nemik commited on
Commit
814cf95
1 Parent(s): db80b82

frostsolutions/frost-vision-v2-google_vit-base-patch16-384-v2024-11-10

Browse files
README.md CHANGED
@@ -26,16 +26,16 @@ model-index:
26
  metrics:
27
  - name: Accuracy
28
  type: accuracy
29
- value: 0.9376760563380282
30
  - name: F1
31
  type: f1
32
- value: 0.8415398388540735
33
  - name: Precision
34
  type: precision
35
- value: 0.8545454545454545
36
  - name: Recall
37
  type: recall
38
- value: 0.8289241622574955
39
  ---
40
 
41
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -45,11 +45,11 @@ should probably proofread and complete it, then remove this comment. -->
45
 
46
  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.
47
  It achieves the following results on the evaluation set:
48
- - Loss: 0.2101
49
- - Accuracy: 0.9377
50
- - F1: 0.8415
51
- - Precision: 0.8545
52
- - Recall: 0.8289
53
 
54
  ## Model description
55
 
 
26
  metrics:
27
  - name: Accuracy
28
  type: accuracy
29
+ value: 0.9274647887323944
30
  - name: F1
31
  type: f1
32
+ value: 0.8186619718309859
33
  - name: Precision
34
  type: precision
35
+ value: 0.8172231985940246
36
  - name: Recall
37
  type: recall
38
+ value: 0.8201058201058201
39
  ---
40
 
41
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
45
 
46
  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.
47
  It achieves the following results on the evaluation set:
48
+ - Loss: 0.1847
49
+ - Accuracy: 0.9275
50
+ - F1: 0.8187
51
+ - Precision: 0.8172
52
+ - Recall: 0.8201
53
 
54
  ## Model description
55
 
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "eval_accuracy": 0.9274647887323944,
4
+ "eval_f1": 0.8186619718309859,
5
+ "eval_loss": 0.18472303450107574,
6
+ "eval_precision": 0.8172231985940246,
7
+ "eval_recall": 0.8201058201058201,
8
+ "eval_runtime": 4.289,
9
+ "eval_samples_per_second": 66.215,
10
+ "eval_steps_per_second": 8.393,
11
+ "total_flos": 7.774335248148726e+18,
12
+ "train_loss": 0.08187752739294595,
13
+ "train_runtime": 1287.2767,
14
+ "train_samples_per_second": 26.428,
15
+ "train_steps_per_second": 1.655
16
+ }
eval_results.json ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "eval_accuracy": 0.9274647887323944,
4
+ "eval_f1": 0.8186619718309859,
5
+ "eval_loss": 0.18472303450107574,
6
+ "eval_precision": 0.8172231985940246,
7
+ "eval_recall": 0.8201058201058201,
8
+ "eval_runtime": 4.289,
9
+ "eval_samples_per_second": 66.215,
10
+ "eval_steps_per_second": 8.393
11
+ }
runs/Nov10_06-44-55_a61ec4b32b41/events.out.tfevents.1731222412.a61ec4b32b41.1365.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dbaca6c6ddfae805d576c7b10c3010d17bfa72a5ec8c6b2f7f27f216084dc00a
3
+ size 560
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 30.0,
3
+ "total_flos": 7.774335248148726e+18,
4
+ "train_loss": 0.08187752739294595,
5
+ "train_runtime": 1287.2767,
6
+ "train_samples_per_second": 26.428,
7
+ "train_steps_per_second": 1.655
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,1785 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.18472303450107574,
3
+ "best_model_checkpoint": "frost-vision-v2-google_vit-base-patch16-384-v2024-11-10/checkpoint-500",
4
+ "epoch": 30.0,
5
+ "eval_steps": 100,
6
+ "global_step": 2130,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.14084507042253522,
13
+ "grad_norm": 1.9532430171966553,
14
+ "learning_rate": 9.389671361502347e-06,
15
+ "loss": 0.7081,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.28169014084507044,
20
+ "grad_norm": 1.4905153512954712,
21
+ "learning_rate": 1.8779342723004694e-05,
22
+ "loss": 0.5966,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.4225352112676056,
27
+ "grad_norm": 1.0629397630691528,
28
+ "learning_rate": 2.8169014084507046e-05,
29
+ "loss": 0.4634,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.5633802816901409,
34
+ "grad_norm": 1.1944268941879272,
35
+ "learning_rate": 3.755868544600939e-05,
36
+ "loss": 0.4103,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.704225352112676,
41
+ "grad_norm": 1.0600364208221436,
42
+ "learning_rate": 4.694835680751174e-05,
43
+ "loss": 0.3456,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.8450704225352113,
48
+ "grad_norm": 0.9409961700439453,
49
+ "learning_rate": 5.633802816901409e-05,
50
+ "loss": 0.2917,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.9859154929577465,
55
+ "grad_norm": 0.9866936206817627,
56
+ "learning_rate": 6.572769953051644e-05,
57
+ "loss": 0.2536,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 1.1267605633802817,
62
+ "grad_norm": 0.8040021657943726,
63
+ "learning_rate": 7.511737089201878e-05,
64
+ "loss": 0.236,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 1.267605633802817,
69
+ "grad_norm": 0.7174156904220581,
70
+ "learning_rate": 8.450704225352113e-05,
71
+ "loss": 0.2305,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 1.408450704225352,
76
+ "grad_norm": 1.468618631362915,
77
+ "learning_rate": 9.389671361502347e-05,
78
+ "loss": 0.2243,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 1.408450704225352,
83
+ "eval_accuracy": 0.9242957746478874,
84
+ "eval_f1": 0.7981220657276995,
85
+ "eval_loss": 0.20880411565303802,
86
+ "eval_precision": 0.8534136546184738,
87
+ "eval_recall": 0.7495590828924162,
88
+ "eval_runtime": 4.7453,
89
+ "eval_samples_per_second": 59.848,
90
+ "eval_steps_per_second": 7.586,
91
+ "step": 100
92
+ },
93
+ {
94
+ "epoch": 1.5492957746478875,
95
+ "grad_norm": 1.7064496278762817,
96
+ "learning_rate": 0.00010328638497652582,
97
+ "loss": 0.2272,
98
+ "step": 110
99
+ },
100
+ {
101
+ "epoch": 1.6901408450704225,
102
+ "grad_norm": 0.8617244362831116,
103
+ "learning_rate": 0.00011267605633802819,
104
+ "loss": 0.2142,
105
+ "step": 120
106
+ },
107
+ {
108
+ "epoch": 1.8309859154929577,
109
+ "grad_norm": 0.8898526430130005,
110
+ "learning_rate": 0.00012206572769953053,
111
+ "loss": 0.2208,
112
+ "step": 130
113
+ },
114
+ {
115
+ "epoch": 1.971830985915493,
116
+ "grad_norm": 1.2177144289016724,
117
+ "learning_rate": 0.00013145539906103288,
118
+ "loss": 0.2147,
119
+ "step": 140
120
+ },
121
+ {
122
+ "epoch": 2.112676056338028,
123
+ "grad_norm": 1.7734003067016602,
124
+ "learning_rate": 0.00014084507042253522,
125
+ "loss": 0.1993,
126
+ "step": 150
127
+ },
128
+ {
129
+ "epoch": 2.2535211267605635,
130
+ "grad_norm": 0.9507364630699158,
131
+ "learning_rate": 0.00015023474178403755,
132
+ "loss": 0.1806,
133
+ "step": 160
134
+ },
135
+ {
136
+ "epoch": 2.3943661971830985,
137
+ "grad_norm": 0.9511057734489441,
138
+ "learning_rate": 0.00015962441314553992,
139
+ "loss": 0.1913,
140
+ "step": 170
141
+ },
142
+ {
143
+ "epoch": 2.535211267605634,
144
+ "grad_norm": 1.419704556465149,
145
+ "learning_rate": 0.00016901408450704225,
146
+ "loss": 0.1949,
147
+ "step": 180
148
+ },
149
+ {
150
+ "epoch": 2.676056338028169,
151
+ "grad_norm": 0.8608860373497009,
152
+ "learning_rate": 0.00017840375586854461,
153
+ "loss": 0.1907,
154
+ "step": 190
155
+ },
156
+ {
157
+ "epoch": 2.816901408450704,
158
+ "grad_norm": 1.1266043186187744,
159
+ "learning_rate": 0.00018779342723004695,
160
+ "loss": 0.2438,
161
+ "step": 200
162
+ },
163
+ {
164
+ "epoch": 2.816901408450704,
165
+ "eval_accuracy": 0.9299295774647888,
166
+ "eval_f1": 0.8102955195424214,
167
+ "eval_loss": 0.18185150623321533,
168
+ "eval_precision": 0.8817427385892116,
169
+ "eval_recall": 0.7495590828924162,
170
+ "eval_runtime": 4.6977,
171
+ "eval_samples_per_second": 60.455,
172
+ "eval_steps_per_second": 7.663,
173
+ "step": 200
174
+ },
175
+ {
176
+ "epoch": 2.9577464788732395,
177
+ "grad_norm": 0.9358466267585754,
178
+ "learning_rate": 0.0001971830985915493,
179
+ "loss": 0.1805,
180
+ "step": 210
181
+ },
182
+ {
183
+ "epoch": 3.0985915492957745,
184
+ "grad_norm": 0.5238935351371765,
185
+ "learning_rate": 0.00019926969222743872,
186
+ "loss": 0.1614,
187
+ "step": 220
188
+ },
189
+ {
190
+ "epoch": 3.23943661971831,
191
+ "grad_norm": 0.8152980208396912,
192
+ "learning_rate": 0.00019822639540949402,
193
+ "loss": 0.1632,
194
+ "step": 230
195
+ },
196
+ {
197
+ "epoch": 3.380281690140845,
198
+ "grad_norm": 0.8670392036437988,
199
+ "learning_rate": 0.0001971830985915493,
200
+ "loss": 0.17,
201
+ "step": 240
202
+ },
203
+ {
204
+ "epoch": 3.52112676056338,
205
+ "grad_norm": 0.719203531742096,
206
+ "learning_rate": 0.0001961398017736046,
207
+ "loss": 0.1843,
208
+ "step": 250
209
+ },
210
+ {
211
+ "epoch": 3.6619718309859155,
212
+ "grad_norm": 1.1024322509765625,
213
+ "learning_rate": 0.0001950965049556599,
214
+ "loss": 0.1843,
215
+ "step": 260
216
+ },
217
+ {
218
+ "epoch": 3.802816901408451,
219
+ "grad_norm": 1.002933382987976,
220
+ "learning_rate": 0.0001940532081377152,
221
+ "loss": 0.1657,
222
+ "step": 270
223
+ },
224
+ {
225
+ "epoch": 3.943661971830986,
226
+ "grad_norm": 0.6921798586845398,
227
+ "learning_rate": 0.0001930099113197705,
228
+ "loss": 0.1881,
229
+ "step": 280
230
+ },
231
+ {
232
+ "epoch": 4.084507042253521,
233
+ "grad_norm": 0.5872893333435059,
234
+ "learning_rate": 0.0001919666145018258,
235
+ "loss": 0.1775,
236
+ "step": 290
237
+ },
238
+ {
239
+ "epoch": 4.225352112676056,
240
+ "grad_norm": 0.7991660833358765,
241
+ "learning_rate": 0.0001909233176838811,
242
+ "loss": 0.1338,
243
+ "step": 300
244
+ },
245
+ {
246
+ "epoch": 4.225352112676056,
247
+ "eval_accuracy": 0.9376760563380282,
248
+ "eval_f1": 0.8448729184925504,
249
+ "eval_loss": 0.1607813537120819,
250
+ "eval_precision": 0.8397212543554007,
251
+ "eval_recall": 0.8500881834215167,
252
+ "eval_runtime": 4.716,
253
+ "eval_samples_per_second": 60.22,
254
+ "eval_steps_per_second": 7.634,
255
+ "step": 300
256
+ },
257
+ {
258
+ "epoch": 4.366197183098592,
259
+ "grad_norm": 0.3925558030605316,
260
+ "learning_rate": 0.00018988002086593636,
261
+ "loss": 0.1397,
262
+ "step": 310
263
+ },
264
+ {
265
+ "epoch": 4.507042253521127,
266
+ "grad_norm": 0.9107651114463806,
267
+ "learning_rate": 0.00018883672404799165,
268
+ "loss": 0.1328,
269
+ "step": 320
270
+ },
271
+ {
272
+ "epoch": 4.647887323943662,
273
+ "grad_norm": 0.6482114791870117,
274
+ "learning_rate": 0.00018779342723004695,
275
+ "loss": 0.1343,
276
+ "step": 330
277
+ },
278
+ {
279
+ "epoch": 4.788732394366197,
280
+ "grad_norm": 0.8836609125137329,
281
+ "learning_rate": 0.00018675013041210225,
282
+ "loss": 0.1527,
283
+ "step": 340
284
+ },
285
+ {
286
+ "epoch": 4.929577464788732,
287
+ "grad_norm": 0.8424317240715027,
288
+ "learning_rate": 0.00018570683359415754,
289
+ "loss": 0.1338,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 5.070422535211268,
294
+ "grad_norm": 0.824109673500061,
295
+ "learning_rate": 0.00018466353677621284,
296
+ "loss": 0.1455,
297
+ "step": 360
298
+ },
299
+ {
300
+ "epoch": 5.211267605633803,
301
+ "grad_norm": 0.7583167552947998,
302
+ "learning_rate": 0.00018362023995826813,
303
+ "loss": 0.1045,
304
+ "step": 370
305
+ },
306
+ {
307
+ "epoch": 5.352112676056338,
308
+ "grad_norm": 0.9834734797477722,
309
+ "learning_rate": 0.00018257694314032343,
310
+ "loss": 0.1297,
311
+ "step": 380
312
+ },
313
+ {
314
+ "epoch": 5.492957746478873,
315
+ "grad_norm": 0.6775982975959778,
316
+ "learning_rate": 0.00018153364632237873,
317
+ "loss": 0.1382,
318
+ "step": 390
319
+ },
320
+ {
321
+ "epoch": 5.633802816901408,
322
+ "grad_norm": 0.4346470236778259,
323
+ "learning_rate": 0.00018049034950443402,
324
+ "loss": 0.1224,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 5.633802816901408,
329
+ "eval_accuracy": 0.9271126760563381,
330
+ "eval_f1": 0.8179419525065963,
331
+ "eval_loss": 0.17346476018428802,
332
+ "eval_precision": 0.8157894736842105,
333
+ "eval_recall": 0.8201058201058201,
334
+ "eval_runtime": 4.3151,
335
+ "eval_samples_per_second": 65.816,
336
+ "eval_steps_per_second": 8.343,
337
+ "step": 400
338
+ },
339
+ {
340
+ "epoch": 5.774647887323944,
341
+ "grad_norm": 0.6250389218330383,
342
+ "learning_rate": 0.00017944705268648932,
343
+ "loss": 0.1535,
344
+ "step": 410
345
+ },
346
+ {
347
+ "epoch": 5.915492957746479,
348
+ "grad_norm": 0.7581302523612976,
349
+ "learning_rate": 0.00017840375586854461,
350
+ "loss": 0.1217,
351
+ "step": 420
352
+ },
353
+ {
354
+ "epoch": 6.056338028169014,
355
+ "grad_norm": 1.0700663328170776,
356
+ "learning_rate": 0.0001773604590505999,
357
+ "loss": 0.1206,
358
+ "step": 430
359
+ },
360
+ {
361
+ "epoch": 6.197183098591549,
362
+ "grad_norm": 0.9656069278717041,
363
+ "learning_rate": 0.0001763171622326552,
364
+ "loss": 0.1063,
365
+ "step": 440
366
+ },
367
+ {
368
+ "epoch": 6.338028169014084,
369
+ "grad_norm": 0.8910039067268372,
370
+ "learning_rate": 0.0001752738654147105,
371
+ "loss": 0.1105,
372
+ "step": 450
373
+ },
374
+ {
375
+ "epoch": 6.47887323943662,
376
+ "grad_norm": 1.0498372316360474,
377
+ "learning_rate": 0.0001742305685967658,
378
+ "loss": 0.1064,
379
+ "step": 460
380
+ },
381
+ {
382
+ "epoch": 6.619718309859155,
383
+ "grad_norm": 0.8038674592971802,
384
+ "learning_rate": 0.00017318727177882107,
385
+ "loss": 0.1048,
386
+ "step": 470
387
+ },
388
+ {
389
+ "epoch": 6.76056338028169,
390
+ "grad_norm": 0.5430222749710083,
391
+ "learning_rate": 0.00017214397496087636,
392
+ "loss": 0.1081,
393
+ "step": 480
394
+ },
395
+ {
396
+ "epoch": 6.901408450704225,
397
+ "grad_norm": 0.6888221502304077,
398
+ "learning_rate": 0.00017110067814293166,
399
+ "loss": 0.1176,
400
+ "step": 490
401
+ },
402
+ {
403
+ "epoch": 7.042253521126761,
404
+ "grad_norm": 0.5913430452346802,
405
+ "learning_rate": 0.00017005738132498696,
406
+ "loss": 0.1065,
407
+ "step": 500
408
+ },
409
+ {
410
+ "epoch": 7.042253521126761,
411
+ "eval_accuracy": 0.9274647887323944,
412
+ "eval_f1": 0.8186619718309859,
413
+ "eval_loss": 0.18472303450107574,
414
+ "eval_precision": 0.8172231985940246,
415
+ "eval_recall": 0.8201058201058201,
416
+ "eval_runtime": 4.2833,
417
+ "eval_samples_per_second": 66.305,
418
+ "eval_steps_per_second": 8.405,
419
+ "step": 500
420
+ },
421
+ {
422
+ "epoch": 7.183098591549296,
423
+ "grad_norm": 0.7211888432502747,
424
+ "learning_rate": 0.00016901408450704225,
425
+ "loss": 0.0907,
426
+ "step": 510
427
+ },
428
+ {
429
+ "epoch": 7.323943661971831,
430
+ "grad_norm": 0.9712802767753601,
431
+ "learning_rate": 0.00016797078768909755,
432
+ "loss": 0.1174,
433
+ "step": 520
434
+ },
435
+ {
436
+ "epoch": 7.464788732394366,
437
+ "grad_norm": 0.5647371411323547,
438
+ "learning_rate": 0.00016692749087115284,
439
+ "loss": 0.1188,
440
+ "step": 530
441
+ },
442
+ {
443
+ "epoch": 7.605633802816901,
444
+ "grad_norm": 0.8738060593605042,
445
+ "learning_rate": 0.00016588419405320814,
446
+ "loss": 0.1261,
447
+ "step": 540
448
+ },
449
+ {
450
+ "epoch": 7.746478873239437,
451
+ "grad_norm": 0.4535655081272125,
452
+ "learning_rate": 0.00016484089723526344,
453
+ "loss": 0.0865,
454
+ "step": 550
455
+ },
456
+ {
457
+ "epoch": 7.887323943661972,
458
+ "grad_norm": 0.4072222411632538,
459
+ "learning_rate": 0.00016379760041731873,
460
+ "loss": 0.0974,
461
+ "step": 560
462
+ },
463
+ {
464
+ "epoch": 8.028169014084508,
465
+ "grad_norm": 0.36978864669799805,
466
+ "learning_rate": 0.00016275430359937403,
467
+ "loss": 0.0989,
468
+ "step": 570
469
+ },
470
+ {
471
+ "epoch": 8.169014084507042,
472
+ "grad_norm": 0.6507691740989685,
473
+ "learning_rate": 0.00016171100678142932,
474
+ "loss": 0.0989,
475
+ "step": 580
476
+ },
477
+ {
478
+ "epoch": 8.309859154929578,
479
+ "grad_norm": 0.8923565149307251,
480
+ "learning_rate": 0.00016066770996348462,
481
+ "loss": 0.0901,
482
+ "step": 590
483
+ },
484
+ {
485
+ "epoch": 8.450704225352112,
486
+ "grad_norm": 0.8466383814811707,
487
+ "learning_rate": 0.00015962441314553992,
488
+ "loss": 0.1008,
489
+ "step": 600
490
+ },
491
+ {
492
+ "epoch": 8.450704225352112,
493
+ "eval_accuracy": 0.9404929577464789,
494
+ "eval_f1": 0.8505747126436781,
495
+ "eval_loss": 0.17097999155521393,
496
+ "eval_precision": 0.8528368794326241,
497
+ "eval_recall": 0.8483245149911817,
498
+ "eval_runtime": 3.9495,
499
+ "eval_samples_per_second": 71.908,
500
+ "eval_steps_per_second": 9.115,
501
+ "step": 600
502
+ },
503
+ {
504
+ "epoch": 8.591549295774648,
505
+ "grad_norm": 0.568720817565918,
506
+ "learning_rate": 0.0001585811163275952,
507
+ "loss": 0.084,
508
+ "step": 610
509
+ },
510
+ {
511
+ "epoch": 8.732394366197184,
512
+ "grad_norm": 0.6025291085243225,
513
+ "learning_rate": 0.0001575378195096505,
514
+ "loss": 0.1059,
515
+ "step": 620
516
+ },
517
+ {
518
+ "epoch": 8.873239436619718,
519
+ "grad_norm": 0.6547625660896301,
520
+ "learning_rate": 0.0001564945226917058,
521
+ "loss": 0.0871,
522
+ "step": 630
523
+ },
524
+ {
525
+ "epoch": 9.014084507042254,
526
+ "grad_norm": 0.46392133831977844,
527
+ "learning_rate": 0.00015545122587376107,
528
+ "loss": 0.0856,
529
+ "step": 640
530
+ },
531
+ {
532
+ "epoch": 9.154929577464788,
533
+ "grad_norm": 0.29296088218688965,
534
+ "learning_rate": 0.00015440792905581637,
535
+ "loss": 0.0883,
536
+ "step": 650
537
+ },
538
+ {
539
+ "epoch": 9.295774647887324,
540
+ "grad_norm": 0.6499446630477905,
541
+ "learning_rate": 0.00015336463223787167,
542
+ "loss": 0.0632,
543
+ "step": 660
544
+ },
545
+ {
546
+ "epoch": 9.43661971830986,
547
+ "grad_norm": 0.27808037400245667,
548
+ "learning_rate": 0.00015232133541992696,
549
+ "loss": 0.0731,
550
+ "step": 670
551
+ },
552
+ {
553
+ "epoch": 9.577464788732394,
554
+ "grad_norm": 0.6585447788238525,
555
+ "learning_rate": 0.00015127803860198226,
556
+ "loss": 0.073,
557
+ "step": 680
558
+ },
559
+ {
560
+ "epoch": 9.71830985915493,
561
+ "grad_norm": 0.5550310611724854,
562
+ "learning_rate": 0.00015023474178403755,
563
+ "loss": 0.0762,
564
+ "step": 690
565
+ },
566
+ {
567
+ "epoch": 9.859154929577464,
568
+ "grad_norm": 0.8603278398513794,
569
+ "learning_rate": 0.00014919144496609285,
570
+ "loss": 0.1005,
571
+ "step": 700
572
+ },
573
+ {
574
+ "epoch": 9.859154929577464,
575
+ "eval_accuracy": 0.9383802816901409,
576
+ "eval_f1": 0.8404740200546946,
577
+ "eval_loss": 0.18234650790691376,
578
+ "eval_precision": 0.869811320754717,
579
+ "eval_recall": 0.8130511463844797,
580
+ "eval_runtime": 4.0053,
581
+ "eval_samples_per_second": 70.906,
582
+ "eval_steps_per_second": 8.988,
583
+ "step": 700
584
+ },
585
+ {
586
+ "epoch": 10.0,
587
+ "grad_norm": 0.5497912168502808,
588
+ "learning_rate": 0.00014814814814814815,
589
+ "loss": 0.0828,
590
+ "step": 710
591
+ },
592
+ {
593
+ "epoch": 10.140845070422536,
594
+ "grad_norm": 0.5562108755111694,
595
+ "learning_rate": 0.00014710485133020344,
596
+ "loss": 0.0771,
597
+ "step": 720
598
+ },
599
+ {
600
+ "epoch": 10.28169014084507,
601
+ "grad_norm": 0.5177081823348999,
602
+ "learning_rate": 0.00014606155451225874,
603
+ "loss": 0.0699,
604
+ "step": 730
605
+ },
606
+ {
607
+ "epoch": 10.422535211267606,
608
+ "grad_norm": 1.1959189176559448,
609
+ "learning_rate": 0.00014501825769431403,
610
+ "loss": 0.0706,
611
+ "step": 740
612
+ },
613
+ {
614
+ "epoch": 10.56338028169014,
615
+ "grad_norm": 0.41302013397216797,
616
+ "learning_rate": 0.00014397496087636933,
617
+ "loss": 0.0562,
618
+ "step": 750
619
+ },
620
+ {
621
+ "epoch": 10.704225352112676,
622
+ "grad_norm": 0.8963214159011841,
623
+ "learning_rate": 0.00014293166405842463,
624
+ "loss": 0.0795,
625
+ "step": 760
626
+ },
627
+ {
628
+ "epoch": 10.845070422535212,
629
+ "grad_norm": 0.5553390979766846,
630
+ "learning_rate": 0.00014188836724047992,
631
+ "loss": 0.0841,
632
+ "step": 770
633
+ },
634
+ {
635
+ "epoch": 10.985915492957746,
636
+ "grad_norm": 0.5120792984962463,
637
+ "learning_rate": 0.00014084507042253522,
638
+ "loss": 0.0571,
639
+ "step": 780
640
+ },
641
+ {
642
+ "epoch": 11.126760563380282,
643
+ "grad_norm": 0.8299622535705566,
644
+ "learning_rate": 0.00013980177360459051,
645
+ "loss": 0.0674,
646
+ "step": 790
647
+ },
648
+ {
649
+ "epoch": 11.267605633802816,
650
+ "grad_norm": 0.8283889889717102,
651
+ "learning_rate": 0.00013875847678664578,
652
+ "loss": 0.0756,
653
+ "step": 800
654
+ },
655
+ {
656
+ "epoch": 11.267605633802816,
657
+ "eval_accuracy": 0.9415492957746479,
658
+ "eval_f1": 0.8520499108734403,
659
+ "eval_loss": 0.1770765483379364,
660
+ "eval_precision": 0.8612612612612612,
661
+ "eval_recall": 0.8430335097001763,
662
+ "eval_runtime": 3.9753,
663
+ "eval_samples_per_second": 71.441,
664
+ "eval_steps_per_second": 9.056,
665
+ "step": 800
666
+ },
667
+ {
668
+ "epoch": 11.408450704225352,
669
+ "grad_norm": 0.7964696288108826,
670
+ "learning_rate": 0.00013771517996870108,
671
+ "loss": 0.0674,
672
+ "step": 810
673
+ },
674
+ {
675
+ "epoch": 11.549295774647888,
676
+ "grad_norm": 0.5819689631462097,
677
+ "learning_rate": 0.00013667188315075638,
678
+ "loss": 0.0744,
679
+ "step": 820
680
+ },
681
+ {
682
+ "epoch": 11.690140845070422,
683
+ "grad_norm": 0.4807089865207672,
684
+ "learning_rate": 0.0001356285863328117,
685
+ "loss": 0.09,
686
+ "step": 830
687
+ },
688
+ {
689
+ "epoch": 11.830985915492958,
690
+ "grad_norm": 0.4166727364063263,
691
+ "learning_rate": 0.000134585289514867,
692
+ "loss": 0.057,
693
+ "step": 840
694
+ },
695
+ {
696
+ "epoch": 11.971830985915492,
697
+ "grad_norm": 0.5973660349845886,
698
+ "learning_rate": 0.0001335419926969223,
699
+ "loss": 0.0678,
700
+ "step": 850
701
+ },
702
+ {
703
+ "epoch": 12.112676056338028,
704
+ "grad_norm": 0.8271676898002625,
705
+ "learning_rate": 0.0001324986958789776,
706
+ "loss": 0.0676,
707
+ "step": 860
708
+ },
709
+ {
710
+ "epoch": 12.253521126760564,
711
+ "grad_norm": 0.4148211181163788,
712
+ "learning_rate": 0.00013145539906103288,
713
+ "loss": 0.0724,
714
+ "step": 870
715
+ },
716
+ {
717
+ "epoch": 12.394366197183098,
718
+ "grad_norm": 0.8119566440582275,
719
+ "learning_rate": 0.00013041210224308818,
720
+ "loss": 0.0513,
721
+ "step": 880
722
+ },
723
+ {
724
+ "epoch": 12.535211267605634,
725
+ "grad_norm": 0.3651484549045563,
726
+ "learning_rate": 0.00012936880542514348,
727
+ "loss": 0.0681,
728
+ "step": 890
729
+ },
730
+ {
731
+ "epoch": 12.676056338028168,
732
+ "grad_norm": 0.5199844241142273,
733
+ "learning_rate": 0.00012832550860719877,
734
+ "loss": 0.0653,
735
+ "step": 900
736
+ },
737
+ {
738
+ "epoch": 12.676056338028168,
739
+ "eval_accuracy": 0.9323943661971831,
740
+ "eval_f1": 0.8309859154929577,
741
+ "eval_loss": 0.1971179097890854,
742
+ "eval_precision": 0.8295254833040422,
743
+ "eval_recall": 0.8324514991181657,
744
+ "eval_runtime": 4.0066,
745
+ "eval_samples_per_second": 70.884,
746
+ "eval_steps_per_second": 8.985,
747
+ "step": 900
748
+ },
749
+ {
750
+ "epoch": 12.816901408450704,
751
+ "grad_norm": 0.5670329332351685,
752
+ "learning_rate": 0.00012728221178925407,
753
+ "loss": 0.0619,
754
+ "step": 910
755
+ },
756
+ {
757
+ "epoch": 12.95774647887324,
758
+ "grad_norm": 0.5461100935935974,
759
+ "learning_rate": 0.00012623891497130936,
760
+ "loss": 0.0411,
761
+ "step": 920
762
+ },
763
+ {
764
+ "epoch": 13.098591549295774,
765
+ "grad_norm": 0.5733127593994141,
766
+ "learning_rate": 0.00012519561815336466,
767
+ "loss": 0.0544,
768
+ "step": 930
769
+ },
770
+ {
771
+ "epoch": 13.23943661971831,
772
+ "grad_norm": 1.001535177230835,
773
+ "learning_rate": 0.00012415232133541993,
774
+ "loss": 0.0723,
775
+ "step": 940
776
+ },
777
+ {
778
+ "epoch": 13.380281690140846,
779
+ "grad_norm": 0.4734277129173279,
780
+ "learning_rate": 0.00012310902451747523,
781
+ "loss": 0.0457,
782
+ "step": 950
783
+ },
784
+ {
785
+ "epoch": 13.52112676056338,
786
+ "grad_norm": 0.41569235920906067,
787
+ "learning_rate": 0.00012206572769953053,
788
+ "loss": 0.0611,
789
+ "step": 960
790
+ },
791
+ {
792
+ "epoch": 13.661971830985916,
793
+ "grad_norm": 0.656200110912323,
794
+ "learning_rate": 0.00012102243088158583,
795
+ "loss": 0.0578,
796
+ "step": 970
797
+ },
798
+ {
799
+ "epoch": 13.80281690140845,
800
+ "grad_norm": 0.4374938905239105,
801
+ "learning_rate": 0.00011997913406364111,
802
+ "loss": 0.0493,
803
+ "step": 980
804
+ },
805
+ {
806
+ "epoch": 13.943661971830986,
807
+ "grad_norm": 0.4476602077484131,
808
+ "learning_rate": 0.00011893583724569641,
809
+ "loss": 0.0645,
810
+ "step": 990
811
+ },
812
+ {
813
+ "epoch": 14.084507042253522,
814
+ "grad_norm": 0.4841475784778595,
815
+ "learning_rate": 0.0001178925404277517,
816
+ "loss": 0.0367,
817
+ "step": 1000
818
+ },
819
+ {
820
+ "epoch": 14.084507042253522,
821
+ "eval_accuracy": 0.9295774647887324,
822
+ "eval_f1": 0.8220640569395018,
823
+ "eval_loss": 0.2123415768146515,
824
+ "eval_precision": 0.829443447037702,
825
+ "eval_recall": 0.8148148148148148,
826
+ "eval_runtime": 3.9803,
827
+ "eval_samples_per_second": 71.351,
828
+ "eval_steps_per_second": 9.045,
829
+ "step": 1000
830
+ },
831
+ {
832
+ "epoch": 14.225352112676056,
833
+ "grad_norm": 0.36940839886665344,
834
+ "learning_rate": 0.000116849243609807,
835
+ "loss": 0.0583,
836
+ "step": 1010
837
+ },
838
+ {
839
+ "epoch": 14.366197183098592,
840
+ "grad_norm": 0.360748291015625,
841
+ "learning_rate": 0.0001158059467918623,
842
+ "loss": 0.0557,
843
+ "step": 1020
844
+ },
845
+ {
846
+ "epoch": 14.507042253521126,
847
+ "grad_norm": 0.5215537548065186,
848
+ "learning_rate": 0.0001147626499739176,
849
+ "loss": 0.0687,
850
+ "step": 1030
851
+ },
852
+ {
853
+ "epoch": 14.647887323943662,
854
+ "grad_norm": 0.504537045955658,
855
+ "learning_rate": 0.00011371935315597289,
856
+ "loss": 0.07,
857
+ "step": 1040
858
+ },
859
+ {
860
+ "epoch": 14.788732394366198,
861
+ "grad_norm": 0.6165650486946106,
862
+ "learning_rate": 0.00011267605633802819,
863
+ "loss": 0.0527,
864
+ "step": 1050
865
+ },
866
+ {
867
+ "epoch": 14.929577464788732,
868
+ "grad_norm": 0.4196755588054657,
869
+ "learning_rate": 0.00011163275952008347,
870
+ "loss": 0.039,
871
+ "step": 1060
872
+ },
873
+ {
874
+ "epoch": 15.070422535211268,
875
+ "grad_norm": 0.789492130279541,
876
+ "learning_rate": 0.00011058946270213876,
877
+ "loss": 0.0518,
878
+ "step": 1070
879
+ },
880
+ {
881
+ "epoch": 15.211267605633802,
882
+ "grad_norm": 0.3040258288383484,
883
+ "learning_rate": 0.00010954616588419406,
884
+ "loss": 0.0426,
885
+ "step": 1080
886
+ },
887
+ {
888
+ "epoch": 15.352112676056338,
889
+ "grad_norm": 0.5353887677192688,
890
+ "learning_rate": 0.00010850286906624936,
891
+ "loss": 0.06,
892
+ "step": 1090
893
+ },
894
+ {
895
+ "epoch": 15.492957746478874,
896
+ "grad_norm": 0.4748779833316803,
897
+ "learning_rate": 0.00010745957224830465,
898
+ "loss": 0.0459,
899
+ "step": 1100
900
+ },
901
+ {
902
+ "epoch": 15.492957746478874,
903
+ "eval_accuracy": 0.9334507042253521,
904
+ "eval_f1": 0.832,
905
+ "eval_loss": 0.20056356489658356,
906
+ "eval_precision": 0.8387096774193549,
907
+ "eval_recall": 0.8253968253968254,
908
+ "eval_runtime": 3.8873,
909
+ "eval_samples_per_second": 73.058,
910
+ "eval_steps_per_second": 9.261,
911
+ "step": 1100
912
+ },
913
+ {
914
+ "epoch": 15.633802816901408,
915
+ "grad_norm": 1.0769853591918945,
916
+ "learning_rate": 0.00010641627543035995,
917
+ "loss": 0.0564,
918
+ "step": 1110
919
+ },
920
+ {
921
+ "epoch": 15.774647887323944,
922
+ "grad_norm": 0.2092517763376236,
923
+ "learning_rate": 0.00010537297861241524,
924
+ "loss": 0.037,
925
+ "step": 1120
926
+ },
927
+ {
928
+ "epoch": 15.915492957746478,
929
+ "grad_norm": 0.4689568877220154,
930
+ "learning_rate": 0.00010432968179447054,
931
+ "loss": 0.0401,
932
+ "step": 1130
933
+ },
934
+ {
935
+ "epoch": 16.056338028169016,
936
+ "grad_norm": 0.2354661524295807,
937
+ "learning_rate": 0.00010328638497652582,
938
+ "loss": 0.0459,
939
+ "step": 1140
940
+ },
941
+ {
942
+ "epoch": 16.197183098591548,
943
+ "grad_norm": 0.11875198781490326,
944
+ "learning_rate": 0.00010224308815858112,
945
+ "loss": 0.0369,
946
+ "step": 1150
947
+ },
948
+ {
949
+ "epoch": 16.338028169014084,
950
+ "grad_norm": 0.5551555156707764,
951
+ "learning_rate": 0.00010119979134063642,
952
+ "loss": 0.034,
953
+ "step": 1160
954
+ },
955
+ {
956
+ "epoch": 16.47887323943662,
957
+ "grad_norm": 0.35518983006477356,
958
+ "learning_rate": 0.00010015649452269171,
959
+ "loss": 0.0523,
960
+ "step": 1170
961
+ },
962
+ {
963
+ "epoch": 16.619718309859156,
964
+ "grad_norm": 0.6546155214309692,
965
+ "learning_rate": 9.911319770474701e-05,
966
+ "loss": 0.0472,
967
+ "step": 1180
968
+ },
969
+ {
970
+ "epoch": 16.760563380281692,
971
+ "grad_norm": 1.1401140689849854,
972
+ "learning_rate": 9.80699008868023e-05,
973
+ "loss": 0.0565,
974
+ "step": 1190
975
+ },
976
+ {
977
+ "epoch": 16.901408450704224,
978
+ "grad_norm": 0.5762555599212646,
979
+ "learning_rate": 9.70266040688576e-05,
980
+ "loss": 0.0559,
981
+ "step": 1200
982
+ },
983
+ {
984
+ "epoch": 16.901408450704224,
985
+ "eval_accuracy": 0.9313380281690141,
986
+ "eval_f1": 0.8232094288304623,
987
+ "eval_loss": 0.20966318249702454,
988
+ "eval_precision": 0.8470149253731343,
989
+ "eval_recall": 0.800705467372134,
990
+ "eval_runtime": 3.9873,
991
+ "eval_samples_per_second": 71.227,
992
+ "eval_steps_per_second": 9.029,
993
+ "step": 1200
994
+ },
995
+ {
996
+ "epoch": 17.04225352112676,
997
+ "grad_norm": 0.5705350637435913,
998
+ "learning_rate": 9.59833072509129e-05,
999
+ "loss": 0.0455,
1000
+ "step": 1210
1001
+ },
1002
+ {
1003
+ "epoch": 17.183098591549296,
1004
+ "grad_norm": 0.08219115436077118,
1005
+ "learning_rate": 9.494001043296818e-05,
1006
+ "loss": 0.029,
1007
+ "step": 1220
1008
+ },
1009
+ {
1010
+ "epoch": 17.323943661971832,
1011
+ "grad_norm": 0.41505807638168335,
1012
+ "learning_rate": 9.389671361502347e-05,
1013
+ "loss": 0.05,
1014
+ "step": 1230
1015
+ },
1016
+ {
1017
+ "epoch": 17.464788732394368,
1018
+ "grad_norm": 0.5585283637046814,
1019
+ "learning_rate": 9.285341679707877e-05,
1020
+ "loss": 0.0424,
1021
+ "step": 1240
1022
+ },
1023
+ {
1024
+ "epoch": 17.6056338028169,
1025
+ "grad_norm": 0.4773949980735779,
1026
+ "learning_rate": 9.181011997913407e-05,
1027
+ "loss": 0.043,
1028
+ "step": 1250
1029
+ },
1030
+ {
1031
+ "epoch": 17.746478873239436,
1032
+ "grad_norm": 0.5886436700820923,
1033
+ "learning_rate": 9.076682316118936e-05,
1034
+ "loss": 0.043,
1035
+ "step": 1260
1036
+ },
1037
+ {
1038
+ "epoch": 17.887323943661972,
1039
+ "grad_norm": 0.45917659997940063,
1040
+ "learning_rate": 8.972352634324466e-05,
1041
+ "loss": 0.0444,
1042
+ "step": 1270
1043
+ },
1044
+ {
1045
+ "epoch": 18.028169014084508,
1046
+ "grad_norm": 0.5295466780662537,
1047
+ "learning_rate": 8.868022952529996e-05,
1048
+ "loss": 0.0385,
1049
+ "step": 1280
1050
+ },
1051
+ {
1052
+ "epoch": 18.169014084507044,
1053
+ "grad_norm": 0.24455229938030243,
1054
+ "learning_rate": 8.763693270735525e-05,
1055
+ "loss": 0.0263,
1056
+ "step": 1290
1057
+ },
1058
+ {
1059
+ "epoch": 18.309859154929576,
1060
+ "grad_norm": 0.34424206614494324,
1061
+ "learning_rate": 8.659363588941053e-05,
1062
+ "loss": 0.0382,
1063
+ "step": 1300
1064
+ },
1065
+ {
1066
+ "epoch": 18.309859154929576,
1067
+ "eval_accuracy": 0.9352112676056338,
1068
+ "eval_f1": 0.8371681415929203,
1069
+ "eval_loss": 0.2055191993713379,
1070
+ "eval_precision": 0.8401420959147424,
1071
+ "eval_recall": 0.8342151675485009,
1072
+ "eval_runtime": 4.0365,
1073
+ "eval_samples_per_second": 70.358,
1074
+ "eval_steps_per_second": 8.919,
1075
+ "step": 1300
1076
+ },
1077
+ {
1078
+ "epoch": 18.450704225352112,
1079
+ "grad_norm": 0.10903073102235794,
1080
+ "learning_rate": 8.555033907146583e-05,
1081
+ "loss": 0.0276,
1082
+ "step": 1310
1083
+ },
1084
+ {
1085
+ "epoch": 18.591549295774648,
1086
+ "grad_norm": 0.4288780987262726,
1087
+ "learning_rate": 8.450704225352113e-05,
1088
+ "loss": 0.0457,
1089
+ "step": 1320
1090
+ },
1091
+ {
1092
+ "epoch": 18.732394366197184,
1093
+ "grad_norm": 0.9041640758514404,
1094
+ "learning_rate": 8.346374543557642e-05,
1095
+ "loss": 0.0404,
1096
+ "step": 1330
1097
+ },
1098
+ {
1099
+ "epoch": 18.87323943661972,
1100
+ "grad_norm": 0.3118440806865692,
1101
+ "learning_rate": 8.242044861763172e-05,
1102
+ "loss": 0.0404,
1103
+ "step": 1340
1104
+ },
1105
+ {
1106
+ "epoch": 19.014084507042252,
1107
+ "grad_norm": 0.39209580421447754,
1108
+ "learning_rate": 8.137715179968701e-05,
1109
+ "loss": 0.029,
1110
+ "step": 1350
1111
+ },
1112
+ {
1113
+ "epoch": 19.154929577464788,
1114
+ "grad_norm": 0.6325451731681824,
1115
+ "learning_rate": 8.033385498174231e-05,
1116
+ "loss": 0.0299,
1117
+ "step": 1360
1118
+ },
1119
+ {
1120
+ "epoch": 19.295774647887324,
1121
+ "grad_norm": 0.33151936531066895,
1122
+ "learning_rate": 7.92905581637976e-05,
1123
+ "loss": 0.0277,
1124
+ "step": 1370
1125
+ },
1126
+ {
1127
+ "epoch": 19.43661971830986,
1128
+ "grad_norm": 0.40647652745246887,
1129
+ "learning_rate": 7.82472613458529e-05,
1130
+ "loss": 0.0285,
1131
+ "step": 1380
1132
+ },
1133
+ {
1134
+ "epoch": 19.577464788732396,
1135
+ "grad_norm": 0.3072868585586548,
1136
+ "learning_rate": 7.720396452790818e-05,
1137
+ "loss": 0.0314,
1138
+ "step": 1390
1139
+ },
1140
+ {
1141
+ "epoch": 19.718309859154928,
1142
+ "grad_norm": 0.6823632121086121,
1143
+ "learning_rate": 7.616066770996348e-05,
1144
+ "loss": 0.0361,
1145
+ "step": 1400
1146
+ },
1147
+ {
1148
+ "epoch": 19.718309859154928,
1149
+ "eval_accuracy": 0.9334507042253521,
1150
+ "eval_f1": 0.8304932735426009,
1151
+ "eval_loss": 0.20700237154960632,
1152
+ "eval_precision": 0.8448905109489051,
1153
+ "eval_recall": 0.8165784832451499,
1154
+ "eval_runtime": 4.3433,
1155
+ "eval_samples_per_second": 65.388,
1156
+ "eval_steps_per_second": 8.289,
1157
+ "step": 1400
1158
+ },
1159
+ {
1160
+ "epoch": 19.859154929577464,
1161
+ "grad_norm": 0.4812707006931305,
1162
+ "learning_rate": 7.511737089201878e-05,
1163
+ "loss": 0.0515,
1164
+ "step": 1410
1165
+ },
1166
+ {
1167
+ "epoch": 20.0,
1168
+ "grad_norm": 0.9425677061080933,
1169
+ "learning_rate": 7.407407407407407e-05,
1170
+ "loss": 0.0408,
1171
+ "step": 1420
1172
+ },
1173
+ {
1174
+ "epoch": 20.140845070422536,
1175
+ "grad_norm": 0.23266366124153137,
1176
+ "learning_rate": 7.303077725612937e-05,
1177
+ "loss": 0.033,
1178
+ "step": 1430
1179
+ },
1180
+ {
1181
+ "epoch": 20.281690140845072,
1182
+ "grad_norm": 0.6943417191505432,
1183
+ "learning_rate": 7.198748043818467e-05,
1184
+ "loss": 0.0371,
1185
+ "step": 1440
1186
+ },
1187
+ {
1188
+ "epoch": 20.422535211267604,
1189
+ "grad_norm": 0.5533664226531982,
1190
+ "learning_rate": 7.094418362023996e-05,
1191
+ "loss": 0.0333,
1192
+ "step": 1450
1193
+ },
1194
+ {
1195
+ "epoch": 20.56338028169014,
1196
+ "grad_norm": 0.3414551913738251,
1197
+ "learning_rate": 6.990088680229526e-05,
1198
+ "loss": 0.0344,
1199
+ "step": 1460
1200
+ },
1201
+ {
1202
+ "epoch": 20.704225352112676,
1203
+ "grad_norm": 0.21747425198554993,
1204
+ "learning_rate": 6.885758998435054e-05,
1205
+ "loss": 0.0507,
1206
+ "step": 1470
1207
+ },
1208
+ {
1209
+ "epoch": 20.845070422535212,
1210
+ "grad_norm": 0.10106892138719559,
1211
+ "learning_rate": 6.781429316640585e-05,
1212
+ "loss": 0.0336,
1213
+ "step": 1480
1214
+ },
1215
+ {
1216
+ "epoch": 20.985915492957748,
1217
+ "grad_norm": 0.11431819200515747,
1218
+ "learning_rate": 6.677099634846115e-05,
1219
+ "loss": 0.0349,
1220
+ "step": 1490
1221
+ },
1222
+ {
1223
+ "epoch": 21.12676056338028,
1224
+ "grad_norm": 0.24515172839164734,
1225
+ "learning_rate": 6.572769953051644e-05,
1226
+ "loss": 0.0358,
1227
+ "step": 1500
1228
+ },
1229
+ {
1230
+ "epoch": 21.12676056338028,
1231
+ "eval_accuracy": 0.9397887323943662,
1232
+ "eval_f1": 0.8458070333633905,
1233
+ "eval_loss": 0.1959269642829895,
1234
+ "eval_precision": 0.8653136531365314,
1235
+ "eval_recall": 0.8271604938271605,
1236
+ "eval_runtime": 4.6741,
1237
+ "eval_samples_per_second": 60.76,
1238
+ "eval_steps_per_second": 7.702,
1239
+ "step": 1500
1240
+ },
1241
+ {
1242
+ "epoch": 21.267605633802816,
1243
+ "grad_norm": 0.1288166046142578,
1244
+ "learning_rate": 6.468440271257174e-05,
1245
+ "loss": 0.0319,
1246
+ "step": 1510
1247
+ },
1248
+ {
1249
+ "epoch": 21.408450704225352,
1250
+ "grad_norm": 0.3846772611141205,
1251
+ "learning_rate": 6.364110589462703e-05,
1252
+ "loss": 0.0292,
1253
+ "step": 1520
1254
+ },
1255
+ {
1256
+ "epoch": 21.549295774647888,
1257
+ "grad_norm": 0.8348280191421509,
1258
+ "learning_rate": 6.259780907668233e-05,
1259
+ "loss": 0.0408,
1260
+ "step": 1530
1261
+ },
1262
+ {
1263
+ "epoch": 21.690140845070424,
1264
+ "grad_norm": 0.24261720478534698,
1265
+ "learning_rate": 6.155451225873761e-05,
1266
+ "loss": 0.0279,
1267
+ "step": 1540
1268
+ },
1269
+ {
1270
+ "epoch": 21.830985915492956,
1271
+ "grad_norm": 0.5973533391952515,
1272
+ "learning_rate": 6.0511215440792915e-05,
1273
+ "loss": 0.0363,
1274
+ "step": 1550
1275
+ },
1276
+ {
1277
+ "epoch": 21.971830985915492,
1278
+ "grad_norm": 0.3994261622428894,
1279
+ "learning_rate": 5.9467918622848205e-05,
1280
+ "loss": 0.0294,
1281
+ "step": 1560
1282
+ },
1283
+ {
1284
+ "epoch": 22.112676056338028,
1285
+ "grad_norm": 0.2752346992492676,
1286
+ "learning_rate": 5.84246218049035e-05,
1287
+ "loss": 0.0355,
1288
+ "step": 1570
1289
+ },
1290
+ {
1291
+ "epoch": 22.253521126760564,
1292
+ "grad_norm": 0.0635056421160698,
1293
+ "learning_rate": 5.73813249869588e-05,
1294
+ "loss": 0.0192,
1295
+ "step": 1580
1296
+ },
1297
+ {
1298
+ "epoch": 22.3943661971831,
1299
+ "grad_norm": 0.4133341908454895,
1300
+ "learning_rate": 5.633802816901409e-05,
1301
+ "loss": 0.0412,
1302
+ "step": 1590
1303
+ },
1304
+ {
1305
+ "epoch": 22.535211267605632,
1306
+ "grad_norm": 0.30646753311157227,
1307
+ "learning_rate": 5.529473135106938e-05,
1308
+ "loss": 0.0382,
1309
+ "step": 1600
1310
+ },
1311
+ {
1312
+ "epoch": 22.535211267605632,
1313
+ "eval_accuracy": 0.9320422535211268,
1314
+ "eval_f1": 0.8269058295964126,
1315
+ "eval_loss": 0.20972707867622375,
1316
+ "eval_precision": 0.8412408759124088,
1317
+ "eval_recall": 0.8130511463844797,
1318
+ "eval_runtime": 4.7317,
1319
+ "eval_samples_per_second": 60.021,
1320
+ "eval_steps_per_second": 7.608,
1321
+ "step": 1600
1322
+ },
1323
+ {
1324
+ "epoch": 22.676056338028168,
1325
+ "grad_norm": 1.3331828117370605,
1326
+ "learning_rate": 5.425143453312468e-05,
1327
+ "loss": 0.0331,
1328
+ "step": 1610
1329
+ },
1330
+ {
1331
+ "epoch": 22.816901408450704,
1332
+ "grad_norm": 0.5091800093650818,
1333
+ "learning_rate": 5.3208137715179974e-05,
1334
+ "loss": 0.0187,
1335
+ "step": 1620
1336
+ },
1337
+ {
1338
+ "epoch": 22.95774647887324,
1339
+ "grad_norm": 0.1486244797706604,
1340
+ "learning_rate": 5.216484089723527e-05,
1341
+ "loss": 0.02,
1342
+ "step": 1630
1343
+ },
1344
+ {
1345
+ "epoch": 23.098591549295776,
1346
+ "grad_norm": 0.07892175763845444,
1347
+ "learning_rate": 5.112154407929056e-05,
1348
+ "loss": 0.0254,
1349
+ "step": 1640
1350
+ },
1351
+ {
1352
+ "epoch": 23.239436619718308,
1353
+ "grad_norm": 0.7007431387901306,
1354
+ "learning_rate": 5.0078247261345856e-05,
1355
+ "loss": 0.0179,
1356
+ "step": 1650
1357
+ },
1358
+ {
1359
+ "epoch": 23.380281690140844,
1360
+ "grad_norm": 0.4025111496448517,
1361
+ "learning_rate": 4.903495044340115e-05,
1362
+ "loss": 0.0256,
1363
+ "step": 1660
1364
+ },
1365
+ {
1366
+ "epoch": 23.52112676056338,
1367
+ "grad_norm": 0.30540022253990173,
1368
+ "learning_rate": 4.799165362545645e-05,
1369
+ "loss": 0.0254,
1370
+ "step": 1670
1371
+ },
1372
+ {
1373
+ "epoch": 23.661971830985916,
1374
+ "grad_norm": 0.45088985562324524,
1375
+ "learning_rate": 4.694835680751174e-05,
1376
+ "loss": 0.0326,
1377
+ "step": 1680
1378
+ },
1379
+ {
1380
+ "epoch": 23.802816901408452,
1381
+ "grad_norm": 0.15087296068668365,
1382
+ "learning_rate": 4.5905059989567033e-05,
1383
+ "loss": 0.0229,
1384
+ "step": 1690
1385
+ },
1386
+ {
1387
+ "epoch": 23.943661971830984,
1388
+ "grad_norm": 0.419587641954422,
1389
+ "learning_rate": 4.486176317162233e-05,
1390
+ "loss": 0.0285,
1391
+ "step": 1700
1392
+ },
1393
+ {
1394
+ "epoch": 23.943661971830984,
1395
+ "eval_accuracy": 0.9415492957746479,
1396
+ "eval_f1": 0.851520572450805,
1397
+ "eval_loss": 0.20160087943077087,
1398
+ "eval_precision": 0.8638838475499092,
1399
+ "eval_recall": 0.8395061728395061,
1400
+ "eval_runtime": 4.681,
1401
+ "eval_samples_per_second": 60.671,
1402
+ "eval_steps_per_second": 7.691,
1403
+ "step": 1700
1404
+ },
1405
+ {
1406
+ "epoch": 24.08450704225352,
1407
+ "grad_norm": 0.2833576202392578,
1408
+ "learning_rate": 4.3818466353677626e-05,
1409
+ "loss": 0.0136,
1410
+ "step": 1710
1411
+ },
1412
+ {
1413
+ "epoch": 24.225352112676056,
1414
+ "grad_norm": 0.13959386944770813,
1415
+ "learning_rate": 4.2775169535732915e-05,
1416
+ "loss": 0.0195,
1417
+ "step": 1720
1418
+ },
1419
+ {
1420
+ "epoch": 24.366197183098592,
1421
+ "grad_norm": 0.27296531200408936,
1422
+ "learning_rate": 4.173187271778821e-05,
1423
+ "loss": 0.0283,
1424
+ "step": 1730
1425
+ },
1426
+ {
1427
+ "epoch": 24.507042253521128,
1428
+ "grad_norm": 0.21670201420783997,
1429
+ "learning_rate": 4.068857589984351e-05,
1430
+ "loss": 0.0196,
1431
+ "step": 1740
1432
+ },
1433
+ {
1434
+ "epoch": 24.647887323943664,
1435
+ "grad_norm": 0.5724736452102661,
1436
+ "learning_rate": 3.96452790818988e-05,
1437
+ "loss": 0.031,
1438
+ "step": 1750
1439
+ },
1440
+ {
1441
+ "epoch": 24.788732394366196,
1442
+ "grad_norm": 0.5493142604827881,
1443
+ "learning_rate": 3.860198226395409e-05,
1444
+ "loss": 0.0265,
1445
+ "step": 1760
1446
+ },
1447
+ {
1448
+ "epoch": 24.929577464788732,
1449
+ "grad_norm": 0.21168898046016693,
1450
+ "learning_rate": 3.755868544600939e-05,
1451
+ "loss": 0.0224,
1452
+ "step": 1770
1453
+ },
1454
+ {
1455
+ "epoch": 25.070422535211268,
1456
+ "grad_norm": 0.5312497019767761,
1457
+ "learning_rate": 3.6515388628064685e-05,
1458
+ "loss": 0.024,
1459
+ "step": 1780
1460
+ },
1461
+ {
1462
+ "epoch": 25.211267605633804,
1463
+ "grad_norm": 0.5924656391143799,
1464
+ "learning_rate": 3.547209181011998e-05,
1465
+ "loss": 0.0151,
1466
+ "step": 1790
1467
+ },
1468
+ {
1469
+ "epoch": 25.352112676056336,
1470
+ "grad_norm": 0.05471556633710861,
1471
+ "learning_rate": 3.442879499217527e-05,
1472
+ "loss": 0.0141,
1473
+ "step": 1800
1474
+ },
1475
+ {
1476
+ "epoch": 25.352112676056336,
1477
+ "eval_accuracy": 0.9366197183098591,
1478
+ "eval_f1": 0.8384201077199281,
1479
+ "eval_loss": 0.2161274552345276,
1480
+ "eval_precision": 0.8537477148080439,
1481
+ "eval_recall": 0.8236331569664903,
1482
+ "eval_runtime": 4.5245,
1483
+ "eval_samples_per_second": 62.77,
1484
+ "eval_steps_per_second": 7.957,
1485
+ "step": 1800
1486
+ },
1487
+ {
1488
+ "epoch": 25.492957746478872,
1489
+ "grad_norm": 0.2500435709953308,
1490
+ "learning_rate": 3.338549817423057e-05,
1491
+ "loss": 0.0177,
1492
+ "step": 1810
1493
+ },
1494
+ {
1495
+ "epoch": 25.633802816901408,
1496
+ "grad_norm": 0.2633652985095978,
1497
+ "learning_rate": 3.234220135628587e-05,
1498
+ "loss": 0.0372,
1499
+ "step": 1820
1500
+ },
1501
+ {
1502
+ "epoch": 25.774647887323944,
1503
+ "grad_norm": 0.2851497232913971,
1504
+ "learning_rate": 3.1298904538341165e-05,
1505
+ "loss": 0.0253,
1506
+ "step": 1830
1507
+ },
1508
+ {
1509
+ "epoch": 25.91549295774648,
1510
+ "grad_norm": 0.13328520953655243,
1511
+ "learning_rate": 3.0255607720396458e-05,
1512
+ "loss": 0.0311,
1513
+ "step": 1840
1514
+ },
1515
+ {
1516
+ "epoch": 26.056338028169016,
1517
+ "grad_norm": 0.5192949771881104,
1518
+ "learning_rate": 2.921231090245175e-05,
1519
+ "loss": 0.0155,
1520
+ "step": 1850
1521
+ },
1522
+ {
1523
+ "epoch": 26.197183098591548,
1524
+ "grad_norm": 0.16768410801887512,
1525
+ "learning_rate": 2.8169014084507046e-05,
1526
+ "loss": 0.0282,
1527
+ "step": 1860
1528
+ },
1529
+ {
1530
+ "epoch": 26.338028169014084,
1531
+ "grad_norm": 0.045551054179668427,
1532
+ "learning_rate": 2.712571726656234e-05,
1533
+ "loss": 0.0194,
1534
+ "step": 1870
1535
+ },
1536
+ {
1537
+ "epoch": 26.47887323943662,
1538
+ "grad_norm": 0.33393070101737976,
1539
+ "learning_rate": 2.6082420448617635e-05,
1540
+ "loss": 0.0139,
1541
+ "step": 1880
1542
+ },
1543
+ {
1544
+ "epoch": 26.619718309859156,
1545
+ "grad_norm": 0.2400021106004715,
1546
+ "learning_rate": 2.5039123630672928e-05,
1547
+ "loss": 0.0111,
1548
+ "step": 1890
1549
+ },
1550
+ {
1551
+ "epoch": 26.760563380281692,
1552
+ "grad_norm": 0.5022717118263245,
1553
+ "learning_rate": 2.3995826812728224e-05,
1554
+ "loss": 0.0179,
1555
+ "step": 1900
1556
+ },
1557
+ {
1558
+ "epoch": 26.760563380281692,
1559
+ "eval_accuracy": 0.9376760563380282,
1560
+ "eval_f1": 0.8426666666666667,
1561
+ "eval_loss": 0.2072737216949463,
1562
+ "eval_precision": 0.8494623655913979,
1563
+ "eval_recall": 0.8359788359788359,
1564
+ "eval_runtime": 4.0574,
1565
+ "eval_samples_per_second": 69.995,
1566
+ "eval_steps_per_second": 8.873,
1567
+ "step": 1900
1568
+ },
1569
+ {
1570
+ "epoch": 26.901408450704224,
1571
+ "grad_norm": 0.41647252440452576,
1572
+ "learning_rate": 2.2952529994783517e-05,
1573
+ "loss": 0.0285,
1574
+ "step": 1910
1575
+ },
1576
+ {
1577
+ "epoch": 27.04225352112676,
1578
+ "grad_norm": 0.3066710829734802,
1579
+ "learning_rate": 2.1909233176838813e-05,
1580
+ "loss": 0.0225,
1581
+ "step": 1920
1582
+ },
1583
+ {
1584
+ "epoch": 27.183098591549296,
1585
+ "grad_norm": 0.09687885642051697,
1586
+ "learning_rate": 2.0865936358894105e-05,
1587
+ "loss": 0.0109,
1588
+ "step": 1930
1589
+ },
1590
+ {
1591
+ "epoch": 27.323943661971832,
1592
+ "grad_norm": 0.25405797362327576,
1593
+ "learning_rate": 1.98226395409494e-05,
1594
+ "loss": 0.0204,
1595
+ "step": 1940
1596
+ },
1597
+ {
1598
+ "epoch": 27.464788732394368,
1599
+ "grad_norm": 0.06933202594518661,
1600
+ "learning_rate": 1.8779342723004694e-05,
1601
+ "loss": 0.0274,
1602
+ "step": 1950
1603
+ },
1604
+ {
1605
+ "epoch": 27.6056338028169,
1606
+ "grad_norm": 0.13852407038211823,
1607
+ "learning_rate": 1.773604590505999e-05,
1608
+ "loss": 0.0194,
1609
+ "step": 1960
1610
+ },
1611
+ {
1612
+ "epoch": 27.746478873239436,
1613
+ "grad_norm": 0.1922004669904709,
1614
+ "learning_rate": 1.6692749087115286e-05,
1615
+ "loss": 0.0228,
1616
+ "step": 1970
1617
+ },
1618
+ {
1619
+ "epoch": 27.887323943661972,
1620
+ "grad_norm": 0.2851165235042572,
1621
+ "learning_rate": 1.5649452269170582e-05,
1622
+ "loss": 0.022,
1623
+ "step": 1980
1624
+ },
1625
+ {
1626
+ "epoch": 28.028169014084508,
1627
+ "grad_norm": 0.30591297149658203,
1628
+ "learning_rate": 1.4606155451225875e-05,
1629
+ "loss": 0.0179,
1630
+ "step": 1990
1631
+ },
1632
+ {
1633
+ "epoch": 28.169014084507044,
1634
+ "grad_norm": 0.9221152067184448,
1635
+ "learning_rate": 1.356285863328117e-05,
1636
+ "loss": 0.0263,
1637
+ "step": 2000
1638
+ },
1639
+ {
1640
+ "epoch": 28.169014084507044,
1641
+ "eval_accuracy": 0.9390845070422535,
1642
+ "eval_f1": 0.8456735057983943,
1643
+ "eval_loss": 0.20973354578018188,
1644
+ "eval_precision": 0.855595667870036,
1645
+ "eval_recall": 0.8359788359788359,
1646
+ "eval_runtime": 3.8982,
1647
+ "eval_samples_per_second": 72.853,
1648
+ "eval_steps_per_second": 9.235,
1649
+ "step": 2000
1650
+ },
1651
+ {
1652
+ "epoch": 28.309859154929576,
1653
+ "grad_norm": 0.2080974280834198,
1654
+ "learning_rate": 1.2519561815336464e-05,
1655
+ "loss": 0.0223,
1656
+ "step": 2010
1657
+ },
1658
+ {
1659
+ "epoch": 28.450704225352112,
1660
+ "grad_norm": 0.265176385641098,
1661
+ "learning_rate": 1.1476264997391758e-05,
1662
+ "loss": 0.0156,
1663
+ "step": 2020
1664
+ },
1665
+ {
1666
+ "epoch": 28.591549295774648,
1667
+ "grad_norm": 0.6490623950958252,
1668
+ "learning_rate": 1.0432968179447053e-05,
1669
+ "loss": 0.0143,
1670
+ "step": 2030
1671
+ },
1672
+ {
1673
+ "epoch": 28.732394366197184,
1674
+ "grad_norm": 0.08013524860143661,
1675
+ "learning_rate": 9.389671361502347e-06,
1676
+ "loss": 0.0175,
1677
+ "step": 2040
1678
+ },
1679
+ {
1680
+ "epoch": 28.87323943661972,
1681
+ "grad_norm": 0.24310840666294098,
1682
+ "learning_rate": 8.346374543557643e-06,
1683
+ "loss": 0.0208,
1684
+ "step": 2050
1685
+ },
1686
+ {
1687
+ "epoch": 29.014084507042252,
1688
+ "grad_norm": 0.21826568245887756,
1689
+ "learning_rate": 7.303077725612938e-06,
1690
+ "loss": 0.0209,
1691
+ "step": 2060
1692
+ },
1693
+ {
1694
+ "epoch": 29.154929577464788,
1695
+ "grad_norm": 0.10844382643699646,
1696
+ "learning_rate": 6.259780907668232e-06,
1697
+ "loss": 0.0122,
1698
+ "step": 2070
1699
+ },
1700
+ {
1701
+ "epoch": 29.295774647887324,
1702
+ "grad_norm": 0.1837497502565384,
1703
+ "learning_rate": 5.216484089723526e-06,
1704
+ "loss": 0.0155,
1705
+ "step": 2080
1706
+ },
1707
+ {
1708
+ "epoch": 29.43661971830986,
1709
+ "grad_norm": 0.8813931941986084,
1710
+ "learning_rate": 4.173187271778822e-06,
1711
+ "loss": 0.0229,
1712
+ "step": 2090
1713
+ },
1714
+ {
1715
+ "epoch": 29.577464788732396,
1716
+ "grad_norm": 0.10835859924554825,
1717
+ "learning_rate": 3.129890453834116e-06,
1718
+ "loss": 0.0191,
1719
+ "step": 2100
1720
+ },
1721
+ {
1722
+ "epoch": 29.577464788732396,
1723
+ "eval_accuracy": 0.9376760563380282,
1724
+ "eval_f1": 0.8415398388540735,
1725
+ "eval_loss": 0.21010643243789673,
1726
+ "eval_precision": 0.8545454545454545,
1727
+ "eval_recall": 0.8289241622574955,
1728
+ "eval_runtime": 4.0827,
1729
+ "eval_samples_per_second": 69.561,
1730
+ "eval_steps_per_second": 8.818,
1731
+ "step": 2100
1732
+ },
1733
+ {
1734
+ "epoch": 29.718309859154928,
1735
+ "grad_norm": 0.5981993079185486,
1736
+ "learning_rate": 2.086593635889411e-06,
1737
+ "loss": 0.024,
1738
+ "step": 2110
1739
+ },
1740
+ {
1741
+ "epoch": 29.859154929577464,
1742
+ "grad_norm": 0.029736319556832314,
1743
+ "learning_rate": 1.0432968179447054e-06,
1744
+ "loss": 0.0147,
1745
+ "step": 2120
1746
+ },
1747
+ {
1748
+ "epoch": 30.0,
1749
+ "grad_norm": 0.1090225800871849,
1750
+ "learning_rate": 0.0,
1751
+ "loss": 0.0163,
1752
+ "step": 2130
1753
+ },
1754
+ {
1755
+ "epoch": 30.0,
1756
+ "step": 2130,
1757
+ "total_flos": 7.774335248148726e+18,
1758
+ "train_loss": 0.08187752739294595,
1759
+ "train_runtime": 1287.2767,
1760
+ "train_samples_per_second": 26.428,
1761
+ "train_steps_per_second": 1.655
1762
+ }
1763
+ ],
1764
+ "logging_steps": 10,
1765
+ "max_steps": 2130,
1766
+ "num_input_tokens_seen": 0,
1767
+ "num_train_epochs": 30,
1768
+ "save_steps": 500,
1769
+ "stateful_callbacks": {
1770
+ "TrainerControl": {
1771
+ "args": {
1772
+ "should_epoch_stop": false,
1773
+ "should_evaluate": false,
1774
+ "should_log": false,
1775
+ "should_save": true,
1776
+ "should_training_stop": true
1777
+ },
1778
+ "attributes": {}
1779
+ }
1780
+ },
1781
+ "total_flos": 7.774335248148726e+18,
1782
+ "train_batch_size": 16,
1783
+ "trial_name": null,
1784
+ "trial_params": null
1785
+ }