+ "step": 412
328
+ },
329
+ {
330
+ "epoch": 8.16,
331
+ "learning_rate": 4.597523219814241e-05,
332
+ "loss": 0.1462,
333
+ "step": 420
334
+ },
335
+ {
336
+ "epoch": 8.35,
337
+ "learning_rate": 4.5691434468524254e-05,
338
+ "loss": 0.124,
339
+ "step": 430
340
+ },
341
+ {
342
+ "epoch": 8.54,
343
+ "learning_rate": 4.540763673890609e-05,
344
+ "loss": 0.1477,
345
+ "step": 440
346
+ },
347
+ {
348
+ "epoch": 8.74,
349
+ "learning_rate": 4.512383900928793e-05,
350
+ "loss": 0.1319,
351
+ "step": 450
352
+ },
353
+ {
354
+ "epoch": 8.93,
355
+ "learning_rate": 4.4840041279669764e-05,
356
+ "loss": 0.1079,
357
+ "step": 460
358
+ },
359
+ {
360
+ "epoch": 8.99,
361
+ "eval_accuracy": 0.7391304347826086,
362
+ "eval_loss": 0.973010778427124,
363
+ "eval_runtime": 0.8122,
364
+ "eval_samples_per_second": 56.638,
365
+ "eval_steps_per_second": 7.388,
366
+ "step": 463
367
+ },
368
+ {
369
+ "epoch": 9.13,
370
+ "learning_rate": 4.45562435500516e-05,
371
+ "loss": 0.1712,
372
+ "step": 470
373
+ },
374
+ {
375
+ "epoch": 9.32,
376
+ "learning_rate": 4.427244582043344e-05,
377
+ "loss": 0.1125,
378
+ "step": 480
379
+ },
380
+ {
381
+ "epoch": 9.51,
382
+ "learning_rate": 4.398864809081527e-05,
383
+ "loss": 0.1858,
384
+ "step": 490
385
+ },
386
+ {
387
+ "epoch": 9.71,
388
+ "learning_rate": 4.3704850361197114e-05,
389
+ "loss": 0.1768,
390
+ "step": 500
391
+ },
392
+ {
393
+ "epoch": 9.9,
394
+ "learning_rate": 4.342105263157895e-05,
395
+ "loss": 0.2703,
396
+ "step": 510
397
+ },
398
+ {
399
+ "epoch": 10.0,
400
+ "eval_accuracy": 0.8043478260869565,
401
+ "eval_loss": 0.7767899632453918,
402
+ "eval_runtime": 0.7982,
403
+ "eval_samples_per_second": 57.63,
404
+ "eval_steps_per_second": 7.517,
405
+ "step": 515
406
+ },
407
+ {
408
+ "epoch": 10.1,
409
+ "learning_rate": 4.313725490196078e-05,
410
+ "loss": 0.0645,
411
+ "step": 520
412
+ },
413
+ {
414
+ "epoch": 10.29,
415
+ "learning_rate": 4.2853457172342624e-05,
416
+ "loss": 0.1355,
417
+ "step": 530
418
+ },
419
+ {
420
+ "epoch": 10.49,
421
+ "learning_rate": 4.2569659442724465e-05,
422
+ "loss": 0.2119,
423
+ "step": 540
424
+ },
425
+ {
426
+ "epoch": 10.68,
427
+ "learning_rate": 4.22858617131063e-05,
428
+ "loss": 0.2935,
429
+ "step": 550
430
+ },
431
+ {
432
+ "epoch": 10.87,
433
+ "learning_rate": 4.200206398348813e-05,
434
+ "loss": 0.178,
435
+ "step": 560
436
+ },
437
+ {
438
+ "epoch": 10.99,
439
+ "eval_accuracy": 0.782608695652174,
440
+ "eval_loss": 0.8520329594612122,
441
+ "eval_runtime": 0.7822,
442
+ "eval_samples_per_second": 58.809,
443
+ "eval_steps_per_second": 7.671,
444
+ "step": 566
445
+ },
446
+ {
447
+ "epoch": 11.07,
448
+ "learning_rate": 4.171826625386997e-05,
449
+ "loss": 0.1782,
450
+ "step": 570
451
+ },
452
+ {
453
+ "epoch": 11.26,
454
+ "learning_rate": 4.143446852425181e-05,
455
+ "loss": 0.1341,
456
+ "step": 580
457
+ },
458
+ {
459
+ "epoch": 11.46,
460
+ "learning_rate": 4.115067079463365e-05,
461
+ "loss": 0.0238,
462
+ "step": 590
463
+ },
464
+ {
465
+ "epoch": 11.65,
466
+ "learning_rate": 4.0866873065015484e-05,
467
+ "loss": 0.1414,
468
+ "step": 600
469
+ },
470
+ {
471
+ "epoch": 11.84,
472
+ "learning_rate": 4.058307533539732e-05,
473
+ "loss": 0.2191,
474
+ "step": 610
475
+ },
476
+ {
477
+ "epoch": 12.0,
478
+ "eval_accuracy": 0.7391304347826086,
479
+ "eval_loss": 1.0049434900283813,
480
+ "eval_runtime": 0.7802,
481
+ "eval_samples_per_second": 58.961,
482
+ "eval_steps_per_second": 7.691,
483
+ "step": 618
484
+ },
485
+ {
486
+ "epoch": 12.04,
487
+ "learning_rate": 4.029927760577915e-05,
488
+ "loss": 0.1217,
489
+ "step": 620
490
+ },
491
+ {
492
+ "epoch": 12.23,
493
+ "learning_rate": 4.001547987616099e-05,
494
+ "loss": 0.1584,
495
+ "step": 630
496
+ },
497
+ {
498
+ "epoch": 12.43,
499
+ "learning_rate": 3.9731682146542834e-05,
500
+ "loss": 0.1076,
501
+ "step": 640
502
+ },
503
+ {
504
+ "epoch": 12.62,
505
+ "learning_rate": 3.944788441692467e-05,
506
+ "loss": 0.0764,
507
+ "step": 650
508
+ },
509
+ {
510
+ "epoch": 12.82,
511
+ "learning_rate": 3.91640866873065e-05,
512
+ "loss": 0.0597,
513
+ "step": 660
514
+ },
515
+ {
516
+ "epoch": 12.99,
517
+ "eval_accuracy": 0.7608695652173914,
518
+ "eval_loss": 0.833352267742157,
519
+ "eval_runtime": 0.7732,
520
+ "eval_samples_per_second": 59.495,
521
+ "eval_steps_per_second": 7.76,
522
+ "step": 669
523
+ },
524
+ {
525
+ "epoch": 13.01,
526
+ "learning_rate": 3.888028895768834e-05,
527
+ "loss": 0.071,
528
+ "step": 670
529
+ },
530
+ {
531
+ "epoch": 13.2,
532
+ "learning_rate": 3.859649122807018e-05,
533
+ "loss": 0.0797,
534
+ "step": 680
535
+ },
536
+ {
537
+ "epoch": 13.4,
538
+ "learning_rate": 3.831269349845202e-05,
539
+ "loss": 0.1046,
540
+ "step": 690
541
+ },
542
+ {
543
+ "epoch": 13.59,
544
+ "learning_rate": 3.802889576883385e-05,
545
+ "loss": 0.1188,
546
+ "step": 700
547
+ },
548
+ {
549
+ "epoch": 13.79,
550
+ "learning_rate": 3.774509803921569e-05,
551
+ "loss": 0.0627,
552
+ "step": 710
553
+ },
554
+ {
555
+ "epoch": 13.98,
556
+ "learning_rate": 3.746130030959752e-05,
557
+ "loss": 0.0881,
558
+ "step": 720
559
+ },
560
+ {
561
+ "epoch": 14.0,
562
+ "eval_accuracy": 0.7608695652173914,
563
+ "eval_loss": 0.9984623193740845,
564
+ "eval_runtime": 0.7912,
565
+ "eval_samples_per_second": 58.141,
566
+ "eval_steps_per_second": 7.584,
567
+ "step": 721
568
+ },
569
+ {
570
+ "epoch": 14.17,
571
+ "learning_rate": 3.7177502579979356e-05,
572
+ "loss": 0.1591,
573
+ "step": 730
574
+ },
575
+ {
576
+ "epoch": 14.37,
577
+ "learning_rate": 3.6893704850361204e-05,
578
+ "loss": 0.1445,
579
+ "step": 740
580
+ },
581
+ {
582
+ "epoch": 14.56,
583
+ "learning_rate": 3.660990712074304e-05,
584
+ "loss": 0.0678,
585
+ "step": 750
586
+ },
587
+ {
588
+ "epoch": 14.76,
589
+ "learning_rate": 3.632610939112487e-05,
590
+ "loss": 0.0799,
591
+ "step": 760
592
+ },
593
+ {
594
+ "epoch": 14.95,
595
+ "learning_rate": 3.6042311661506706e-05,
596
+ "loss": 0.1265,
597
+ "step": 770
598
+ },
599
+ {
600
+ "epoch": 14.99,
601
+ "eval_accuracy": 0.8043478260869565,
602
+ "eval_loss": 0.9443035125732422,
603
+ "eval_runtime": 0.8091,
604
+ "eval_samples_per_second": 56.851,
605
+ "eval_steps_per_second": 7.415,
606
+ "step": 772
607
+ },
608
+ {
609
+ "epoch": 15.15,
610
+ "learning_rate": 3.575851393188854e-05,
611
+ "loss": 0.1045,
612
+ "step": 780
613
+ },
614
+ {
615
+ "epoch": 15.34,
616
+ "learning_rate": 3.547471620227039e-05,
617
+ "loss": 0.1274,
618
+ "step": 790
619
+ },
620
+ {
621
+ "epoch": 15.53,
622
+ "learning_rate": 3.519091847265222e-05,
623
+ "loss": 0.105,
624
+ "step": 800
625
+ },
626
+ {
627
+ "epoch": 15.73,
628
+ "learning_rate": 3.490712074303406e-05,
629
+ "loss": 0.1277,
630
+ "step": 810
631
+ },
632
+ {
633
+ "epoch": 15.92,
634
+ "learning_rate": 3.462332301341589e-05,
635
+ "loss": 0.0696,
636
+ "step": 820
637
+ },
638
+ {
639
+ "epoch": 16.0,
640
+ "eval_accuracy": 0.8260869565217391,
641
+ "eval_loss": 0.9877853393554688,
642
+ "eval_runtime": 0.808,
643
+ "eval_samples_per_second": 56.928,
644
+ "eval_steps_per_second": 7.425,
645
+ "step": 824
646
+ },
647
+ {
648
+ "epoch": 16.12,
649
+ "learning_rate": 3.4339525283797725e-05,
650
+ "loss": 0.1507,
651
+ "step": 830
652
+ },
653
+ {
654
+ "epoch": 16.31,
655
+ "learning_rate": 3.405572755417957e-05,
656
+ "loss": 0.1025,
657
+ "step": 840
658
+ },
659
+ {
660
+ "epoch": 16.5,
661
+ "learning_rate": 3.377192982456141e-05,
662
+ "loss": 0.0113,
663
+ "step": 850
664
+ },
665
+ {
666
+ "epoch": 16.7,
667
+ "learning_rate": 3.348813209494324e-05,
668
+ "loss": 0.0393,
669
+ "step": 860
670
+ },
671
+ {
672
+ "epoch": 16.89,
673
+ "learning_rate": 3.3204334365325076e-05,
674
+ "loss": 0.1198,
675
+ "step": 870
676
+ },
677
+ {
678
+ "epoch": 16.99,
679
+ "eval_accuracy": 0.8043478260869565,
680
+ "eval_loss": 0.8784380555152893,
681
+ "eval_runtime": 0.8375,
682
+ "eval_samples_per_second": 54.923,
683
+ "eval_steps_per_second": 7.164,
684
+ "step": 875
685
+ },
686
+ {
687
+ "epoch": 17.09,
688
+ "learning_rate": 3.292053663570691e-05,
689
+ "loss": 0.1388,
690
+ "step": 880
691
+ },
692
+ {
693
+ "epoch": 17.28,
694
+ "learning_rate": 3.263673890608876e-05,
695
+ "loss": 0.1075,
696
+ "step": 890
697
+ },
698
+ {
699
+ "epoch": 17.48,
700
+ "learning_rate": 3.235294117647059e-05,
701
+ "loss": 0.0449,
702
+ "step": 900
703
+ },
704
+ {
705
+ "epoch": 17.67,
706
+ "learning_rate": 3.2069143446852426e-05,
707
+ "loss": 0.0916,
708
+ "step": 910
709
+ },
710
+ {
711
+ "epoch": 17.86,
712
+ "learning_rate": 3.178534571723426e-05,
713
+ "loss": 0.1484,
714
+ "step": 920
715
+ },
716
+ {
717
+ "epoch": 18.0,
718
+ "eval_accuracy": 0.7608695652173914,
719
+ "eval_loss": 0.9595004320144653,
720
+ "eval_runtime": 0.7957,
721
+ "eval_samples_per_second": 57.814,
722
+ "eval_steps_per_second": 7.541,
723
+ "step": 927
724
+ },
725
+ {
726
+ "epoch": 18.06,
727
+ "learning_rate": 3.1501547987616095e-05,
728
+ "loss": 0.0764,
729
+ "step": 930
730
+ },
731
+ {
732
+ "epoch": 18.25,
733
+ "learning_rate": 3.121775025799794e-05,
734
+ "loss": 0.0681,
735
+ "step": 940
736
+ },
737
+ {
738
+ "epoch": 18.45,
739
+ "learning_rate": 3.093395252837978e-05,
740
+ "loss": 0.064,
741
+ "step": 950
742
+ },
743
+ {
744
+ "epoch": 18.64,
745
+ "learning_rate": 3.065015479876161e-05,
746
+ "loss": 0.067,
747
+ "step": 960
748
+ },
749
+ {
750
+ "epoch": 18.83,
751
+ "learning_rate": 3.0366357069143445e-05,
752
+ "loss": 0.2887,
753
+ "step": 970
754
+ },
755
+ {
756
+ "epoch": 18.99,
757
+ "eval_accuracy": 0.8043478260869565,
758
+ "eval_loss": 1.0562622547149658,
759
+ "eval_runtime": 0.8164,
760
+ "eval_samples_per_second": 56.345,
761
+ "eval_steps_per_second": 7.349,
762
+ "step": 978
763
+ },
764
+ {
765
+ "epoch": 19.03,
766
+ "learning_rate": 3.0082559339525283e-05,
767
+ "loss": 0.1123,
768
+ "step": 980
769
+ },
770
+ {
771
+ "epoch": 19.22,
772
+ "learning_rate": 2.9798761609907124e-05,
773
+ "loss": 0.0329,
774
+ "step": 990
775
+ },
776
+ {
777
+ "epoch": 19.42,
778
+ "learning_rate": 2.9514963880288958e-05,
779
+ "loss": 0.0561,
780
+ "step": 1000
781
+ },
782
+ {
783
+ "epoch": 19.61,
784
+ "learning_rate": 2.9231166150670796e-05,
785
+ "loss": 0.1581,
786
+ "step": 1010
787
+ },
788
+ {
789
+ "epoch": 19.81,
790
+ "learning_rate": 2.894736842105263e-05,
791
+ "loss": 0.1128,
792
+ "step": 1020
793
+ },
794
+ {
795
+ "epoch": 20.0,
796
+ "learning_rate": 2.8663570691434468e-05,
797
+ "loss": 0.1423,
798
+ "step": 1030
799
+ },
800
+ {
801
+ "epoch": 20.0,
802
+ "eval_accuracy": 0.8043478260869565,
803
+ "eval_loss": 0.8549535870552063,
804
+ "eval_runtime": 0.8176,
805
+ "eval_samples_per_second": 56.263,
806
+ "eval_steps_per_second": 7.339,
807
+ "step": 1030
808
+ },
809
+ {
810
+ "epoch": 20.19,
811
+ "learning_rate": 2.837977296181631e-05,
812
+ "loss": 0.0869,
813
+ "step": 1040
814
+ },
815
+ {
816
+ "epoch": 20.39,
817
+ "learning_rate": 2.8095975232198143e-05,
818
+ "loss": 0.0625,
819
+ "step": 1050
820
+ },
821
+ {
822
+ "epoch": 20.58,
823
+ "learning_rate": 2.781217750257998e-05,
824
+ "loss": 0.136,
825
+ "step": 1060
826
+ },
827
+ {
828
+ "epoch": 20.78,
829
+ "learning_rate": 2.7528379772961815e-05,
830
+ "loss": 0.2042,
831
+ "step": 1070
832
+ },
833
+ {
834
+ "epoch": 20.97,
835
+ "learning_rate": 2.7244582043343656e-05,
836
+ "loss": 0.083,
837
+ "step": 1080
838
+ },
839
+ {
840
+ "epoch": 20.99,
841
+ "eval_accuracy": 0.782608695652174,
842
+ "eval_loss": 0.9092756509780884,
843
+ "eval_runtime": 0.8048,
844
+ "eval_samples_per_second": 57.154,
845
+ "eval_steps_per_second": 7.455,
846
+ "step": 1081
847
+ },
848
+ {
849
+ "epoch": 21.17,
850
+ "learning_rate": 2.696078431372549e-05,
851
+ "loss": 0.0607,
852
+ "step": 1090
853
+ },
854
+ {
855
+ "epoch": 21.36,
856
+ "learning_rate": 2.6676986584107328e-05,
857
+ "loss": 0.0796,
858
+ "step": 1100
859
+ },
860
+ {
861
+ "epoch": 21.55,
862
+ "learning_rate": 2.6393188854489165e-05,
863
+ "loss": 0.0582,
864
+ "step": 1110
865
+ },
866
+ {
867
+ "epoch": 21.75,
868
+ "learning_rate": 2.6109391124871003e-05,
869
+ "loss": 0.0274,
870
+ "step": 1120
871
+ },
872
+ {
873
+ "epoch": 21.94,
874
+ "learning_rate": 2.582559339525284e-05,
875
+ "loss": 0.0695,
876
+ "step": 1130
877
+ },
878
+ {
879
+ "epoch": 22.0,
880
+ "eval_accuracy": 0.6739130434782609,
881
+ "eval_loss": 1.2757601737976074,
882
+ "eval_runtime": 0.8066,
883
+ "eval_samples_per_second": 57.031,
884
+ "eval_steps_per_second": 7.439,
885
+ "step": 1133
886
+ },
887
+ {
888
+ "epoch": 22.14,
889
+ "learning_rate": 2.5541795665634675e-05,
890
+ "loss": 0.0193,
891
+ "step": 1140
892
+ },
893
+ {
894
+ "epoch": 22.33,
895
+ "learning_rate": 2.5257997936016512e-05,
896
+ "loss": 0.0856,
897
+ "step": 1150
898
+ },
899
+ {
900
+ "epoch": 22.52,
901
+ "learning_rate": 2.497420020639835e-05,
902
+ "loss": 0.0969,
903
+ "step": 1160
904
+ },
905
+ {
906
+ "epoch": 22.72,
907
+ "learning_rate": 2.4690402476780188e-05,
908
+ "loss": 0.1214,
909
+ "step": 1170
910
+ },
911
+ {
912
+ "epoch": 22.91,
913
+ "learning_rate": 2.4406604747162022e-05,
914
+ "loss": 0.0285,
915
+ "step": 1180
916
+ },
917
+ {
918
+ "epoch": 22.99,
919
+ "eval_accuracy": 0.7608695652173914,
920
+ "eval_loss": 1.0851949453353882,
921
+ "eval_runtime": 0.8033,
922
+ "eval_samples_per_second": 57.263,
923
+ "eval_steps_per_second": 7.469,
924
+ "step": 1184
925
+ },
926
+ {
927
+ "epoch": 23.11,
928
+ "learning_rate": 2.412280701754386e-05,
929
+ "loss": 0.1273,
930
+ "step": 1190
931
+ },
932
+ {
933
+ "epoch": 23.3,
934
+ "learning_rate": 2.3839009287925697e-05,
935
+ "loss": 0.0777,
936
+ "step": 1200
937
+ },
938
+ {
939
+ "epoch": 23.5,
940
+ "learning_rate": 2.3555211558307535e-05,
941
+ "loss": 0.0507,
942
+ "step": 1210
943
+ },
944
+ {
945
+ "epoch": 23.69,
946
+ "learning_rate": 2.3271413828689372e-05,
947
+ "loss": 0.1113,
948
+ "step": 1220
949
+ },
950
+ {
951
+ "epoch": 23.88,
952
+ "learning_rate": 2.2987616099071207e-05,
953
+ "loss": 0.0132,
954
+ "step": 1230
955
+ },
956
+ {
957
+ "epoch": 24.0,
958
+ "eval_accuracy": 0.6956521739130435,
959
+ "eval_loss": 1.3340740203857422,
960
+ "eval_runtime": 0.8554,
961
+ "eval_samples_per_second": 53.774,
962
+ "eval_steps_per_second": 7.014,
963
+ "step": 1236
964
+ },
965
+ {
966
+ "epoch": 24.08,
967
+ "learning_rate": 2.2703818369453044e-05,
968
+ "loss": 0.054,
969
+ "step": 1240
970
+ },
971
+ {
972
+ "epoch": 24.27,
973
+ "learning_rate": 2.2420020639834882e-05,
974
+ "loss": 0.0702,
975
+ "step": 1250
976
+ },
977
+ {
978
+ "epoch": 24.47,
979
+ "learning_rate": 2.213622291021672e-05,
980
+ "loss": 0.0461,
981
+ "step": 1260
982
+ },
983
+ {
984
+ "epoch": 24.66,
985
+ "learning_rate": 2.1852425180598557e-05,
986
+ "loss": 0.0837,
987
+ "step": 1270
988
+ },
989
+ {
990
+ "epoch": 24.85,
991
+ "learning_rate": 2.156862745098039e-05,
992
+ "loss": 0.0957,
993
+ "step": 1280
994
+ },
995
+ {
996
+ "epoch": 24.99,
997
+ "eval_accuracy": 0.7391304347826086,
998
+ "eval_loss": 1.1964709758758545,
999
+ "eval_runtime": 0.8164,
1000
+ "eval_samples_per_second": 56.344,
1001
+ "eval_steps_per_second": 7.349,
1002
+ "step": 1287
1003
+ },
1004
+ {
1005
+ "epoch": 25.05,
1006
+ "learning_rate": 2.1284829721362232e-05,
1007
+ "loss": 0.0968,
1008
+ "step": 1290
1009
+ },
1010
+ {
1011
+ "epoch": 25.24,
1012
+ "learning_rate": 2.1001031991744067e-05,
1013
+ "loss": 0.1526,
1014
+ "step": 1300
1015
+ },
1016
+ {
1017
+ "epoch": 25.44,
1018
+ "learning_rate": 2.0717234262125904e-05,
1019
+ "loss": 0.0653,
1020
+ "step": 1310
1021
+ },
1022
+ {
1023
+ "epoch": 25.63,
1024
+ "learning_rate": 2.0433436532507742e-05,
1025
+ "loss": 0.1046,
1026
+ "step": 1320
1027
+ },
1028
+ {
1029
+ "epoch": 25.83,
1030
+ "learning_rate": 2.0149638802889576e-05,
1031
+ "loss": 0.0633,
1032
+ "step": 1330
1033
+ },
1034
+ {
1035
+ "epoch": 26.0,
1036
+ "eval_accuracy": 0.7608695652173914,
1037
+ "eval_loss": 1.1199151277542114,
1038
+ "eval_runtime": 0.8717,
1039
+ "eval_samples_per_second": 52.769,
1040
+ "eval_steps_per_second": 6.883,
1041
+ "step": 1339
1042
+ },
1043
+ {
1044
+ "epoch": 26.02,
1045
+ "learning_rate": 1.9865841073271417e-05,
1046
+ "loss": 0.008,
1047
+ "step": 1340
1048
+ },
1049
+ {
1050
+ "epoch": 26.21,
1051
+ "learning_rate": 1.958204334365325e-05,
1052
+ "loss": 0.123,
1053
+ "step": 1350
1054
+ },
1055
+ {
1056
+ "epoch": 26.41,
1057
+ "learning_rate": 1.929824561403509e-05,
1058
+ "loss": 0.0715,
1059
+ "step": 1360
1060
+ },
1061
+ {
1062
+ "epoch": 26.6,
1063
+ "learning_rate": 1.9014447884416927e-05,
1064
+ "loss": 0.0765,
1065
+ "step": 1370
1066
+ },
1067
+ {
1068
+ "epoch": 26.8,
1069
+ "learning_rate": 1.873065015479876e-05,
1070
+ "loss": 0.0991,
1071
+ "step": 1380
1072
+ },
1073
+ {
1074
+ "epoch": 26.99,
1075
+ "learning_rate": 1.8446852425180602e-05,
1076
+ "loss": 0.0705,
1077
+ "step": 1390
1078
+ },
1079
+ {
1080
+ "epoch": 26.99,
1081
+ "eval_accuracy": 0.8043478260869565,
1082
+ "eval_loss": 1.0550963878631592,
1083
+ "eval_runtime": 0.8012,
1084
+ "eval_samples_per_second": 57.414,
1085
+ "eval_steps_per_second": 7.489,
1086
+ "step": 1390
1087
+ },
1088
+ {
1089
+ "epoch": 27.18,
1090
+ "learning_rate": 1.8163054695562436e-05,
1091
+ "loss": 0.0501,
1092
+ "step": 1400
1093
+ },
1094
+ {
1095
+ "epoch": 27.38,
1096
+ "learning_rate": 1.787925696594427e-05,
1097
+ "loss": 0.0841,
1098
+ "step": 1410
1099
+ },
1100
+ {
1101
+ "epoch": 27.57,
1102
+ "learning_rate": 1.759545923632611e-05,
1103
+ "loss": 0.0257,
1104
+ "step": 1420
1105
+ },
1106
+ {
1107
+ "epoch": 27.77,
1108
+ "learning_rate": 1.7311661506707946e-05,
1109
+ "loss": 0.0027,
1110
+ "step": 1430
1111
+ },
1112
+ {
1113
+ "epoch": 27.96,
1114
+ "learning_rate": 1.7027863777089787e-05,
1115
+ "loss": 0.0564,
1116
+ "step": 1440
1117
+ },
1118
+ {
1119
+ "epoch": 28.0,
1120
+ "eval_accuracy": 0.7391304347826086,
1121
+ "eval_loss": 1.4332396984100342,
1122
+ "eval_runtime": 0.8092,
1123
+ "eval_samples_per_second": 56.846,
1124
+ "eval_steps_per_second": 7.415,
1125
+ "step": 1442
1126
+ },
1127
+ {
1128
+ "epoch": 28.16,
1129
+ "learning_rate": 1.674406604747162e-05,
1130
+ "loss": 0.1458,
1131
+ "step": 1450
1132
+ },
1133
+ {
1134
+ "epoch": 28.35,
1135
+ "learning_rate": 1.6460268317853455e-05,
1136
+ "loss": 0.0965,
1137
+ "step": 1460
1138
+ },
1139
+ {
1140
+ "epoch": 28.54,
1141
+ "learning_rate": 1.6176470588235296e-05,
1142
+ "loss": 0.0443,
1143
+ "step": 1470
1144
+ },
1145
+ {
1146
+ "epoch": 28.74,
1147
+ "learning_rate": 1.589267285861713e-05,
1148
+ "loss": 0.0277,
1149
+ "step": 1480
1150
+ },
1151
+ {
1152
+ "epoch": 28.93,
1153
+ "learning_rate": 1.560887512899897e-05,
1154
+ "loss": 0.0798,
1155
+ "step": 1490
1156
+ },
1157
+ {
1158
+ "epoch": 28.99,
1159
+ "eval_accuracy": 0.7391304347826086,
1160
+ "eval_loss": 1.3854602575302124,
1161
+ "eval_runtime": 0.7817,
1162
+ "eval_samples_per_second": 58.847,
1163
+ "eval_steps_per_second": 7.676,
1164
+ "step": 1493
1165
+ },
1166
+ {
1167
+ "epoch": 29.13,
1168
+ "learning_rate": 1.5325077399380806e-05,
1169
+ "loss": 0.0701,
1170
+ "step": 1500
1171
+ },
1172
+ {
1173
+ "epoch": 29.32,
1174
+ "learning_rate": 1.5041279669762642e-05,
1175
+ "loss": 0.0716,
1176
+ "step": 1510
1177
+ },
1178
+ {
1179
+ "epoch": 29.51,
1180
+ "learning_rate": 1.4757481940144479e-05,
1181
+ "loss": 0.0463,
1182
+ "step": 1520
1183
+ },
1184
+ {
1185
+ "epoch": 29.71,
1186
+ "learning_rate": 1.4473684210526315e-05,
1187
+ "loss": 0.0273,
1188
+ "step": 1530
1189
+ },
1190
+ {
1191
+ "epoch": 29.9,
1192
+ "learning_rate": 1.4189886480908154e-05,
1193
+ "loss": 0.0326,
1194
+ "step": 1540
1195
+ },
1196
+ {
1197
+ "epoch": 30.0,
1198
+ "eval_accuracy": 0.8043478260869565,
1199
+ "eval_loss": 1.0533674955368042,
1200
+ "eval_runtime": 0.8902,
1201
+ "eval_samples_per_second": 51.674,
1202
+ "eval_steps_per_second": 6.74,
1203
+ "step": 1545
1204
+ },
1205
+ {
1206
+ "epoch": 30.1,
1207
+ "learning_rate": 1.390608875128999e-05,
1208
+ "loss": 0.0074,
1209
+ "step": 1550
1210
+ },
1211
+ {
1212
+ "epoch": 30.29,
1213
+ "learning_rate": 1.3622291021671828e-05,
1214
+ "loss": 0.0241,
1215
+ "step": 1560
1216
+ },
1217
+ {
1218
+ "epoch": 30.49,
1219
+ "learning_rate": 1.3338493292053664e-05,
1220
+ "loss": 0.089,
1221
+ "step": 1570
1222
+ },
1223
+ {
1224
+ "epoch": 30.68,
1225
+ "learning_rate": 1.3054695562435501e-05,
1226
+ "loss": 0.0401,
1227
+ "step": 1580
1228
+ },
1229
+ {
1230
+ "epoch": 30.87,
1231
+ "learning_rate": 1.2770897832817337e-05,
1232
+ "loss": 0.092,
1233
+ "step": 1590
1234
+ },
1235
+ {
1236
+ "epoch": 30.99,
1237
+ "eval_accuracy": 0.7608695652173914,
1238
+ "eval_loss": 1.17453932762146,
1239
+ "eval_runtime": 0.8884,
1240
+ "eval_samples_per_second": 51.778,
1241
+ "eval_steps_per_second": 6.754,
1242
+ "step": 1596
1243
+ },
1244
+ {
1245
+ "epoch": 31.07,
1246
+ "learning_rate": 1.2487100103199175e-05,
1247
+ "loss": 0.0336,
1248
+ "step": 1600
1249
+ },
1250
+ {
1251
+ "epoch": 31.26,
1252
+ "learning_rate": 1.2203302373581011e-05,
1253
+ "loss": 0.0161,
1254
+ "step": 1610
1255
+ },
1256
+ {
1257
+ "epoch": 31.46,
1258
+ "learning_rate": 1.1919504643962849e-05,
1259
+ "loss": 0.0428,
1260
+ "step": 1620
1261
+ },
1262
+ {
1263
+ "epoch": 31.65,
1264
+ "learning_rate": 1.1635706914344686e-05,
1265
+ "loss": 0.0379,
1266
+ "step": 1630
1267
+ },
1268
+ {
1269
+ "epoch": 31.84,
1270
+ "learning_rate": 1.1351909184726522e-05,
1271
+ "loss": 0.1243,
1272
+ "step": 1640
1273
+ },
1274
+ {
1275
+ "epoch": 32.0,
1276
+ "eval_accuracy": 0.8043478260869565,
1277
+ "eval_loss": 1.1341125965118408,
1278
+ "eval_runtime": 0.8712,
1279
+ "eval_samples_per_second": 52.801,
1280
+ "eval_steps_per_second": 6.887,
1281
+ "step": 1648
1282
+ },
1283
+ {
1284
+ "epoch": 32.04,
1285
+ "learning_rate": 1.106811145510836e-05,
1286
+ "loss": 0.0214,
1287
+ "step": 1650
1288
+ },
1289
+ {
1290
+ "epoch": 32.23,
1291
+ "learning_rate": 1.0784313725490196e-05,
1292
+ "loss": 0.0496,
1293
+ "step": 1660
1294
+ },
1295
+ {
1296
+ "epoch": 32.43,
1297
+ "learning_rate": 1.0500515995872033e-05,
1298
+ "loss": 0.0483,
1299
+ "step": 1670
1300
+ },
1301
+ {
1302
+ "epoch": 32.62,
1303
+ "learning_rate": 1.0216718266253871e-05,
1304
+ "loss": 0.0087,
1305
+ "step": 1680
1306
+ },
1307
+ {
1308
+ "epoch": 32.82,
1309
+ "learning_rate": 9.932920536635709e-06,
1310
+ "loss": 0.062,
1311
+ "step": 1690
1312
+ },
1313
+ {
1314
+ "epoch": 32.99,
1315
+ "eval_accuracy": 0.782608695652174,
1316
+ "eval_loss": 1.2647993564605713,
1317
+ "eval_runtime": 0.8321,
1318
+ "eval_samples_per_second": 55.281,
1319
+ "eval_steps_per_second": 7.211,
1320
+ "step": 1699
1321
+ },
1322
+ {
1323
+ "epoch": 33.01,
1324
+ "learning_rate": 9.649122807017545e-06,
1325
+ "loss": 0.1248,
1326
+ "step": 1700
1327
+ },
1328
+ {
1329
+ "epoch": 33.2,
1330
+ "learning_rate": 9.36532507739938e-06,
1331
+ "loss": 0.0765,
1332
+ "step": 1710
1333
+ },
1334
+ {
1335
+ "epoch": 33.4,
1336
+ "learning_rate": 9.081527347781218e-06,
1337
+ "loss": 0.0796,
1338
+ "step": 1720
1339
+ },
1340
+ {
1341
+ "epoch": 33.59,
1342
+ "learning_rate": 8.797729618163056e-06,
1343
+ "loss": 0.0274,
1344
+ "step": 1730
1345
+ },
1346
+ {
1347
+ "epoch": 33.79,
1348
+ "learning_rate": 8.513931888544893e-06,
1349
+ "loss": 0.0525,
1350
+ "step": 1740
1351
+ },
1352
+ {
1353
+ "epoch": 33.98,
1354
+ "learning_rate": 8.230134158926728e-06,
1355
+ "loss": 0.0941,
1356
+ "step": 1750
1357
+ },
1358
+ {
1359
+ "epoch": 34.0,
1360
+ "eval_accuracy": 0.782608695652174,
1361
+ "eval_loss": 1.1236172914505005,
1362
+ "eval_runtime": 0.8314,
1363
+ "eval_samples_per_second": 55.331,
1364
+ "eval_steps_per_second": 7.217,
1365
+ "step": 1751
1366
+ },
1367
+ {
1368
+ "epoch": 34.17,
1369
+ "learning_rate": 7.946336429308565e-06,
1370
+ "loss": 0.0313,
1371
+ "step": 1760
1372
+ },
1373
+ {
1374
+ "epoch": 34.37,
1375
+ "learning_rate": 7.662538699690403e-06,
1376
+ "loss": 0.046,
1377
+ "step": 1770
1378
+ },
1379
+ {
1380
+ "epoch": 34.56,
1381
+ "learning_rate": 7.3787409700722396e-06,
1382
+ "loss": 0.0394,
1383
+ "step": 1780
1384
+ },
1385
+ {
1386
+ "epoch": 34.76,
1387
+ "learning_rate": 7.094943240454077e-06,
1388
+ "loss": 0.0209,
1389
+ "step": 1790
1390
+ },
1391
+ {
1392
+ "epoch": 34.95,
1393
+ "learning_rate": 6.811145510835914e-06,
1394
+ "loss": 0.0119,
1395
+ "step": 1800
1396
+ },
1397
+ {
1398
+ "epoch": 34.99,
1399
+ "eval_accuracy": 0.8043478260869565,
1400
+ "eval_loss": 1.1302545070648193,
1401
+ "eval_runtime": 0.8257,
1402
+ "eval_samples_per_second": 55.711,
1403
+ "eval_steps_per_second": 7.267,
1404
+ "step": 1802
1405
+ },
1406
+ {
1407
+ "epoch": 35.15,
1408
+ "learning_rate": 6.527347781217751e-06,
1409
+ "loss": 0.0719,
1410
+ "step": 1810
1411
+ },
1412
+ {
1413
+ "epoch": 35.34,
1414
+ "learning_rate": 6.2435500515995875e-06,
1415
+ "loss": 0.0428,
1416
+ "step": 1820
1417
+ },
1418
+ {
1419
+ "epoch": 35.53,
1420
+ "learning_rate": 5.959752321981424e-06,
1421
+ "loss": 0.0183,
1422
+ "step": 1830
1423
+ },
1424
+ {
1425
+ "epoch": 35.73,
1426
+ "learning_rate": 5.675954592363261e-06,
1427
+ "loss": 0.027,
1428
+ "step": 1840
1429
+ },
1430
+ {
1431
+ "epoch": 35.92,
1432
+ "learning_rate": 5.392156862745098e-06,
1433
+ "loss": 0.044,
1434
+ "step": 1850
1435
+ },
1436
+ {
1437
+ "epoch": 36.0,
1438
+ "eval_accuracy": 0.782608695652174,
1439
+ "eval_loss": 1.1848399639129639,
1440
+ "eval_runtime": 0.8214,
1441
+ "eval_samples_per_second": 56.0,
1442
+ "eval_steps_per_second": 7.304,
1443
+ "step": 1854
1444
+ },
1445
+ {
1446
+ "epoch": 36.12,
1447
+ "learning_rate": 5.1083591331269355e-06,
1448
+ "loss": 0.0312,
1449
+ "step": 1860
1450
+ },
1451
+ {
1452
+ "epoch": 36.31,
1453
+ "learning_rate": 4.824561403508772e-06,
1454
+ "loss": 0.0166,
1455
+ "step": 1870
1456
+ },
1457
+ {
1458
+ "epoch": 36.5,
1459
+ "learning_rate": 4.540763673890609e-06,
1460
+ "loss": 0.0951,
1461
+ "step": 1880
1462
+ },
1463
+ {
1464
+ "epoch": 36.7,
1465
+ "learning_rate": 4.256965944272447e-06,
1466
+ "loss": 0.1538,
1467
+ "step": 1890
1468
+ },
1469
+ {
1470
+ "epoch": 36.89,
1471
+ "learning_rate": 3.973168214654283e-06,
1472
+ "loss": 0.0073,
1473
+ "step": 1900
1474
+ },
1475
+ {
1476
+ "epoch": 36.99,
1477
+ "eval_accuracy": 0.7608695652173914,
1478
+ "eval_loss": 1.1795865297317505,
1479
+ "eval_runtime": 0.8136,
1480
+ "eval_samples_per_second": 56.539,
1481
+ "eval_steps_per_second": 7.375,
1482
+ "step": 1905
1483
+ },
1484
+ {
1485
+ "epoch": 37.09,
1486
+ "learning_rate": 3.6893704850361198e-06,
1487
+ "loss": 0.0617,
1488
+ "step": 1910
1489
+ },
1490
+ {
1491
+ "epoch": 37.28,
1492
+ "learning_rate": 3.405572755417957e-06,
1493
+ "loss": 0.0482,
1494
+ "step": 1920
1495
+ },
1496
+ {
1497
+ "epoch": 37.48,
1498
+ "learning_rate": 3.1217750257997938e-06,
1499
+ "loss": 0.0452,
1500
+ "step": 1930
1501
+ },
1502
+ {
1503
+ "epoch": 37.67,
1504
+ "learning_rate": 2.8379772961816305e-06,
1505
+ "loss": 0.0511,
1506
+ "step": 1940
1507
+ },
1508
+ {
1509
+ "epoch": 37.86,
1510
+ "learning_rate": 2.5541795665634677e-06,
1511
+ "loss": 0.0149,
1512
+ "step": 1950
1513
+ },
1514
+ {
1515
+ "epoch": 38.0,
1516
+ "eval_accuracy": 0.782608695652174,
1517
+ "eval_loss": 1.2491459846496582,
1518
+ "eval_runtime": 0.8526,
1519
+ "eval_samples_per_second": 53.952,
1520
+ "eval_steps_per_second": 7.037,
1521
+ "step": 1957
1522
+ },
1523
+ {
1524
+ "epoch": 38.06,
1525
+ "learning_rate": 2.2703818369453045e-06,
1526
+ "loss": 0.0527,
1527
+ "step": 1960
1528
+ },
1529
+ {
1530
+ "epoch": 38.25,
1531
+ "learning_rate": 1.9865841073271413e-06,
1532
+ "loss": 0.0693,
1533
+ "step": 1970
1534
+ },
1535
+ {
1536
+ "epoch": 38.45,
1537
+ "learning_rate": 1.7027863777089785e-06,
1538
+ "loss": 0.0776,
1539
+ "step": 1980
1540
+ },
1541
+ {
1542
+ "epoch": 38.64,
1543
+ "learning_rate": 1.4189886480908153e-06,
1544
+ "loss": 0.0413,
1545
+ "step": 1990
1546
+ },
1547
+ {
1548
+ "epoch": 38.83,
1549
+ "learning_rate": 1.1351909184726523e-06,
1550
+ "loss": 0.0194,
1551
+ "step": 2000
1552
+ },
1553
+ {
1554
+ "epoch": 38.99,
1555
+ "eval_accuracy": 0.782608695652174,
1556
+ "eval_loss": 1.1812182664871216,
1557
+ "eval_runtime": 0.8162,
1558
+ "eval_samples_per_second": 56.361,
1559
+ "eval_steps_per_second": 7.351,
1560
+ "step": 2008
1561
+ },
1562
+ {
1563
+ "epoch": 39.03,
1564
+ "learning_rate": 8.513931888544892e-07,
1565
+ "loss": 0.0202,
1566
+ "step": 2010
1567
+ },
1568
+ {
1569
+ "epoch": 39.22,
1570
+ "learning_rate": 5.675954592363261e-07,
1571
+ "loss": 0.0311,
1572
+ "step": 2020
1573
+ },
1574
+ {
1575
+ "epoch": 39.42,
1576
+ "learning_rate": 2.8379772961816306e-07,
1577
+ "loss": 0.037,
1578
+ "step": 2030
1579
+ },
1580
+ {
1581
+ "epoch": 39.61,
1582
+ "learning_rate": 0.0,
1583
+ "loss": 0.0577,
1584
+ "step": 2040
1585
+ },
1586
+ {
1587
+ "epoch": 39.61,
1588
+ "eval_accuracy": 0.7608695652173914,
1589
+ "eval_loss": 1.177699327468872,
1590
+ "eval_runtime": 0.8677,
1591
+ "eval_samples_per_second": 53.015,
1592
+ "eval_steps_per_second": 6.915,
1593
+ "step": 2040
1594
+ },
1595
+ {
1596
+ "epoch": 39.61,
1597
+ "step": 2040,
1598
+ "total_flos": 2.5142726714989363e+18,
1599
+ "train_loss": 0.19644491698172892,
1600
+ "train_runtime": 666.0117,
1601
+ "train_samples_per_second": 49.188,
1602
+ "train_steps_per_second": 3.063
1603
+ }
1604
+ ],
1605
+ "logging_steps": 10,
1606
+ "max_steps": 2040,
1607
+ "num_input_tokens_seen": 0,
1608
+ "num_train_epochs": 40,
1609
+ "save_steps": 500,
1610
+ "total_flos": 2.5142726714989363e+18,
1611
+ "train_batch_size": 8,
1612
+ "trial_name": null,
1613
+ "trial_params": null
1614
+ }