lombardata commited on
Commit
b5fb034
1 Parent(s): 831ac89

🍻 cheers

Browse files
README.md CHANGED
@@ -1,6 +1,11 @@
1
  ---
2
- base_model: lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze
 
 
 
3
  tags:
 
 
4
  - generated_from_trainer
5
  metrics:
6
  - accuracy
@@ -14,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
14
 
15
  # DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze
16
 
17
- This model is a fine-tuned version of [lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze](https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze) on the None dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.1190
20
- - F1 Micro: 0.8245
21
- - F1 Macro: 0.7104
22
- - Roc Auc: 0.8839
23
- - Accuracy: 0.3216
24
  - Learning Rate: 0.0000
25
 
26
  ## Model description
 
1
  ---
2
+ language:
3
+ - eng
4
+ license: apache-2.0
5
+ base_model: facebook/dinov2-large
6
  tags:
7
+ - multilabel-image-classification
8
+ - multilabel
9
  - generated_from_trainer
10
  metrics:
11
  - accuracy
 
19
 
20
  # DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze
21
 
22
+ This model is a fine-tuned version of [facebook/dinov2-large](https://huggingface.co/facebook/dinov2-large) on the multilabel_complete_dataset dataset.
23
  It achieves the following results on the evaluation set:
24
+ - Loss: 0.1181
25
+ - F1 Micro: 0.8219
26
+ - F1 Macro: 0.7131
27
+ - Roc Auc: 0.8797
28
+ - Accuracy: 0.3214
29
  - Learning Rate: 0.0000
30
 
31
  ## Model description
all_results.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 127.0,
3
+ "eval_accuracy": 0.3214161749392572,
4
+ "eval_f1_macro": 0.7130919408806389,
5
+ "eval_f1_micro": 0.8219422171312205,
6
+ "eval_loss": 0.11805282533168793,
7
+ "eval_roc_auc": 0.8796827550466734,
8
+ "eval_runtime": 695.0536,
9
+ "eval_samples_per_second": 4.145,
10
+ "eval_steps_per_second": 0.131,
11
+ "learning_rate": 1.0000000000000002e-07,
12
+ "train_loss": 0.0713793940258306,
13
+ "train_runtime": 211000.3165,
14
+ "train_samples_per_second": 6.162,
15
+ "train_steps_per_second": 0.193
16
+ }
config.json CHANGED
@@ -1,5 +1,5 @@
1
  {
2
- "_name_or_path": "lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze",
3
  "apply_layernorm": true,
4
  "architectures": [
5
  "NewheadDinov2ForImageClassification"
 
1
  {
2
+ "_name_or_path": "facebook/dinov2-large2024_04_07",
3
  "apply_layernorm": true,
4
  "architectures": [
5
  "NewheadDinov2ForImageClassification"
eval_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 127.0,
3
+ "eval_accuracy": 0.3214161749392572,
4
+ "eval_f1_macro": 0.7130919408806389,
5
+ "eval_f1_micro": 0.8219422171312205,
6
+ "eval_loss": 0.11805282533168793,
7
+ "eval_roc_auc": 0.8796827550466734,
8
+ "eval_runtime": 695.0536,
9
+ "eval_samples_per_second": 4.145,
10
+ "eval_steps_per_second": 0.131,
11
+ "learning_rate": 1.0000000000000002e-07
12
+ }
runs/Apr07_03-28-54_datavisu2/events.out.tfevents.1712665220.datavisu2.1419.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1c1a6e6501359d66fd41f5bcee69980c7be659f04660e50c5dc68d20439a004f
3
+ size 634
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 127.0,
3
+ "learning_rate": 1.0000000000000002e-07,
4
+ "train_loss": 0.0713793940258306,
5
+ "train_runtime": 211000.3165,
6
+ "train_samples_per_second": 6.162,
7
+ "train_steps_per_second": 0.193
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,2090 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.11870068311691284,
3
+ "best_model_checkpoint": "/home1/datawork/mcontini/models/multilabel/huggingface/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze/checkpoint-31707",
4
+ "epoch": 127.0,
5
+ "eval_steps": 500,
6
+ "global_step": 34417,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "eval_accuracy": 0.2206588881262869,
14
+ "eval_f1_macro": 0.49067608571654225,
15
+ "eval_f1_micro": 0.7368702869126769,
16
+ "eval_loss": 0.1679287850856781,
17
+ "eval_roc_auc": 0.8187531778602468,
18
+ "eval_runtime": 714.8817,
19
+ "eval_samples_per_second": 4.076,
20
+ "eval_steps_per_second": 0.129,
21
+ "learning_rate": 0.001,
22
+ "step": 271
23
+ },
24
+ {
25
+ "epoch": 1.85,
26
+ "learning_rate": 0.001,
27
+ "loss": 0.2713,
28
+ "step": 500
29
+ },
30
+ {
31
+ "epoch": 2.0,
32
+ "eval_accuracy": 0.2515442690459849,
33
+ "eval_f1_macro": 0.538945165530625,
34
+ "eval_f1_micro": 0.761357152262652,
35
+ "eval_loss": 0.15397264063358307,
36
+ "eval_roc_auc": 0.8356163050775759,
37
+ "eval_runtime": 720.5749,
38
+ "eval_samples_per_second": 4.044,
39
+ "eval_steps_per_second": 0.128,
40
+ "learning_rate": 0.001,
41
+ "step": 542
42
+ },
43
+ {
44
+ "epoch": 3.0,
45
+ "eval_accuracy": 0.2525737817433082,
46
+ "eval_f1_macro": 0.6053847883354035,
47
+ "eval_f1_micro": 0.7728152001031681,
48
+ "eval_loss": 0.14765480160713196,
49
+ "eval_roc_auc": 0.8471568306864959,
50
+ "eval_runtime": 730.8961,
51
+ "eval_samples_per_second": 3.987,
52
+ "eval_steps_per_second": 0.126,
53
+ "learning_rate": 0.001,
54
+ "step": 813
55
+ },
56
+ {
57
+ "epoch": 3.69,
58
+ "learning_rate": 0.001,
59
+ "loss": 0.1679,
60
+ "step": 1000
61
+ },
62
+ {
63
+ "epoch": 4.0,
64
+ "eval_accuracy": 0.2594371997254633,
65
+ "eval_f1_macro": 0.5847686766758632,
66
+ "eval_f1_micro": 0.7755244755244756,
67
+ "eval_loss": 0.1577611267566681,
68
+ "eval_roc_auc": 0.8442033715582856,
69
+ "eval_runtime": 714.9053,
70
+ "eval_samples_per_second": 4.076,
71
+ "eval_steps_per_second": 0.129,
72
+ "learning_rate": 0.001,
73
+ "step": 1084
74
+ },
75
+ {
76
+ "epoch": 5.0,
77
+ "eval_accuracy": 0.2618393960192176,
78
+ "eval_f1_macro": 0.6124919180480791,
79
+ "eval_f1_micro": 0.7818521347933113,
80
+ "eval_loss": 0.1426197737455368,
81
+ "eval_roc_auc": 0.8555100487696016,
82
+ "eval_runtime": 733.6285,
83
+ "eval_samples_per_second": 3.972,
84
+ "eval_steps_per_second": 0.125,
85
+ "learning_rate": 0.001,
86
+ "step": 1355
87
+ },
88
+ {
89
+ "epoch": 5.54,
90
+ "learning_rate": 0.001,
91
+ "loss": 0.1598,
92
+ "step": 1500
93
+ },
94
+ {
95
+ "epoch": 6.0,
96
+ "eval_accuracy": 0.25497597803706246,
97
+ "eval_f1_macro": 0.6238861296316591,
98
+ "eval_f1_micro": 0.7822487732024749,
99
+ "eval_loss": 0.1422213762998581,
100
+ "eval_roc_auc": 0.854243572334213,
101
+ "eval_runtime": 724.219,
102
+ "eval_samples_per_second": 4.024,
103
+ "eval_steps_per_second": 0.127,
104
+ "learning_rate": 0.001,
105
+ "step": 1626
106
+ },
107
+ {
108
+ "epoch": 7.0,
109
+ "eval_accuracy": 0.255662319835278,
110
+ "eval_f1_macro": 0.6319974141584176,
111
+ "eval_f1_micro": 0.7825118828416049,
112
+ "eval_loss": 0.1426122486591339,
113
+ "eval_roc_auc": 0.8534283984871422,
114
+ "eval_runtime": 746.9062,
115
+ "eval_samples_per_second": 3.901,
116
+ "eval_steps_per_second": 0.123,
117
+ "learning_rate": 0.001,
118
+ "step": 1897
119
+ },
120
+ {
121
+ "epoch": 7.38,
122
+ "learning_rate": 0.001,
123
+ "loss": 0.1571,
124
+ "step": 2000
125
+ },
126
+ {
127
+ "epoch": 8.0,
128
+ "eval_accuracy": 0.26286890871654084,
129
+ "eval_f1_macro": 0.622325808061766,
130
+ "eval_f1_micro": 0.7755977927651747,
131
+ "eval_loss": 0.1528300642967224,
132
+ "eval_roc_auc": 0.8437122124534273,
133
+ "eval_runtime": 750.7491,
134
+ "eval_samples_per_second": 3.881,
135
+ "eval_steps_per_second": 0.123,
136
+ "learning_rate": 0.001,
137
+ "step": 2168
138
+ },
139
+ {
140
+ "epoch": 9.0,
141
+ "eval_accuracy": 0.24811256005490734,
142
+ "eval_f1_macro": 0.6413065292545327,
143
+ "eval_f1_micro": 0.7796035913942063,
144
+ "eval_loss": 0.1438213288784027,
145
+ "eval_roc_auc": 0.8548646774451706,
146
+ "eval_runtime": 702.3842,
147
+ "eval_samples_per_second": 4.149,
148
+ "eval_steps_per_second": 0.131,
149
+ "learning_rate": 0.001,
150
+ "step": 2439
151
+ },
152
+ {
153
+ "epoch": 9.23,
154
+ "learning_rate": 0.001,
155
+ "loss": 0.1554,
156
+ "step": 2500
157
+ },
158
+ {
159
+ "epoch": 10.0,
160
+ "eval_accuracy": 0.26973232669869596,
161
+ "eval_f1_macro": 0.6289481397755302,
162
+ "eval_f1_micro": 0.7888898226741743,
163
+ "eval_loss": 0.14050152897834778,
164
+ "eval_roc_auc": 0.8621391370400399,
165
+ "eval_runtime": 707.8611,
166
+ "eval_samples_per_second": 4.117,
167
+ "eval_steps_per_second": 0.13,
168
+ "learning_rate": 0.001,
169
+ "step": 2710
170
+ },
171
+ {
172
+ "epoch": 11.0,
173
+ "eval_accuracy": 0.26835964310226496,
174
+ "eval_f1_macro": 0.6222474437475864,
175
+ "eval_f1_micro": 0.789769130122821,
176
+ "eval_loss": 0.14092855155467987,
177
+ "eval_roc_auc": 0.8613792306841266,
178
+ "eval_runtime": 711.2194,
179
+ "eval_samples_per_second": 4.097,
180
+ "eval_steps_per_second": 0.129,
181
+ "learning_rate": 0.001,
182
+ "step": 2981
183
+ },
184
+ {
185
+ "epoch": 11.07,
186
+ "learning_rate": 0.001,
187
+ "loss": 0.1536,
188
+ "step": 3000
189
+ },
190
+ {
191
+ "epoch": 12.0,
192
+ "eval_accuracy": 0.272477693891558,
193
+ "eval_f1_macro": 0.6165764997376745,
194
+ "eval_f1_micro": 0.7862856154611094,
195
+ "eval_loss": 0.1391688883304596,
196
+ "eval_roc_auc": 0.8528316876814077,
197
+ "eval_runtime": 716.1193,
198
+ "eval_samples_per_second": 4.069,
199
+ "eval_steps_per_second": 0.128,
200
+ "learning_rate": 0.001,
201
+ "step": 3252
202
+ },
203
+ {
204
+ "epoch": 12.92,
205
+ "learning_rate": 0.001,
206
+ "loss": 0.1526,
207
+ "step": 3500
208
+ },
209
+ {
210
+ "epoch": 13.0,
211
+ "eval_accuracy": 0.2625257378174331,
212
+ "eval_f1_macro": 0.6418972133310931,
213
+ "eval_f1_micro": 0.7877202761222827,
214
+ "eval_loss": 0.13992685079574585,
215
+ "eval_roc_auc": 0.8558891516745222,
216
+ "eval_runtime": 710.6314,
217
+ "eval_samples_per_second": 4.101,
218
+ "eval_steps_per_second": 0.129,
219
+ "learning_rate": 0.001,
220
+ "step": 3523
221
+ },
222
+ {
223
+ "epoch": 14.0,
224
+ "eval_accuracy": 0.26492793411118737,
225
+ "eval_f1_macro": 0.632569342193695,
226
+ "eval_f1_micro": 0.7859690345319514,
227
+ "eval_loss": 0.1437946856021881,
228
+ "eval_roc_auc": 0.8609422021280538,
229
+ "eval_runtime": 694.5265,
230
+ "eval_samples_per_second": 4.196,
231
+ "eval_steps_per_second": 0.132,
232
+ "learning_rate": 0.001,
233
+ "step": 3794
234
+ },
235
+ {
236
+ "epoch": 14.76,
237
+ "learning_rate": 0.001,
238
+ "loss": 0.1535,
239
+ "step": 4000
240
+ },
241
+ {
242
+ "epoch": 15.0,
243
+ "eval_accuracy": 0.2735072065888813,
244
+ "eval_f1_macro": 0.6499317144862637,
245
+ "eval_f1_micro": 0.7930450968779085,
246
+ "eval_loss": 0.13769365847110748,
247
+ "eval_roc_auc": 0.8625195818341326,
248
+ "eval_runtime": 690.6675,
249
+ "eval_samples_per_second": 4.219,
250
+ "eval_steps_per_second": 0.133,
251
+ "learning_rate": 0.001,
252
+ "step": 4065
253
+ },
254
+ {
255
+ "epoch": 16.0,
256
+ "eval_accuracy": 0.2676733013040494,
257
+ "eval_f1_macro": 0.6435362275021823,
258
+ "eval_f1_micro": 0.7868312757201646,
259
+ "eval_loss": 0.13966824114322662,
260
+ "eval_roc_auc": 0.8526053782518377,
261
+ "eval_runtime": 692.8928,
262
+ "eval_samples_per_second": 4.206,
263
+ "eval_steps_per_second": 0.133,
264
+ "learning_rate": 0.001,
265
+ "step": 4336
266
+ },
267
+ {
268
+ "epoch": 16.61,
269
+ "learning_rate": 0.001,
270
+ "loss": 0.1517,
271
+ "step": 4500
272
+ },
273
+ {
274
+ "epoch": 17.0,
275
+ "eval_accuracy": 0.26458476321207963,
276
+ "eval_f1_macro": 0.6400976985447264,
277
+ "eval_f1_micro": 0.7928011464216472,
278
+ "eval_loss": 0.1382310837507248,
279
+ "eval_roc_auc": 0.8634350422987906,
280
+ "eval_runtime": 684.8353,
281
+ "eval_samples_per_second": 4.255,
282
+ "eval_steps_per_second": 0.134,
283
+ "learning_rate": 0.001,
284
+ "step": 4607
285
+ },
286
+ {
287
+ "epoch": 18.0,
288
+ "eval_accuracy": 0.26835964310226496,
289
+ "eval_f1_macro": 0.6285887470600311,
290
+ "eval_f1_micro": 0.7912449392712552,
291
+ "eval_loss": 0.1392030268907547,
292
+ "eval_roc_auc": 0.8624142094617062,
293
+ "eval_runtime": 684.0603,
294
+ "eval_samples_per_second": 4.26,
295
+ "eval_steps_per_second": 0.134,
296
+ "learning_rate": 0.001,
297
+ "step": 4878
298
+ },
299
+ {
300
+ "epoch": 18.45,
301
+ "learning_rate": 0.001,
302
+ "loss": 0.1524,
303
+ "step": 5000
304
+ },
305
+ {
306
+ "epoch": 19.0,
307
+ "eval_accuracy": 0.26355525051475637,
308
+ "eval_f1_macro": 0.6182634857374021,
309
+ "eval_f1_micro": 0.7874116344434035,
310
+ "eval_loss": 0.1391826868057251,
311
+ "eval_roc_auc": 0.8575979217492725,
312
+ "eval_runtime": 683.6418,
313
+ "eval_samples_per_second": 4.262,
314
+ "eval_steps_per_second": 0.135,
315
+ "learning_rate": 0.001,
316
+ "step": 5149
317
+ },
318
+ {
319
+ "epoch": 20.0,
320
+ "eval_accuracy": 0.25978037062457104,
321
+ "eval_f1_macro": 0.6286353323757679,
322
+ "eval_f1_micro": 0.7878349022447502,
323
+ "eval_loss": 0.13860712945461273,
324
+ "eval_roc_auc": 0.857768912450985,
325
+ "eval_runtime": 691.5194,
326
+ "eval_samples_per_second": 4.214,
327
+ "eval_steps_per_second": 0.133,
328
+ "learning_rate": 0.001,
329
+ "step": 5420
330
+ },
331
+ {
332
+ "epoch": 20.3,
333
+ "learning_rate": 0.001,
334
+ "loss": 0.1527,
335
+ "step": 5500
336
+ },
337
+ {
338
+ "epoch": 21.0,
339
+ "eval_accuracy": 0.2601235415236788,
340
+ "eval_f1_macro": 0.6408380159549439,
341
+ "eval_f1_micro": 0.7879620486841541,
342
+ "eval_loss": 0.13738416135311127,
343
+ "eval_roc_auc": 0.8556654235498968,
344
+ "eval_runtime": 685.1425,
345
+ "eval_samples_per_second": 4.253,
346
+ "eval_steps_per_second": 0.134,
347
+ "learning_rate": 0.001,
348
+ "step": 5691
349
+ },
350
+ {
351
+ "epoch": 22.0,
352
+ "eval_accuracy": 0.2704186684969115,
353
+ "eval_f1_macro": 0.6476322873247978,
354
+ "eval_f1_micro": 0.7897032412554519,
355
+ "eval_loss": 0.13765838742256165,
356
+ "eval_roc_auc": 0.8577058498082936,
357
+ "eval_runtime": 697.7818,
358
+ "eval_samples_per_second": 4.176,
359
+ "eval_steps_per_second": 0.132,
360
+ "learning_rate": 0.001,
361
+ "step": 5962
362
+ },
363
+ {
364
+ "epoch": 22.14,
365
+ "learning_rate": 0.001,
366
+ "loss": 0.1513,
367
+ "step": 6000
368
+ },
369
+ {
370
+ "epoch": 23.0,
371
+ "eval_accuracy": 0.26973232669869596,
372
+ "eval_f1_macro": 0.6442899469566483,
373
+ "eval_f1_micro": 0.7955080753701211,
374
+ "eval_loss": 0.13733763992786407,
375
+ "eval_roc_auc": 0.865520436446917,
376
+ "eval_runtime": 694.652,
377
+ "eval_samples_per_second": 4.195,
378
+ "eval_steps_per_second": 0.132,
379
+ "learning_rate": 0.001,
380
+ "step": 6233
381
+ },
382
+ {
383
+ "epoch": 23.99,
384
+ "learning_rate": 0.001,
385
+ "loss": 0.1514,
386
+ "step": 6500
387
+ },
388
+ {
389
+ "epoch": 24.0,
390
+ "eval_accuracy": 0.2656142759094029,
391
+ "eval_f1_macro": 0.6476856744170203,
392
+ "eval_f1_micro": 0.7877039652128988,
393
+ "eval_loss": 0.15933051705360413,
394
+ "eval_roc_auc": 0.854747407398644,
395
+ "eval_runtime": 693.7997,
396
+ "eval_samples_per_second": 4.2,
397
+ "eval_steps_per_second": 0.133,
398
+ "learning_rate": 0.001,
399
+ "step": 6504
400
+ },
401
+ {
402
+ "epoch": 25.0,
403
+ "eval_accuracy": 0.2656142759094029,
404
+ "eval_f1_macro": 0.6476685424002145,
405
+ "eval_f1_micro": 0.7909313518534156,
406
+ "eval_loss": 0.1371144950389862,
407
+ "eval_roc_auc": 0.8619333499582761,
408
+ "eval_runtime": 697.1431,
409
+ "eval_samples_per_second": 4.18,
410
+ "eval_steps_per_second": 0.132,
411
+ "learning_rate": 0.001,
412
+ "step": 6775
413
+ },
414
+ {
415
+ "epoch": 25.83,
416
+ "learning_rate": 0.001,
417
+ "loss": 0.1513,
418
+ "step": 7000
419
+ },
420
+ {
421
+ "epoch": 26.0,
422
+ "eval_accuracy": 0.26664378860672616,
423
+ "eval_f1_macro": 0.6272636530018297,
424
+ "eval_f1_micro": 0.7871246489522575,
425
+ "eval_loss": 0.13742324709892273,
426
+ "eval_roc_auc": 0.8534693919460551,
427
+ "eval_runtime": 694.4047,
428
+ "eval_samples_per_second": 4.196,
429
+ "eval_steps_per_second": 0.132,
430
+ "learning_rate": 0.001,
431
+ "step": 7046
432
+ },
433
+ {
434
+ "epoch": 27.0,
435
+ "eval_accuracy": 0.26458476321207963,
436
+ "eval_f1_macro": 0.6470102889103709,
437
+ "eval_f1_micro": 0.7933764066578238,
438
+ "eval_loss": 0.13733525574207306,
439
+ "eval_roc_auc": 0.8595147784088812,
440
+ "eval_runtime": 705.8388,
441
+ "eval_samples_per_second": 4.128,
442
+ "eval_steps_per_second": 0.13,
443
+ "learning_rate": 0.001,
444
+ "step": 7317
445
+ },
446
+ {
447
+ "epoch": 27.68,
448
+ "learning_rate": 0.001,
449
+ "loss": 0.1508,
450
+ "step": 7500
451
+ },
452
+ {
453
+ "epoch": 28.0,
454
+ "eval_accuracy": 0.2735072065888813,
455
+ "eval_f1_macro": 0.652320369204269,
456
+ "eval_f1_micro": 0.7932826525791349,
457
+ "eval_loss": 0.13527436554431915,
458
+ "eval_roc_auc": 0.8584141813020159,
459
+ "eval_runtime": 689.3753,
460
+ "eval_samples_per_second": 4.227,
461
+ "eval_steps_per_second": 0.133,
462
+ "learning_rate": 0.001,
463
+ "step": 7588
464
+ },
465
+ {
466
+ "epoch": 29.0,
467
+ "eval_accuracy": 0.27762525737817434,
468
+ "eval_f1_macro": 0.652221916857101,
469
+ "eval_f1_micro": 0.7959942533592496,
470
+ "eval_loss": 0.1361834853887558,
471
+ "eval_roc_auc": 0.864468679187198,
472
+ "eval_runtime": 711.9932,
473
+ "eval_samples_per_second": 4.093,
474
+ "eval_steps_per_second": 0.129,
475
+ "learning_rate": 0.001,
476
+ "step": 7859
477
+ },
478
+ {
479
+ "epoch": 29.52,
480
+ "learning_rate": 0.001,
481
+ "loss": 0.1506,
482
+ "step": 8000
483
+ },
484
+ {
485
+ "epoch": 30.0,
486
+ "eval_accuracy": 0.25051475634866166,
487
+ "eval_f1_macro": 0.6283472874420969,
488
+ "eval_f1_micro": 0.784903397164032,
489
+ "eval_loss": 0.13839198648929596,
490
+ "eval_roc_auc": 0.8546034947804995,
491
+ "eval_runtime": 711.3244,
492
+ "eval_samples_per_second": 4.097,
493
+ "eval_steps_per_second": 0.129,
494
+ "learning_rate": 0.001,
495
+ "step": 8130
496
+ },
497
+ {
498
+ "epoch": 31.0,
499
+ "eval_accuracy": 0.2717913520933425,
500
+ "eval_f1_macro": 0.6629936811365008,
501
+ "eval_f1_micro": 0.7963528413910094,
502
+ "eval_loss": 0.13423041999340057,
503
+ "eval_roc_auc": 0.8636329317556207,
504
+ "eval_runtime": 717.6936,
505
+ "eval_samples_per_second": 4.06,
506
+ "eval_steps_per_second": 0.128,
507
+ "learning_rate": 0.001,
508
+ "step": 8401
509
+ },
510
+ {
511
+ "epoch": 31.37,
512
+ "learning_rate": 0.001,
513
+ "loss": 0.151,
514
+ "step": 8500
515
+ },
516
+ {
517
+ "epoch": 32.0,
518
+ "eval_accuracy": 0.2717913520933425,
519
+ "eval_f1_macro": 0.655637689548565,
520
+ "eval_f1_micro": 0.7967574308875494,
521
+ "eval_loss": 0.13658006489276886,
522
+ "eval_roc_auc": 0.8695589279972756,
523
+ "eval_runtime": 709.164,
524
+ "eval_samples_per_second": 4.109,
525
+ "eval_steps_per_second": 0.13,
526
+ "learning_rate": 0.001,
527
+ "step": 8672
528
+ },
529
+ {
530
+ "epoch": 33.0,
531
+ "eval_accuracy": 0.2824296499656829,
532
+ "eval_f1_macro": 0.6635088812977026,
533
+ "eval_f1_micro": 0.7984512261126608,
534
+ "eval_loss": 0.13591675460338593,
535
+ "eval_roc_auc": 0.8700680281449957,
536
+ "eval_runtime": 711.6984,
537
+ "eval_samples_per_second": 4.094,
538
+ "eval_steps_per_second": 0.129,
539
+ "learning_rate": 0.001,
540
+ "step": 8943
541
+ },
542
+ {
543
+ "epoch": 33.21,
544
+ "learning_rate": 0.001,
545
+ "loss": 0.1507,
546
+ "step": 9000
547
+ },
548
+ {
549
+ "epoch": 34.0,
550
+ "eval_accuracy": 0.28140013726835966,
551
+ "eval_f1_macro": 0.640009703489063,
552
+ "eval_f1_micro": 0.7998982533491605,
553
+ "eval_loss": 0.13349105417728424,
554
+ "eval_roc_auc": 0.8656642978283616,
555
+ "eval_runtime": 702.7082,
556
+ "eval_samples_per_second": 4.147,
557
+ "eval_steps_per_second": 0.131,
558
+ "learning_rate": 0.001,
559
+ "step": 9214
560
+ },
561
+ {
562
+ "epoch": 35.0,
563
+ "eval_accuracy": 0.272477693891558,
564
+ "eval_f1_macro": 0.6519975762807232,
565
+ "eval_f1_micro": 0.7962736584748978,
566
+ "eval_loss": 0.13431623578071594,
567
+ "eval_roc_auc": 0.8653056916528975,
568
+ "eval_runtime": 695.9796,
569
+ "eval_samples_per_second": 4.187,
570
+ "eval_steps_per_second": 0.132,
571
+ "learning_rate": 0.001,
572
+ "step": 9485
573
+ },
574
+ {
575
+ "epoch": 35.06,
576
+ "learning_rate": 0.001,
577
+ "loss": 0.1495,
578
+ "step": 9500
579
+ },
580
+ {
581
+ "epoch": 36.0,
582
+ "eval_accuracy": 0.26355525051475637,
583
+ "eval_f1_macro": 0.6451443559854592,
584
+ "eval_f1_micro": 0.7924289154590393,
585
+ "eval_loss": 0.14291881024837494,
586
+ "eval_roc_auc": 0.8625517456725394,
587
+ "eval_runtime": 695.1357,
588
+ "eval_samples_per_second": 4.192,
589
+ "eval_steps_per_second": 0.132,
590
+ "learning_rate": 0.001,
591
+ "step": 9756
592
+ },
593
+ {
594
+ "epoch": 36.9,
595
+ "learning_rate": 0.001,
596
+ "loss": 0.1496,
597
+ "step": 10000
598
+ },
599
+ {
600
+ "epoch": 37.0,
601
+ "eval_accuracy": 0.2731640356897735,
602
+ "eval_f1_macro": 0.653137873575441,
603
+ "eval_f1_micro": 0.7981288539230278,
604
+ "eval_loss": 0.13305164873600006,
605
+ "eval_roc_auc": 0.8638193078136003,
606
+ "eval_runtime": 697.8225,
607
+ "eval_samples_per_second": 4.176,
608
+ "eval_steps_per_second": 0.132,
609
+ "learning_rate": 0.001,
610
+ "step": 10027
611
+ },
612
+ {
613
+ "epoch": 38.0,
614
+ "eval_accuracy": 0.26835964310226496,
615
+ "eval_f1_macro": 0.6306015382070221,
616
+ "eval_f1_micro": 0.7938126806051335,
617
+ "eval_loss": 0.13497667014598846,
618
+ "eval_roc_auc": 0.8616735348011272,
619
+ "eval_runtime": 691.2148,
620
+ "eval_samples_per_second": 4.216,
621
+ "eval_steps_per_second": 0.133,
622
+ "learning_rate": 0.001,
623
+ "step": 10298
624
+ },
625
+ {
626
+ "epoch": 38.75,
627
+ "learning_rate": 0.001,
628
+ "loss": 0.1503,
629
+ "step": 10500
630
+ },
631
+ {
632
+ "epoch": 39.0,
633
+ "eval_accuracy": 0.2800274536719286,
634
+ "eval_f1_macro": 0.6464711728800093,
635
+ "eval_f1_micro": 0.7983622472668946,
636
+ "eval_loss": 0.1351996511220932,
637
+ "eval_roc_auc": 0.866066786187883,
638
+ "eval_runtime": 704.3219,
639
+ "eval_samples_per_second": 4.137,
640
+ "eval_steps_per_second": 0.131,
641
+ "learning_rate": 0.001,
642
+ "step": 10569
643
+ },
644
+ {
645
+ "epoch": 40.0,
646
+ "eval_accuracy": 0.27282086479066575,
647
+ "eval_f1_macro": 0.6271011637782951,
648
+ "eval_f1_micro": 0.792462504807076,
649
+ "eval_loss": 0.13469766080379486,
650
+ "eval_roc_auc": 0.8594016621216213,
651
+ "eval_runtime": 698.796,
652
+ "eval_samples_per_second": 4.17,
653
+ "eval_steps_per_second": 0.132,
654
+ "learning_rate": 0.001,
655
+ "step": 10840
656
+ },
657
+ {
658
+ "epoch": 40.59,
659
+ "learning_rate": 0.001,
660
+ "loss": 0.1505,
661
+ "step": 11000
662
+ },
663
+ {
664
+ "epoch": 41.0,
665
+ "eval_accuracy": 0.2721345229924502,
666
+ "eval_f1_macro": 0.6600650686410704,
667
+ "eval_f1_micro": 0.7934581450398106,
668
+ "eval_loss": 0.13396936655044556,
669
+ "eval_roc_auc": 0.8579358774851804,
670
+ "eval_runtime": 697.9239,
671
+ "eval_samples_per_second": 4.175,
672
+ "eval_steps_per_second": 0.132,
673
+ "learning_rate": 0.001,
674
+ "step": 11111
675
+ },
676
+ {
677
+ "epoch": 42.0,
678
+ "eval_accuracy": 0.27110501029512696,
679
+ "eval_f1_macro": 0.6636104085895331,
680
+ "eval_f1_micro": 0.7982560108364375,
681
+ "eval_loss": 0.13215309381484985,
682
+ "eval_roc_auc": 0.8652014818411935,
683
+ "eval_runtime": 682.8814,
684
+ "eval_samples_per_second": 4.267,
685
+ "eval_steps_per_second": 0.135,
686
+ "learning_rate": 0.001,
687
+ "step": 11382
688
+ },
689
+ {
690
+ "epoch": 42.44,
691
+ "learning_rate": 0.001,
692
+ "loss": 0.1491,
693
+ "step": 11500
694
+ },
695
+ {
696
+ "epoch": 43.0,
697
+ "eval_accuracy": 0.2735072065888813,
698
+ "eval_f1_macro": 0.6493239813433992,
699
+ "eval_f1_micro": 0.7948717948717948,
700
+ "eval_loss": 0.13603103160858154,
701
+ "eval_roc_auc": 0.8634709756212143,
702
+ "eval_runtime": 691.0843,
703
+ "eval_samples_per_second": 4.217,
704
+ "eval_steps_per_second": 0.133,
705
+ "learning_rate": 0.001,
706
+ "step": 11653
707
+ },
708
+ {
709
+ "epoch": 44.0,
710
+ "eval_accuracy": 0.28140013726835966,
711
+ "eval_f1_macro": 0.6400431467798345,
712
+ "eval_f1_micro": 0.7955270207066627,
713
+ "eval_loss": 0.13608315587043762,
714
+ "eval_roc_auc": 0.8624513565815523,
715
+ "eval_runtime": 684.569,
716
+ "eval_samples_per_second": 4.257,
717
+ "eval_steps_per_second": 0.134,
718
+ "learning_rate": 0.001,
719
+ "step": 11924
720
+ },
721
+ {
722
+ "epoch": 44.28,
723
+ "learning_rate": 0.001,
724
+ "loss": 0.1507,
725
+ "step": 12000
726
+ },
727
+ {
728
+ "epoch": 45.0,
729
+ "eval_accuracy": 0.28140013726835966,
730
+ "eval_f1_macro": 0.6424034446314527,
731
+ "eval_f1_micro": 0.7970817780794026,
732
+ "eval_loss": 0.13283775746822357,
733
+ "eval_roc_auc": 0.863982631455529,
734
+ "eval_runtime": 683.5579,
735
+ "eval_samples_per_second": 4.263,
736
+ "eval_steps_per_second": 0.135,
737
+ "learning_rate": 0.001,
738
+ "step": 12195
739
+ },
740
+ {
741
+ "epoch": 46.0,
742
+ "eval_accuracy": 0.2786547700754976,
743
+ "eval_f1_macro": 0.6468829157126516,
744
+ "eval_f1_micro": 0.7938707872422847,
745
+ "eval_loss": 0.13275618851184845,
746
+ "eval_roc_auc": 0.8581338694154104,
747
+ "eval_runtime": 680.6581,
748
+ "eval_samples_per_second": 4.281,
749
+ "eval_steps_per_second": 0.135,
750
+ "learning_rate": 0.001,
751
+ "step": 12466
752
+ },
753
+ {
754
+ "epoch": 46.13,
755
+ "learning_rate": 0.001,
756
+ "loss": 0.1495,
757
+ "step": 12500
758
+ },
759
+ {
760
+ "epoch": 47.0,
761
+ "eval_accuracy": 0.27522306108442,
762
+ "eval_f1_macro": 0.6351226382655659,
763
+ "eval_f1_micro": 0.7977320453590928,
764
+ "eval_loss": 0.133217915892601,
765
+ "eval_roc_auc": 0.8671532629726034,
766
+ "eval_runtime": 679.0852,
767
+ "eval_samples_per_second": 4.291,
768
+ "eval_steps_per_second": 0.135,
769
+ "learning_rate": 0.001,
770
+ "step": 12737
771
+ },
772
+ {
773
+ "epoch": 47.97,
774
+ "learning_rate": 0.001,
775
+ "loss": 0.1498,
776
+ "step": 13000
777
+ },
778
+ {
779
+ "epoch": 48.0,
780
+ "eval_accuracy": 0.2817433081674674,
781
+ "eval_f1_macro": 0.6490013214958164,
782
+ "eval_f1_micro": 0.8012753282711751,
783
+ "eval_loss": 0.1325378566980362,
784
+ "eval_roc_auc": 0.8694327770935429,
785
+ "eval_runtime": 680.9219,
786
+ "eval_samples_per_second": 4.279,
787
+ "eval_steps_per_second": 0.135,
788
+ "learning_rate": 0.001,
789
+ "step": 13008
790
+ },
791
+ {
792
+ "epoch": 49.0,
793
+ "eval_accuracy": 0.2882635552505148,
794
+ "eval_f1_macro": 0.6738484837965685,
795
+ "eval_f1_micro": 0.8061649892618015,
796
+ "eval_loss": 0.12826864421367645,
797
+ "eval_roc_auc": 0.8710295165867058,
798
+ "eval_runtime": 681.5467,
799
+ "eval_samples_per_second": 4.276,
800
+ "eval_steps_per_second": 0.135,
801
+ "learning_rate": 0.0001,
802
+ "step": 13279
803
+ },
804
+ {
805
+ "epoch": 49.82,
806
+ "learning_rate": 0.0001,
807
+ "loss": 0.1416,
808
+ "step": 13500
809
+ },
810
+ {
811
+ "epoch": 50.0,
812
+ "eval_accuracy": 0.2872340425531915,
813
+ "eval_f1_macro": 0.6734470768849717,
814
+ "eval_f1_micro": 0.8086902026321288,
815
+ "eval_loss": 0.12865300476551056,
816
+ "eval_roc_auc": 0.8747272405310621,
817
+ "eval_runtime": 680.1838,
818
+ "eval_samples_per_second": 4.284,
819
+ "eval_steps_per_second": 0.135,
820
+ "learning_rate": 0.0001,
821
+ "step": 13550
822
+ },
823
+ {
824
+ "epoch": 51.0,
825
+ "eval_accuracy": 0.2899794097460535,
826
+ "eval_f1_macro": 0.6713719689925478,
827
+ "eval_f1_micro": 0.8067049484884311,
828
+ "eval_loss": 0.12803621590137482,
829
+ "eval_roc_auc": 0.870555059108937,
830
+ "eval_runtime": 682.2484,
831
+ "eval_samples_per_second": 4.271,
832
+ "eval_steps_per_second": 0.135,
833
+ "learning_rate": 0.0001,
834
+ "step": 13821
835
+ },
836
+ {
837
+ "epoch": 51.66,
838
+ "learning_rate": 0.0001,
839
+ "loss": 0.1387,
840
+ "step": 14000
841
+ },
842
+ {
843
+ "epoch": 52.0,
844
+ "eval_accuracy": 0.2899794097460535,
845
+ "eval_f1_macro": 0.6744045610607882,
846
+ "eval_f1_micro": 0.8067475584492453,
847
+ "eval_loss": 0.12618477642536163,
848
+ "eval_roc_auc": 0.8701989061466676,
849
+ "eval_runtime": 687.4248,
850
+ "eval_samples_per_second": 4.239,
851
+ "eval_steps_per_second": 0.134,
852
+ "learning_rate": 0.0001,
853
+ "step": 14092
854
+ },
855
+ {
856
+ "epoch": 53.0,
857
+ "eval_accuracy": 0.2910089224433768,
858
+ "eval_f1_macro": 0.6763777606257594,
859
+ "eval_f1_micro": 0.8094476254631189,
860
+ "eval_loss": 0.1262361854314804,
861
+ "eval_roc_auc": 0.8728950289492735,
862
+ "eval_runtime": 697.5202,
863
+ "eval_samples_per_second": 4.178,
864
+ "eval_steps_per_second": 0.132,
865
+ "learning_rate": 0.0001,
866
+ "step": 14363
867
+ },
868
+ {
869
+ "epoch": 53.51,
870
+ "learning_rate": 0.0001,
871
+ "loss": 0.1356,
872
+ "step": 14500
873
+ },
874
+ {
875
+ "epoch": 54.0,
876
+ "eval_accuracy": 0.2947838023335621,
877
+ "eval_f1_macro": 0.6743662487872923,
878
+ "eval_f1_micro": 0.809105193867584,
879
+ "eval_loss": 0.12573254108428955,
880
+ "eval_roc_auc": 0.8701859740035948,
881
+ "eval_runtime": 693.4876,
882
+ "eval_samples_per_second": 4.202,
883
+ "eval_steps_per_second": 0.133,
884
+ "learning_rate": 0.0001,
885
+ "step": 14634
886
+ },
887
+ {
888
+ "epoch": 55.0,
889
+ "eval_accuracy": 0.2947838023335621,
890
+ "eval_f1_macro": 0.681373956781595,
891
+ "eval_f1_micro": 0.8106019238039233,
892
+ "eval_loss": 0.12566907703876495,
893
+ "eval_roc_auc": 0.8742077383085138,
894
+ "eval_runtime": 700.1378,
895
+ "eval_samples_per_second": 4.162,
896
+ "eval_steps_per_second": 0.131,
897
+ "learning_rate": 0.0001,
898
+ "step": 14905
899
+ },
900
+ {
901
+ "epoch": 55.35,
902
+ "learning_rate": 0.0001,
903
+ "loss": 0.1348,
904
+ "step": 15000
905
+ },
906
+ {
907
+ "epoch": 56.0,
908
+ "eval_accuracy": 0.3009608785175017,
909
+ "eval_f1_macro": 0.6772158941721765,
910
+ "eval_f1_micro": 0.8107721439091101,
911
+ "eval_loss": 0.12600058317184448,
912
+ "eval_roc_auc": 0.873783436948714,
913
+ "eval_runtime": 679.772,
914
+ "eval_samples_per_second": 4.287,
915
+ "eval_steps_per_second": 0.135,
916
+ "learning_rate": 0.0001,
917
+ "step": 15176
918
+ },
919
+ {
920
+ "epoch": 57.0,
921
+ "eval_accuracy": 0.29855868222374743,
922
+ "eval_f1_macro": 0.6806264224832896,
923
+ "eval_f1_micro": 0.8128972900635664,
924
+ "eval_loss": 0.12499917298555374,
925
+ "eval_roc_auc": 0.8767727269259479,
926
+ "eval_runtime": 685.6404,
927
+ "eval_samples_per_second": 4.25,
928
+ "eval_steps_per_second": 0.134,
929
+ "learning_rate": 0.0001,
930
+ "step": 15447
931
+ },
932
+ {
933
+ "epoch": 57.2,
934
+ "learning_rate": 0.0001,
935
+ "loss": 0.135,
936
+ "step": 15500
937
+ },
938
+ {
939
+ "epoch": 58.0,
940
+ "eval_accuracy": 0.3081674673987646,
941
+ "eval_f1_macro": 0.6858667117679004,
942
+ "eval_f1_micro": 0.8141957160856784,
943
+ "eval_loss": 0.12423347681760788,
944
+ "eval_roc_auc": 0.8762249584407599,
945
+ "eval_runtime": 700.338,
946
+ "eval_samples_per_second": 4.161,
947
+ "eval_steps_per_second": 0.131,
948
+ "learning_rate": 0.0001,
949
+ "step": 15718
950
+ },
951
+ {
952
+ "epoch": 59.0,
953
+ "eval_accuracy": 0.3026767330130405,
954
+ "eval_f1_macro": 0.6869701850967913,
955
+ "eval_f1_micro": 0.8124398308986648,
956
+ "eval_loss": 0.12452811747789383,
957
+ "eval_roc_auc": 0.8762597292024611,
958
+ "eval_runtime": 691.5065,
959
+ "eval_samples_per_second": 4.214,
960
+ "eval_steps_per_second": 0.133,
961
+ "learning_rate": 0.0001,
962
+ "step": 15989
963
+ },
964
+ {
965
+ "epoch": 59.04,
966
+ "learning_rate": 0.0001,
967
+ "loss": 0.1334,
968
+ "step": 16000
969
+ },
970
+ {
971
+ "epoch": 60.0,
972
+ "eval_accuracy": 0.3030199039121482,
973
+ "eval_f1_macro": 0.6853545892350839,
974
+ "eval_f1_micro": 0.8137861803580391,
975
+ "eval_loss": 0.1242317408323288,
976
+ "eval_roc_auc": 0.8772149949469599,
977
+ "eval_runtime": 700.2391,
978
+ "eval_samples_per_second": 4.161,
979
+ "eval_steps_per_second": 0.131,
980
+ "learning_rate": 0.0001,
981
+ "step": 16260
982
+ },
983
+ {
984
+ "epoch": 60.89,
985
+ "learning_rate": 0.0001,
986
+ "loss": 0.1335,
987
+ "step": 16500
988
+ },
989
+ {
990
+ "epoch": 61.0,
991
+ "eval_accuracy": 0.3064516129032258,
992
+ "eval_f1_macro": 0.6889451502143565,
993
+ "eval_f1_micro": 0.8139671855279764,
994
+ "eval_loss": 0.12397264689207077,
995
+ "eval_roc_auc": 0.8756026651507299,
996
+ "eval_runtime": 691.2006,
997
+ "eval_samples_per_second": 4.216,
998
+ "eval_steps_per_second": 0.133,
999
+ "learning_rate": 0.0001,
1000
+ "step": 16531
1001
+ },
1002
+ {
1003
+ "epoch": 62.0,
1004
+ "eval_accuracy": 0.3016472203157172,
1005
+ "eval_f1_macro": 0.6808873228168837,
1006
+ "eval_f1_micro": 0.8152336604024614,
1007
+ "eval_loss": 0.12486530840396881,
1008
+ "eval_roc_auc": 0.8798070104112615,
1009
+ "eval_runtime": 696.5149,
1010
+ "eval_samples_per_second": 4.184,
1011
+ "eval_steps_per_second": 0.132,
1012
+ "learning_rate": 0.0001,
1013
+ "step": 16802
1014
+ },
1015
+ {
1016
+ "epoch": 62.73,
1017
+ "learning_rate": 0.0001,
1018
+ "loss": 0.1308,
1019
+ "step": 17000
1020
+ },
1021
+ {
1022
+ "epoch": 63.0,
1023
+ "eval_accuracy": 0.30679478380233355,
1024
+ "eval_f1_macro": 0.6848490082628171,
1025
+ "eval_f1_micro": 0.814602720114531,
1026
+ "eval_loss": 0.1233312338590622,
1027
+ "eval_roc_auc": 0.8756913615144166,
1028
+ "eval_runtime": 687.8446,
1029
+ "eval_samples_per_second": 4.236,
1030
+ "eval_steps_per_second": 0.134,
1031
+ "learning_rate": 0.0001,
1032
+ "step": 17073
1033
+ },
1034
+ {
1035
+ "epoch": 64.0,
1036
+ "eval_accuracy": 0.3057652711050103,
1037
+ "eval_f1_macro": 0.6908433037124228,
1038
+ "eval_f1_micro": 0.8151443922095367,
1039
+ "eval_loss": 0.12344498932361603,
1040
+ "eval_roc_auc": 0.8769218021539817,
1041
+ "eval_runtime": 702.1676,
1042
+ "eval_samples_per_second": 4.15,
1043
+ "eval_steps_per_second": 0.131,
1044
+ "learning_rate": 0.0001,
1045
+ "step": 17344
1046
+ },
1047
+ {
1048
+ "epoch": 64.58,
1049
+ "learning_rate": 0.0001,
1050
+ "loss": 0.1326,
1051
+ "step": 17500
1052
+ },
1053
+ {
1054
+ "epoch": 65.0,
1055
+ "eval_accuracy": 0.303363074811256,
1056
+ "eval_f1_macro": 0.6812168886607934,
1057
+ "eval_f1_micro": 0.8124392614188533,
1058
+ "eval_loss": 0.12330327183008194,
1059
+ "eval_roc_auc": 0.8734876628507912,
1060
+ "eval_runtime": 680.7902,
1061
+ "eval_samples_per_second": 4.28,
1062
+ "eval_steps_per_second": 0.135,
1063
+ "learning_rate": 0.0001,
1064
+ "step": 17615
1065
+ },
1066
+ {
1067
+ "epoch": 66.0,
1068
+ "eval_accuracy": 0.3026767330130405,
1069
+ "eval_f1_macro": 0.6878311502092693,
1070
+ "eval_f1_micro": 0.8144785071642787,
1071
+ "eval_loss": 0.12320297956466675,
1072
+ "eval_roc_auc": 0.8788401544128625,
1073
+ "eval_runtime": 690.4713,
1074
+ "eval_samples_per_second": 4.22,
1075
+ "eval_steps_per_second": 0.133,
1076
+ "learning_rate": 0.0001,
1077
+ "step": 17886
1078
+ },
1079
+ {
1080
+ "epoch": 66.42,
1081
+ "learning_rate": 0.0001,
1082
+ "loss": 0.1306,
1083
+ "step": 18000
1084
+ },
1085
+ {
1086
+ "epoch": 67.0,
1087
+ "eval_accuracy": 0.3074811256005491,
1088
+ "eval_f1_macro": 0.6857441975499322,
1089
+ "eval_f1_micro": 0.8115152031343156,
1090
+ "eval_loss": 0.1227714866399765,
1091
+ "eval_roc_auc": 0.8706963391854371,
1092
+ "eval_runtime": 709.7385,
1093
+ "eval_samples_per_second": 4.106,
1094
+ "eval_steps_per_second": 0.13,
1095
+ "learning_rate": 0.0001,
1096
+ "step": 18157
1097
+ },
1098
+ {
1099
+ "epoch": 68.0,
1100
+ "eval_accuracy": 0.3074811256005491,
1101
+ "eval_f1_macro": 0.6913054019394733,
1102
+ "eval_f1_micro": 0.8153315962007229,
1103
+ "eval_loss": 0.12259615212678909,
1104
+ "eval_roc_auc": 0.8766517898688044,
1105
+ "eval_runtime": 686.725,
1106
+ "eval_samples_per_second": 4.243,
1107
+ "eval_steps_per_second": 0.134,
1108
+ "learning_rate": 0.0001,
1109
+ "step": 18428
1110
+ },
1111
+ {
1112
+ "epoch": 68.27,
1113
+ "learning_rate": 0.0001,
1114
+ "loss": 0.1299,
1115
+ "step": 18500
1116
+ },
1117
+ {
1118
+ "epoch": 69.0,
1119
+ "eval_accuracy": 0.30851063829787234,
1120
+ "eval_f1_macro": 0.676353464691654,
1121
+ "eval_f1_micro": 0.8143055965585593,
1122
+ "eval_loss": 0.12271784245967865,
1123
+ "eval_roc_auc": 0.8750562129363642,
1124
+ "eval_runtime": 693.9986,
1125
+ "eval_samples_per_second": 4.199,
1126
+ "eval_steps_per_second": 0.133,
1127
+ "learning_rate": 0.0001,
1128
+ "step": 18699
1129
+ },
1130
+ {
1131
+ "epoch": 70.0,
1132
+ "eval_accuracy": 0.31056966369251887,
1133
+ "eval_f1_macro": 0.6999234521712909,
1134
+ "eval_f1_micro": 0.81868109179502,
1135
+ "eval_loss": 0.12295936793088913,
1136
+ "eval_roc_auc": 0.8837914226495347,
1137
+ "eval_runtime": 690.298,
1138
+ "eval_samples_per_second": 4.221,
1139
+ "eval_steps_per_second": 0.133,
1140
+ "learning_rate": 0.0001,
1141
+ "step": 18970
1142
+ },
1143
+ {
1144
+ "epoch": 70.11,
1145
+ "learning_rate": 0.0001,
1146
+ "loss": 0.1295,
1147
+ "step": 19000
1148
+ },
1149
+ {
1150
+ "epoch": 71.0,
1151
+ "eval_accuracy": 0.30679478380233355,
1152
+ "eval_f1_macro": 0.6893382907323766,
1153
+ "eval_f1_micro": 0.8152834008097165,
1154
+ "eval_loss": 0.12247700244188309,
1155
+ "eval_roc_auc": 0.8756051952189744,
1156
+ "eval_runtime": 708.7396,
1157
+ "eval_samples_per_second": 4.112,
1158
+ "eval_steps_per_second": 0.13,
1159
+ "learning_rate": 0.0001,
1160
+ "step": 19241
1161
+ },
1162
+ {
1163
+ "epoch": 71.96,
1164
+ "learning_rate": 0.0001,
1165
+ "loss": 0.1289,
1166
+ "step": 19500
1167
+ },
1168
+ {
1169
+ "epoch": 72.0,
1170
+ "eval_accuracy": 0.30370624571036375,
1171
+ "eval_f1_macro": 0.6867583154614194,
1172
+ "eval_f1_micro": 0.815056734916049,
1173
+ "eval_loss": 0.12231950461864471,
1174
+ "eval_roc_auc": 0.877607106850003,
1175
+ "eval_runtime": 677.1006,
1176
+ "eval_samples_per_second": 4.304,
1177
+ "eval_steps_per_second": 0.136,
1178
+ "learning_rate": 0.0001,
1179
+ "step": 19512
1180
+ },
1181
+ {
1182
+ "epoch": 73.0,
1183
+ "eval_accuracy": 0.30542210020590255,
1184
+ "eval_f1_macro": 0.6917904729736315,
1185
+ "eval_f1_micro": 0.8165095327886026,
1186
+ "eval_loss": 0.12229206413030624,
1187
+ "eval_roc_auc": 0.8781856460477259,
1188
+ "eval_runtime": 692.4475,
1189
+ "eval_samples_per_second": 4.208,
1190
+ "eval_steps_per_second": 0.133,
1191
+ "learning_rate": 0.0001,
1192
+ "step": 19783
1193
+ },
1194
+ {
1195
+ "epoch": 73.8,
1196
+ "learning_rate": 0.0001,
1197
+ "loss": 0.1279,
1198
+ "step": 20000
1199
+ },
1200
+ {
1201
+ "epoch": 74.0,
1202
+ "eval_accuracy": 0.30542210020590255,
1203
+ "eval_f1_macro": 0.6855916835036744,
1204
+ "eval_f1_micro": 0.8142863173892853,
1205
+ "eval_loss": 0.12248736619949341,
1206
+ "eval_roc_auc": 0.8747476347871186,
1207
+ "eval_runtime": 687.0839,
1208
+ "eval_samples_per_second": 4.241,
1209
+ "eval_steps_per_second": 0.134,
1210
+ "learning_rate": 0.0001,
1211
+ "step": 20054
1212
+ },
1213
+ {
1214
+ "epoch": 75.0,
1215
+ "eval_accuracy": 0.31022649279341113,
1216
+ "eval_f1_macro": 0.6878236573831232,
1217
+ "eval_f1_micro": 0.8167385749591589,
1218
+ "eval_loss": 0.12206920981407166,
1219
+ "eval_roc_auc": 0.8784119298589457,
1220
+ "eval_runtime": 681.8157,
1221
+ "eval_samples_per_second": 4.274,
1222
+ "eval_steps_per_second": 0.135,
1223
+ "learning_rate": 0.0001,
1224
+ "step": 20325
1225
+ },
1226
+ {
1227
+ "epoch": 75.65,
1228
+ "learning_rate": 0.0001,
1229
+ "loss": 0.1276,
1230
+ "step": 20500
1231
+ },
1232
+ {
1233
+ "epoch": 76.0,
1234
+ "eval_accuracy": 0.31674673987645846,
1235
+ "eval_f1_macro": 0.6963741846422794,
1236
+ "eval_f1_micro": 0.8190190440471725,
1237
+ "eval_loss": 0.12172180414199829,
1238
+ "eval_roc_auc": 0.8812250665245784,
1239
+ "eval_runtime": 672.3327,
1240
+ "eval_samples_per_second": 4.334,
1241
+ "eval_steps_per_second": 0.137,
1242
+ "learning_rate": 0.0001,
1243
+ "step": 20596
1244
+ },
1245
+ {
1246
+ "epoch": 77.0,
1247
+ "eval_accuracy": 0.31022649279341113,
1248
+ "eval_f1_macro": 0.6940319651093022,
1249
+ "eval_f1_micro": 0.8179309191268713,
1250
+ "eval_loss": 0.12170004099607468,
1251
+ "eval_roc_auc": 0.8795832054315861,
1252
+ "eval_runtime": 683.2195,
1253
+ "eval_samples_per_second": 4.265,
1254
+ "eval_steps_per_second": 0.135,
1255
+ "learning_rate": 0.0001,
1256
+ "step": 20867
1257
+ },
1258
+ {
1259
+ "epoch": 77.49,
1260
+ "learning_rate": 0.0001,
1261
+ "loss": 0.1274,
1262
+ "step": 21000
1263
+ },
1264
+ {
1265
+ "epoch": 78.0,
1266
+ "eval_accuracy": 0.3081674673987646,
1267
+ "eval_f1_macro": 0.6858950802374723,
1268
+ "eval_f1_micro": 0.8143468859965235,
1269
+ "eval_loss": 0.12156965583562851,
1270
+ "eval_roc_auc": 0.8735390327661822,
1271
+ "eval_runtime": 682.8416,
1272
+ "eval_samples_per_second": 4.267,
1273
+ "eval_steps_per_second": 0.135,
1274
+ "learning_rate": 0.0001,
1275
+ "step": 21138
1276
+ },
1277
+ {
1278
+ "epoch": 79.0,
1279
+ "eval_accuracy": 0.31468771448181193,
1280
+ "eval_f1_macro": 0.6944979922976369,
1281
+ "eval_f1_micro": 0.8164588948787063,
1282
+ "eval_loss": 0.1215372309088707,
1283
+ "eval_roc_auc": 0.8766491829455099,
1284
+ "eval_runtime": 683.544,
1285
+ "eval_samples_per_second": 4.263,
1286
+ "eval_steps_per_second": 0.135,
1287
+ "learning_rate": 0.0001,
1288
+ "step": 21409
1289
+ },
1290
+ {
1291
+ "epoch": 79.34,
1292
+ "learning_rate": 0.0001,
1293
+ "loss": 0.1269,
1294
+ "step": 21500
1295
+ },
1296
+ {
1297
+ "epoch": 80.0,
1298
+ "eval_accuracy": 0.31468771448181193,
1299
+ "eval_f1_macro": 0.6999268076002686,
1300
+ "eval_f1_micro": 0.8192881937183724,
1301
+ "eval_loss": 0.12135831266641617,
1302
+ "eval_roc_auc": 0.8802967507718976,
1303
+ "eval_runtime": 676.6048,
1304
+ "eval_samples_per_second": 4.307,
1305
+ "eval_steps_per_second": 0.136,
1306
+ "learning_rate": 0.0001,
1307
+ "step": 21680
1308
+ },
1309
+ {
1310
+ "epoch": 81.0,
1311
+ "eval_accuracy": 0.3112560054907344,
1312
+ "eval_f1_macro": 0.697411996468389,
1313
+ "eval_f1_micro": 0.81943004106691,
1314
+ "eval_loss": 0.12142007052898407,
1315
+ "eval_roc_auc": 0.8828322337109193,
1316
+ "eval_runtime": 681.4379,
1317
+ "eval_samples_per_second": 4.276,
1318
+ "eval_steps_per_second": 0.135,
1319
+ "learning_rate": 0.0001,
1320
+ "step": 21951
1321
+ },
1322
+ {
1323
+ "epoch": 81.18,
1324
+ "learning_rate": 0.0001,
1325
+ "loss": 0.1259,
1326
+ "step": 22000
1327
+ },
1328
+ {
1329
+ "epoch": 82.0,
1330
+ "eval_accuracy": 0.31022649279341113,
1331
+ "eval_f1_macro": 0.695644238424867,
1332
+ "eval_f1_micro": 0.817148370317547,
1333
+ "eval_loss": 0.12121300399303436,
1334
+ "eval_roc_auc": 0.8782139123103657,
1335
+ "eval_runtime": 687.8898,
1336
+ "eval_samples_per_second": 4.236,
1337
+ "eval_steps_per_second": 0.134,
1338
+ "learning_rate": 0.0001,
1339
+ "step": 22222
1340
+ },
1341
+ {
1342
+ "epoch": 83.0,
1343
+ "eval_accuracy": 0.31228551818805766,
1344
+ "eval_f1_macro": 0.6969593372601354,
1345
+ "eval_f1_micro": 0.8189676877885018,
1346
+ "eval_loss": 0.12076255679130554,
1347
+ "eval_roc_auc": 0.8791048759964009,
1348
+ "eval_runtime": 683.6587,
1349
+ "eval_samples_per_second": 4.262,
1350
+ "eval_steps_per_second": 0.135,
1351
+ "learning_rate": 0.0001,
1352
+ "step": 22493
1353
+ },
1354
+ {
1355
+ "epoch": 83.03,
1356
+ "learning_rate": 0.0001,
1357
+ "loss": 0.1258,
1358
+ "step": 22500
1359
+ },
1360
+ {
1361
+ "epoch": 84.0,
1362
+ "eval_accuracy": 0.31537405628002746,
1363
+ "eval_f1_macro": 0.699678101503351,
1364
+ "eval_f1_micro": 0.8203842940685045,
1365
+ "eval_loss": 0.12093241512775421,
1366
+ "eval_roc_auc": 0.8812674531093634,
1367
+ "eval_runtime": 674.7538,
1368
+ "eval_samples_per_second": 4.319,
1369
+ "eval_steps_per_second": 0.136,
1370
+ "learning_rate": 0.0001,
1371
+ "step": 22764
1372
+ },
1373
+ {
1374
+ "epoch": 84.87,
1375
+ "learning_rate": 0.0001,
1376
+ "loss": 0.1251,
1377
+ "step": 23000
1378
+ },
1379
+ {
1380
+ "epoch": 85.0,
1381
+ "eval_accuracy": 0.3064516129032258,
1382
+ "eval_f1_macro": 0.6934546509096284,
1383
+ "eval_f1_micro": 0.8163317114448911,
1384
+ "eval_loss": 0.12106911092996597,
1385
+ "eval_roc_auc": 0.875220724757365,
1386
+ "eval_runtime": 682.5674,
1387
+ "eval_samples_per_second": 4.269,
1388
+ "eval_steps_per_second": 0.135,
1389
+ "learning_rate": 0.0001,
1390
+ "step": 23035
1391
+ },
1392
+ {
1393
+ "epoch": 86.0,
1394
+ "eval_accuracy": 0.31537405628002746,
1395
+ "eval_f1_macro": 0.6971612900401489,
1396
+ "eval_f1_micro": 0.8200962947456564,
1397
+ "eval_loss": 0.12031004577875137,
1398
+ "eval_roc_auc": 0.8804175853556411,
1399
+ "eval_runtime": 685.5285,
1400
+ "eval_samples_per_second": 4.251,
1401
+ "eval_steps_per_second": 0.134,
1402
+ "learning_rate": 0.0001,
1403
+ "step": 23306
1404
+ },
1405
+ {
1406
+ "epoch": 86.72,
1407
+ "learning_rate": 0.0001,
1408
+ "loss": 0.1251,
1409
+ "step": 23500
1410
+ },
1411
+ {
1412
+ "epoch": 87.0,
1413
+ "eval_accuracy": 0.3150308853809197,
1414
+ "eval_f1_macro": 0.6946815591935039,
1415
+ "eval_f1_micro": 0.8182238085240395,
1416
+ "eval_loss": 0.12081247568130493,
1417
+ "eval_roc_auc": 0.8785070282482941,
1418
+ "eval_runtime": 683.1403,
1419
+ "eval_samples_per_second": 4.266,
1420
+ "eval_steps_per_second": 0.135,
1421
+ "learning_rate": 0.0001,
1422
+ "step": 23577
1423
+ },
1424
+ {
1425
+ "epoch": 88.0,
1426
+ "eval_accuracy": 0.31537405628002746,
1427
+ "eval_f1_macro": 0.6936867859001314,
1428
+ "eval_f1_micro": 0.8180521768308028,
1429
+ "eval_loss": 0.12136104702949524,
1430
+ "eval_roc_auc": 0.8787500393049534,
1431
+ "eval_runtime": 674.7599,
1432
+ "eval_samples_per_second": 4.319,
1433
+ "eval_steps_per_second": 0.136,
1434
+ "learning_rate": 0.0001,
1435
+ "step": 23848
1436
+ },
1437
+ {
1438
+ "epoch": 88.56,
1439
+ "learning_rate": 0.0001,
1440
+ "loss": 0.1246,
1441
+ "step": 24000
1442
+ },
1443
+ {
1444
+ "epoch": 89.0,
1445
+ "eval_accuracy": 0.31056966369251887,
1446
+ "eval_f1_macro": 0.6953152626493269,
1447
+ "eval_f1_micro": 0.8201318053981447,
1448
+ "eval_loss": 0.12058280408382416,
1449
+ "eval_roc_auc": 0.8796589952705502,
1450
+ "eval_runtime": 682.7847,
1451
+ "eval_samples_per_second": 4.268,
1452
+ "eval_steps_per_second": 0.135,
1453
+ "learning_rate": 0.0001,
1454
+ "step": 24119
1455
+ },
1456
+ {
1457
+ "epoch": 90.0,
1458
+ "eval_accuracy": 0.3164035689773507,
1459
+ "eval_f1_macro": 0.696027478931947,
1460
+ "eval_f1_micro": 0.8213629530649741,
1461
+ "eval_loss": 0.12102781236171722,
1462
+ "eval_roc_auc": 0.8819102943655497,
1463
+ "eval_runtime": 686.1462,
1464
+ "eval_samples_per_second": 4.247,
1465
+ "eval_steps_per_second": 0.134,
1466
+ "learning_rate": 0.0001,
1467
+ "step": 24390
1468
+ },
1469
+ {
1470
+ "epoch": 90.41,
1471
+ "learning_rate": 0.0001,
1472
+ "loss": 0.1239,
1473
+ "step": 24500
1474
+ },
1475
+ {
1476
+ "epoch": 91.0,
1477
+ "eval_accuracy": 0.31537405628002746,
1478
+ "eval_f1_macro": 0.7005956031917913,
1479
+ "eval_f1_micro": 0.8201800293070964,
1480
+ "eval_loss": 0.1198815330862999,
1481
+ "eval_roc_auc": 0.8804638695161929,
1482
+ "eval_runtime": 683.8668,
1483
+ "eval_samples_per_second": 4.261,
1484
+ "eval_steps_per_second": 0.135,
1485
+ "learning_rate": 0.0001,
1486
+ "step": 24661
1487
+ },
1488
+ {
1489
+ "epoch": 92.0,
1490
+ "eval_accuracy": 0.316060398078243,
1491
+ "eval_f1_macro": 0.703909588701163,
1492
+ "eval_f1_micro": 0.8221735718121359,
1493
+ "eval_loss": 0.12077653408050537,
1494
+ "eval_roc_auc": 0.8856479260631466,
1495
+ "eval_runtime": 672.8479,
1496
+ "eval_samples_per_second": 4.331,
1497
+ "eval_steps_per_second": 0.137,
1498
+ "learning_rate": 0.0001,
1499
+ "step": 24932
1500
+ },
1501
+ {
1502
+ "epoch": 92.25,
1503
+ "learning_rate": 0.0001,
1504
+ "loss": 0.1238,
1505
+ "step": 25000
1506
+ },
1507
+ {
1508
+ "epoch": 93.0,
1509
+ "eval_accuracy": 0.31331503088538093,
1510
+ "eval_f1_macro": 0.7003774810286753,
1511
+ "eval_f1_micro": 0.8199147228492601,
1512
+ "eval_loss": 0.12042003870010376,
1513
+ "eval_roc_auc": 0.8807814566143944,
1514
+ "eval_runtime": 693.4839,
1515
+ "eval_samples_per_second": 4.202,
1516
+ "eval_steps_per_second": 0.133,
1517
+ "learning_rate": 0.0001,
1518
+ "step": 25203
1519
+ },
1520
+ {
1521
+ "epoch": 94.0,
1522
+ "eval_accuracy": 0.3143445435827042,
1523
+ "eval_f1_macro": 0.7036330518306199,
1524
+ "eval_f1_micro": 0.8230411686586985,
1525
+ "eval_loss": 0.1200033500790596,
1526
+ "eval_roc_auc": 0.8847080200565651,
1527
+ "eval_runtime": 700.664,
1528
+ "eval_samples_per_second": 4.159,
1529
+ "eval_steps_per_second": 0.131,
1530
+ "learning_rate": 0.0001,
1531
+ "step": 25474
1532
+ },
1533
+ {
1534
+ "epoch": 94.1,
1535
+ "learning_rate": 0.0001,
1536
+ "loss": 0.1237,
1537
+ "step": 25500
1538
+ },
1539
+ {
1540
+ "epoch": 95.0,
1541
+ "eval_accuracy": 0.318805765271105,
1542
+ "eval_f1_macro": 0.7068704040733741,
1543
+ "eval_f1_micro": 0.8209067379143359,
1544
+ "eval_loss": 0.12061866372823715,
1545
+ "eval_roc_auc": 0.8817328760198262,
1546
+ "eval_runtime": 701.3444,
1547
+ "eval_samples_per_second": 4.155,
1548
+ "eval_steps_per_second": 0.131,
1549
+ "learning_rate": 0.0001,
1550
+ "step": 25745
1551
+ },
1552
+ {
1553
+ "epoch": 95.94,
1554
+ "learning_rate": 0.0001,
1555
+ "loss": 0.1234,
1556
+ "step": 26000
1557
+ },
1558
+ {
1559
+ "epoch": 96.0,
1560
+ "eval_accuracy": 0.31468771448181193,
1561
+ "eval_f1_macro": 0.7059911076490224,
1562
+ "eval_f1_micro": 0.8222408026755854,
1563
+ "eval_loss": 0.12005680054426193,
1564
+ "eval_roc_auc": 0.8820439842373163,
1565
+ "eval_runtime": 697.7134,
1566
+ "eval_samples_per_second": 4.177,
1567
+ "eval_steps_per_second": 0.132,
1568
+ "learning_rate": 0.0001,
1569
+ "step": 26016
1570
+ },
1571
+ {
1572
+ "epoch": 97.0,
1573
+ "eval_accuracy": 0.30919698009608787,
1574
+ "eval_f1_macro": 0.7073615141822082,
1575
+ "eval_f1_micro": 0.8207813798836243,
1576
+ "eval_loss": 0.12039094418287277,
1577
+ "eval_roc_auc": 0.882995480497798,
1578
+ "eval_runtime": 701.6443,
1579
+ "eval_samples_per_second": 4.153,
1580
+ "eval_steps_per_second": 0.131,
1581
+ "learning_rate": 0.0001,
1582
+ "step": 26287
1583
+ },
1584
+ {
1585
+ "epoch": 97.79,
1586
+ "learning_rate": 1e-05,
1587
+ "loss": 0.1215,
1588
+ "step": 26500
1589
+ },
1590
+ {
1591
+ "epoch": 98.0,
1592
+ "eval_accuracy": 0.318805765271105,
1593
+ "eval_f1_macro": 0.7125305658957566,
1594
+ "eval_f1_micro": 0.8240848103362597,
1595
+ "eval_loss": 0.12003627419471741,
1596
+ "eval_roc_auc": 0.8859474440267361,
1597
+ "eval_runtime": 703.1401,
1598
+ "eval_samples_per_second": 4.144,
1599
+ "eval_steps_per_second": 0.131,
1600
+ "learning_rate": 1e-05,
1601
+ "step": 26558
1602
+ },
1603
+ {
1604
+ "epoch": 99.0,
1605
+ "eval_accuracy": 0.31708991077556625,
1606
+ "eval_f1_macro": 0.7126519915135826,
1607
+ "eval_f1_micro": 0.8246739805423309,
1608
+ "eval_loss": 0.11952651292085648,
1609
+ "eval_roc_auc": 0.8863652921241748,
1610
+ "eval_runtime": 704.3933,
1611
+ "eval_samples_per_second": 4.137,
1612
+ "eval_steps_per_second": 0.131,
1613
+ "learning_rate": 1e-05,
1614
+ "step": 26829
1615
+ },
1616
+ {
1617
+ "epoch": 99.63,
1618
+ "learning_rate": 1e-05,
1619
+ "loss": 0.1208,
1620
+ "step": 27000
1621
+ },
1622
+ {
1623
+ "epoch": 100.0,
1624
+ "eval_accuracy": 0.3164035689773507,
1625
+ "eval_f1_macro": 0.7076695321633534,
1626
+ "eval_f1_micro": 0.8225050234427328,
1627
+ "eval_loss": 0.11920821666717529,
1628
+ "eval_roc_auc": 0.8817868272401618,
1629
+ "eval_runtime": 693.6368,
1630
+ "eval_samples_per_second": 4.201,
1631
+ "eval_steps_per_second": 0.133,
1632
+ "learning_rate": 1e-05,
1633
+ "step": 27100
1634
+ },
1635
+ {
1636
+ "epoch": 101.0,
1637
+ "eval_accuracy": 0.31708991077556625,
1638
+ "eval_f1_macro": 0.706035453076238,
1639
+ "eval_f1_micro": 0.8232053422370618,
1640
+ "eval_loss": 0.11927199363708496,
1641
+ "eval_roc_auc": 0.8830828349723125,
1642
+ "eval_runtime": 710.5245,
1643
+ "eval_samples_per_second": 4.101,
1644
+ "eval_steps_per_second": 0.129,
1645
+ "learning_rate": 1e-05,
1646
+ "step": 27371
1647
+ },
1648
+ {
1649
+ "epoch": 101.48,
1650
+ "learning_rate": 1e-05,
1651
+ "loss": 0.1195,
1652
+ "step": 27500
1653
+ },
1654
+ {
1655
+ "epoch": 102.0,
1656
+ "eval_accuracy": 0.31846259437199725,
1657
+ "eval_f1_macro": 0.710532612125401,
1658
+ "eval_f1_micro": 0.8237933039793969,
1659
+ "eval_loss": 0.11972019821405411,
1660
+ "eval_roc_auc": 0.8848494282248152,
1661
+ "eval_runtime": 702.1024,
1662
+ "eval_samples_per_second": 4.15,
1663
+ "eval_steps_per_second": 0.131,
1664
+ "learning_rate": 1e-05,
1665
+ "step": 27642
1666
+ },
1667
+ {
1668
+ "epoch": 103.0,
1669
+ "eval_accuracy": 0.31400137268359646,
1670
+ "eval_f1_macro": 0.7075672175938867,
1671
+ "eval_f1_micro": 0.8215544737283866,
1672
+ "eval_loss": 0.11914487928152084,
1673
+ "eval_roc_auc": 0.880506256100978,
1674
+ "eval_runtime": 705.9918,
1675
+ "eval_samples_per_second": 4.128,
1676
+ "eval_steps_per_second": 0.13,
1677
+ "learning_rate": 1e-05,
1678
+ "step": 27913
1679
+ },
1680
+ {
1681
+ "epoch": 103.32,
1682
+ "learning_rate": 1e-05,
1683
+ "loss": 0.1197,
1684
+ "step": 28000
1685
+ },
1686
+ {
1687
+ "epoch": 104.0,
1688
+ "eval_accuracy": 0.32017844886753605,
1689
+ "eval_f1_macro": 0.7063140405914397,
1690
+ "eval_f1_micro": 0.8238771177621446,
1691
+ "eval_loss": 0.11925092339515686,
1692
+ "eval_roc_auc": 0.8842515548583585,
1693
+ "eval_runtime": 693.831,
1694
+ "eval_samples_per_second": 4.2,
1695
+ "eval_steps_per_second": 0.133,
1696
+ "learning_rate": 1e-05,
1697
+ "step": 28184
1698
+ },
1699
+ {
1700
+ "epoch": 105.0,
1701
+ "eval_accuracy": 0.3126286890871654,
1702
+ "eval_f1_macro": 0.7070687169886732,
1703
+ "eval_f1_micro": 0.8213039640169827,
1704
+ "eval_loss": 0.11896480619907379,
1705
+ "eval_roc_auc": 0.8798775351669367,
1706
+ "eval_runtime": 704.4845,
1707
+ "eval_samples_per_second": 4.136,
1708
+ "eval_steps_per_second": 0.131,
1709
+ "learning_rate": 1e-05,
1710
+ "step": 28455
1711
+ },
1712
+ {
1713
+ "epoch": 105.17,
1714
+ "learning_rate": 1e-05,
1715
+ "loss": 0.1189,
1716
+ "step": 28500
1717
+ },
1718
+ {
1719
+ "epoch": 106.0,
1720
+ "eval_accuracy": 0.32017844886753605,
1721
+ "eval_f1_macro": 0.7060565909099816,
1722
+ "eval_f1_micro": 0.8232792762746488,
1723
+ "eval_loss": 0.11903885006904602,
1724
+ "eval_roc_auc": 0.8834646985106276,
1725
+ "eval_runtime": 702.4127,
1726
+ "eval_samples_per_second": 4.149,
1727
+ "eval_steps_per_second": 0.131,
1728
+ "learning_rate": 1e-05,
1729
+ "step": 28726
1730
+ },
1731
+ {
1732
+ "epoch": 107.0,
1733
+ "eval_accuracy": 0.3164035689773507,
1734
+ "eval_f1_macro": 0.7038085792556672,
1735
+ "eval_f1_micro": 0.822364833689862,
1736
+ "eval_loss": 0.1193847730755806,
1737
+ "eval_roc_auc": 0.8811169591371075,
1738
+ "eval_runtime": 704.7837,
1739
+ "eval_samples_per_second": 4.135,
1740
+ "eval_steps_per_second": 0.131,
1741
+ "learning_rate": 1e-05,
1742
+ "step": 28997
1743
+ },
1744
+ {
1745
+ "epoch": 107.01,
1746
+ "learning_rate": 1e-05,
1747
+ "loss": 0.1194,
1748
+ "step": 29000
1749
+ },
1750
+ {
1751
+ "epoch": 108.0,
1752
+ "eval_accuracy": 0.3191489361702128,
1753
+ "eval_f1_macro": 0.7110378528873161,
1754
+ "eval_f1_micro": 0.8232445520581114,
1755
+ "eval_loss": 0.11912781000137329,
1756
+ "eval_roc_auc": 0.883028832515277,
1757
+ "eval_runtime": 712.626,
1758
+ "eval_samples_per_second": 4.089,
1759
+ "eval_steps_per_second": 0.129,
1760
+ "learning_rate": 1e-05,
1761
+ "step": 29268
1762
+ },
1763
+ {
1764
+ "epoch": 108.86,
1765
+ "learning_rate": 1e-05,
1766
+ "loss": 0.1187,
1767
+ "step": 29500
1768
+ },
1769
+ {
1770
+ "epoch": 109.0,
1771
+ "eval_accuracy": 0.317433081674674,
1772
+ "eval_f1_macro": 0.7101233565606322,
1773
+ "eval_f1_micro": 0.8229869639937963,
1774
+ "eval_loss": 0.11885793507099152,
1775
+ "eval_roc_auc": 0.8816620950806512,
1776
+ "eval_runtime": 704.8781,
1777
+ "eval_samples_per_second": 4.134,
1778
+ "eval_steps_per_second": 0.131,
1779
+ "learning_rate": 1e-05,
1780
+ "step": 29539
1781
+ },
1782
+ {
1783
+ "epoch": 110.0,
1784
+ "eval_accuracy": 0.316060398078243,
1785
+ "eval_f1_macro": 0.7043927351077296,
1786
+ "eval_f1_micro": 0.8223593100843511,
1787
+ "eval_loss": 0.11920594424009323,
1788
+ "eval_roc_auc": 0.8809626700624849,
1789
+ "eval_runtime": 698.225,
1790
+ "eval_samples_per_second": 4.173,
1791
+ "eval_steps_per_second": 0.132,
1792
+ "learning_rate": 1e-05,
1793
+ "step": 29810
1794
+ },
1795
+ {
1796
+ "epoch": 110.7,
1797
+ "learning_rate": 1e-05,
1798
+ "loss": 0.1185,
1799
+ "step": 30000
1800
+ },
1801
+ {
1802
+ "epoch": 111.0,
1803
+ "eval_accuracy": 0.317433081674674,
1804
+ "eval_f1_macro": 0.7082791431366098,
1805
+ "eval_f1_micro": 0.8226405643444505,
1806
+ "eval_loss": 0.1191883385181427,
1807
+ "eval_roc_auc": 0.8827318446199321,
1808
+ "eval_runtime": 708.9619,
1809
+ "eval_samples_per_second": 4.11,
1810
+ "eval_steps_per_second": 0.13,
1811
+ "learning_rate": 1e-05,
1812
+ "step": 30081
1813
+ },
1814
+ {
1815
+ "epoch": 112.0,
1816
+ "eval_accuracy": 0.3205216197666438,
1817
+ "eval_f1_macro": 0.7093446393354451,
1818
+ "eval_f1_micro": 0.8239410221167062,
1819
+ "eval_loss": 0.11904006451368332,
1820
+ "eval_roc_auc": 0.8841101210717583,
1821
+ "eval_runtime": 709.1728,
1822
+ "eval_samples_per_second": 4.109,
1823
+ "eval_steps_per_second": 0.13,
1824
+ "learning_rate": 1e-05,
1825
+ "step": 30352
1826
+ },
1827
+ {
1828
+ "epoch": 112.55,
1829
+ "learning_rate": 1e-05,
1830
+ "loss": 0.119,
1831
+ "step": 30500
1832
+ },
1833
+ {
1834
+ "epoch": 113.0,
1835
+ "eval_accuracy": 0.31708991077556625,
1836
+ "eval_f1_macro": 0.708022770004008,
1837
+ "eval_f1_micro": 0.8232849960526862,
1838
+ "eval_loss": 0.11945341527462006,
1839
+ "eval_roc_auc": 0.8844778642879284,
1840
+ "eval_runtime": 708.0786,
1841
+ "eval_samples_per_second": 4.115,
1842
+ "eval_steps_per_second": 0.13,
1843
+ "learning_rate": 1e-05,
1844
+ "step": 30623
1845
+ },
1846
+ {
1847
+ "epoch": 114.0,
1848
+ "eval_accuracy": 0.3181194234728895,
1849
+ "eval_f1_macro": 0.7061751648131201,
1850
+ "eval_f1_micro": 0.82203889216264,
1851
+ "eval_loss": 0.11900585889816284,
1852
+ "eval_roc_auc": 0.8798916554890817,
1853
+ "eval_runtime": 701.8292,
1854
+ "eval_samples_per_second": 4.152,
1855
+ "eval_steps_per_second": 0.131,
1856
+ "learning_rate": 1e-05,
1857
+ "step": 30894
1858
+ },
1859
+ {
1860
+ "epoch": 114.39,
1861
+ "learning_rate": 1e-05,
1862
+ "loss": 0.1182,
1863
+ "step": 31000
1864
+ },
1865
+ {
1866
+ "epoch": 115.0,
1867
+ "eval_accuracy": 0.317433081674674,
1868
+ "eval_f1_macro": 0.7081036238174518,
1869
+ "eval_f1_micro": 0.8228996779994145,
1870
+ "eval_loss": 0.11918609589338303,
1871
+ "eval_roc_auc": 0.8823203985481548,
1872
+ "eval_runtime": 711.6073,
1873
+ "eval_samples_per_second": 4.095,
1874
+ "eval_steps_per_second": 0.129,
1875
+ "learning_rate": 1e-05,
1876
+ "step": 31165
1877
+ },
1878
+ {
1879
+ "epoch": 116.0,
1880
+ "eval_accuracy": 0.32498284145504464,
1881
+ "eval_f1_macro": 0.7128134624321966,
1882
+ "eval_f1_micro": 0.8255601659751037,
1883
+ "eval_loss": 0.11899947375059128,
1884
+ "eval_roc_auc": 0.8861634026181779,
1885
+ "eval_runtime": 692.0205,
1886
+ "eval_samples_per_second": 4.211,
1887
+ "eval_steps_per_second": 0.133,
1888
+ "learning_rate": 1.0000000000000002e-06,
1889
+ "step": 31436
1890
+ },
1891
+ {
1892
+ "epoch": 116.24,
1893
+ "learning_rate": 1.0000000000000002e-06,
1894
+ "loss": 0.1191,
1895
+ "step": 31500
1896
+ },
1897
+ {
1898
+ "epoch": 117.0,
1899
+ "eval_accuracy": 0.31708991077556625,
1900
+ "eval_f1_macro": 0.7103576429034303,
1901
+ "eval_f1_micro": 0.8231059020510674,
1902
+ "eval_loss": 0.11870068311691284,
1903
+ "eval_roc_auc": 0.8821442452365535,
1904
+ "eval_runtime": 704.8278,
1905
+ "eval_samples_per_second": 4.134,
1906
+ "eval_steps_per_second": 0.131,
1907
+ "learning_rate": 1.0000000000000002e-06,
1908
+ "step": 31707
1909
+ },
1910
+ {
1911
+ "epoch": 118.0,
1912
+ "eval_accuracy": 0.31983527796842826,
1913
+ "eval_f1_macro": 0.7061058223171173,
1914
+ "eval_f1_micro": 0.8235982284616027,
1915
+ "eval_loss": 0.11885705590248108,
1916
+ "eval_roc_auc": 0.8829722487532972,
1917
+ "eval_runtime": 702.381,
1918
+ "eval_samples_per_second": 4.149,
1919
+ "eval_steps_per_second": 0.131,
1920
+ "learning_rate": 1.0000000000000002e-06,
1921
+ "step": 31978
1922
+ },
1923
+ {
1924
+ "epoch": 118.08,
1925
+ "learning_rate": 1.0000000000000002e-06,
1926
+ "loss": 0.1179,
1927
+ "step": 32000
1928
+ },
1929
+ {
1930
+ "epoch": 119.0,
1931
+ "eval_accuracy": 0.3181194234728895,
1932
+ "eval_f1_macro": 0.7080350612891082,
1933
+ "eval_f1_micro": 0.8233181609387398,
1934
+ "eval_loss": 0.11888550966978073,
1935
+ "eval_roc_auc": 0.882981257702253,
1936
+ "eval_runtime": 695.5592,
1937
+ "eval_samples_per_second": 4.189,
1938
+ "eval_steps_per_second": 0.132,
1939
+ "learning_rate": 1.0000000000000002e-06,
1940
+ "step": 32249
1941
+ },
1942
+ {
1943
+ "epoch": 119.93,
1944
+ "learning_rate": 1.0000000000000002e-06,
1945
+ "loss": 0.1176,
1946
+ "step": 32500
1947
+ },
1948
+ {
1949
+ "epoch": 120.0,
1950
+ "eval_accuracy": 0.31846259437199725,
1951
+ "eval_f1_macro": 0.710061474614843,
1952
+ "eval_f1_micro": 0.8238579526508836,
1953
+ "eval_loss": 0.11897823214530945,
1954
+ "eval_roc_auc": 0.883849117735537,
1955
+ "eval_runtime": 696.3539,
1956
+ "eval_samples_per_second": 4.185,
1957
+ "eval_steps_per_second": 0.132,
1958
+ "learning_rate": 1.0000000000000002e-06,
1959
+ "step": 32520
1960
+ },
1961
+ {
1962
+ "epoch": 121.0,
1963
+ "eval_accuracy": 0.3208647906657515,
1964
+ "eval_f1_macro": 0.7127667221711917,
1965
+ "eval_f1_micro": 0.8254309453929147,
1966
+ "eval_loss": 0.11947356164455414,
1967
+ "eval_roc_auc": 0.8872447167930093,
1968
+ "eval_runtime": 693.3862,
1969
+ "eval_samples_per_second": 4.203,
1970
+ "eval_steps_per_second": 0.133,
1971
+ "learning_rate": 1.0000000000000002e-06,
1972
+ "step": 32791
1973
+ },
1974
+ {
1975
+ "epoch": 121.77,
1976
+ "learning_rate": 1.0000000000000002e-06,
1977
+ "loss": 0.1175,
1978
+ "step": 33000
1979
+ },
1980
+ {
1981
+ "epoch": 122.0,
1982
+ "eval_accuracy": 0.31537405628002746,
1983
+ "eval_f1_macro": 0.7048262553703367,
1984
+ "eval_f1_micro": 0.8222669349429913,
1985
+ "eval_loss": 0.11920282989740372,
1986
+ "eval_roc_auc": 0.8812519652796962,
1987
+ "eval_runtime": 684.1671,
1988
+ "eval_samples_per_second": 4.259,
1989
+ "eval_steps_per_second": 0.134,
1990
+ "learning_rate": 1.0000000000000002e-06,
1991
+ "step": 33062
1992
+ },
1993
+ {
1994
+ "epoch": 123.0,
1995
+ "eval_accuracy": 0.3212079615648593,
1996
+ "eval_f1_macro": 0.7154256602927447,
1997
+ "eval_f1_micro": 0.8255195344970907,
1998
+ "eval_loss": 0.1192421168088913,
1999
+ "eval_roc_auc": 0.8856336776492516,
2000
+ "eval_runtime": 707.8447,
2001
+ "eval_samples_per_second": 4.117,
2002
+ "eval_steps_per_second": 0.13,
2003
+ "learning_rate": 1.0000000000000002e-06,
2004
+ "step": 33333
2005
+ },
2006
+ {
2007
+ "epoch": 123.62,
2008
+ "learning_rate": 1.0000000000000002e-07,
2009
+ "loss": 0.1176,
2010
+ "step": 33500
2011
+ },
2012
+ {
2013
+ "epoch": 124.0,
2014
+ "eval_accuracy": 0.3208647906657515,
2015
+ "eval_f1_macro": 0.7109465417900758,
2016
+ "eval_f1_micro": 0.8238532110091744,
2017
+ "eval_loss": 0.11888780444860458,
2018
+ "eval_roc_auc": 0.8836948286609143,
2019
+ "eval_runtime": 715.7035,
2020
+ "eval_samples_per_second": 4.072,
2021
+ "eval_steps_per_second": 0.129,
2022
+ "learning_rate": 1.0000000000000002e-07,
2023
+ "step": 33604
2024
+ },
2025
+ {
2026
+ "epoch": 125.0,
2027
+ "eval_accuracy": 0.3225806451612903,
2028
+ "eval_f1_macro": 0.7101964285625825,
2029
+ "eval_f1_micro": 0.8251958006998834,
2030
+ "eval_loss": 0.11894452571868896,
2031
+ "eval_roc_auc": 0.8847439533789889,
2032
+ "eval_runtime": 706.7139,
2033
+ "eval_samples_per_second": 4.123,
2034
+ "eval_steps_per_second": 0.13,
2035
+ "learning_rate": 1.0000000000000002e-07,
2036
+ "step": 33875
2037
+ },
2038
+ {
2039
+ "epoch": 125.46,
2040
+ "learning_rate": 1.0000000000000002e-07,
2041
+ "loss": 0.1179,
2042
+ "step": 34000
2043
+ },
2044
+ {
2045
+ "epoch": 126.0,
2046
+ "eval_accuracy": 0.3164035689773507,
2047
+ "eval_f1_macro": 0.702507166276965,
2048
+ "eval_f1_micro": 0.8206187305066173,
2049
+ "eval_loss": 0.11890433728694916,
2050
+ "eval_roc_auc": 0.8787088152808908,
2051
+ "eval_runtime": 705.4083,
2052
+ "eval_samples_per_second": 4.131,
2053
+ "eval_steps_per_second": 0.13,
2054
+ "learning_rate": 1.0000000000000002e-07,
2055
+ "step": 34146
2056
+ },
2057
+ {
2058
+ "epoch": 127.0,
2059
+ "eval_accuracy": 0.32155113246396705,
2060
+ "eval_f1_macro": 0.7103986965520878,
2061
+ "eval_f1_micro": 0.8245357813477989,
2062
+ "eval_loss": 0.11901579052209854,
2063
+ "eval_roc_auc": 0.8838838116421884,
2064
+ "eval_runtime": 704.1405,
2065
+ "eval_samples_per_second": 4.138,
2066
+ "eval_steps_per_second": 0.131,
2067
+ "learning_rate": 1.0000000000000002e-07,
2068
+ "step": 34417
2069
+ },
2070
+ {
2071
+ "epoch": 127.0,
2072
+ "learning_rate": 1.0000000000000002e-07,
2073
+ "step": 34417,
2074
+ "total_flos": 1.6323301632319567e+21,
2075
+ "train_loss": 0.0713793940258306,
2076
+ "train_runtime": 211000.3165,
2077
+ "train_samples_per_second": 6.162,
2078
+ "train_steps_per_second": 0.193
2079
+ }
2080
+ ],
2081
+ "logging_steps": 500,
2082
+ "max_steps": 40650,
2083
+ "num_input_tokens_seen": 0,
2084
+ "num_train_epochs": 150,
2085
+ "save_steps": 500,
2086
+ "total_flos": 1.6323301632319567e+21,
2087
+ "train_batch_size": 32,
2088
+ "trial_name": null,
2089
+ "trial_params": null
2090
+ }