nemik commited on
Commit
188d300
·
verified ·
1 Parent(s): e27896e

End of training

Browse files
README.md CHANGED
@@ -25,16 +25,16 @@ model-index:
25
  metrics:
26
  - name: Accuracy
27
  type: accuracy
28
- value: 0.972
29
  - name: F1
30
  type: f1
31
- value: 0.93007769145394
32
  - name: Precision
33
  type: precision
34
- value: 0.9352678571428571
35
  - name: Recall
36
  type: recall
37
- value: 0.9249448123620309
38
  ---
39
 
40
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -44,11 +44,11 @@ should probably proofread and complete it, then remove this comment. -->
44
 
45
  This model is a fine-tuned version of [google/vit-base-patch16-384](https://huggingface.co/google/vit-base-patch16-384) on the webdataset dataset.
46
  It achieves the following results on the evaluation set:
47
- - Loss: 0.0948
48
- - Accuracy: 0.972
49
- - F1: 0.9301
50
- - Precision: 0.9353
51
- - Recall: 0.9249
52
 
53
  ## Model description
54
 
 
25
  metrics:
26
  - name: Accuracy
27
  type: accuracy
28
+ value: 0.9746666666666667
29
  - name: F1
30
  type: f1
31
+ value: 0.9372937293729373
32
  - name: Precision
33
  type: precision
34
+ value: 0.9342105263157895
35
  - name: Recall
36
  type: recall
37
+ value: 0.9403973509933775
38
  ---
39
 
40
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
44
 
45
  This model is a fine-tuned version of [google/vit-base-patch16-384](https://huggingface.co/google/vit-base-patch16-384) on the webdataset dataset.
46
  It achieves the following results on the evaluation set:
47
+ - Loss: 0.0795
48
+ - Accuracy: 0.9747
49
+ - F1: 0.9373
50
+ - Precision: 0.9342
51
+ - Recall: 0.9404
52
 
53
  ## Model description
54
 
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "eval_accuracy": 0.9746666666666667,
4
+ "eval_f1": 0.9372937293729373,
5
+ "eval_loss": 0.07946290820837021,
6
+ "eval_precision": 0.9342105263157895,
7
+ "eval_recall": 0.9403973509933775,
8
+ "eval_runtime": 3.4614,
9
+ "eval_samples_per_second": 65.003,
10
+ "eval_steps_per_second": 8.378,
11
+ "total_flos": 6.170107339800576e+18,
12
+ "train_loss": 0.032361031031748005,
13
+ "train_runtime": 1047.1267,
14
+ "train_samples_per_second": 25.785,
15
+ "train_steps_per_second": 1.633
16
+ }
eval_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "eval_accuracy": 0.9746666666666667,
4
+ "eval_f1": 0.9372937293729373,
5
+ "eval_loss": 0.07946290820837021,
6
+ "eval_precision": 0.9342105263157895,
7
+ "eval_recall": 0.9403973509933775,
8
+ "eval_runtime": 3.4614,
9
+ "eval_samples_per_second": 65.003,
10
+ "eval_steps_per_second": 8.378
11
+ }
runs/Jul26_00-15-08_fa9f47f2515c/events.out.tfevents.1721954620.fa9f47f2515c.437.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7c97ed4313031c2cde976057b6e0f56ecce94366c3bacf7e8e7574d800f2fed3
3
+ size 560
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "total_flos": 6.170107339800576e+18,
4
+ "train_loss": 0.032361031031748005,
5
+ "train_runtime": 1047.1267,
6
+ "train_samples_per_second": 25.785,
7
+ "train_steps_per_second": 1.633
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1443 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.07946290820837021,
3
+ "best_model_checkpoint": "vit-base-patch16-384-finetuned_v2024-7-25-frost/checkpoint-500",
4
+ "epoch": 30.0,
5
+ "eval_steps": 100,
6
+ "global_step": 1710,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.17543859649122806,
13
+ "grad_norm": 0.33591213822364807,
14
+ "learning_rate": 1.1695906432748537e-05,
15
+ "loss": 0.0492,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.3508771929824561,
20
+ "grad_norm": 0.31977248191833496,
21
+ "learning_rate": 2.3391812865497074e-05,
22
+ "loss": 0.0587,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.5263157894736842,
27
+ "grad_norm": 0.3621572256088257,
28
+ "learning_rate": 3.508771929824561e-05,
29
+ "loss": 0.0383,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.7017543859649122,
34
+ "grad_norm": 0.23749703168869019,
35
+ "learning_rate": 4.678362573099415e-05,
36
+ "loss": 0.034,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.8771929824561403,
41
+ "grad_norm": 0.6749160289764404,
42
+ "learning_rate": 5.847953216374269e-05,
43
+ "loss": 0.0534,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 1.0526315789473684,
48
+ "grad_norm": 0.5174930691719055,
49
+ "learning_rate": 7.017543859649122e-05,
50
+ "loss": 0.0703,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 1.2280701754385965,
55
+ "grad_norm": 0.645950436592102,
56
+ "learning_rate": 8.187134502923976e-05,
57
+ "loss": 0.0445,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 1.4035087719298245,
62
+ "grad_norm": 0.12476111948490143,
63
+ "learning_rate": 9.35672514619883e-05,
64
+ "loss": 0.0488,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 1.5789473684210527,
69
+ "grad_norm": 0.13328130543231964,
70
+ "learning_rate": 0.00010526315789473685,
71
+ "loss": 0.028,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 1.7543859649122808,
76
+ "grad_norm": 0.7791523933410645,
77
+ "learning_rate": 0.00011695906432748539,
78
+ "loss": 0.0624,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 1.7543859649122808,
83
+ "eval_accuracy": 0.9866666666666667,
84
+ "eval_f1": 0.9665178571428571,
85
+ "eval_loss": 0.045776307582855225,
86
+ "eval_precision": 0.9774266365688488,
87
+ "eval_recall": 0.9558498896247241,
88
+ "eval_runtime": 3.3187,
89
+ "eval_samples_per_second": 67.797,
90
+ "eval_steps_per_second": 8.738,
91
+ "step": 100
92
+ },
93
+ {
94
+ "epoch": 1.9298245614035088,
95
+ "grad_norm": 1.0299451351165771,
96
+ "learning_rate": 0.0001286549707602339,
97
+ "loss": 0.0689,
98
+ "step": 110
99
+ },
100
+ {
101
+ "epoch": 2.1052631578947367,
102
+ "grad_norm": 0.5722321271896362,
103
+ "learning_rate": 0.00014035087719298245,
104
+ "loss": 0.0509,
105
+ "step": 120
106
+ },
107
+ {
108
+ "epoch": 2.280701754385965,
109
+ "grad_norm": 0.4698758125305176,
110
+ "learning_rate": 0.00015204678362573098,
111
+ "loss": 0.0377,
112
+ "step": 130
113
+ },
114
+ {
115
+ "epoch": 2.456140350877193,
116
+ "grad_norm": 0.47861894965171814,
117
+ "learning_rate": 0.00016374269005847952,
118
+ "loss": 0.0608,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 2.6315789473684212,
123
+ "grad_norm": 0.847069263458252,
124
+ "learning_rate": 0.00017543859649122806,
125
+ "loss": 0.0545,
126
+ "step": 150
127
+ },
128
+ {
129
+ "epoch": 2.807017543859649,
130
+ "grad_norm": 0.5486211180686951,
131
+ "learning_rate": 0.0001871345029239766,
132
+ "loss": 0.053,
133
+ "step": 160
134
+ },
135
+ {
136
+ "epoch": 2.982456140350877,
137
+ "grad_norm": 0.8176277875900269,
138
+ "learning_rate": 0.00019883040935672513,
139
+ "loss": 0.0632,
140
+ "step": 170
141
+ },
142
+ {
143
+ "epoch": 3.1578947368421053,
144
+ "grad_norm": 0.20802472531795502,
145
+ "learning_rate": 0.00019883040935672513,
146
+ "loss": 0.0808,
147
+ "step": 180
148
+ },
149
+ {
150
+ "epoch": 3.3333333333333335,
151
+ "grad_norm": 0.915105402469635,
152
+ "learning_rate": 0.00019753086419753085,
153
+ "loss": 0.0839,
154
+ "step": 190
155
+ },
156
+ {
157
+ "epoch": 3.5087719298245617,
158
+ "grad_norm": 1.2987483739852905,
159
+ "learning_rate": 0.00019623131903833657,
160
+ "loss": 0.0729,
161
+ "step": 200
162
+ },
163
+ {
164
+ "epoch": 3.5087719298245617,
165
+ "eval_accuracy": 0.9688888888888889,
166
+ "eval_f1": 0.9220489977728286,
167
+ "eval_loss": 0.09421814233064651,
168
+ "eval_precision": 0.9303370786516854,
169
+ "eval_recall": 0.9139072847682119,
170
+ "eval_runtime": 4.053,
171
+ "eval_samples_per_second": 55.514,
172
+ "eval_steps_per_second": 7.155,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 3.6842105263157894,
177
+ "grad_norm": 0.4004723131656647,
178
+ "learning_rate": 0.0001949317738791423,
179
+ "loss": 0.0778,
180
+ "step": 210
181
+ },
182
+ {
183
+ "epoch": 3.8596491228070176,
184
+ "grad_norm": 0.8877297043800354,
185
+ "learning_rate": 0.00019363222871994802,
186
+ "loss": 0.059,
187
+ "step": 220
188
+ },
189
+ {
190
+ "epoch": 4.035087719298246,
191
+ "grad_norm": 0.39221012592315674,
192
+ "learning_rate": 0.00019233268356075374,
193
+ "loss": 0.0813,
194
+ "step": 230
195
+ },
196
+ {
197
+ "epoch": 4.2105263157894735,
198
+ "grad_norm": 0.5537380576133728,
199
+ "learning_rate": 0.00019103313840155946,
200
+ "loss": 0.075,
201
+ "step": 240
202
+ },
203
+ {
204
+ "epoch": 4.385964912280702,
205
+ "grad_norm": 0.7138140201568604,
206
+ "learning_rate": 0.00018973359324236518,
207
+ "loss": 0.0534,
208
+ "step": 250
209
+ },
210
+ {
211
+ "epoch": 4.56140350877193,
212
+ "grad_norm": 0.724746584892273,
213
+ "learning_rate": 0.0001884340480831709,
214
+ "loss": 0.0854,
215
+ "step": 260
216
+ },
217
+ {
218
+ "epoch": 4.7368421052631575,
219
+ "grad_norm": 0.7256239652633667,
220
+ "learning_rate": 0.0001871345029239766,
221
+ "loss": 0.0689,
222
+ "step": 270
223
+ },
224
+ {
225
+ "epoch": 4.912280701754386,
226
+ "grad_norm": 0.4805387854576111,
227
+ "learning_rate": 0.00018583495776478232,
228
+ "loss": 0.0508,
229
+ "step": 280
230
+ },
231
+ {
232
+ "epoch": 5.087719298245614,
233
+ "grad_norm": 1.4754490852355957,
234
+ "learning_rate": 0.00018453541260558804,
235
+ "loss": 0.0549,
236
+ "step": 290
237
+ },
238
+ {
239
+ "epoch": 5.2631578947368425,
240
+ "grad_norm": 0.6193797588348389,
241
+ "learning_rate": 0.00018323586744639376,
242
+ "loss": 0.0566,
243
+ "step": 300
244
+ },
245
+ {
246
+ "epoch": 5.2631578947368425,
247
+ "eval_accuracy": 0.972,
248
+ "eval_f1": 0.9311475409836065,
249
+ "eval_loss": 0.08017320185899734,
250
+ "eval_precision": 0.922077922077922,
251
+ "eval_recall": 0.9403973509933775,
252
+ "eval_runtime": 4.0432,
253
+ "eval_samples_per_second": 55.649,
254
+ "eval_steps_per_second": 7.173,
255
+ "step": 300
256
+ },
257
+ {
258
+ "epoch": 5.43859649122807,
259
+ "grad_norm": 0.4263758957386017,
260
+ "learning_rate": 0.00018193632228719948,
261
+ "loss": 0.0529,
262
+ "step": 310
263
+ },
264
+ {
265
+ "epoch": 5.614035087719298,
266
+ "grad_norm": 0.6219185590744019,
267
+ "learning_rate": 0.0001806367771280052,
268
+ "loss": 0.0534,
269
+ "step": 320
270
+ },
271
+ {
272
+ "epoch": 5.7894736842105265,
273
+ "grad_norm": 1.2227765321731567,
274
+ "learning_rate": 0.00017933723196881092,
275
+ "loss": 0.0784,
276
+ "step": 330
277
+ },
278
+ {
279
+ "epoch": 5.964912280701754,
280
+ "grad_norm": 0.7235316038131714,
281
+ "learning_rate": 0.00017803768680961664,
282
+ "loss": 0.0576,
283
+ "step": 340
284
+ },
285
+ {
286
+ "epoch": 6.140350877192983,
287
+ "grad_norm": 0.3069102466106415,
288
+ "learning_rate": 0.00017673814165042236,
289
+ "loss": 0.0666,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 6.315789473684211,
294
+ "grad_norm": 0.48301374912261963,
295
+ "learning_rate": 0.00017543859649122806,
296
+ "loss": 0.0599,
297
+ "step": 360
298
+ },
299
+ {
300
+ "epoch": 6.491228070175438,
301
+ "grad_norm": 0.3786466419696808,
302
+ "learning_rate": 0.00017413905133203378,
303
+ "loss": 0.0687,
304
+ "step": 370
305
+ },
306
+ {
307
+ "epoch": 6.666666666666667,
308
+ "grad_norm": 0.7827033400535583,
309
+ "learning_rate": 0.0001728395061728395,
310
+ "loss": 0.0654,
311
+ "step": 380
312
+ },
313
+ {
314
+ "epoch": 6.842105263157895,
315
+ "grad_norm": 0.3216448128223419,
316
+ "learning_rate": 0.00017153996101364522,
317
+ "loss": 0.0457,
318
+ "step": 390
319
+ },
320
+ {
321
+ "epoch": 7.017543859649122,
322
+ "grad_norm": 0.41389861702919006,
323
+ "learning_rate": 0.00017024041585445094,
324
+ "loss": 0.051,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 7.017543859649122,
329
+ "eval_accuracy": 0.9631111111111111,
330
+ "eval_f1": 0.906636670416198,
331
+ "eval_loss": 0.09651228040456772,
332
+ "eval_precision": 0.9243119266055045,
333
+ "eval_recall": 0.8896247240618101,
334
+ "eval_runtime": 3.9145,
335
+ "eval_samples_per_second": 57.479,
336
+ "eval_steps_per_second": 7.408,
337
+ "step": 400
338
+ },
339
+ {
340
+ "epoch": 7.192982456140351,
341
+ "grad_norm": 0.3265860080718994,
342
+ "learning_rate": 0.00016894087069525666,
343
+ "loss": 0.046,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 7.368421052631579,
348
+ "grad_norm": 0.34316837787628174,
349
+ "learning_rate": 0.00016764132553606238,
350
+ "loss": 0.0492,
351
+ "step": 420
352
+ },
353
+ {
354
+ "epoch": 7.543859649122807,
355
+ "grad_norm": 0.6148121953010559,
356
+ "learning_rate": 0.0001663417803768681,
357
+ "loss": 0.0414,
358
+ "step": 430
359
+ },
360
+ {
361
+ "epoch": 7.719298245614035,
362
+ "grad_norm": 0.17521312832832336,
363
+ "learning_rate": 0.00016504223521767383,
364
+ "loss": 0.0408,
365
+ "step": 440
366
+ },
367
+ {
368
+ "epoch": 7.894736842105263,
369
+ "grad_norm": 0.7820538878440857,
370
+ "learning_rate": 0.00016374269005847952,
371
+ "loss": 0.059,
372
+ "step": 450
373
+ },
374
+ {
375
+ "epoch": 8.070175438596491,
376
+ "grad_norm": 0.3714459538459778,
377
+ "learning_rate": 0.00016244314489928524,
378
+ "loss": 0.0316,
379
+ "step": 460
380
+ },
381
+ {
382
+ "epoch": 8.24561403508772,
383
+ "grad_norm": 0.1257082223892212,
384
+ "learning_rate": 0.00016114359974009096,
385
+ "loss": 0.0366,
386
+ "step": 470
387
+ },
388
+ {
389
+ "epoch": 8.421052631578947,
390
+ "grad_norm": 0.20228251814842224,
391
+ "learning_rate": 0.00015984405458089668,
392
+ "loss": 0.0379,
393
+ "step": 480
394
+ },
395
+ {
396
+ "epoch": 8.596491228070175,
397
+ "grad_norm": 0.9517484307289124,
398
+ "learning_rate": 0.0001585445094217024,
399
+ "loss": 0.0569,
400
+ "step": 490
401
+ },
402
+ {
403
+ "epoch": 8.771929824561404,
404
+ "grad_norm": 0.47640401124954224,
405
+ "learning_rate": 0.00015724496426250813,
406
+ "loss": 0.0686,
407
+ "step": 500
408
+ },
409
+ {
410
+ "epoch": 8.771929824561404,
411
+ "eval_accuracy": 0.9746666666666667,
412
+ "eval_f1": 0.9372937293729373,
413
+ "eval_loss": 0.07946290820837021,
414
+ "eval_precision": 0.9342105263157895,
415
+ "eval_recall": 0.9403973509933775,
416
+ "eval_runtime": 3.9023,
417
+ "eval_samples_per_second": 57.659,
418
+ "eval_steps_per_second": 7.432,
419
+ "step": 500
420
+ },
421
+ {
422
+ "epoch": 8.947368421052632,
423
+ "grad_norm": 0.4718121588230133,
424
+ "learning_rate": 0.00015594541910331385,
425
+ "loss": 0.0396,
426
+ "step": 510
427
+ },
428
+ {
429
+ "epoch": 9.12280701754386,
430
+ "grad_norm": 0.4709373116493225,
431
+ "learning_rate": 0.00015464587394411957,
432
+ "loss": 0.0464,
433
+ "step": 520
434
+ },
435
+ {
436
+ "epoch": 9.298245614035087,
437
+ "grad_norm": 0.5387071371078491,
438
+ "learning_rate": 0.00015334632878492526,
439
+ "loss": 0.045,
440
+ "step": 530
441
+ },
442
+ {
443
+ "epoch": 9.473684210526315,
444
+ "grad_norm": 0.541229784488678,
445
+ "learning_rate": 0.00015204678362573098,
446
+ "loss": 0.06,
447
+ "step": 540
448
+ },
449
+ {
450
+ "epoch": 9.649122807017545,
451
+ "grad_norm": 0.4231584370136261,
452
+ "learning_rate": 0.0001507472384665367,
453
+ "loss": 0.0579,
454
+ "step": 550
455
+ },
456
+ {
457
+ "epoch": 9.824561403508772,
458
+ "grad_norm": 0.37044307589530945,
459
+ "learning_rate": 0.00014944769330734243,
460
+ "loss": 0.041,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 10.0,
465
+ "grad_norm": 0.9554832577705383,
466
+ "learning_rate": 0.00014814814814814815,
467
+ "loss": 0.0405,
468
+ "step": 570
469
+ },
470
+ {
471
+ "epoch": 10.175438596491228,
472
+ "grad_norm": 0.3497917354106903,
473
+ "learning_rate": 0.00014684860298895387,
474
+ "loss": 0.0308,
475
+ "step": 580
476
+ },
477
+ {
478
+ "epoch": 10.350877192982455,
479
+ "grad_norm": 0.725246250629425,
480
+ "learning_rate": 0.0001455490578297596,
481
+ "loss": 0.0471,
482
+ "step": 590
483
+ },
484
+ {
485
+ "epoch": 10.526315789473685,
486
+ "grad_norm": 0.14522100985050201,
487
+ "learning_rate": 0.0001442495126705653,
488
+ "loss": 0.0271,
489
+ "step": 600
490
+ },
491
+ {
492
+ "epoch": 10.526315789473685,
493
+ "eval_accuracy": 0.9693333333333334,
494
+ "eval_f1": 0.9239250275633959,
495
+ "eval_loss": 0.09354085475206375,
496
+ "eval_precision": 0.9229074889867841,
497
+ "eval_recall": 0.9249448123620309,
498
+ "eval_runtime": 3.4058,
499
+ "eval_samples_per_second": 66.064,
500
+ "eval_steps_per_second": 8.515,
501
+ "step": 600
502
+ },
503
+ {
504
+ "epoch": 10.701754385964913,
505
+ "grad_norm": 0.11523815244436264,
506
+ "learning_rate": 0.00014294996751137103,
507
+ "loss": 0.0226,
508
+ "step": 610
509
+ },
510
+ {
511
+ "epoch": 10.87719298245614,
512
+ "grad_norm": 0.06203211098909378,
513
+ "learning_rate": 0.00014165042235217672,
514
+ "loss": 0.0452,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 11.052631578947368,
519
+ "grad_norm": 0.2138066440820694,
520
+ "learning_rate": 0.00014035087719298245,
521
+ "loss": 0.0389,
522
+ "step": 630
523
+ },
524
+ {
525
+ "epoch": 11.228070175438596,
526
+ "grad_norm": 0.4954126179218292,
527
+ "learning_rate": 0.00013905133203378817,
528
+ "loss": 0.0315,
529
+ "step": 640
530
+ },
531
+ {
532
+ "epoch": 11.403508771929825,
533
+ "grad_norm": 0.24581827223300934,
534
+ "learning_rate": 0.0001377517868745939,
535
+ "loss": 0.0404,
536
+ "step": 650
537
+ },
538
+ {
539
+ "epoch": 11.578947368421053,
540
+ "grad_norm": 0.35176780819892883,
541
+ "learning_rate": 0.0001364522417153996,
542
+ "loss": 0.0206,
543
+ "step": 660
544
+ },
545
+ {
546
+ "epoch": 11.75438596491228,
547
+ "grad_norm": 0.20645667612552643,
548
+ "learning_rate": 0.00013515269655620533,
549
+ "loss": 0.0348,
550
+ "step": 670
551
+ },
552
+ {
553
+ "epoch": 11.929824561403509,
554
+ "grad_norm": 0.27070721983909607,
555
+ "learning_rate": 0.00013385315139701105,
556
+ "loss": 0.0256,
557
+ "step": 680
558
+ },
559
+ {
560
+ "epoch": 12.105263157894736,
561
+ "grad_norm": 0.353639155626297,
562
+ "learning_rate": 0.00013255360623781677,
563
+ "loss": 0.0377,
564
+ "step": 690
565
+ },
566
+ {
567
+ "epoch": 12.280701754385966,
568
+ "grad_norm": 0.3672114312648773,
569
+ "learning_rate": 0.0001312540610786225,
570
+ "loss": 0.0273,
571
+ "step": 700
572
+ },
573
+ {
574
+ "epoch": 12.280701754385966,
575
+ "eval_accuracy": 0.9715555555555555,
576
+ "eval_f1": 0.9299781181619255,
577
+ "eval_loss": 0.0975123792886734,
578
+ "eval_precision": 0.9219088937093276,
579
+ "eval_recall": 0.9381898454746137,
580
+ "eval_runtime": 3.2728,
581
+ "eval_samples_per_second": 68.749,
582
+ "eval_steps_per_second": 8.861,
583
+ "step": 700
584
+ },
585
+ {
586
+ "epoch": 12.456140350877194,
587
+ "grad_norm": 0.8677971363067627,
588
+ "learning_rate": 0.0001299545159194282,
589
+ "loss": 0.0307,
590
+ "step": 710
591
+ },
592
+ {
593
+ "epoch": 12.631578947368421,
594
+ "grad_norm": 0.5751360654830933,
595
+ "learning_rate": 0.0001286549707602339,
596
+ "loss": 0.0246,
597
+ "step": 720
598
+ },
599
+ {
600
+ "epoch": 12.807017543859649,
601
+ "grad_norm": 0.4053627848625183,
602
+ "learning_rate": 0.00012735542560103963,
603
+ "loss": 0.0448,
604
+ "step": 730
605
+ },
606
+ {
607
+ "epoch": 12.982456140350877,
608
+ "grad_norm": 0.21051475405693054,
609
+ "learning_rate": 0.00012605588044184535,
610
+ "loss": 0.0339,
611
+ "step": 740
612
+ },
613
+ {
614
+ "epoch": 13.157894736842104,
615
+ "grad_norm": 0.3146640360355377,
616
+ "learning_rate": 0.00012475633528265107,
617
+ "loss": 0.0321,
618
+ "step": 750
619
+ },
620
+ {
621
+ "epoch": 13.333333333333334,
622
+ "grad_norm": 0.3593007028102875,
623
+ "learning_rate": 0.0001234567901234568,
624
+ "loss": 0.03,
625
+ "step": 760
626
+ },
627
+ {
628
+ "epoch": 13.508771929824562,
629
+ "grad_norm": 0.4649389982223511,
630
+ "learning_rate": 0.00012215724496426251,
631
+ "loss": 0.0499,
632
+ "step": 770
633
+ },
634
+ {
635
+ "epoch": 13.68421052631579,
636
+ "grad_norm": 0.3773185908794403,
637
+ "learning_rate": 0.00012085769980506822,
638
+ "loss": 0.0384,
639
+ "step": 780
640
+ },
641
+ {
642
+ "epoch": 13.859649122807017,
643
+ "grad_norm": 0.1289927214384079,
644
+ "learning_rate": 0.00011955815464587394,
645
+ "loss": 0.0258,
646
+ "step": 790
647
+ },
648
+ {
649
+ "epoch": 14.035087719298245,
650
+ "grad_norm": 0.8369467854499817,
651
+ "learning_rate": 0.00011825860948667966,
652
+ "loss": 0.0445,
653
+ "step": 800
654
+ },
655
+ {
656
+ "epoch": 14.035087719298245,
657
+ "eval_accuracy": 0.9697777777777777,
658
+ "eval_f1": 0.9247787610619469,
659
+ "eval_loss": 0.09097114205360413,
660
+ "eval_precision": 0.926829268292683,
661
+ "eval_recall": 0.9227373068432672,
662
+ "eval_runtime": 3.3935,
663
+ "eval_samples_per_second": 66.303,
664
+ "eval_steps_per_second": 8.546,
665
+ "step": 800
666
+ },
667
+ {
668
+ "epoch": 14.210526315789474,
669
+ "grad_norm": 0.19551879167556763,
670
+ "learning_rate": 0.00011695906432748539,
671
+ "loss": 0.0199,
672
+ "step": 810
673
+ },
674
+ {
675
+ "epoch": 14.385964912280702,
676
+ "grad_norm": 0.2516753673553467,
677
+ "learning_rate": 0.00011565951916829109,
678
+ "loss": 0.0267,
679
+ "step": 820
680
+ },
681
+ {
682
+ "epoch": 14.56140350877193,
683
+ "grad_norm": 0.21096718311309814,
684
+ "learning_rate": 0.00011435997400909681,
685
+ "loss": 0.0158,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 14.736842105263158,
690
+ "grad_norm": 0.15884388983249664,
691
+ "learning_rate": 0.00011306042884990253,
692
+ "loss": 0.0286,
693
+ "step": 840
694
+ },
695
+ {
696
+ "epoch": 14.912280701754385,
697
+ "grad_norm": 0.063226118683815,
698
+ "learning_rate": 0.00011176088369070826,
699
+ "loss": 0.021,
700
+ "step": 850
701
+ },
702
+ {
703
+ "epoch": 15.087719298245615,
704
+ "grad_norm": 0.3069198429584503,
705
+ "learning_rate": 0.00011046133853151396,
706
+ "loss": 0.0266,
707
+ "step": 860
708
+ },
709
+ {
710
+ "epoch": 15.263157894736842,
711
+ "grad_norm": 0.17305561900138855,
712
+ "learning_rate": 0.00010916179337231968,
713
+ "loss": 0.039,
714
+ "step": 870
715
+ },
716
+ {
717
+ "epoch": 15.43859649122807,
718
+ "grad_norm": 0.3221393823623657,
719
+ "learning_rate": 0.0001078622482131254,
720
+ "loss": 0.0274,
721
+ "step": 880
722
+ },
723
+ {
724
+ "epoch": 15.614035087719298,
725
+ "grad_norm": 0.4601212739944458,
726
+ "learning_rate": 0.00010656270305393113,
727
+ "loss": 0.0272,
728
+ "step": 890
729
+ },
730
+ {
731
+ "epoch": 15.789473684210526,
732
+ "grad_norm": 0.4427635073661804,
733
+ "learning_rate": 0.00010526315789473685,
734
+ "loss": 0.0217,
735
+ "step": 900
736
+ },
737
+ {
738
+ "epoch": 15.789473684210526,
739
+ "eval_accuracy": 0.9697777777777777,
740
+ "eval_f1": 0.9242761692650334,
741
+ "eval_loss": 0.09419577568769455,
742
+ "eval_precision": 0.9325842696629213,
743
+ "eval_recall": 0.9161147902869757,
744
+ "eval_runtime": 3.4224,
745
+ "eval_samples_per_second": 65.743,
746
+ "eval_steps_per_second": 8.473,
747
+ "step": 900
748
+ },
749
+ {
750
+ "epoch": 15.964912280701755,
751
+ "grad_norm": 0.04713751748204231,
752
+ "learning_rate": 0.00010396361273554256,
753
+ "loss": 0.0109,
754
+ "step": 910
755
+ },
756
+ {
757
+ "epoch": 16.140350877192983,
758
+ "grad_norm": 0.4792565703392029,
759
+ "learning_rate": 0.00010266406757634828,
760
+ "loss": 0.0218,
761
+ "step": 920
762
+ },
763
+ {
764
+ "epoch": 16.31578947368421,
765
+ "grad_norm": 0.46472465991973877,
766
+ "learning_rate": 0.000101364522417154,
767
+ "loss": 0.034,
768
+ "step": 930
769
+ },
770
+ {
771
+ "epoch": 16.49122807017544,
772
+ "grad_norm": 0.45568612217903137,
773
+ "learning_rate": 0.00010006497725795972,
774
+ "loss": 0.0293,
775
+ "step": 940
776
+ },
777
+ {
778
+ "epoch": 16.666666666666668,
779
+ "grad_norm": 0.23771442472934723,
780
+ "learning_rate": 9.876543209876543e-05,
781
+ "loss": 0.0287,
782
+ "step": 950
783
+ },
784
+ {
785
+ "epoch": 16.842105263157894,
786
+ "grad_norm": 0.42116573452949524,
787
+ "learning_rate": 9.746588693957115e-05,
788
+ "loss": 0.019,
789
+ "step": 960
790
+ },
791
+ {
792
+ "epoch": 17.017543859649123,
793
+ "grad_norm": 0.15334849059581757,
794
+ "learning_rate": 9.616634178037687e-05,
795
+ "loss": 0.0307,
796
+ "step": 970
797
+ },
798
+ {
799
+ "epoch": 17.19298245614035,
800
+ "grad_norm": 0.26137152314186096,
801
+ "learning_rate": 9.486679662118259e-05,
802
+ "loss": 0.0265,
803
+ "step": 980
804
+ },
805
+ {
806
+ "epoch": 17.36842105263158,
807
+ "grad_norm": 0.38436606526374817,
808
+ "learning_rate": 9.35672514619883e-05,
809
+ "loss": 0.0205,
810
+ "step": 990
811
+ },
812
+ {
813
+ "epoch": 17.54385964912281,
814
+ "grad_norm": 0.23191127181053162,
815
+ "learning_rate": 9.226770630279402e-05,
816
+ "loss": 0.0257,
817
+ "step": 1000
818
+ },
819
+ {
820
+ "epoch": 17.54385964912281,
821
+ "eval_accuracy": 0.9684444444444444,
822
+ "eval_f1": 0.921023359288098,
823
+ "eval_loss": 0.09062980860471725,
824
+ "eval_precision": 0.9282511210762332,
825
+ "eval_recall": 0.9139072847682119,
826
+ "eval_runtime": 4.7001,
827
+ "eval_samples_per_second": 47.872,
828
+ "eval_steps_per_second": 6.17,
829
+ "step": 1000
830
+ },
831
+ {
832
+ "epoch": 17.719298245614034,
833
+ "grad_norm": 0.46758100390434265,
834
+ "learning_rate": 9.096816114359974e-05,
835
+ "loss": 0.0332,
836
+ "step": 1010
837
+ },
838
+ {
839
+ "epoch": 17.894736842105264,
840
+ "grad_norm": 0.2081506997346878,
841
+ "learning_rate": 8.966861598440546e-05,
842
+ "loss": 0.0189,
843
+ "step": 1020
844
+ },
845
+ {
846
+ "epoch": 18.07017543859649,
847
+ "grad_norm": 0.15531718730926514,
848
+ "learning_rate": 8.836907082521118e-05,
849
+ "loss": 0.0326,
850
+ "step": 1030
851
+ },
852
+ {
853
+ "epoch": 18.24561403508772,
854
+ "grad_norm": 0.3089030683040619,
855
+ "learning_rate": 8.706952566601689e-05,
856
+ "loss": 0.0134,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 18.42105263157895,
861
+ "grad_norm": 0.23171091079711914,
862
+ "learning_rate": 8.576998050682261e-05,
863
+ "loss": 0.0368,
864
+ "step": 1050
865
+ },
866
+ {
867
+ "epoch": 18.596491228070175,
868
+ "grad_norm": 0.20561914145946503,
869
+ "learning_rate": 8.447043534762833e-05,
870
+ "loss": 0.027,
871
+ "step": 1060
872
+ },
873
+ {
874
+ "epoch": 18.771929824561404,
875
+ "grad_norm": 0.268462598323822,
876
+ "learning_rate": 8.317089018843405e-05,
877
+ "loss": 0.0261,
878
+ "step": 1070
879
+ },
880
+ {
881
+ "epoch": 18.94736842105263,
882
+ "grad_norm": 0.6412032246589661,
883
+ "learning_rate": 8.187134502923976e-05,
884
+ "loss": 0.0203,
885
+ "step": 1080
886
+ },
887
+ {
888
+ "epoch": 19.12280701754386,
889
+ "grad_norm": 0.28043922781944275,
890
+ "learning_rate": 8.057179987004548e-05,
891
+ "loss": 0.0276,
892
+ "step": 1090
893
+ },
894
+ {
895
+ "epoch": 19.29824561403509,
896
+ "grad_norm": 0.23408719897270203,
897
+ "learning_rate": 7.92722547108512e-05,
898
+ "loss": 0.0188,
899
+ "step": 1100
900
+ },
901
+ {
902
+ "epoch": 19.29824561403509,
903
+ "eval_accuracy": 0.9675555555555555,
904
+ "eval_f1": 0.9180695847362513,
905
+ "eval_loss": 0.10280066728591919,
906
+ "eval_precision": 0.9337899543378996,
907
+ "eval_recall": 0.9028697571743929,
908
+ "eval_runtime": 3.5953,
909
+ "eval_samples_per_second": 62.582,
910
+ "eval_steps_per_second": 8.066,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 19.473684210526315,
915
+ "grad_norm": 0.30617180466651917,
916
+ "learning_rate": 7.797270955165692e-05,
917
+ "loss": 0.0244,
918
+ "step": 1110
919
+ },
920
+ {
921
+ "epoch": 19.649122807017545,
922
+ "grad_norm": 0.15706390142440796,
923
+ "learning_rate": 7.667316439246263e-05,
924
+ "loss": 0.0112,
925
+ "step": 1120
926
+ },
927
+ {
928
+ "epoch": 19.82456140350877,
929
+ "grad_norm": 0.19885846972465515,
930
+ "learning_rate": 7.537361923326835e-05,
931
+ "loss": 0.0242,
932
+ "step": 1130
933
+ },
934
+ {
935
+ "epoch": 20.0,
936
+ "grad_norm": 0.05422697588801384,
937
+ "learning_rate": 7.407407407407407e-05,
938
+ "loss": 0.0148,
939
+ "step": 1140
940
+ },
941
+ {
942
+ "epoch": 20.17543859649123,
943
+ "grad_norm": 0.024661371484398842,
944
+ "learning_rate": 7.27745289148798e-05,
945
+ "loss": 0.0113,
946
+ "step": 1150
947
+ },
948
+ {
949
+ "epoch": 20.350877192982455,
950
+ "grad_norm": 0.24859677255153656,
951
+ "learning_rate": 7.147498375568552e-05,
952
+ "loss": 0.0149,
953
+ "step": 1160
954
+ },
955
+ {
956
+ "epoch": 20.526315789473685,
957
+ "grad_norm": 0.041024014353752136,
958
+ "learning_rate": 7.017543859649122e-05,
959
+ "loss": 0.0212,
960
+ "step": 1170
961
+ },
962
+ {
963
+ "epoch": 20.70175438596491,
964
+ "grad_norm": 0.04569957032799721,
965
+ "learning_rate": 6.887589343729694e-05,
966
+ "loss": 0.0175,
967
+ "step": 1180
968
+ },
969
+ {
970
+ "epoch": 20.87719298245614,
971
+ "grad_norm": 0.074521504342556,
972
+ "learning_rate": 6.757634827810267e-05,
973
+ "loss": 0.0216,
974
+ "step": 1190
975
+ },
976
+ {
977
+ "epoch": 21.05263157894737,
978
+ "grad_norm": 0.24125021696090698,
979
+ "learning_rate": 6.627680311890839e-05,
980
+ "loss": 0.0196,
981
+ "step": 1200
982
+ },
983
+ {
984
+ "epoch": 21.05263157894737,
985
+ "eval_accuracy": 0.9697777777777777,
986
+ "eval_f1": 0.9244444444444444,
987
+ "eval_loss": 0.10203476250171661,
988
+ "eval_precision": 0.930648769574944,
989
+ "eval_recall": 0.9183222958057395,
990
+ "eval_runtime": 3.9346,
991
+ "eval_samples_per_second": 57.185,
992
+ "eval_steps_per_second": 7.37,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 21.228070175438596,
997
+ "grad_norm": 0.25364601612091064,
998
+ "learning_rate": 6.49772579597141e-05,
999
+ "loss": 0.0123,
1000
+ "step": 1210
1001
+ },
1002
+ {
1003
+ "epoch": 21.403508771929825,
1004
+ "grad_norm": 0.652316153049469,
1005
+ "learning_rate": 6.367771280051981e-05,
1006
+ "loss": 0.0186,
1007
+ "step": 1220
1008
+ },
1009
+ {
1010
+ "epoch": 21.57894736842105,
1011
+ "grad_norm": 0.27665695548057556,
1012
+ "learning_rate": 6.237816764132554e-05,
1013
+ "loss": 0.0161,
1014
+ "step": 1230
1015
+ },
1016
+ {
1017
+ "epoch": 21.75438596491228,
1018
+ "grad_norm": 0.6138045787811279,
1019
+ "learning_rate": 6.107862248213126e-05,
1020
+ "loss": 0.0227,
1021
+ "step": 1240
1022
+ },
1023
+ {
1024
+ "epoch": 21.92982456140351,
1025
+ "grad_norm": 0.07879935950040817,
1026
+ "learning_rate": 5.977907732293697e-05,
1027
+ "loss": 0.0195,
1028
+ "step": 1250
1029
+ },
1030
+ {
1031
+ "epoch": 22.105263157894736,
1032
+ "grad_norm": 0.08955523371696472,
1033
+ "learning_rate": 5.847953216374269e-05,
1034
+ "loss": 0.0311,
1035
+ "step": 1260
1036
+ },
1037
+ {
1038
+ "epoch": 22.280701754385966,
1039
+ "grad_norm": 0.2407924085855484,
1040
+ "learning_rate": 5.717998700454841e-05,
1041
+ "loss": 0.0129,
1042
+ "step": 1270
1043
+ },
1044
+ {
1045
+ "epoch": 22.45614035087719,
1046
+ "grad_norm": 0.06635931879281998,
1047
+ "learning_rate": 5.588044184535413e-05,
1048
+ "loss": 0.0151,
1049
+ "step": 1280
1050
+ },
1051
+ {
1052
+ "epoch": 22.63157894736842,
1053
+ "grad_norm": 0.33943337202072144,
1054
+ "learning_rate": 5.458089668615984e-05,
1055
+ "loss": 0.0111,
1056
+ "step": 1290
1057
+ },
1058
+ {
1059
+ "epoch": 22.80701754385965,
1060
+ "grad_norm": 0.1347126066684723,
1061
+ "learning_rate": 5.328135152696556e-05,
1062
+ "loss": 0.025,
1063
+ "step": 1300
1064
+ },
1065
+ {
1066
+ "epoch": 22.80701754385965,
1067
+ "eval_accuracy": 0.9702222222222222,
1068
+ "eval_f1": 0.9258028792912514,
1069
+ "eval_loss": 0.10051768273115158,
1070
+ "eval_precision": 0.9288888888888889,
1071
+ "eval_recall": 0.9227373068432672,
1072
+ "eval_runtime": 3.7925,
1073
+ "eval_samples_per_second": 59.328,
1074
+ "eval_steps_per_second": 7.647,
1075
+ "step": 1300
1076
+ },
1077
+ {
1078
+ "epoch": 22.982456140350877,
1079
+ "grad_norm": 0.636250913143158,
1080
+ "learning_rate": 5.198180636777128e-05,
1081
+ "loss": 0.0192,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 23.157894736842106,
1086
+ "grad_norm": 0.04535467177629471,
1087
+ "learning_rate": 5.0682261208577e-05,
1088
+ "loss": 0.0056,
1089
+ "step": 1320
1090
+ },
1091
+ {
1092
+ "epoch": 23.333333333333332,
1093
+ "grad_norm": 0.5852391123771667,
1094
+ "learning_rate": 4.938271604938271e-05,
1095
+ "loss": 0.0134,
1096
+ "step": 1330
1097
+ },
1098
+ {
1099
+ "epoch": 23.50877192982456,
1100
+ "grad_norm": 0.187445729970932,
1101
+ "learning_rate": 4.8083170890188434e-05,
1102
+ "loss": 0.011,
1103
+ "step": 1340
1104
+ },
1105
+ {
1106
+ "epoch": 23.68421052631579,
1107
+ "grad_norm": 0.06503254175186157,
1108
+ "learning_rate": 4.678362573099415e-05,
1109
+ "loss": 0.0071,
1110
+ "step": 1350
1111
+ },
1112
+ {
1113
+ "epoch": 23.859649122807017,
1114
+ "grad_norm": 0.1949852555990219,
1115
+ "learning_rate": 4.548408057179987e-05,
1116
+ "loss": 0.0159,
1117
+ "step": 1360
1118
+ },
1119
+ {
1120
+ "epoch": 24.035087719298247,
1121
+ "grad_norm": 0.39588192105293274,
1122
+ "learning_rate": 4.418453541260559e-05,
1123
+ "loss": 0.021,
1124
+ "step": 1370
1125
+ },
1126
+ {
1127
+ "epoch": 24.210526315789473,
1128
+ "grad_norm": 0.39399102330207825,
1129
+ "learning_rate": 4.2884990253411305e-05,
1130
+ "loss": 0.0062,
1131
+ "step": 1380
1132
+ },
1133
+ {
1134
+ "epoch": 24.385964912280702,
1135
+ "grad_norm": 0.07124770432710648,
1136
+ "learning_rate": 4.1585445094217026e-05,
1137
+ "loss": 0.0103,
1138
+ "step": 1390
1139
+ },
1140
+ {
1141
+ "epoch": 24.56140350877193,
1142
+ "grad_norm": 0.03512560948729515,
1143
+ "learning_rate": 4.028589993502274e-05,
1144
+ "loss": 0.009,
1145
+ "step": 1400
1146
+ },
1147
+ {
1148
+ "epoch": 24.56140350877193,
1149
+ "eval_accuracy": 0.9728888888888889,
1150
+ "eval_f1": 0.9324473975636766,
1151
+ "eval_loss": 0.0976104736328125,
1152
+ "eval_precision": 0.9355555555555556,
1153
+ "eval_recall": 0.9293598233995585,
1154
+ "eval_runtime": 4.2464,
1155
+ "eval_samples_per_second": 52.986,
1156
+ "eval_steps_per_second": 6.829,
1157
+ "step": 1400
1158
+ },
1159
+ {
1160
+ "epoch": 24.736842105263158,
1161
+ "grad_norm": 0.1529558002948761,
1162
+ "learning_rate": 3.898635477582846e-05,
1163
+ "loss": 0.0163,
1164
+ "step": 1410
1165
+ },
1166
+ {
1167
+ "epoch": 24.912280701754387,
1168
+ "grad_norm": 0.01818114146590233,
1169
+ "learning_rate": 3.7686809616634176e-05,
1170
+ "loss": 0.007,
1171
+ "step": 1420
1172
+ },
1173
+ {
1174
+ "epoch": 25.087719298245613,
1175
+ "grad_norm": 0.028448551893234253,
1176
+ "learning_rate": 3.63872644574399e-05,
1177
+ "loss": 0.0102,
1178
+ "step": 1430
1179
+ },
1180
+ {
1181
+ "epoch": 25.263157894736842,
1182
+ "grad_norm": 0.46592578291893005,
1183
+ "learning_rate": 3.508771929824561e-05,
1184
+ "loss": 0.0117,
1185
+ "step": 1440
1186
+ },
1187
+ {
1188
+ "epoch": 25.43859649122807,
1189
+ "grad_norm": 0.28624555468559265,
1190
+ "learning_rate": 3.378817413905133e-05,
1191
+ "loss": 0.0098,
1192
+ "step": 1450
1193
+ },
1194
+ {
1195
+ "epoch": 25.614035087719298,
1196
+ "grad_norm": 0.29074254631996155,
1197
+ "learning_rate": 3.248862897985705e-05,
1198
+ "loss": 0.014,
1199
+ "step": 1460
1200
+ },
1201
+ {
1202
+ "epoch": 25.789473684210527,
1203
+ "grad_norm": 0.10731395334005356,
1204
+ "learning_rate": 3.118908382066277e-05,
1205
+ "loss": 0.0135,
1206
+ "step": 1470
1207
+ },
1208
+ {
1209
+ "epoch": 25.964912280701753,
1210
+ "grad_norm": 0.348628431558609,
1211
+ "learning_rate": 2.9889538661468486e-05,
1212
+ "loss": 0.0146,
1213
+ "step": 1480
1214
+ },
1215
+ {
1216
+ "epoch": 26.140350877192983,
1217
+ "grad_norm": 0.1509704291820526,
1218
+ "learning_rate": 2.8589993502274203e-05,
1219
+ "loss": 0.0122,
1220
+ "step": 1490
1221
+ },
1222
+ {
1223
+ "epoch": 26.31578947368421,
1224
+ "grad_norm": 0.18676963448524475,
1225
+ "learning_rate": 2.729044834307992e-05,
1226
+ "loss": 0.0184,
1227
+ "step": 1500
1228
+ },
1229
+ {
1230
+ "epoch": 26.31578947368421,
1231
+ "eval_accuracy": 0.9715555555555555,
1232
+ "eval_f1": 0.9290465631929047,
1233
+ "eval_loss": 0.09868455678224564,
1234
+ "eval_precision": 0.933184855233853,
1235
+ "eval_recall": 0.9249448123620309,
1236
+ "eval_runtime": 4.3646,
1237
+ "eval_samples_per_second": 51.551,
1238
+ "eval_steps_per_second": 6.644,
1239
+ "step": 1500
1240
+ },
1241
+ {
1242
+ "epoch": 26.49122807017544,
1243
+ "grad_norm": 0.08011528104543686,
1244
+ "learning_rate": 2.599090318388564e-05,
1245
+ "loss": 0.0114,
1246
+ "step": 1510
1247
+ },
1248
+ {
1249
+ "epoch": 26.666666666666668,
1250
+ "grad_norm": 0.02512381784617901,
1251
+ "learning_rate": 2.4691358024691357e-05,
1252
+ "loss": 0.0061,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 26.842105263157894,
1257
+ "grad_norm": 0.014133486896753311,
1258
+ "learning_rate": 2.3391812865497074e-05,
1259
+ "loss": 0.0085,
1260
+ "step": 1530
1261
+ },
1262
+ {
1263
+ "epoch": 27.017543859649123,
1264
+ "grad_norm": 0.1167803630232811,
1265
+ "learning_rate": 2.2092267706302795e-05,
1266
+ "loss": 0.0136,
1267
+ "step": 1540
1268
+ },
1269
+ {
1270
+ "epoch": 27.19298245614035,
1271
+ "grad_norm": 0.1177089735865593,
1272
+ "learning_rate": 2.0792722547108513e-05,
1273
+ "loss": 0.0108,
1274
+ "step": 1550
1275
+ },
1276
+ {
1277
+ "epoch": 27.36842105263158,
1278
+ "grad_norm": 0.049749329686164856,
1279
+ "learning_rate": 1.949317738791423e-05,
1280
+ "loss": 0.0082,
1281
+ "step": 1560
1282
+ },
1283
+ {
1284
+ "epoch": 27.54385964912281,
1285
+ "grad_norm": 0.027572443708777428,
1286
+ "learning_rate": 1.819363222871995e-05,
1287
+ "loss": 0.0109,
1288
+ "step": 1570
1289
+ },
1290
+ {
1291
+ "epoch": 27.719298245614034,
1292
+ "grad_norm": 0.05256260186433792,
1293
+ "learning_rate": 1.6894087069525666e-05,
1294
+ "loss": 0.0085,
1295
+ "step": 1580
1296
+ },
1297
+ {
1298
+ "epoch": 27.894736842105264,
1299
+ "grad_norm": 0.2057517021894455,
1300
+ "learning_rate": 1.5594541910331384e-05,
1301
+ "loss": 0.0129,
1302
+ "step": 1590
1303
+ },
1304
+ {
1305
+ "epoch": 28.07017543859649,
1306
+ "grad_norm": 0.014588022604584694,
1307
+ "learning_rate": 1.4294996751137102e-05,
1308
+ "loss": 0.0048,
1309
+ "step": 1600
1310
+ },
1311
+ {
1312
+ "epoch": 28.07017543859649,
1313
+ "eval_accuracy": 0.972,
1314
+ "eval_f1": 0.93007769145394,
1315
+ "eval_loss": 0.09578772634267807,
1316
+ "eval_precision": 0.9352678571428571,
1317
+ "eval_recall": 0.9249448123620309,
1318
+ "eval_runtime": 3.2602,
1319
+ "eval_samples_per_second": 69.014,
1320
+ "eval_steps_per_second": 8.895,
1321
+ "step": 1600
1322
+ },
1323
+ {
1324
+ "epoch": 28.24561403508772,
1325
+ "grad_norm": 0.01958753541111946,
1326
+ "learning_rate": 1.299545159194282e-05,
1327
+ "loss": 0.0187,
1328
+ "step": 1610
1329
+ },
1330
+ {
1331
+ "epoch": 28.42105263157895,
1332
+ "grad_norm": 0.24536529183387756,
1333
+ "learning_rate": 1.1695906432748537e-05,
1334
+ "loss": 0.0063,
1335
+ "step": 1620
1336
+ },
1337
+ {
1338
+ "epoch": 28.596491228070175,
1339
+ "grad_norm": 0.1079196035861969,
1340
+ "learning_rate": 1.0396361273554257e-05,
1341
+ "loss": 0.0095,
1342
+ "step": 1630
1343
+ },
1344
+ {
1345
+ "epoch": 28.771929824561404,
1346
+ "grad_norm": 0.3536442518234253,
1347
+ "learning_rate": 9.096816114359974e-06,
1348
+ "loss": 0.0065,
1349
+ "step": 1640
1350
+ },
1351
+ {
1352
+ "epoch": 28.94736842105263,
1353
+ "grad_norm": 0.034931574016809464,
1354
+ "learning_rate": 7.797270955165692e-06,
1355
+ "loss": 0.0116,
1356
+ "step": 1650
1357
+ },
1358
+ {
1359
+ "epoch": 29.12280701754386,
1360
+ "grad_norm": 0.21458140015602112,
1361
+ "learning_rate": 6.49772579597141e-06,
1362
+ "loss": 0.0092,
1363
+ "step": 1660
1364
+ },
1365
+ {
1366
+ "epoch": 29.29824561403509,
1367
+ "grad_norm": 0.2249114066362381,
1368
+ "learning_rate": 5.198180636777128e-06,
1369
+ "loss": 0.0119,
1370
+ "step": 1670
1371
+ },
1372
+ {
1373
+ "epoch": 29.473684210526315,
1374
+ "grad_norm": 0.14802134037017822,
1375
+ "learning_rate": 3.898635477582846e-06,
1376
+ "loss": 0.0078,
1377
+ "step": 1680
1378
+ },
1379
+ {
1380
+ "epoch": 29.649122807017545,
1381
+ "grad_norm": 0.35757485032081604,
1382
+ "learning_rate": 2.599090318388564e-06,
1383
+ "loss": 0.0173,
1384
+ "step": 1690
1385
+ },
1386
+ {
1387
+ "epoch": 29.82456140350877,
1388
+ "grad_norm": 0.05533298850059509,
1389
+ "learning_rate": 1.299545159194282e-06,
1390
+ "loss": 0.0072,
1391
+ "step": 1700
1392
+ },
1393
+ {
1394
+ "epoch": 29.82456140350877,
1395
+ "eval_accuracy": 0.972,
1396
+ "eval_f1": 0.93007769145394,
1397
+ "eval_loss": 0.09479136019945145,
1398
+ "eval_precision": 0.9352678571428571,
1399
+ "eval_recall": 0.9249448123620309,
1400
+ "eval_runtime": 3.2997,
1401
+ "eval_samples_per_second": 68.189,
1402
+ "eval_steps_per_second": 8.789,
1403
+ "step": 1700
1404
+ },
1405
+ {
1406
+ "epoch": 30.0,
1407
+ "grad_norm": 0.050033435225486755,
1408
+ "learning_rate": 0.0,
1409
+ "loss": 0.0053,
1410
+ "step": 1710
1411
+ },
1412
+ {
1413
+ "epoch": 30.0,
1414
+ "step": 1710,
1415
+ "total_flos": 6.170107339800576e+18,
1416
+ "train_loss": 0.032361031031748005,
1417
+ "train_runtime": 1047.1267,
1418
+ "train_samples_per_second": 25.785,
1419
+ "train_steps_per_second": 1.633
1420
+ }
1421
+ ],
1422
+ "logging_steps": 10,
1423
+ "max_steps": 1710,
1424
+ "num_input_tokens_seen": 0,
1425
+ "num_train_epochs": 30,
1426
+ "save_steps": 500,
1427
+ "stateful_callbacks": {
1428
+ "TrainerControl": {
1429
+ "args": {
1430
+ "should_epoch_stop": false,
1431
+ "should_evaluate": false,
1432
+ "should_log": false,
1433
+ "should_save": true,
1434
+ "should_training_stop": true
1435
+ },
1436
+ "attributes": {}
1437
+ }
1438
+ },
1439
+ "total_flos": 6.170107339800576e+18,
1440
+ "train_batch_size": 16,
1441
+ "trial_name": null,
1442
+ "trial_params": null
1443
+ }