Hugging Face's logo

faridkarimli
/
SWIN_finetuned_constant

Image Classification
PyTorch
swinv2
vision
Generated from Trainer
Model card Files Community
faridkarimli commited on
Commit
d3be7c3
•
1 Parent(s): f6d8597

Training in progress, epoch 42

Browse files
Files changed (9) hide show
  1. README.md +131 -0
  2. all_results.json +12 -0
  3. config.json +0 -0
  4. eval_results.json +8 -0
  5. preprocessor_config.json +22 -0
  6. pytorch_model.bin +3 -0
  7. train_results.json +7 -0
  8. trainer_state.json +1756 -0
  9. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,131 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: microsoft/swinv2-base-patch4-window12-192-22k
4
+ tags:
5
+ - image-classification
6
+ - vision
7
+ - generated_from_trainer
8
+ metrics:
9
+ - accuracy
10
+ model-index:
11
+ - name: kaggle22
12
+ results: []
13
+ ---
14
+
15
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
16
+ should probably proofread and complete it, then remove this comment. -->
17
+
18
+ # kaggle22
19
+
20
+ This model is a fine-tuned version of [microsoft/swinv2-base-patch4-window12-192-22k](https://huggingface.co/microsoft/swinv2-base-patch4-window12-192-22k) on an unknown dataset.
21
+ It achieves the following results on the evaluation set:
22
+ - Loss: 2.1484
23
+ - Accuracy: 0.7132
24
+
25
+ ## Model description
26
+
27
+ More information needed
28
+
29
+ ## Intended uses & limitations
30
+
31
+ More information needed
32
+
33
+ ## Training and evaluation data
34
+
35
+ More information needed
36
+
37
+ ## Training procedure
38
+
39
+ ### Training hyperparameters
40
+
41
+ The following hyperparameters were used during training:
42
+ - learning_rate: 0.005
43
+ - train_batch_size: 512
44
+ - eval_batch_size: 16
45
+ - seed: 42
46
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
47
+ - lr_scheduler_type: cosine
48
+ - num_epochs: 70.0
49
+
50
+ ### Training results
51
+
52
+ | Training Loss | Epoch | Step | Accuracy | Validation Loss |
53
+ |:-------------:|:-----:|:-----:|:--------:|:---------------:|
54
+ | 9.5189 | 1.0 | 1313 | 0.0002 | 9.4097 |
55
+ | 9.1635 | 2.0 | 2626 | 0.0006 | 9.0418 |
56
+ | 8.3432 | 3.0 | 3939 | 0.0065 | 7.8454 |
57
+ | 6.9913 | 4.0 | 5252 | 0.0489 | 6.3118 |
58
+ | 5.5048 | 5.0 | 6565 | 0.1423 | 4.9493 |
59
+ | 4.6895 | 6.0 | 7878 | 0.2450 | 3.9601 |
60
+ | 3.8881 | 7.0 | 9191 | 0.3136 | 3.4186 |
61
+ | 3.391 | 8.0 | 10504 | 0.3766 | 2.9798 |
62
+ | 3.0887 | 9.0 | 11817 | 0.4221 | 2.7054 |
63
+ | 2.7935 | 10.0 | 13130 | 0.4552 | 2.5013 |
64
+ | 2.5629 | 11.0 | 14443 | 0.4804 | 2.3581 |
65
+ | 2.3777 | 12.0 | 15756 | 0.4809 | 2.3543 |
66
+ | 2.2264 | 13.0 | 17069 | 0.5179 | 2.1632 |
67
+ | 2.0932 | 14.0 | 18382 | 0.5219 | 2.1362 |
68
+ | 1.9667 | 15.0 | 19695 | 0.5591 | 1.9567 |
69
+ | 1.8788 | 16.0 | 21008 | 0.5610 | 1.9347 |
70
+ | 1.7705 | 17.0 | 22321 | 0.5684 | 1.9483 |
71
+ | 1.7089 | 18.0 | 23634 | 0.5791 | 1.8928 |
72
+ | 1.6068 | 19.0 | 24947 | 0.5855 | 1.8435 |
73
+ | 1.5572 | 20.0 | 26260 | 0.5880 | 1.8408 |
74
+ | 1.4938 | 21.0 | 27573 | 0.6110 | 1.7413 |
75
+ | 1.4182 | 22.0 | 28886 | 0.6155 | 1.7196 |
76
+ | 1.3784 | 23.0 | 30199 | 0.6238 | 1.7105 |
77
+ | 1.3578 | 24.0 | 31512 | 0.6176 | 1.7759 |
78
+ | 1.2763 | 25.0 | 32825 | 0.6219 | 1.7365 |
79
+ | 1.2484 | 26.0 | 34138 | 0.6199 | 1.7483 |
80
+ | 1.1936 | 27.0 | 35451 | 0.6314 | 1.7003 |
81
+ | 1.1499 | 28.0 | 36764 | 0.6247 | 1.7399 |
82
+ | 1.1418 | 29.0 | 38077 | 0.6317 | 1.7091 |
83
+ | 1.0895 | 30.0 | 39390 | 0.6383 | 1.7166 |
84
+ | 1.0706 | 31.0 | 40703 | 0.6374 | 1.7384 |
85
+ | 1.0541 | 32.0 | 42016 | 0.6409 | 1.7336 |
86
+ | 1.0013 | 33.0 | 43329 | 0.6451 | 1.7185 |
87
+ | 0.9811 | 34.0 | 44642 | 0.6479 | 1.7246 |
88
+ | 0.9447 | 35.0 | 45955 | 0.6540 | 1.7245 |
89
+ | 0.6587 | 36.0 | 47268 | 0.7019 | 1.5849 |
90
+ | 0.6044 | 37.0 | 48581 | 0.7062 | 1.6146 |
91
+ | 0.572 | 38.0 | 49894 | 0.7081 | 1.6583 |
92
+ | 0.545 | 39.0 | 51207 | 0.7087 | 1.6993 |
93
+ | 0.5341 | 40.0 | 52520 | 0.7106 | 1.7078 |
94
+ | 0.5284 | 41.0 | 53833 | 0.7105 | 1.7241 |
95
+ | 0.5186 | 42.0 | 55146 | 0.7112 | 1.7408 |
96
+ | 0.506 | 43.0 | 56459 | 0.7106 | 1.7487 |
97
+ | 0.5043 | 44.0 | 57772 | 0.7109 | 1.7547 |
98
+ | 0.5094 | 45.0 | 59085 | 0.7111 | 1.7536 |
99
+ | 0.5547 | 46.0 | 60398 | 0.7069 | 1.7074 |
100
+ | 0.5391 | 47.0 | 61711 | 0.7090 | 1.7401 |
101
+ | 0.5253 | 48.0 | 63024 | 0.7093 | 1.7770 |
102
+ | 0.5066 | 49.0 | 64337 | 0.7102 | 1.8135 |
103
+ | 0.495 | 50.0 | 65650 | 0.7110 | 1.8452 |
104
+ | 0.4813 | 51.0 | 66963 | 0.7107 | 1.8846 |
105
+ | 0.4704 | 52.0 | 68276 | 0.7124 | 1.8989 |
106
+ | 0.4689 | 53.0 | 69589 | 0.7132 | 1.9311 |
107
+ | 0.4611 | 54.0 | 70902 | 0.7131 | 1.9354 |
108
+ | 0.4547 | 55.0 | 72215 | 0.7133 | 1.9741 |
109
+ | 0.4481 | 56.0 | 73528 | 0.7131 | 1.9899 |
110
+ | 0.4709 | 57.0 | 74841 | 1.9412 | 0.7104 |
111
+ | 0.4647 | 58.0 | 76154 | 1.9707 | 0.7098 |
112
+ | 0.4566 | 59.0 | 77467 | 2.0151 | 0.7116 |
113
+ | 0.4511 | 60.0 | 78780 | 2.0363 | 0.7114 |
114
+ | 0.4423 | 61.0 | 80093 | 2.0710 | 0.7112 |
115
+ | 0.4356 | 62.0 | 81406 | 2.0611 | 0.7116 |
116
+ | 0.4272 | 63.0 | 82719 | 2.0891 | 0.7118 |
117
+ | 0.4254 | 64.0 | 84032 | 2.0879 | 0.7124 |
118
+ | 0.4221 | 65.0 | 85345 | 2.1167 | 0.7131 |
119
+ | 0.4189 | 66.0 | 86658 | 2.1363 | 0.7129 |
120
+ | 0.4219 | 67.0 | 87971 | 2.1355 | 0.7130 |
121
+ | 0.4149 | 68.0 | 89284 | 2.1466 | 0.7132 |
122
+ | 0.4125 | 69.0 | 90597 | 2.1478 | 0.7131 |
123
+ | 0.4162 | 70.0 | 91910 | 2.1484 | 0.7132 |
124
+
125
+
126
+ ### Framework versions
127
+
128
+ - Transformers 4.33.3
129
+ - Pytorch 2.1.2
130
+ - Datasets 2.16.1
131
+ - Tokenizers 0.13.3
all_results.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 70.0,
3
+ "eval_accuracy": 0.7132267571670984,
4
+ "eval_loss": 2.1484131813049316,
5
+ "eval_runtime": 2588.9498,
6
+ "eval_samples_per_second": 64.874,
7
+ "eval_steps_per_second": 4.055,
8
+ "train_loss": 0.08673005339085323,
9
+ "train_runtime": 125919.3774,
10
+ "train_samples_per_second": 373.471,
11
+ "train_steps_per_second": 0.73
12
+ }
config.json ADDED
The diff for this file is too large to render. See raw diff
 
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 70.0,
3
+ "eval_accuracy": 0.7132267571670984,
4
+ "eval_loss": 2.1484131813049316,
5
+ "eval_runtime": 2588.9498,
6
+ "eval_samples_per_second": 64.874,
7
+ "eval_steps_per_second": 4.055
8
+ }
preprocessor_config.json ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_normalize": true,
3
+ "do_rescale": true,
4
+ "do_resize": true,
5
+ "image_mean": [
6
+ 0.485,
7
+ 0.456,
8
+ 0.406
9
+ ],
10
+ "image_processor_type": "ViTImageProcessor",
11
+ "image_std": [
12
+ 0.229,
13
+ 0.224,
14
+ 0.225
15
+ ],
16
+ "resample": 3,
17
+ "rescale_factor": 0.00392156862745098,
18
+ "size": {
19
+ "height": 192,
20
+ "width": 192
21
+ }
22
+ }
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:127f2c4261b6b0b38d4c3bfb81c74dec23f5a9fdb99dc6e714bae07d79d78940
3
+ size 411298414
train_results.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 70.0,
3
+ "train_loss": 0.08673005339085323,
4
+ "train_runtime": 125919.3774,
5
+ "train_samples_per_second": 373.471,
6
+ "train_steps_per_second": 0.73
7
+ }
trainer_state.json ADDED
@@ -0,0 +1,1756 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 70.0,
5
+ "eval_steps": 500,
6
+ "global_step": 91910,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.38,
13
+ "learning_rate": 0.004999955274106163,
14
+ "loss": 9.6382,
15
+ "step": 500
16
+ },
17
+ {
18
+ "epoch": 0.76,
19
+ "learning_rate": 0.004999821098024974,
20
+ "loss": 9.5189,
21
+ "step": 1000
22
+ },
23
+ {
24
+ "epoch": 1.0,
25
+ "eval_accuracy": 0.00022625107915810782,
26
+ "eval_loss": 9.40969181060791,
27
+ "eval_runtime": 6168.4952,
28
+ "eval_samples_per_second": 27.228,
29
+ "eval_steps_per_second": 1.702,
30
+ "step": 1313
31
+ },
32
+ {
33
+ "epoch": 1.14,
34
+ "learning_rate": 0.0049995974765573515,
35
+ "loss": 9.363,
36
+ "step": 1500
37
+ },
38
+ {
39
+ "epoch": 1.52,
40
+ "learning_rate": 0.00499928441770463,
41
+ "loss": 9.2724,
42
+ "step": 2000
43
+ },
44
+ {
45
+ "epoch": 1.9,
46
+ "learning_rate": 0.00499888193266828,
47
+ "loss": 9.1635,
48
+ "step": 2500
49
+ },
50
+ {
51
+ "epoch": 2.0,
52
+ "eval_accuracy": 0.0006430293828704117,
53
+ "eval_loss": 9.041797637939453,
54
+ "eval_runtime": 4719.6922,
55
+ "eval_samples_per_second": 35.586,
56
+ "eval_steps_per_second": 2.224,
57
+ "step": 2626
58
+ },
59
+ {
60
+ "epoch": 2.28,
61
+ "learning_rate": 0.004998390035849502,
62
+ "loss": 8.8459,
63
+ "step": 3000
64
+ },
65
+ {
66
+ "epoch": 2.67,
67
+ "learning_rate": 0.0049978087448487194,
68
+ "loss": 8.3432,
69
+ "step": 3500
70
+ },
71
+ {
72
+ "epoch": 3.0,
73
+ "eval_accuracy": 0.006466017683308029,
74
+ "eval_loss": 7.845377445220947,
75
+ "eval_runtime": 2624.5595,
76
+ "eval_samples_per_second": 63.994,
77
+ "eval_steps_per_second": 4.0,
78
+ "step": 3939
79
+ },
80
+ {
81
+ "epoch": 3.05,
82
+ "learning_rate": 0.004997138080464938,
83
+ "loss": 7.8952,
84
+ "step": 4000
85
+ },
86
+ {
87
+ "epoch": 3.43,
88
+ "learning_rate": 0.004996378066695007,
89
+ "loss": 7.4467,
90
+ "step": 4500
91
+ },
92
+ {
93
+ "epoch": 3.81,
94
+ "learning_rate": 0.004995528730732766,
95
+ "loss": 6.9913,
96
+ "step": 5000
97
+ },
98
+ {
99
+ "epoch": 4.0,
100
+ "eval_accuracy": 0.04887023309815129,
101
+ "eval_loss": 6.311844825744629,
102
+ "eval_runtime": 2630.4579,
103
+ "eval_samples_per_second": 63.85,
104
+ "eval_steps_per_second": 3.991,
105
+ "step": 5252
106
+ },
107
+ {
108
+ "epoch": 4.19,
109
+ "learning_rate": 0.0049945901029680604,
110
+ "loss": 6.4357,
111
+ "step": 5500
112
+ },
113
+ {
114
+ "epoch": 4.57,
115
+ "learning_rate": 0.004993562216985665,
116
+ "loss": 5.9619,
117
+ "step": 6000
118
+ },
119
+ {
120
+ "epoch": 4.95,
121
+ "learning_rate": 0.004992445109564073,
122
+ "loss": 5.5048,
123
+ "step": 6500
124
+ },
125
+ {
126
+ "epoch": 5.0,
127
+ "eval_accuracy": 0.1423000208389152,
128
+ "eval_loss": 4.949326515197754,
129
+ "eval_runtime": 2650.8854,
130
+ "eval_samples_per_second": 63.358,
131
+ "eval_steps_per_second": 3.96,
132
+ "step": 6565
133
+ },
134
+ {
135
+ "epoch": 5.33,
136
+ "learning_rate": 0.004991238820674189,
137
+ "loss": 4.982,
138
+ "step": 7000
139
+ },
140
+ {
141
+ "epoch": 5.71,
142
+ "learning_rate": 0.00498994339347789,
143
+ "loss": 4.6895,
144
+ "step": 7500
145
+ },
146
+ {
147
+ "epoch": 6.0,
148
+ "eval_accuracy": 0.24502396475246346,
149
+ "eval_loss": 3.960129499435425,
150
+ "eval_runtime": 2617.6761,
151
+ "eval_samples_per_second": 64.162,
152
+ "eval_steps_per_second": 4.01,
153
+ "step": 7878
154
+ },
155
+ {
156
+ "epoch": 6.09,
157
+ "learning_rate": 0.004988558874326489,
158
+ "loss": 4.3479,
159
+ "step": 8000
160
+ },
161
+ {
162
+ "epoch": 6.47,
163
+ "learning_rate": 0.004987085312759071,
164
+ "loss": 3.9972,
165
+ "step": 8500
166
+ },
167
+ {
168
+ "epoch": 6.85,
169
+ "learning_rate": 0.004985522761500722,
170
+ "loss": 3.8881,
171
+ "step": 9000
172
+ },
173
+ {
174
+ "epoch": 7.0,
175
+ "eval_accuracy": 0.31358994968890475,
176
+ "eval_loss": 3.4186415672302246,
177
+ "eval_runtime": 2616.4405,
178
+ "eval_samples_per_second": 64.192,
179
+ "eval_steps_per_second": 4.012,
180
+ "step": 9191
181
+ },
182
+ {
183
+ "epoch": 7.24,
184
+ "learning_rate": 0.004983871276460643,
185
+ "loss": 3.6025,
186
+ "step": 9500
187
+ },
188
+ {
189
+ "epoch": 7.62,
190
+ "learning_rate": 0.0049821309167301505,
191
+ "loss": 3.4698,
192
+ "step": 10000
193
+ },
194
+ {
195
+ "epoch": 8.0,
196
+ "learning_rate": 0.004980301744580561,
197
+ "loss": 3.391,
198
+ "step": 10500
199
+ },
200
+ {
201
+ "epoch": 8.0,
202
+ "eval_accuracy": 0.37655919740406657,
203
+ "eval_loss": 2.9798192977905273,
204
+ "eval_runtime": 2644.3802,
205
+ "eval_samples_per_second": 63.514,
206
+ "eval_steps_per_second": 3.97,
207
+ "step": 10504
208
+ },
209
+ {
210
+ "epoch": 8.38,
211
+ "learning_rate": 0.004978383825460959,
212
+ "loss": 3.1113,
213
+ "step": 11000
214
+ },
215
+ {
216
+ "epoch": 8.76,
217
+ "learning_rate": 0.004976377227995865,
218
+ "loss": 3.0887,
219
+ "step": 11500
220
+ },
221
+ {
222
+ "epoch": 9.0,
223
+ "eval_accuracy": 0.422124973951356,
224
+ "eval_loss": 2.705430030822754,
225
+ "eval_runtime": 2622.7588,
226
+ "eval_samples_per_second": 64.038,
227
+ "eval_steps_per_second": 4.003,
228
+ "step": 11817
229
+ },
230
+ {
231
+ "epoch": 9.14,
232
+ "learning_rate": 0.00497428202398277,
233
+ "loss": 2.925,
234
+ "step": 12000
235
+ },
236
+ {
237
+ "epoch": 9.52,
238
+ "learning_rate": 0.004972098288389572,
239
+ "loss": 2.7977,
240
+ "step": 12500
241
+ },
242
+ {
243
+ "epoch": 9.9,
244
+ "learning_rate": 0.004969826099351891,
245
+ "loss": 2.7935,
246
+ "step": 13000
247
+ },
248
+ {
249
+ "epoch": 10.0,
250
+ "eval_accuracy": 0.45515167753267244,
251
+ "eval_loss": 2.501314401626587,
252
+ "eval_runtime": 2679.1466,
253
+ "eval_samples_per_second": 62.69,
254
+ "eval_steps_per_second": 3.918,
255
+ "step": 13130
256
+ },
257
+ {
258
+ "epoch": 10.28,
259
+ "learning_rate": 0.004967465538170278,
260
+ "loss": 2.5817,
261
+ "step": 13500
262
+ },
263
+ {
264
+ "epoch": 10.66,
265
+ "learning_rate": 0.004965016689307298,
266
+ "loss": 2.5629,
267
+ "step": 14000
268
+ },
269
+ {
270
+ "epoch": 11.0,
271
+ "eval_accuracy": 0.4803846268345688,
272
+ "eval_loss": 2.3581297397613525,
273
+ "eval_runtime": 2615.9499,
274
+ "eval_samples_per_second": 64.204,
275
+ "eval_steps_per_second": 4.013,
276
+ "step": 14443
277
+ },
278
+ {
279
+ "epoch": 11.04,
280
+ "learning_rate": 0.004962479640384515,
281
+ "loss": 2.5549,
282
+ "step": 14500
283
+ },
284
+ {
285
+ "epoch": 11.42,
286
+ "learning_rate": 0.004959854482179353,
287
+ "loss": 2.3342,
288
+ "step": 15000
289
+ },
290
+ {
291
+ "epoch": 11.81,
292
+ "learning_rate": 0.004957141308621851,
293
+ "loss": 2.3777,
294
+ "step": 15500
295
+ },
296
+ {
297
+ "epoch": 12.0,
298
+ "eval_accuracy": 0.48086689887172157,
299
+ "eval_loss": 2.3543388843536377,
300
+ "eval_runtime": 2613.6118,
301
+ "eval_samples_per_second": 64.262,
302
+ "eval_steps_per_second": 4.017,
303
+ "step": 15756
304
+ },
305
+ {
306
+ "epoch": 12.19,
307
+ "learning_rate": 0.004954340216791298,
308
+ "loss": 2.1546,
309
+ "step": 16000
310
+ },
311
+ {
312
+ "epoch": 12.57,
313
+ "learning_rate": 0.004951451306912763,
314
+ "loss": 2.2162,
315
+ "step": 16500
316
+ },
317
+ {
318
+ "epoch": 12.95,
319
+ "learning_rate": 0.0049484746823535075,
320
+ "loss": 2.2264,
321
+ "step": 17000
322
+ },
323
+ {
324
+ "epoch": 13.0,
325
+ "eval_accuracy": 0.5178887201929088,
326
+ "eval_loss": 2.163175344467163,
327
+ "eval_runtime": 5970.7626,
328
+ "eval_samples_per_second": 28.13,
329
+ "eval_steps_per_second": 1.758,
330
+ "step": 17069
331
+ },
332
+ {
333
+ "epoch": 13.33,
334
+ "learning_rate": 0.004945410449619286,
335
+ "loss": 2.035,
336
+ "step": 17500
337
+ },
338
+ {
339
+ "epoch": 13.71,
340
+ "learning_rate": 0.0049422587183505385,
341
+ "loss": 2.0932,
342
+ "step": 18000
343
+ },
344
+ {
345
+ "epoch": 14.0,
346
+ "eval_accuracy": 0.5219076538358489,
347
+ "eval_loss": 2.136173725128174,
348
+ "eval_runtime": 2759.3303,
349
+ "eval_samples_per_second": 60.868,
350
+ "eval_steps_per_second": 3.805,
351
+ "step": 18382
352
+ },
353
+ {
354
+ "epoch": 14.09,
355
+ "learning_rate": 0.004939019601318461,
356
+ "loss": 2.0545,
357
+ "step": 18500
358
+ },
359
+ {
360
+ "epoch": 14.47,
361
+ "learning_rate": 0.004935693214420979,
362
+ "loss": 1.9224,
363
+ "step": 19000
364
+ },
365
+ {
366
+ "epoch": 14.85,
367
+ "learning_rate": 0.004932279676678594,
368
+ "loss": 1.9667,
369
+ "step": 19500
370
+ },
371
+ {
372
+ "epoch": 15.0,
373
+ "eval_accuracy": 0.5591200023815903,
374
+ "eval_loss": 1.956689715385437,
375
+ "eval_runtime": 2781.1993,
376
+ "eval_samples_per_second": 60.389,
377
+ "eval_steps_per_second": 3.775,
378
+ "step": 19695
379
+ },
380
+ {
381
+ "epoch": 15.23,
382
+ "learning_rate": 0.004928779110230126,
383
+ "loss": 1.8635,
384
+ "step": 20000
385
+ },
386
+ {
387
+ "epoch": 15.61,
388
+ "learning_rate": 0.004925191640328348,
389
+ "loss": 1.8396,
390
+ "step": 20500
391
+ },
392
+ {
393
+ "epoch": 15.99,
394
+ "learning_rate": 0.004921517395335496,
395
+ "loss": 1.8788,
396
+ "step": 21000
397
+ },
398
+ {
399
+ "epoch": 16.0,
400
+ "eval_accuracy": 0.5609538269179244,
401
+ "eval_loss": 1.9347097873687744,
402
+ "eval_runtime": 2696.6196,
403
+ "eval_samples_per_second": 62.284,
404
+ "eval_steps_per_second": 3.893,
405
+ "step": 21008
406
+ },
407
+ {
408
+ "epoch": 16.37,
409
+ "learning_rate": 0.004917756506718685,
410
+ "loss": 1.7225,
411
+ "step": 21500
412
+ },
413
+ {
414
+ "epoch": 16.76,
415
+ "learning_rate": 0.004913909109045198,
416
+ "loss": 1.7705,
417
+ "step": 22000
418
+ },
419
+ {
420
+ "epoch": 17.0,
421
+ "eval_accuracy": 0.5684439284332112,
422
+ "eval_loss": 1.948270559310913,
423
+ "eval_runtime": 2619.7155,
424
+ "eval_samples_per_second": 64.112,
425
+ "eval_steps_per_second": 4.007,
426
+ "step": 22321
427
+ },
428
+ {
429
+ "epoch": 17.14,
430
+ "learning_rate": 0.004909975339977676,
431
+ "loss": 1.7215,
432
+ "step": 22500
433
+ },
434
+ {
435
+ "epoch": 17.52,
436
+ "learning_rate": 0.004905955340269188,
437
+ "loss": 1.6354,
438
+ "step": 23000
439
+ },
440
+ {
441
+ "epoch": 17.9,
442
+ "learning_rate": 0.004901849253758198,
443
+ "loss": 1.7089,
444
+ "step": 23500
445
+ },
446
+ {
447
+ "epoch": 18.0,
448
+ "eval_accuracy": 0.5790598672263404,
449
+ "eval_loss": 1.8927983045578003,
450
+ "eval_runtime": 2625.6089,
451
+ "eval_samples_per_second": 63.968,
452
+ "eval_steps_per_second": 3.998,
453
+ "step": 23634
454
+ },
455
+ {
456
+ "epoch": 18.28,
457
+ "learning_rate": 0.004897657227363419,
458
+ "loss": 1.579,
459
+ "step": 24000
460
+ },
461
+ {
462
+ "epoch": 18.66,
463
+ "learning_rate": 0.004893379411078551,
464
+ "loss": 1.6068,
465
+ "step": 24500
466
+ },
467
+ {
468
+ "epoch": 19.0,
469
+ "eval_accuracy": 0.5855497008127177,
470
+ "eval_loss": 1.843549370765686,
471
+ "eval_runtime": 2690.4067,
472
+ "eval_samples_per_second": 62.427,
473
+ "eval_steps_per_second": 3.902,
474
+ "step": 24947
475
+ },
476
+ {
477
+ "epoch": 19.04,
478
+ "learning_rate": 0.004889015957966921,
479
+ "loss": 1.6159,
480
+ "step": 25000
481
+ },
482
+ {
483
+ "epoch": 19.42,
484
+ "learning_rate": 0.004884567024156001,
485
+ "loss": 1.4949,
486
+ "step": 25500
487
+ },
488
+ {
489
+ "epoch": 19.8,
490
+ "learning_rate": 0.004880032768831825,
491
+ "loss": 1.5572,
492
+ "step": 26000
493
+ },
494
+ {
495
+ "epoch": 20.0,
496
+ "eval_accuracy": 0.5879670149742491,
497
+ "eval_loss": 1.840835690498352,
498
+ "eval_runtime": 2628.5762,
499
+ "eval_samples_per_second": 63.896,
500
+ "eval_steps_per_second": 3.994,
501
+ "step": 26260
502
+ },
503
+ {
504
+ "epoch": 20.18,
505
+ "learning_rate": 0.004875413354233289,
506
+ "loss": 1.4928,
507
+ "step": 26500
508
+ },
509
+ {
510
+ "epoch": 20.56,
511
+ "learning_rate": 0.0048707089456463525,
512
+ "loss": 1.4466,
513
+ "step": 27000
514
+ },
515
+ {
516
+ "epoch": 20.94,
517
+ "learning_rate": 0.004865919711398117,
518
+ "loss": 1.4938,
519
+ "step": 27500
520
+ },
521
+ {
522
+ "epoch": 21.0,
523
+ "eval_accuracy": 0.6109672233634009,
524
+ "eval_loss": 1.7412927150726318,
525
+ "eval_runtime": 2614.3175,
526
+ "eval_samples_per_second": 64.244,
527
+ "eval_steps_per_second": 4.016,
528
+ "step": 27573
529
+ },
530
+ {
531
+ "epoch": 21.33,
532
+ "learning_rate": 0.0048610458228508105,
533
+ "loss": 1.3796,
534
+ "step": 28000
535
+ },
536
+ {
537
+ "epoch": 21.71,
538
+ "learning_rate": 0.004856087454395648,
539
+ "loss": 1.4182,
540
+ "step": 28500
541
+ },
542
+ {
543
+ "epoch": 22.0,
544
+ "eval_accuracy": 0.6154565210919591,
545
+ "eval_loss": 1.7196179628372192,
546
+ "eval_runtime": 2640.6742,
547
+ "eval_samples_per_second": 63.603,
548
+ "eval_steps_per_second": 3.975,
549
+ "step": 28886
550
+ },
551
+ {
552
+ "epoch": 22.09,
553
+ "learning_rate": 0.0048510447834466,
554
+ "loss": 1.4096,
555
+ "step": 29000
556
+ },
557
+ {
558
+ "epoch": 22.47,
559
+ "learning_rate": 0.0048459179904340385,
560
+ "loss": 1.3308,
561
+ "step": 29500
562
+ },
563
+ {
564
+ "epoch": 22.85,
565
+ "learning_rate": 0.004840707258798284,
566
+ "loss": 1.3784,
567
+ "step": 30000
568
+ },
569
+ {
570
+ "epoch": 23.0,
571
+ "eval_accuracy": 0.6238218570450418,
572
+ "eval_loss": 1.7105211019515991,
573
+ "eval_runtime": 2885.4428,
574
+ "eval_samples_per_second": 58.208,
575
+ "eval_steps_per_second": 3.638,
576
+ "step": 30199
577
+ },
578
+ {
579
+ "epoch": 23.23,
580
+ "learning_rate": 0.004835412774983039,
581
+ "loss": 1.3104,
582
+ "step": 30500
583
+ },
584
+ {
585
+ "epoch": 23.61,
586
+ "learning_rate": 0.004830034728428723,
587
+ "loss": 1.301,
588
+ "step": 31000
589
+ },
590
+ {
591
+ "epoch": 23.99,
592
+ "learning_rate": 0.004824573311565684,
593
+ "loss": 1.3578,
594
+ "step": 31500
595
+ },
596
+ {
597
+ "epoch": 24.0,
598
+ "eval_accuracy": 0.6176059063439612,
599
+ "eval_loss": 1.7758694887161255,
600
+ "eval_runtime": 2617.3025,
601
+ "eval_samples_per_second": 64.171,
602
+ "eval_steps_per_second": 4.011,
603
+ "step": 31512
604
+ },
605
+ {
606
+ "epoch": 24.37,
607
+ "learning_rate": 0.004819028719807326,
608
+ "loss": 1.2176,
609
+ "step": 32000
610
+ },
611
+ {
612
+ "epoch": 24.75,
613
+ "learning_rate": 0.004813401151543105,
614
+ "loss": 1.2763,
615
+ "step": 32500
616
+ },
617
+ {
618
+ "epoch": 25.0,
619
+ "eval_accuracy": 0.6218749069691286,
620
+ "eval_loss": 1.7364755868911743,
621
+ "eval_runtime": 5925.4842,
622
+ "eval_samples_per_second": 28.345,
623
+ "eval_steps_per_second": 1.772,
624
+ "step": 32825
625
+ },
626
+ {
627
+ "epoch": 25.13,
628
+ "learning_rate": 0.004807690808131437,
629
+ "loss": 1.2551,
630
+ "step": 33000
631
+ },
632
+ {
633
+ "epoch": 25.51,
634
+ "learning_rate": 0.004801897893892493,
635
+ "loss": 1.1861,
636
+ "step": 33500
637
+ },
638
+ {
639
+ "epoch": 25.89,
640
+ "learning_rate": 0.004796022616100888,
641
+ "loss": 1.2484,
642
+ "step": 34000
643
+ },
644
+ {
645
+ "epoch": 26.0,
646
+ "eval_accuracy": 0.6199220029174481,
647
+ "eval_loss": 1.7483400106430054,
648
+ "eval_runtime": 2684.7282,
649
+ "eval_samples_per_second": 62.559,
650
+ "eval_steps_per_second": 3.91,
651
+ "step": 34138
652
+ },
653
+ {
654
+ "epoch": 26.28,
655
+ "learning_rate": 0.004790065184978262,
656
+ "loss": 1.1726,
657
+ "step": 34500
658
+ },
659
+ {
660
+ "epoch": 26.66,
661
+ "learning_rate": 0.004784025813685761,
662
+ "loss": 1.1936,
663
+ "step": 35000
664
+ },
665
+ {
666
+ "epoch": 27.0,
667
+ "eval_accuracy": 0.6313953142210711,
668
+ "eval_loss": 1.700264573097229,
669
+ "eval_runtime": 2640.722,
670
+ "eval_samples_per_second": 63.602,
671
+ "eval_steps_per_second": 3.975,
672
+ "step": 35451
673
+ },
674
+ {
675
+ "epoch": 27.04,
676
+ "learning_rate": 0.004777904718316407,
677
+ "loss": 1.2231,
678
+ "step": 35500
679
+ },
680
+ {
681
+ "epoch": 27.42,
682
+ "learning_rate": 0.00477170211788737,
683
+ "loss": 1.1121,
684
+ "step": 36000
685
+ },
686
+ {
687
+ "epoch": 27.8,
688
+ "learning_rate": 0.00476541823433213,
689
+ "loss": 1.1499,
690
+ "step": 36500
691
+ },
692
+ {
693
+ "epoch": 28.0,
694
+ "eval_accuracy": 0.6247209073859069,
695
+ "eval_loss": 1.7398767471313477,
696
+ "eval_runtime": 2691.471,
697
+ "eval_samples_per_second": 62.403,
698
+ "eval_steps_per_second": 3.9,
699
+ "step": 36764
700
+ },
701
+ {
702
+ "epoch": 28.18,
703
+ "learning_rate": 0.004759053292492532,
704
+ "loss": 1.1329,
705
+ "step": 37000
706
+ },
707
+ {
708
+ "epoch": 28.56,
709
+ "learning_rate": 0.004752607520110748,
710
+ "loss": 1.1121,
711
+ "step": 37500
712
+ },
713
+ {
714
+ "epoch": 28.94,
715
+ "learning_rate": 0.004746081147821121,
716
+ "loss": 1.1418,
717
+ "step": 38000
718
+ },
719
+ {
720
+ "epoch": 29.0,
721
+ "eval_accuracy": 0.6316751510821351,
722
+ "eval_loss": 1.7091107368469238,
723
+ "eval_runtime": 2632.2426,
724
+ "eval_samples_per_second": 63.807,
725
+ "eval_steps_per_second": 3.988,
726
+ "step": 38077
727
+ },
728
+ {
729
+ "epoch": 29.32,
730
+ "learning_rate": 0.0047394744091419205,
731
+ "loss": 1.0493,
732
+ "step": 38500
733
+ },
734
+ {
735
+ "epoch": 29.7,
736
+ "learning_rate": 0.004732787540466979,
737
+ "loss": 1.0895,
738
+ "step": 39000
739
+ },
740
+ {
741
+ "epoch": 30.0,
742
+ "eval_accuracy": 0.6383078800869281,
743
+ "eval_loss": 1.7165930271148682,
744
+ "eval_runtime": 2645.0605,
745
+ "eval_samples_per_second": 63.498,
746
+ "eval_steps_per_second": 3.969,
747
+ "step": 39390
748
+ },
749
+ {
750
+ "epoch": 30.08,
751
+ "learning_rate": 0.004726020781057241,
752
+ "loss": 1.0905,
753
+ "step": 39500
754
+ },
755
+ {
756
+ "epoch": 30.46,
757
+ "learning_rate": 0.004719174373032195,
758
+ "loss": 1.0191,
759
+ "step": 40000
760
+ },
761
+ {
762
+ "epoch": 30.85,
763
+ "learning_rate": 0.0047122485613612175,
764
+ "loss": 1.0706,
765
+ "step": 40500
766
+ },
767
+ {
768
+ "epoch": 31.0,
769
+ "eval_accuracy": 0.6373611979399244,
770
+ "eval_loss": 1.7384090423583984,
771
+ "eval_runtime": 2658.4828,
772
+ "eval_samples_per_second": 63.177,
773
+ "eval_steps_per_second": 3.949,
774
+ "step": 40703
775
+ },
776
+ {
777
+ "epoch": 31.23,
778
+ "learning_rate": 0.004705243593854802,
779
+ "loss": 1.0205,
780
+ "step": 41000
781
+ },
782
+ {
783
+ "epoch": 31.61,
784
+ "learning_rate": 0.004698159721155694,
785
+ "loss": 1.0072,
786
+ "step": 41500
787
+ },
788
+ {
789
+ "epoch": 31.99,
790
+ "learning_rate": 0.004690997196729925,
791
+ "loss": 1.0541,
792
+ "step": 42000
793
+ },
794
+ {
795
+ "epoch": 32.0,
796
+ "eval_accuracy": 0.640868089666875,
797
+ "eval_loss": 1.7335774898529053,
798
+ "eval_runtime": 2595.2091,
799
+ "eval_samples_per_second": 64.717,
800
+ "eval_steps_per_second": 4.045,
801
+ "step": 42016
802
+ },
803
+ {
804
+ "epoch": 32.37,
805
+ "learning_rate": 0.004683756276857741,
806
+ "loss": 0.9544,
807
+ "step": 42500
808
+ },
809
+ {
810
+ "epoch": 32.75,
811
+ "learning_rate": 0.004676437220624432,
812
+ "loss": 1.0013,
813
+ "step": 43000
814
+ },
815
+ {
816
+ "epoch": 33.0,
817
+ "eval_accuracy": 0.6451489982435772,
818
+ "eval_loss": 1.7185142040252686,
819
+ "eval_runtime": 2605.1367,
820
+ "eval_samples_per_second": 64.471,
821
+ "eval_steps_per_second": 4.03,
822
+ "step": 43329
823
+ },
824
+ {
825
+ "epoch": 33.13,
826
+ "learning_rate": 0.004669040289911064,
827
+ "loss": 0.9986,
828
+ "step": 43500
829
+ },
830
+ {
831
+ "epoch": 33.51,
832
+ "learning_rate": 0.004661565749385107,
833
+ "loss": 0.9399,
834
+ "step": 44000
835
+ },
836
+ {
837
+ "epoch": 33.89,
838
+ "learning_rate": 0.004654013866490966,
839
+ "loss": 0.9811,
840
+ "step": 44500
841
+ },
842
+ {
843
+ "epoch": 34.0,
844
+ "eval_accuracy": 0.6478640111934745,
845
+ "eval_loss": 1.7246044874191284,
846
+ "eval_runtime": 2595.1453,
847
+ "eval_samples_per_second": 64.719,
848
+ "eval_steps_per_second": 4.045,
849
+ "step": 44642
850
+ },
851
+ {
852
+ "epoch": 34.27,
853
+ "learning_rate": 0.004646384911440411,
854
+ "loss": 0.914,
855
+ "step": 45000
856
+ },
857
+ {
858
+ "epoch": 34.65,
859
+ "learning_rate": 0.004638679157202909,
860
+ "loss": 0.9447,
861
+ "step": 45500
862
+ },
863
+ {
864
+ "epoch": 35.0,
865
+ "eval_accuracy": 0.6540263761126492,
866
+ "eval_loss": 1.7245060205459595,
867
+ "eval_runtime": 2605.4725,
868
+ "eval_samples_per_second": 64.462,
869
+ "eval_steps_per_second": 4.029,
870
+ "step": 45955
871
+ },
872
+ {
873
+ "epoch": 35.03,
874
+ "learning_rate": 0.0005810494104962897,
875
+ "loss": 0.8424,
876
+ "step": 46000
877
+ },
878
+ {
879
+ "epoch": 35.42,
880
+ "learning_rate": 0.0005391326120998649,
881
+ "loss": 0.692,
882
+ "step": 46500
883
+ },
884
+ {
885
+ "epoch": 35.8,
886
+ "learning_rate": 0.0004986016371937035,
887
+ "loss": 0.6587,
888
+ "step": 47000
889
+ },
890
+ {
891
+ "epoch": 36.0,
892
+ "eval_accuracy": 0.7018546634515197,
893
+ "eval_loss": 1.5848796367645264,
894
+ "eval_runtime": 4946.5461,
895
+ "eval_samples_per_second": 33.954,
896
+ "eval_steps_per_second": 2.122,
897
+ "step": 47268
898
+ },
899
+ {
900
+ "epoch": 36.18,
901
+ "learning_rate": 0.0004594851306405101,
902
+ "loss": 0.6234,
903
+ "step": 47500
904
+ },
905
+ {
906
+ "epoch": 36.56,
907
+ "learning_rate": 0.0004218107376415739,
908
+ "loss": 0.6053,
909
+ "step": 48000
910
+ },
911
+ {
912
+ "epoch": 36.94,
913
+ "learning_rate": 0.00038560508419879514,
914
+ "loss": 0.6044,
915
+ "step": 48500
916
+ },
917
+ {
918
+ "epoch": 37.0,
919
+ "eval_accuracy": 0.706195111785895,
920
+ "eval_loss": 1.6146132946014404,
921
+ "eval_runtime": 2680.1847,
922
+ "eval_samples_per_second": 62.665,
923
+ "eval_steps_per_second": 3.917,
924
+ "step": 48581
925
+ },
926
+ {
927
+ "epoch": 37.32,
928
+ "learning_rate": 0.0003508937582970281,
929
+ "loss": 0.5763,
930
+ "step": 49000
931
+ },
932
+ {
933
+ "epoch": 37.7,
934
+ "learning_rate": 0.0003177012918200253,
935
+ "loss": 0.572,
936
+ "step": 49500
937
+ },
938
+ {
939
+ "epoch": 38.0,
940
+ "eval_accuracy": 0.7081241999345063,
941
+ "eval_loss": 1.6583054065704346,
942
+ "eval_runtime": 2714.09,
943
+ "eval_samples_per_second": 61.883,
944
+ "eval_steps_per_second": 3.868,
945
+ "step": 49894
946
+ },
947
+ {
948
+ "epoch": 38.08,
949
+ "learning_rate": 0.0002860511432127716,
950
+ "loss": 0.5693,
951
+ "step": 50000
952
+ },
953
+ {
954
+ "epoch": 38.46,
955
+ "learning_rate": 0.00025596568090246547,
956
+ "loss": 0.5492,
957
+ "step": 50500
958
+ },
959
+ {
960
+ "epoch": 38.84,
961
+ "learning_rate": 0.00022746616748985177,
962
+ "loss": 0.545,
963
+ "step": 51000
964
+ },
965
+ {
966
+ "epoch": 39.0,
967
+ "eval_accuracy": 0.7086898276324015,
968
+ "eval_loss": 1.6993166208267212,
969
+ "eval_runtime": 8190.9805,
970
+ "eval_samples_per_second": 20.505,
971
+ "eval_steps_per_second": 1.282,
972
+ "step": 51207
973
+ },
974
+ {
975
+ "epoch": 39.22,
976
+ "learning_rate": 0.0002005727447220948,
977
+ "loss": 0.5409,
978
+ "step": 51500
979
+ },
980
+ {
981
+ "epoch": 39.6,
982
+ "learning_rate": 0.00017530441925779654,
983
+ "loss": 0.5351,
984
+ "step": 52000
985
+ },
986
+ {
987
+ "epoch": 39.98,
988
+ "learning_rate": 0.00015167904923422986,
989
+ "loss": 0.5341,
990
+ "step": 52500
991
+ },
992
+ {
993
+ "epoch": 40.0,
994
+ "eval_accuracy": 0.7105593760233396,
995
+ "eval_loss": 1.7077844142913818,
996
+ "eval_runtime": 33225.1715,
997
+ "eval_samples_per_second": 5.055,
998
+ "eval_steps_per_second": 0.316,
999
+ "step": 52520
1000
+ },
1001
+ {
1002
+ "epoch": 40.37,
1003
+ "learning_rate": 0.0001297133316462798,
1004
+ "loss": 0.5184,
1005
+ "step": 53000
1006
+ },
1007
+ {
1008
+ "epoch": 40.75,
1009
+ "learning_rate": 0.00010942279054600657,
1010
+ "loss": 0.5284,
1011
+ "step": 53500
1012
+ },
1013
+ {
1014
+ "epoch": 41.0,
1015
+ "eval_accuracy": 0.7104641124110624,
1016
+ "eval_loss": 1.7240744829177856,
1017
+ "eval_runtime": 3401.4075,
1018
+ "eval_samples_per_second": 49.378,
1019
+ "eval_steps_per_second": 3.086,
1020
+ "step": 53833
1021
+ },
1022
+ {
1023
+ "epoch": 41.13,
1024
+ "learning_rate": 9.082176607117799e-05,
1025
+ "loss": 0.5239,
1026
+ "step": 54000
1027
+ },
1028
+ {
1029
+ "epoch": 41.51,
1030
+ "learning_rate": 7.392340431051969e-05,
1031
+ "loss": 0.5152,
1032
+ "step": 54500
1033
+ },
1034
+ {
1035
+ "epoch": 41.89,
1036
+ "learning_rate": 5.87396480128477e-05,
1037
+ "loss": 0.5186,
1038
+ "step": 55000
1039
+ },
1040
+ {
1041
+ "epoch": 42.0,
1042
+ "eval_accuracy": 0.7111666815516061,
1043
+ "eval_loss": 1.7408066987991333,
1044
+ "eval_runtime": 4964.1178,
1045
+ "eval_samples_per_second": 33.834,
1046
+ "eval_steps_per_second": 2.115,
1047
+ "step": 55146
1048
+ },
1049
+ {
1050
+ "epoch": 42.27,
1051
+ "learning_rate": 4.528122814664942e-05,
1052
+ "loss": 0.5113,
1053
+ "step": 55500
1054
+ },
1055
+ {
1056
+ "epoch": 42.65,
1057
+ "learning_rate": 3.35576563160786e-05,
1058
+ "loss": 0.506,
1059
+ "step": 56000
1060
+ },
1061
+ {
1062
+ "epoch": 43.0,
1063
+ "eval_accuracy": 0.7105772379506415,
1064
+ "eval_loss": 1.7486644983291626,
1065
+ "eval_runtime": 6355.0593,
1066
+ "eval_samples_per_second": 26.429,
1067
+ "eval_steps_per_second": 1.652,
1068
+ "step": 56459
1069
+ },
1070
+ {
1071
+ "epoch": 43.03,
1072
+ "learning_rate": 2.357721803872476e-05,
1073
+ "loss": 0.5137,
1074
+ "step": 56500
1075
+ },
1076
+ {
1077
+ "epoch": 43.41,
1078
+ "learning_rate": 1.5346966889905334e-05,
1079
+ "loss": 0.5062,
1080
+ "step": 57000
1081
+ },
1082
+ {
1083
+ "epoch": 43.79,
1084
+ "learning_rate": 8.872719517621642e-06,
1085
+ "loss": 0.5043,
1086
+ "step": 57500
1087
+ },
1088
+ {
1089
+ "epoch": 44.0,
1090
+ "eval_accuracy": 0.7108689827632402,
1091
+ "eval_loss": 1.7547060251235962,
1092
+ "eval_runtime": 4684.6858,
1093
+ "eval_samples_per_second": 35.852,
1094
+ "eval_steps_per_second": 2.241,
1095
+ "step": 57772
1096
+ },
1097
+ {
1098
+ "epoch": 44.17,
1099
+ "learning_rate": 4.159051531700819e-06,
1100
+ "loss": 0.5084,
1101
+ "step": 58000
1102
+ },
1103
+ {
1104
+ "epoch": 44.55,
1105
+ "learning_rate": 1.2092942700298037e-06,
1106
+ "loss": 0.5043,
1107
+ "step": 58500
1108
+ },
1109
+ {
1110
+ "epoch": 44.94,
1111
+ "learning_rate": 2.5532444165032865e-08,
1112
+ "loss": 0.5094,
1113
+ "step": 59000
1114
+ },
1115
+ {
1116
+ "epoch": 45.0,
1117
+ "eval_accuracy": 0.711053556012027,
1118
+ "eval_loss": 1.7535769939422607,
1119
+ "eval_runtime": 2625.3482,
1120
+ "eval_samples_per_second": 63.974,
1121
+ "eval_steps_per_second": 3.999,
1122
+ "step": 59085
1123
+ },
1124
+ {
1125
+ "epoch": 45.32,
1126
+ "learning_rate": 0.0007032209920746526,
1127
+ "loss": 0.5386,
1128
+ "step": 59500
1129
+ },
1130
+ {
1131
+ "epoch": 45.7,
1132
+ "learning_rate": 0.0006689210221610515,
1133
+ "loss": 0.5547,
1134
+ "step": 60000
1135
+ },
1136
+ {
1137
+ "epoch": 46.0,
1138
+ "eval_accuracy": 0.7068619570718347,
1139
+ "eval_loss": 1.7073622941970825,
1140
+ "eval_runtime": 6607.5308,
1141
+ "eval_samples_per_second": 25.419,
1142
+ "eval_steps_per_second": 1.589,
1143
+ "step": 60398
1144
+ },
1145
+ {
1146
+ "epoch": 46.08,
1147
+ "learning_rate": 0.0006353490004690119,
1148
+ "loss": 0.553,
1149
+ "step": 60500
1150
+ },
1151
+ {
1152
+ "epoch": 46.46,
1153
+ "learning_rate": 0.0006025182736066154,
1154
+ "loss": 0.5329,
1155
+ "step": 61000
1156
+ },
1157
+ {
1158
+ "epoch": 46.84,
1159
+ "learning_rate": 0.000570441893479062,
1160
+ "loss": 0.5391,
1161
+ "step": 61500
1162
+ },
1163
+ {
1164
+ "epoch": 47.0,
1165
+ "eval_accuracy": 0.7089637105176982,
1166
+ "eval_loss": 1.740086317062378,
1167
+ "eval_runtime": 5830.6066,
1168
+ "eval_samples_per_second": 28.806,
1169
+ "eval_steps_per_second": 1.8,
1170
+ "step": 61711
1171
+ },
1172
+ {
1173
+ "epoch": 47.22,
1174
+ "learning_rate": 0.0005391326120998649,
1175
+ "loss": 0.5284,
1176
+ "step": 62000
1177
+ },
1178
+ {
1179
+ "epoch": 47.6,
1180
+ "learning_rate": 0.0005086028765212663,
1181
+ "loss": 0.5223,
1182
+ "step": 62500
1183
+ },
1184
+ {
1185
+ "epoch": 47.98,
1186
+ "learning_rate": 0.00047886482388589615,
1187
+ "loss": 0.5253,
1188
+ "step": 63000
1189
+ },
1190
+ {
1191
+ "epoch": 48.0,
1192
+ "eval_accuracy": 0.7092852252091334,
1193
+ "eval_loss": 1.7769759893417358,
1194
+ "eval_runtime": 2617.8614,
1195
+ "eval_samples_per_second": 64.157,
1196
+ "eval_steps_per_second": 4.01,
1197
+ "step": 63024
1198
+ },
1199
+ {
1200
+ "epoch": 48.36,
1201
+ "learning_rate": 0.0004499302766016297,
1202
+ "loss": 0.5041,
1203
+ "step": 63500
1204
+ },
1205
+ {
1206
+ "epoch": 48.74,
1207
+ "learning_rate": 0.0004218107376415739,
1208
+ "loss": 0.5066,
1209
+ "step": 64000
1210
+ },
1211
+ {
1212
+ "epoch": 49.0,
1213
+ "eval_accuracy": 0.7102319073561371,
1214
+ "eval_loss": 1.813459873199463,
1215
+ "eval_runtime": 2608.5743,
1216
+ "eval_samples_per_second": 64.386,
1217
+ "eval_steps_per_second": 4.024,
1218
+ "step": 64337
1219
+ },
1220
+ {
1221
+ "epoch": 49.12,
1222
+ "learning_rate": 0.00039451738597103935,
1223
+ "loss": 0.5057,
1224
+ "step": 64500
1225
+ },
1226
+ {
1227
+ "epoch": 49.5,
1228
+ "learning_rate": 0.00036806107210333034,
1229
+ "loss": 0.4897,
1230
+ "step": 65000
1231
+ },
1232
+ {
1233
+ "epoch": 49.89,
1234
+ "learning_rate": 0.00034245231378610494,
1235
+ "loss": 0.495,
1236
+ "step": 65500
1237
+ },
1238
+ {
1239
+ "epoch": 50.0,
1240
+ "eval_accuracy": 0.7110476020362597,
1241
+ "eval_loss": 1.8451627492904663,
1242
+ "eval_runtime": 2646.4398,
1243
+ "eval_samples_per_second": 63.465,
1244
+ "eval_steps_per_second": 3.967,
1245
+ "step": 65650
1246
+ },
1247
+ {
1248
+ "epoch": 50.27,
1249
+ "learning_rate": 0.0003177012918200253,
1250
+ "loss": 0.4867,
1251
+ "step": 66000
1252
+ },
1253
+ {
1254
+ "epoch": 50.65,
1255
+ "learning_rate": 0.0002938178460113711,
1256
+ "loss": 0.4813,
1257
+ "step": 66500
1258
+ },
1259
+ {
1260
+ "epoch": 51.0,
1261
+ "eval_accuracy": 0.7107320413205919,
1262
+ "eval_loss": 1.8846126794815063,
1263
+ "eval_runtime": 5243.8791,
1264
+ "eval_samples_per_second": 32.029,
1265
+ "eval_steps_per_second": 2.002,
1266
+ "step": 66963
1267
+ },
1268
+ {
1269
+ "epoch": 51.03,
1270
+ "learning_rate": 0.00027081147126020445,
1271
+ "loss": 0.4829,
1272
+ "step": 67000
1273
+ },
1274
+ {
1275
+ "epoch": 51.41,
1276
+ "learning_rate": 0.00024869131378566,
1277
+ "loss": 0.466,
1278
+ "step": 67500
1279
+ },
1280
+ {
1281
+ "epoch": 51.79,
1282
+ "learning_rate": 0.00022746616748985177,
1283
+ "loss": 0.4704,
1284
+ "step": 68000
1285
+ },
1286
+ {
1287
+ "epoch": 52.0,
1288
+ "eval_accuracy": 0.7123693846566045,
1289
+ "eval_loss": 1.8989039659500122,
1290
+ "eval_runtime": 2607.2421,
1291
+ "eval_samples_per_second": 64.419,
1292
+ "eval_steps_per_second": 4.026,
1293
+ "step": 68276
1294
+ },
1295
+ {
1296
+ "epoch": 52.17,
1297
+ "learning_rate": 0.00020714447046184416,
1298
+ "loss": 0.47,
1299
+ "step": 68500
1300
+ },
1301
+ {
1302
+ "epoch": 52.55,
1303
+ "learning_rate": 0.0001877343016230762,
1304
+ "loss": 0.465,
1305
+ "step": 69000
1306
+ },
1307
+ {
1308
+ "epoch": 52.93,
1309
+ "learning_rate": 0.0001692433775155744,
1310
+ "loss": 0.4689,
1311
+ "step": 69500
1312
+ },
1313
+ {
1314
+ "epoch": 53.0,
1315
+ "eval_accuracy": 0.7131850793367271,
1316
+ "eval_loss": 1.931063175201416,
1317
+ "eval_runtime": 2614.8439,
1318
+ "eval_samples_per_second": 64.231,
1319
+ "eval_steps_per_second": 4.015,
1320
+ "step": 69589
1321
+ },
1322
+ {
1323
+ "epoch": 53.31,
1324
+ "learning_rate": 0.00015167904923422986,
1325
+ "loss": 0.4572,
1326
+ "step": 70000
1327
+ },
1328
+ {
1329
+ "epoch": 53.69,
1330
+ "learning_rate": 0.0001350482995043595,
1331
+ "loss": 0.4611,
1332
+ "step": 70500
1333
+ },
1334
+ {
1335
+ "epoch": 54.0,
1336
+ "eval_accuracy": 0.7130659998213807,
1337
+ "eval_loss": 1.93539559841156,
1338
+ "eval_runtime": 2604.5499,
1339
+ "eval_samples_per_second": 64.485,
1340
+ "eval_steps_per_second": 4.031,
1341
+ "step": 70902
1342
+ },
1343
+ {
1344
+ "epoch": 54.07,
1345
+ "learning_rate": 0.00011935773990571164,
1346
+ "loss": 0.4579,
1347
+ "step": 71000
1348
+ },
1349
+ {
1350
+ "epoch": 54.46,
1351
+ "learning_rate": 0.00010461360824402521,
1352
+ "loss": 0.4554,
1353
+ "step": 71500
1354
+ },
1355
+ {
1356
+ "epoch": 54.84,
1357
+ "learning_rate": 9.082176607117799e-05,
1358
+ "loss": 0.4547,
1359
+ "step": 72000
1360
+ },
1361
+ {
1362
+ "epoch": 55.0,
1363
+ "eval_accuracy": 0.7133160668036082,
1364
+ "eval_loss": 1.974073052406311,
1365
+ "eval_runtime": 2574.37,
1366
+ "eval_samples_per_second": 65.241,
1367
+ "eval_steps_per_second": 4.078,
1368
+ "step": 72215
1369
+ },
1370
+ {
1371
+ "epoch": 55.22,
1372
+ "learning_rate": 7.798769635491976e-05,
1373
+ "loss": 0.4497,
1374
+ "step": 72500
1375
+ },
1376
+ {
1377
+ "epoch": 55.6,
1378
+ "learning_rate": 6.611650129910801e-05,
1379
+ "loss": 0.4458,
1380
+ "step": 73000
1381
+ },
1382
+ {
1383
+ "epoch": 55.98,
1384
+ "learning_rate": 5.521290031532178e-05,
1385
+ "loss": 0.4481,
1386
+ "step": 73500
1387
+ },
1388
+ {
1389
+ "epoch": 56.0,
1390
+ "eval_accuracy": 0.713071953797148,
1391
+ "eval_loss": 1.9898993968963623,
1392
+ "eval_runtime": 2622.4841,
1393
+ "eval_samples_per_second": 64.044,
1394
+ "eval_steps_per_second": 4.003,
1395
+ "step": 73528
1396
+ },
1397
+ {
1398
+ "epoch": 56.36,
1399
+ "learning_rate": 0.00045401414969093167,
1400
+ "loss": 0.4607,
1401
+ "step": 74000
1402
+ },
1403
+ {
1404
+ "epoch": 56.74,
1405
+ "learning_rate": 0.0004297610695623985,
1406
+ "loss": 0.4709,
1407
+ "step": 74500
1408
+ },
1409
+ {
1410
+ "epoch": 57.0,
1411
+ "eval_accuracy": 0.7104283885564586,
1412
+ "eval_loss": 1.9411875009536743,
1413
+ "eval_runtime": 3269.1549,
1414
+ "eval_samples_per_second": 51.376,
1415
+ "eval_steps_per_second": 3.211,
1416
+ "step": 74841
1417
+ },
1418
+ {
1419
+ "epoch": 57.12,
1420
+ "learning_rate": 0.0004061126673698029,
1421
+ "loss": 0.4736,
1422
+ "step": 75000
1423
+ },
1424
+ {
1425
+ "epoch": 57.5,
1426
+ "learning_rate": 0.0003830758503675777,
1427
+ "loss": 0.4601,
1428
+ "step": 75500
1429
+ },
1430
+ {
1431
+ "epoch": 57.88,
1432
+ "learning_rate": 0.00036065734717759815,
1433
+ "loss": 0.4647,
1434
+ "step": 76000
1435
+ },
1436
+ {
1437
+ "epoch": 58.0,
1438
+ "eval_accuracy": 0.7097913131493555,
1439
+ "eval_loss": 1.9707293510437012,
1440
+ "eval_runtime": 2595.9104,
1441
+ "eval_samples_per_second": 64.7,
1442
+ "eval_steps_per_second": 4.044,
1443
+ "step": 76154
1444
+ },
1445
+ {
1446
+ "epoch": 58.26,
1447
+ "learning_rate": 0.00033886370582388,
1448
+ "loss": 0.4562,
1449
+ "step": 76500
1450
+ },
1451
+ {
1452
+ "epoch": 58.64,
1453
+ "learning_rate": 0.0003177012918200253,
1454
+ "loss": 0.4566,
1455
+ "step": 77000
1456
+ },
1457
+ {
1458
+ "epoch": 59.0,
1459
+ "eval_accuracy": 0.7116251376856896,
1460
+ "eval_loss": 2.0150792598724365,
1461
+ "eval_runtime": 2589.1949,
1462
+ "eval_samples_per_second": 64.868,
1463
+ "eval_steps_per_second": 4.055,
1464
+ "step": 77467
1465
+ },
1466
+ {
1467
+ "epoch": 59.03,
1468
+ "learning_rate": 0.0002971762863099739,
1469
+ "loss": 0.4535,
1470
+ "step": 77500
1471
+ },
1472
+ {
1473
+ "epoch": 59.41,
1474
+ "learning_rate": 0.0002772946842626087,
1475
+ "loss": 0.448,
1476
+ "step": 78000
1477
+ },
1478
+ {
1479
+ "epoch": 59.79,
1480
+ "learning_rate": 0.000258062292720736,
1481
+ "loss": 0.4511,
1482
+ "step": 78500
1483
+ },
1484
+ {
1485
+ "epoch": 60.0,
1486
+ "eval_accuracy": 0.7113929326307642,
1487
+ "eval_loss": 2.0363075733184814,
1488
+ "eval_runtime": 2640.6575,
1489
+ "eval_samples_per_second": 63.603,
1490
+ "eval_steps_per_second": 3.976,
1491
+ "step": 78780
1492
+ },
1493
+ {
1494
+ "epoch": 60.17,
1495
+ "learning_rate": 0.00023948472910495817,
1496
+ "loss": 0.4446,
1497
+ "step": 79000
1498
+ },
1499
+ {
1500
+ "epoch": 60.55,
1501
+ "learning_rate": 0.0002215674195729317,
1502
+ "loss": 0.4403,
1503
+ "step": 79500
1504
+ },
1505
+ {
1506
+ "epoch": 60.93,
1507
+ "learning_rate": 0.0002043155974344868,
1508
+ "loss": 0.4423,
1509
+ "step": 80000
1510
+ },
1511
+ {
1512
+ "epoch": 61.0,
1513
+ "eval_accuracy": 0.7112143133577447,
1514
+ "eval_loss": 2.0710268020629883,
1515
+ "eval_runtime": 2593.0457,
1516
+ "eval_samples_per_second": 64.771,
1517
+ "eval_steps_per_second": 4.049,
1518
+ "step": 80093
1519
+ },
1520
+ {
1521
+ "epoch": 61.31,
1522
+ "learning_rate": 0.0001877343016230762,
1523
+ "loss": 0.4304,
1524
+ "step": 80500
1525
+ },
1526
+ {
1527
+ "epoch": 61.69,
1528
+ "learning_rate": 0.00017182837522399742,
1529
+ "loss": 0.4356,
1530
+ "step": 81000
1531
+ },
1532
+ {
1533
+ "epoch": 62.0,
1534
+ "eval_accuracy": 0.7116251376856896,
1535
+ "eval_loss": 2.061058521270752,
1536
+ "eval_runtime": 2582.6346,
1537
+ "eval_samples_per_second": 65.032,
1538
+ "eval_steps_per_second": 4.065,
1539
+ "step": 81406
1540
+ },
1541
+ {
1542
+ "epoch": 62.07,
1543
+ "learning_rate": 0.00015660246405981793,
1544
+ "loss": 0.4332,
1545
+ "step": 81500
1546
+ },
1547
+ {
1548
+ "epoch": 62.45,
1549
+ "learning_rate": 0.00014206101533341926,
1550
+ "loss": 0.43,
1551
+ "step": 82000
1552
+ },
1553
+ {
1554
+ "epoch": 62.83,
1555
+ "learning_rate": 0.0001282082763290507,
1556
+ "loss": 0.4272,
1557
+ "step": 82500
1558
+ },
1559
+ {
1560
+ "epoch": 63.0,
1561
+ "eval_accuracy": 0.7117918490071745,
1562
+ "eval_loss": 2.089148998260498,
1563
+ "eval_runtime": 2586.4124,
1564
+ "eval_samples_per_second": 64.937,
1565
+ "eval_steps_per_second": 4.059,
1566
+ "step": 82719
1567
+ },
1568
+ {
1569
+ "epoch": 63.21,
1570
+ "learning_rate": 0.00011504829317177723,
1571
+ "loss": 0.4259,
1572
+ "step": 83000
1573
+ },
1574
+ {
1575
+ "epoch": 63.59,
1576
+ "learning_rate": 0.00010258490964568406,
1577
+ "loss": 0.4271,
1578
+ "step": 83500
1579
+ },
1580
+ {
1581
+ "epoch": 63.98,
1582
+ "learning_rate": 9.082176607117799e-05,
1583
+ "loss": 0.4254,
1584
+ "step": 84000
1585
+ },
1586
+ {
1587
+ "epoch": 64.0,
1588
+ "eval_accuracy": 0.7123812926081391,
1589
+ "eval_loss": 2.0879101753234863,
1590
+ "eval_runtime": 2610.5128,
1591
+ "eval_samples_per_second": 64.338,
1592
+ "eval_steps_per_second": 4.021,
1593
+ "step": 84032
1594
+ },
1595
+ {
1596
+ "epoch": 64.36,
1597
+ "learning_rate": 7.976229824171872e-05,
1598
+ "loss": 0.4249,
1599
+ "step": 84500
1600
+ },
1601
+ {
1602
+ "epoch": 64.74,
1603
+ "learning_rate": 6.940973642028724e-05,
1604
+ "loss": 0.4221,
1605
+ "step": 85000
1606
+ },
1607
+ {
1608
+ "epoch": 65.0,
1609
+ "eval_accuracy": 0.7130957697002174,
1610
+ "eval_loss": 2.1167192459106445,
1611
+ "eval_runtime": 2625.6926,
1612
+ "eval_samples_per_second": 63.966,
1613
+ "eval_steps_per_second": 3.998,
1614
+ "step": 85345
1615
+ },
1616
+ {
1617
+ "epoch": 65.12,
1618
+ "learning_rate": 5.976710439588779e-05,
1619
+ "loss": 0.4227,
1620
+ "step": 85500
1621
+ },
1622
+ {
1623
+ "epoch": 65.5,
1624
+ "learning_rate": 5.083721860035529e-05,
1625
+ "loss": 0.421,
1626
+ "step": 86000
1627
+ },
1628
+ {
1629
+ "epoch": 65.88,
1630
+ "learning_rate": 4.262268728572882e-05,
1631
+ "loss": 0.4189,
1632
+ "step": 86500
1633
+ },
1634
+ {
1635
+ "epoch": 66.0,
1636
+ "eval_accuracy": 0.7129111964514304,
1637
+ "eval_loss": 2.1362690925598145,
1638
+ "eval_runtime": 2775.8163,
1639
+ "eval_samples_per_second": 60.507,
1640
+ "eval_steps_per_second": 3.782,
1641
+ "step": 86658
1642
+ },
1643
+ {
1644
+ "epoch": 66.26,
1645
+ "learning_rate": 3.512590976242902e-05,
1646
+ "loss": 0.4192,
1647
+ "step": 87000
1648
+ },
1649
+ {
1650
+ "epoch": 66.64,
1651
+ "learning_rate": 2.834907569846612e-05,
1652
+ "loss": 0.4219,
1653
+ "step": 87500
1654
+ },
1655
+ {
1656
+ "epoch": 67.0,
1657
+ "eval_accuracy": 0.712958828257569,
1658
+ "eval_loss": 2.1355273723602295,
1659
+ "eval_runtime": 2581.0691,
1660
+ "eval_samples_per_second": 65.072,
1661
+ "eval_steps_per_second": 4.067,
1662
+ "step": 87971
1663
+ },
1664
+ {
1665
+ "epoch": 67.02,
1666
+ "learning_rate": 2.2294164479878255e-05,
1667
+ "loss": 0.4197,
1668
+ "step": 88000
1669
+ },
1670
+ {
1671
+ "epoch": 67.4,
1672
+ "learning_rate": 1.6962944632589183e-05,
1673
+ "loss": 0.4157,
1674
+ "step": 88500
1675
+ },
1676
+ {
1677
+ "epoch": 67.78,
1678
+ "learning_rate": 1.2356973305856779e-05,
1679
+ "loss": 0.4149,
1680
+ "step": 89000
1681
+ },
1682
+ {
1683
+ "epoch": 68.0,
1684
+ "eval_accuracy": 0.7132267571670984,
1685
+ "eval_loss": 2.1466283798217773,
1686
+ "eval_runtime": 2769.7004,
1687
+ "eval_samples_per_second": 60.64,
1688
+ "eval_steps_per_second": 3.79,
1689
+ "step": 89284
1690
+ },
1691
+ {
1692
+ "epoch": 68.16,
1693
+ "learning_rate": 8.477595817457106e-06,
1694
+ "loss": 0.4138,
1695
+ "step": 89500
1696
+ },
1697
+ {
1698
+ "epoch": 68.55,
1699
+ "learning_rate": 5.325945260744314e-06,
1700
+ "loss": 0.4147,
1701
+ "step": 90000
1702
+ },
1703
+ {
1704
+ "epoch": 68.93,
1705
+ "learning_rate": 2.90294217369369e-06,
1706
+ "loss": 0.4125,
1707
+ "step": 90500
1708
+ },
1709
+ {
1710
+ "epoch": 69.0,
1711
+ "eval_accuracy": 0.7131434015063559,
1712
+ "eval_loss": 2.1478145122528076,
1713
+ "eval_runtime": 2585.5905,
1714
+ "eval_samples_per_second": 64.958,
1715
+ "eval_steps_per_second": 4.06,
1716
+ "step": 90597
1717
+ },
1718
+ {
1719
+ "epoch": 69.31,
1720
+ "learning_rate": 1.2092942700298037e-06,
1721
+ "loss": 0.4146,
1722
+ "step": 91000
1723
+ },
1724
+ {
1725
+ "epoch": 69.69,
1726
+ "learning_rate": 2.4549623251851615e-07,
1727
+ "loss": 0.4162,
1728
+ "step": 91500
1729
+ },
1730
+ {
1731
+ "epoch": 70.0,
1732
+ "eval_accuracy": 0.7132267571670984,
1733
+ "eval_loss": 2.1484131813049316,
1734
+ "eval_runtime": 2581.5568,
1735
+ "eval_samples_per_second": 65.06,
1736
+ "eval_steps_per_second": 4.067,
1737
+ "step": 91910
1738
+ },
1739
+ {
1740
+ "epoch": 70.0,
1741
+ "step": 91910,
1742
+ "total_flos": 3.2073215097814647e+21,
1743
+ "train_loss": 0.08673005339085323,
1744
+ "train_runtime": 125919.3774,
1745
+ "train_samples_per_second": 373.471,
1746
+ "train_steps_per_second": 0.73
1747
+ }
1748
+ ],
1749
+ "logging_steps": 500,
1750
+ "max_steps": 91910,
1751
+ "num_train_epochs": 70,
1752
+ "save_steps": 500,
1753
+ "total_flos": 3.2073215097814647e+21,
1754
+ "trial_name": null,
1755
+ "trial_params": null
1756
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b2613814eab86550f4aada3be0340e3ef353a3d32cfe167fa7096a3258ca9bc
3
+ size 4728