amauriciogonzalez commited on
Commit
0eefe61
1 Parent(s): 8454fbd

End of training

Browse files
Files changed (5) hide show
  1. README.md +3 -3
  2. all_results.json +11 -11
  3. eval_results.json +6 -6
  4. train_results.json +6 -6
  5. trainer_state.json +3356 -38
README.md CHANGED
@@ -23,7 +23,7 @@ model-index:
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
- value: 0.8957671957671958
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -33,8 +33,8 @@ should probably proofread and complete it, then remove this comment. -->
33
 
34
  This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on the imagefolder dataset.
35
  It achieves the following results on the evaluation set:
36
- - Loss: 0.4414
37
- - Accuracy: 0.8958
38
 
39
  ## Model description
40
 
 
23
  metrics:
24
  - name: Accuracy
25
  type: accuracy
26
+ value: 0.8968253968253969
27
  ---
28
 
29
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
33
 
34
  This model is a fine-tuned version of [facebook/dinov2-base](https://huggingface.co/facebook/dinov2-base) on the imagefolder dataset.
35
  It achieves the following results on the evaluation set:
36
+ - Loss: 0.4424
37
+ - Accuracy: 0.8968
38
 
39
  ## Model description
40
 
all_results.json CHANGED
@@ -1,13 +1,13 @@
1
  {
2
- "epoch": 1.0,
3
- "eval_accuracy": 0.8634920634920635,
4
- "eval_loss": 0.44639232754707336,
5
- "eval_runtime": 28.206,
6
- "eval_samples_per_second": 67.007,
7
- "eval_steps_per_second": 1.064,
8
- "total_flos": 1.0941516662243328e+18,
9
- "train_loss": 1.2471325397491455,
10
- "train_runtime": 397.8778,
11
- "train_samples_per_second": 26.918,
12
- "train_steps_per_second": 0.106
13
  }
 
1
  {
2
+ "epoch": 49.55223880597015,
3
+ "eval_accuracy": 0.8968253968253969,
4
+ "eval_loss": 0.44243764877319336,
5
+ "eval_runtime": 16.2594,
6
+ "eval_samples_per_second": 116.241,
7
+ "eval_steps_per_second": 3.69,
8
+ "total_flos": 5.421822883117793e+19,
9
+ "train_loss": 0.33880508939903903,
10
+ "train_runtime": 14225.4065,
11
+ "train_samples_per_second": 37.644,
12
+ "train_steps_per_second": 0.292
13
  }
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 1.0,
3
- "eval_accuracy": 0.8634920634920635,
4
- "eval_loss": 0.44639232754707336,
5
- "eval_runtime": 28.206,
6
- "eval_samples_per_second": 67.007,
7
- "eval_steps_per_second": 1.064
8
  }
 
1
  {
2
+ "epoch": 49.55223880597015,
3
+ "eval_accuracy": 0.8968253968253969,
4
+ "eval_loss": 0.44243764877319336,
5
+ "eval_runtime": 16.2594,
6
+ "eval_samples_per_second": 116.241,
7
+ "eval_steps_per_second": 3.69
8
  }
train_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
- "epoch": 1.0,
3
- "total_flos": 1.0941516662243328e+18,
4
- "train_loss": 1.2471325397491455,
5
- "train_runtime": 397.8778,
6
- "train_samples_per_second": 26.918,
7
- "train_steps_per_second": 0.106
8
  }
 
1
  {
2
+ "epoch": 49.55223880597015,
3
+ "total_flos": 5.421822883117793e+19,
4
+ "train_loss": 0.33880508939903903,
5
+ "train_runtime": 14225.4065,
6
+ "train_samples_per_second": 37.644,
7
+ "train_steps_per_second": 0.292
8
  }
trainer_state.json CHANGED
@@ -1,64 +1,3382 @@
1
  {
2
- "best_metric": 0.8634920634920635,
3
- "best_model_checkpoint": "dinov2-base-finetuned-har/checkpoint-42",
4
- "epoch": 1.0,
5
  "eval_steps": 500,
6
- "global_step": 42,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.23809523809523808,
13
- "grad_norm": 44.80964660644531,
14
- "learning_rate": 4.324324324324325e-05,
15
- "loss": 2.2183,
16
  "step": 10
17
  },
18
  {
19
- "epoch": 0.47619047619047616,
20
- "grad_norm": 20.761051177978516,
21
- "learning_rate": 2.9729729729729733e-05,
22
- "loss": 1.1574,
23
  "step": 20
24
  },
25
  {
26
- "epoch": 0.7142857142857143,
27
- "grad_norm": 28.136566162109375,
28
- "learning_rate": 1.6216216216216218e-05,
29
- "loss": 0.8864,
30
  "step": 30
31
  },
32
  {
33
- "epoch": 0.9523809523809523,
34
- "grad_norm": 15.380955696105957,
35
- "learning_rate": 2.702702702702703e-06,
36
- "loss": 0.8084,
37
  "step": 40
38
  },
39
  {
40
- "epoch": 1.0,
41
- "eval_accuracy": 0.8634920634920635,
42
- "eval_loss": 0.44639232754707336,
43
- "eval_runtime": 28.1795,
44
- "eval_samples_per_second": 67.07,
45
- "eval_steps_per_second": 1.065,
46
- "step": 42
47
  },
48
  {
49
- "epoch": 1.0,
50
- "step": 42,
51
- "total_flos": 1.0941516662243328e+18,
52
- "train_loss": 1.2471325397491455,
53
- "train_runtime": 397.8778,
54
- "train_samples_per_second": 26.918,
55
- "train_steps_per_second": 0.106
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
56
  }
57
  ],
58
  "logging_steps": 10,
59
- "max_steps": 42,
60
  "num_input_tokens_seen": 0,
61
- "num_train_epochs": 1,
62
  "save_steps": 500,
63
  "stateful_callbacks": {
64
  "TrainerControl": {
@@ -72,8 +3390,8 @@
72
  "attributes": {}
73
  }
74
  },
75
- "total_flos": 1.0941516662243328e+18,
76
- "train_batch_size": 64,
77
  "trial_name": null,
78
  "trial_params": null
79
  }
 
1
  {
2
+ "best_metric": 0.8968253968253969,
3
+ "best_model_checkpoint": "dinov2-base-finetuned-har/checkpoint-4103",
4
+ "epoch": 49.55223880597015,
5
  "eval_steps": 500,
6
+ "global_step": 4150,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.11940298507462686,
13
+ "grad_norm": 37.050025939941406,
14
+ "learning_rate": 1.2048192771084338e-06,
15
+ "loss": 3.2333,
16
  "step": 10
17
  },
18
  {
19
+ "epoch": 0.23880597014925373,
20
+ "grad_norm": 39.30177307128906,
21
+ "learning_rate": 2.4096385542168676e-06,
22
+ "loss": 2.9768,
23
  "step": 20
24
  },
25
  {
26
+ "epoch": 0.3582089552238806,
27
+ "grad_norm": 35.40634536743164,
28
+ "learning_rate": 3.614457831325301e-06,
29
+ "loss": 2.4932,
30
  "step": 30
31
  },
32
  {
33
+ "epoch": 0.47761194029850745,
34
+ "grad_norm": 44.90218734741211,
35
+ "learning_rate": 4.819277108433735e-06,
36
+ "loss": 1.9544,
37
  "step": 40
38
  },
39
  {
40
+ "epoch": 0.5970149253731343,
41
+ "grad_norm": 36.5886116027832,
42
+ "learning_rate": 6.024096385542169e-06,
43
+ "loss": 1.514,
44
+ "step": 50
 
 
45
  },
46
  {
47
+ "epoch": 0.7164179104477612,
48
+ "grad_norm": 35.85557556152344,
49
+ "learning_rate": 7.228915662650602e-06,
50
+ "loss": 1.2005,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.835820895522388,
55
+ "grad_norm": 33.366024017333984,
56
+ "learning_rate": 8.433734939759036e-06,
57
+ "loss": 1.0445,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.9552238805970149,
62
+ "grad_norm": 34.10721969604492,
63
+ "learning_rate": 9.63855421686747e-06,
64
+ "loss": 0.9155,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.991044776119403,
69
+ "eval_accuracy": 0.8052910052910053,
70
+ "eval_loss": 0.6204278469085693,
71
+ "eval_runtime": 51.558,
72
+ "eval_samples_per_second": 36.658,
73
+ "eval_steps_per_second": 1.164,
74
+ "step": 83
75
+ },
76
+ {
77
+ "epoch": 1.0746268656716418,
78
+ "grad_norm": 38.803226470947266,
79
+ "learning_rate": 1.0843373493975904e-05,
80
+ "loss": 0.8863,
81
+ "step": 90
82
+ },
83
+ {
84
+ "epoch": 1.1940298507462686,
85
+ "grad_norm": 35.97611999511719,
86
+ "learning_rate": 1.2048192771084338e-05,
87
+ "loss": 0.8112,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 1.3134328358208955,
92
+ "grad_norm": 35.07186508178711,
93
+ "learning_rate": 1.3253012048192772e-05,
94
+ "loss": 0.7862,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 1.4328358208955223,
99
+ "grad_norm": 41.96965789794922,
100
+ "learning_rate": 1.4457831325301205e-05,
101
+ "loss": 0.8454,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 1.5522388059701493,
106
+ "grad_norm": 42.4810676574707,
107
+ "learning_rate": 1.566265060240964e-05,
108
+ "loss": 0.8309,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 1.671641791044776,
113
+ "grad_norm": 28.388389587402344,
114
+ "learning_rate": 1.6867469879518073e-05,
115
+ "loss": 0.754,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 1.7910447761194028,
120
+ "grad_norm": 35.19267654418945,
121
+ "learning_rate": 1.8072289156626505e-05,
122
+ "loss": 0.7247,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 1.9104477611940298,
127
+ "grad_norm": 34.15055847167969,
128
+ "learning_rate": 1.927710843373494e-05,
129
+ "loss": 0.749,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 1.9940298507462688,
134
+ "eval_accuracy": 0.8666666666666667,
135
+ "eval_loss": 0.44325119256973267,
136
+ "eval_runtime": 27.443,
137
+ "eval_samples_per_second": 68.87,
138
+ "eval_steps_per_second": 2.186,
139
+ "step": 167
140
+ },
141
+ {
142
+ "epoch": 2.029850746268657,
143
+ "grad_norm": 27.7845516204834,
144
+ "learning_rate": 2.0481927710843373e-05,
145
+ "loss": 0.7159,
146
+ "step": 170
147
+ },
148
+ {
149
+ "epoch": 2.1492537313432836,
150
+ "grad_norm": 32.60486602783203,
151
+ "learning_rate": 2.168674698795181e-05,
152
+ "loss": 0.6907,
153
+ "step": 180
154
+ },
155
+ {
156
+ "epoch": 2.2686567164179103,
157
+ "grad_norm": 34.04874038696289,
158
+ "learning_rate": 2.289156626506024e-05,
159
+ "loss": 0.7388,
160
+ "step": 190
161
+ },
162
+ {
163
+ "epoch": 2.388059701492537,
164
+ "grad_norm": 44.46305465698242,
165
+ "learning_rate": 2.4096385542168677e-05,
166
+ "loss": 0.7852,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 2.5074626865671643,
171
+ "grad_norm": 25.580005645751953,
172
+ "learning_rate": 2.530120481927711e-05,
173
+ "loss": 0.6814,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 2.626865671641791,
178
+ "grad_norm": 38.6807861328125,
179
+ "learning_rate": 2.6506024096385545e-05,
180
+ "loss": 0.7193,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 2.746268656716418,
185
+ "grad_norm": 29.82606315612793,
186
+ "learning_rate": 2.7710843373493977e-05,
187
+ "loss": 0.7395,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 2.8656716417910446,
192
+ "grad_norm": 27.26519203186035,
193
+ "learning_rate": 2.891566265060241e-05,
194
+ "loss": 0.7069,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 2.9850746268656714,
199
+ "grad_norm": 50.2726936340332,
200
+ "learning_rate": 3.012048192771085e-05,
201
+ "loss": 0.8197,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 2.997014925373134,
206
+ "eval_accuracy": 0.8571428571428571,
207
+ "eval_loss": 0.4825701117515564,
208
+ "eval_runtime": 27.1447,
209
+ "eval_samples_per_second": 69.627,
210
+ "eval_steps_per_second": 2.21,
211
+ "step": 251
212
+ },
213
+ {
214
+ "epoch": 3.1044776119402986,
215
+ "grad_norm": 32.886810302734375,
216
+ "learning_rate": 3.132530120481928e-05,
217
+ "loss": 0.719,
218
+ "step": 260
219
+ },
220
+ {
221
+ "epoch": 3.2238805970149254,
222
+ "grad_norm": 25.671785354614258,
223
+ "learning_rate": 3.253012048192771e-05,
224
+ "loss": 0.6762,
225
+ "step": 270
226
+ },
227
+ {
228
+ "epoch": 3.343283582089552,
229
+ "grad_norm": 33.79856491088867,
230
+ "learning_rate": 3.3734939759036146e-05,
231
+ "loss": 0.6348,
232
+ "step": 280
233
+ },
234
+ {
235
+ "epoch": 3.4626865671641793,
236
+ "grad_norm": 25.73578453063965,
237
+ "learning_rate": 3.4939759036144585e-05,
238
+ "loss": 0.6784,
239
+ "step": 290
240
+ },
241
+ {
242
+ "epoch": 3.582089552238806,
243
+ "grad_norm": 40.199222564697266,
244
+ "learning_rate": 3.614457831325301e-05,
245
+ "loss": 0.6659,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 3.701492537313433,
250
+ "grad_norm": 30.21065902709961,
251
+ "learning_rate": 3.734939759036144e-05,
252
+ "loss": 0.7042,
253
+ "step": 310
254
+ },
255
+ {
256
+ "epoch": 3.8208955223880596,
257
+ "grad_norm": 32.785037994384766,
258
+ "learning_rate": 3.855421686746988e-05,
259
+ "loss": 0.7352,
260
+ "step": 320
261
+ },
262
+ {
263
+ "epoch": 3.9402985074626864,
264
+ "grad_norm": 27.440940856933594,
265
+ "learning_rate": 3.9759036144578314e-05,
266
+ "loss": 0.6854,
267
+ "step": 330
268
+ },
269
+ {
270
+ "epoch": 4.0,
271
+ "eval_accuracy": 0.8724867724867725,
272
+ "eval_loss": 0.42430025339126587,
273
+ "eval_runtime": 32.3022,
274
+ "eval_samples_per_second": 58.51,
275
+ "eval_steps_per_second": 1.857,
276
+ "step": 335
277
+ },
278
+ {
279
+ "epoch": 4.059701492537314,
280
+ "grad_norm": 28.7294921875,
281
+ "learning_rate": 4.0963855421686746e-05,
282
+ "loss": 0.6307,
283
+ "step": 340
284
+ },
285
+ {
286
+ "epoch": 4.17910447761194,
287
+ "grad_norm": 29.991897583007812,
288
+ "learning_rate": 4.2168674698795186e-05,
289
+ "loss": 0.6336,
290
+ "step": 350
291
+ },
292
+ {
293
+ "epoch": 4.298507462686567,
294
+ "grad_norm": 27.699886322021484,
295
+ "learning_rate": 4.337349397590362e-05,
296
+ "loss": 0.7364,
297
+ "step": 360
298
+ },
299
+ {
300
+ "epoch": 4.417910447761194,
301
+ "grad_norm": 64.75654602050781,
302
+ "learning_rate": 4.457831325301205e-05,
303
+ "loss": 0.6794,
304
+ "step": 370
305
+ },
306
+ {
307
+ "epoch": 4.537313432835821,
308
+ "grad_norm": 28.235536575317383,
309
+ "learning_rate": 4.578313253012048e-05,
310
+ "loss": 0.6682,
311
+ "step": 380
312
+ },
313
+ {
314
+ "epoch": 4.656716417910448,
315
+ "grad_norm": 30.604808807373047,
316
+ "learning_rate": 4.698795180722892e-05,
317
+ "loss": 0.6224,
318
+ "step": 390
319
+ },
320
+ {
321
+ "epoch": 4.776119402985074,
322
+ "grad_norm": 26.9096736907959,
323
+ "learning_rate": 4.8192771084337354e-05,
324
+ "loss": 0.6643,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 4.895522388059701,
329
+ "grad_norm": 45.720359802246094,
330
+ "learning_rate": 4.9397590361445786e-05,
331
+ "loss": 0.7058,
332
+ "step": 410
333
+ },
334
+ {
335
+ "epoch": 4.991044776119403,
336
+ "eval_accuracy": 0.8592592592592593,
337
+ "eval_loss": 0.434872031211853,
338
+ "eval_runtime": 21.0424,
339
+ "eval_samples_per_second": 89.818,
340
+ "eval_steps_per_second": 2.851,
341
+ "step": 418
342
+ },
343
+ {
344
+ "epoch": 5.014925373134329,
345
+ "grad_norm": 25.985727310180664,
346
+ "learning_rate": 4.99330655957162e-05,
347
+ "loss": 0.7109,
348
+ "step": 420
349
+ },
350
+ {
351
+ "epoch": 5.134328358208955,
352
+ "grad_norm": 38.45668029785156,
353
+ "learning_rate": 4.97991967871486e-05,
354
+ "loss": 0.6265,
355
+ "step": 430
356
+ },
357
+ {
358
+ "epoch": 5.253731343283582,
359
+ "grad_norm": 38.37868881225586,
360
+ "learning_rate": 4.9665327978580993e-05,
361
+ "loss": 0.6318,
362
+ "step": 440
363
+ },
364
+ {
365
+ "epoch": 5.373134328358209,
366
+ "grad_norm": 23.872426986694336,
367
+ "learning_rate": 4.953145917001339e-05,
368
+ "loss": 0.5749,
369
+ "step": 450
370
+ },
371
+ {
372
+ "epoch": 5.492537313432836,
373
+ "grad_norm": 30.97602653503418,
374
+ "learning_rate": 4.9397590361445786e-05,
375
+ "loss": 0.629,
376
+ "step": 460
377
+ },
378
+ {
379
+ "epoch": 5.611940298507463,
380
+ "grad_norm": 54.77509689331055,
381
+ "learning_rate": 4.926372155287818e-05,
382
+ "loss": 0.6286,
383
+ "step": 470
384
+ },
385
+ {
386
+ "epoch": 5.731343283582089,
387
+ "grad_norm": 58.17964553833008,
388
+ "learning_rate": 4.912985274431058e-05,
389
+ "loss": 0.6289,
390
+ "step": 480
391
+ },
392
+ {
393
+ "epoch": 5.850746268656716,
394
+ "grad_norm": 25.494422912597656,
395
+ "learning_rate": 4.8995983935742975e-05,
396
+ "loss": 0.6603,
397
+ "step": 490
398
+ },
399
+ {
400
+ "epoch": 5.970149253731344,
401
+ "grad_norm": 31.255512237548828,
402
+ "learning_rate": 4.886211512717537e-05,
403
+ "loss": 0.6717,
404
+ "step": 500
405
+ },
406
+ {
407
+ "epoch": 5.994029850746268,
408
+ "eval_accuracy": 0.8433862433862434,
409
+ "eval_loss": 0.49840009212493896,
410
+ "eval_runtime": 22.0815,
411
+ "eval_samples_per_second": 85.592,
412
+ "eval_steps_per_second": 2.717,
413
+ "step": 502
414
+ },
415
+ {
416
+ "epoch": 6.08955223880597,
417
+ "grad_norm": 25.194740295410156,
418
+ "learning_rate": 4.872824631860777e-05,
419
+ "loss": 0.5848,
420
+ "step": 510
421
+ },
422
+ {
423
+ "epoch": 6.208955223880597,
424
+ "grad_norm": 53.04393005371094,
425
+ "learning_rate": 4.8594377510040165e-05,
426
+ "loss": 0.6182,
427
+ "step": 520
428
+ },
429
+ {
430
+ "epoch": 6.3283582089552235,
431
+ "grad_norm": 35.287010192871094,
432
+ "learning_rate": 4.8460508701472554e-05,
433
+ "loss": 0.6342,
434
+ "step": 530
435
+ },
436
+ {
437
+ "epoch": 6.447761194029851,
438
+ "grad_norm": 37.487796783447266,
439
+ "learning_rate": 4.832663989290496e-05,
440
+ "loss": 0.6249,
441
+ "step": 540
442
+ },
443
+ {
444
+ "epoch": 6.567164179104478,
445
+ "grad_norm": 20.985584259033203,
446
+ "learning_rate": 4.8192771084337354e-05,
447
+ "loss": 0.6296,
448
+ "step": 550
449
+ },
450
+ {
451
+ "epoch": 6.686567164179104,
452
+ "grad_norm": 25.937698364257812,
453
+ "learning_rate": 4.8058902275769744e-05,
454
+ "loss": 0.5756,
455
+ "step": 560
456
+ },
457
+ {
458
+ "epoch": 6.8059701492537314,
459
+ "grad_norm": 24.166778564453125,
460
+ "learning_rate": 4.792503346720215e-05,
461
+ "loss": 0.6639,
462
+ "step": 570
463
+ },
464
+ {
465
+ "epoch": 6.925373134328359,
466
+ "grad_norm": 17.558452606201172,
467
+ "learning_rate": 4.779116465863454e-05,
468
+ "loss": 0.6544,
469
+ "step": 580
470
+ },
471
+ {
472
+ "epoch": 6.997014925373135,
473
+ "eval_accuracy": 0.8544973544973545,
474
+ "eval_loss": 0.47304412722587585,
475
+ "eval_runtime": 24.5582,
476
+ "eval_samples_per_second": 76.96,
477
+ "eval_steps_per_second": 2.443,
478
+ "step": 586
479
+ },
480
+ {
481
+ "epoch": 7.044776119402985,
482
+ "grad_norm": 25.491958618164062,
483
+ "learning_rate": 4.765729585006693e-05,
484
+ "loss": 0.5738,
485
+ "step": 590
486
+ },
487
+ {
488
+ "epoch": 7.164179104477612,
489
+ "grad_norm": 22.450450897216797,
490
+ "learning_rate": 4.7523427041499336e-05,
491
+ "loss": 0.5737,
492
+ "step": 600
493
+ },
494
+ {
495
+ "epoch": 7.2835820895522385,
496
+ "grad_norm": 25.659818649291992,
497
+ "learning_rate": 4.738955823293173e-05,
498
+ "loss": 0.5931,
499
+ "step": 610
500
+ },
501
+ {
502
+ "epoch": 7.402985074626866,
503
+ "grad_norm": 32.22180938720703,
504
+ "learning_rate": 4.725568942436412e-05,
505
+ "loss": 0.5595,
506
+ "step": 620
507
+ },
508
+ {
509
+ "epoch": 7.522388059701493,
510
+ "grad_norm": 23.480642318725586,
511
+ "learning_rate": 4.712182061579652e-05,
512
+ "loss": 0.5705,
513
+ "step": 630
514
+ },
515
+ {
516
+ "epoch": 7.641791044776119,
517
+ "grad_norm": 23.681175231933594,
518
+ "learning_rate": 4.698795180722892e-05,
519
+ "loss": 0.5688,
520
+ "step": 640
521
+ },
522
+ {
523
+ "epoch": 7.7611940298507465,
524
+ "grad_norm": 24.127212524414062,
525
+ "learning_rate": 4.685408299866131e-05,
526
+ "loss": 0.606,
527
+ "step": 650
528
+ },
529
+ {
530
+ "epoch": 7.880597014925373,
531
+ "grad_norm": 23.920379638671875,
532
+ "learning_rate": 4.672021419009371e-05,
533
+ "loss": 0.5718,
534
+ "step": 660
535
+ },
536
+ {
537
+ "epoch": 8.0,
538
+ "grad_norm": 30.988821029663086,
539
+ "learning_rate": 4.658634538152611e-05,
540
+ "loss": 0.5846,
541
+ "step": 670
542
+ },
543
+ {
544
+ "epoch": 8.0,
545
+ "eval_accuracy": 0.8629629629629629,
546
+ "eval_loss": 0.4630894362926483,
547
+ "eval_runtime": 22.8566,
548
+ "eval_samples_per_second": 82.69,
549
+ "eval_steps_per_second": 2.625,
550
+ "step": 670
551
+ },
552
+ {
553
+ "epoch": 8.119402985074627,
554
+ "grad_norm": 23.982349395751953,
555
+ "learning_rate": 4.64524765729585e-05,
556
+ "loss": 0.5197,
557
+ "step": 680
558
+ },
559
+ {
560
+ "epoch": 8.238805970149254,
561
+ "grad_norm": 26.285791397094727,
562
+ "learning_rate": 4.63186077643909e-05,
563
+ "loss": 0.5108,
564
+ "step": 690
565
+ },
566
+ {
567
+ "epoch": 8.35820895522388,
568
+ "grad_norm": 18.24249839782715,
569
+ "learning_rate": 4.61847389558233e-05,
570
+ "loss": 0.5251,
571
+ "step": 700
572
+ },
573
+ {
574
+ "epoch": 8.477611940298507,
575
+ "grad_norm": 26.010107040405273,
576
+ "learning_rate": 4.605087014725569e-05,
577
+ "loss": 0.6023,
578
+ "step": 710
579
+ },
580
+ {
581
+ "epoch": 8.597014925373134,
582
+ "grad_norm": 28.695484161376953,
583
+ "learning_rate": 4.5917001338688086e-05,
584
+ "loss": 0.507,
585
+ "step": 720
586
+ },
587
+ {
588
+ "epoch": 8.716417910447761,
589
+ "grad_norm": 17.62828254699707,
590
+ "learning_rate": 4.578313253012048e-05,
591
+ "loss": 0.4988,
592
+ "step": 730
593
+ },
594
+ {
595
+ "epoch": 8.835820895522389,
596
+ "grad_norm": 35.14638900756836,
597
+ "learning_rate": 4.564926372155288e-05,
598
+ "loss": 0.5591,
599
+ "step": 740
600
+ },
601
+ {
602
+ "epoch": 8.955223880597014,
603
+ "grad_norm": 30.526409149169922,
604
+ "learning_rate": 4.5515394912985275e-05,
605
+ "loss": 0.5207,
606
+ "step": 750
607
+ },
608
+ {
609
+ "epoch": 8.991044776119402,
610
+ "eval_accuracy": 0.8751322751322751,
611
+ "eval_loss": 0.4072401523590088,
612
+ "eval_runtime": 21.1198,
613
+ "eval_samples_per_second": 89.489,
614
+ "eval_steps_per_second": 2.841,
615
+ "step": 753
616
+ },
617
+ {
618
+ "epoch": 9.074626865671641,
619
+ "grad_norm": 20.028371810913086,
620
+ "learning_rate": 4.538152610441767e-05,
621
+ "loss": 0.4773,
622
+ "step": 760
623
+ },
624
+ {
625
+ "epoch": 9.194029850746269,
626
+ "grad_norm": 21.046306610107422,
627
+ "learning_rate": 4.524765729585007e-05,
628
+ "loss": 0.5069,
629
+ "step": 770
630
+ },
631
+ {
632
+ "epoch": 9.313432835820896,
633
+ "grad_norm": 21.227153778076172,
634
+ "learning_rate": 4.5113788487282465e-05,
635
+ "loss": 0.462,
636
+ "step": 780
637
+ },
638
+ {
639
+ "epoch": 9.432835820895523,
640
+ "grad_norm": 27.245750427246094,
641
+ "learning_rate": 4.497991967871486e-05,
642
+ "loss": 0.4378,
643
+ "step": 790
644
+ },
645
+ {
646
+ "epoch": 9.552238805970148,
647
+ "grad_norm": 12.378539085388184,
648
+ "learning_rate": 4.484605087014726e-05,
649
+ "loss": 0.4878,
650
+ "step": 800
651
+ },
652
+ {
653
+ "epoch": 9.671641791044776,
654
+ "grad_norm": 19.924983978271484,
655
+ "learning_rate": 4.4712182061579654e-05,
656
+ "loss": 0.5377,
657
+ "step": 810
658
+ },
659
+ {
660
+ "epoch": 9.791044776119403,
661
+ "grad_norm": 31.331254959106445,
662
+ "learning_rate": 4.457831325301205e-05,
663
+ "loss": 0.5306,
664
+ "step": 820
665
+ },
666
+ {
667
+ "epoch": 9.91044776119403,
668
+ "grad_norm": 25.974903106689453,
669
+ "learning_rate": 4.4444444444444447e-05,
670
+ "loss": 0.4977,
671
+ "step": 830
672
+ },
673
+ {
674
+ "epoch": 9.99402985074627,
675
+ "eval_accuracy": 0.8608465608465609,
676
+ "eval_loss": 0.46266716718673706,
677
+ "eval_runtime": 34.1921,
678
+ "eval_samples_per_second": 55.276,
679
+ "eval_steps_per_second": 1.755,
680
+ "step": 837
681
+ },
682
+ {
683
+ "epoch": 10.029850746268657,
684
+ "grad_norm": 26.603260040283203,
685
+ "learning_rate": 4.431057563587684e-05,
686
+ "loss": 0.4794,
687
+ "step": 840
688
+ },
689
+ {
690
+ "epoch": 10.149253731343283,
691
+ "grad_norm": 28.95216178894043,
692
+ "learning_rate": 4.417670682730924e-05,
693
+ "loss": 0.443,
694
+ "step": 850
695
+ },
696
+ {
697
+ "epoch": 10.26865671641791,
698
+ "grad_norm": 16.765554428100586,
699
+ "learning_rate": 4.4042838018741636e-05,
700
+ "loss": 0.4752,
701
+ "step": 860
702
+ },
703
+ {
704
+ "epoch": 10.388059701492537,
705
+ "grad_norm": 25.070423126220703,
706
+ "learning_rate": 4.390896921017403e-05,
707
+ "loss": 0.5519,
708
+ "step": 870
709
+ },
710
+ {
711
+ "epoch": 10.507462686567164,
712
+ "grad_norm": 18.10160255432129,
713
+ "learning_rate": 4.377510040160643e-05,
714
+ "loss": 0.4328,
715
+ "step": 880
716
+ },
717
+ {
718
+ "epoch": 10.626865671641792,
719
+ "grad_norm": 32.88917922973633,
720
+ "learning_rate": 4.3641231593038825e-05,
721
+ "loss": 0.4394,
722
+ "step": 890
723
+ },
724
+ {
725
+ "epoch": 10.746268656716419,
726
+ "grad_norm": 23.60790252685547,
727
+ "learning_rate": 4.350736278447122e-05,
728
+ "loss": 0.5036,
729
+ "step": 900
730
+ },
731
+ {
732
+ "epoch": 10.865671641791044,
733
+ "grad_norm": 18.67464828491211,
734
+ "learning_rate": 4.337349397590362e-05,
735
+ "loss": 0.4576,
736
+ "step": 910
737
+ },
738
+ {
739
+ "epoch": 10.985074626865671,
740
+ "grad_norm": 20.74434471130371,
741
+ "learning_rate": 4.3239625167336014e-05,
742
+ "loss": 0.4974,
743
+ "step": 920
744
+ },
745
+ {
746
+ "epoch": 10.997014925373135,
747
+ "eval_accuracy": 0.8661375661375661,
748
+ "eval_loss": 0.4599558711051941,
749
+ "eval_runtime": 23.4296,
750
+ "eval_samples_per_second": 80.667,
751
+ "eval_steps_per_second": 2.561,
752
+ "step": 921
753
+ },
754
+ {
755
+ "epoch": 11.104477611940299,
756
+ "grad_norm": 17.247478485107422,
757
+ "learning_rate": 4.3105756358768404e-05,
758
+ "loss": 0.4217,
759
+ "step": 930
760
+ },
761
+ {
762
+ "epoch": 11.223880597014926,
763
+ "grad_norm": 24.32207489013672,
764
+ "learning_rate": 4.297188755020081e-05,
765
+ "loss": 0.426,
766
+ "step": 940
767
+ },
768
+ {
769
+ "epoch": 11.343283582089553,
770
+ "grad_norm": 19.61281394958496,
771
+ "learning_rate": 4.2838018741633203e-05,
772
+ "loss": 0.4266,
773
+ "step": 950
774
+ },
775
+ {
776
+ "epoch": 11.462686567164178,
777
+ "grad_norm": 34.41626739501953,
778
+ "learning_rate": 4.270414993306559e-05,
779
+ "loss": 0.4094,
780
+ "step": 960
781
+ },
782
+ {
783
+ "epoch": 11.582089552238806,
784
+ "grad_norm": 17.059947967529297,
785
+ "learning_rate": 4.2570281124497996e-05,
786
+ "loss": 0.4498,
787
+ "step": 970
788
+ },
789
+ {
790
+ "epoch": 11.701492537313433,
791
+ "grad_norm": 14.022907257080078,
792
+ "learning_rate": 4.243641231593039e-05,
793
+ "loss": 0.4787,
794
+ "step": 980
795
+ },
796
+ {
797
+ "epoch": 11.82089552238806,
798
+ "grad_norm": 19.341840744018555,
799
+ "learning_rate": 4.230254350736278e-05,
800
+ "loss": 0.5272,
801
+ "step": 990
802
+ },
803
+ {
804
+ "epoch": 11.940298507462687,
805
+ "grad_norm": 23.444053649902344,
806
+ "learning_rate": 4.2168674698795186e-05,
807
+ "loss": 0.4502,
808
+ "step": 1000
809
+ },
810
+ {
811
+ "epoch": 12.0,
812
+ "eval_accuracy": 0.8724867724867725,
813
+ "eval_loss": 0.45481687784194946,
814
+ "eval_runtime": 21.0538,
815
+ "eval_samples_per_second": 89.77,
816
+ "eval_steps_per_second": 2.85,
817
+ "step": 1005
818
+ },
819
+ {
820
+ "epoch": 12.059701492537313,
821
+ "grad_norm": 14.740572929382324,
822
+ "learning_rate": 4.203480589022758e-05,
823
+ "loss": 0.4053,
824
+ "step": 1010
825
+ },
826
+ {
827
+ "epoch": 12.17910447761194,
828
+ "grad_norm": 21.706453323364258,
829
+ "learning_rate": 4.190093708165997e-05,
830
+ "loss": 0.4259,
831
+ "step": 1020
832
+ },
833
+ {
834
+ "epoch": 12.298507462686567,
835
+ "grad_norm": 17.945667266845703,
836
+ "learning_rate": 4.176706827309237e-05,
837
+ "loss": 0.4064,
838
+ "step": 1030
839
+ },
840
+ {
841
+ "epoch": 12.417910447761194,
842
+ "grad_norm": 18.671659469604492,
843
+ "learning_rate": 4.163319946452477e-05,
844
+ "loss": 0.4299,
845
+ "step": 1040
846
+ },
847
+ {
848
+ "epoch": 12.537313432835822,
849
+ "grad_norm": 19.7445068359375,
850
+ "learning_rate": 4.149933065595716e-05,
851
+ "loss": 0.4439,
852
+ "step": 1050
853
+ },
854
+ {
855
+ "epoch": 12.656716417910447,
856
+ "grad_norm": 20.6423282623291,
857
+ "learning_rate": 4.136546184738956e-05,
858
+ "loss": 0.4782,
859
+ "step": 1060
860
+ },
861
+ {
862
+ "epoch": 12.776119402985074,
863
+ "grad_norm": 20.769142150878906,
864
+ "learning_rate": 4.123159303882196e-05,
865
+ "loss": 0.4127,
866
+ "step": 1070
867
+ },
868
+ {
869
+ "epoch": 12.895522388059701,
870
+ "grad_norm": 31.262954711914062,
871
+ "learning_rate": 4.109772423025435e-05,
872
+ "loss": 0.4051,
873
+ "step": 1080
874
+ },
875
+ {
876
+ "epoch": 12.991044776119402,
877
+ "eval_accuracy": 0.870899470899471,
878
+ "eval_loss": 0.44037628173828125,
879
+ "eval_runtime": 20.5272,
880
+ "eval_samples_per_second": 92.073,
881
+ "eval_steps_per_second": 2.923,
882
+ "step": 1088
883
+ },
884
+ {
885
+ "epoch": 13.014925373134329,
886
+ "grad_norm": 16.262481689453125,
887
+ "learning_rate": 4.0963855421686746e-05,
888
+ "loss": 0.4364,
889
+ "step": 1090
890
+ },
891
+ {
892
+ "epoch": 13.134328358208956,
893
+ "grad_norm": 11.120735168457031,
894
+ "learning_rate": 4.082998661311915e-05,
895
+ "loss": 0.3634,
896
+ "step": 1100
897
+ },
898
+ {
899
+ "epoch": 13.253731343283581,
900
+ "grad_norm": 17.504487991333008,
901
+ "learning_rate": 4.069611780455154e-05,
902
+ "loss": 0.3874,
903
+ "step": 1110
904
+ },
905
+ {
906
+ "epoch": 13.373134328358208,
907
+ "grad_norm": 27.072677612304688,
908
+ "learning_rate": 4.0562248995983936e-05,
909
+ "loss": 0.433,
910
+ "step": 1120
911
+ },
912
+ {
913
+ "epoch": 13.492537313432836,
914
+ "grad_norm": 21.18752670288086,
915
+ "learning_rate": 4.042838018741633e-05,
916
+ "loss": 0.423,
917
+ "step": 1130
918
+ },
919
+ {
920
+ "epoch": 13.611940298507463,
921
+ "grad_norm": 15.781579971313477,
922
+ "learning_rate": 4.029451137884873e-05,
923
+ "loss": 0.4167,
924
+ "step": 1140
925
+ },
926
+ {
927
+ "epoch": 13.73134328358209,
928
+ "grad_norm": 21.992441177368164,
929
+ "learning_rate": 4.0160642570281125e-05,
930
+ "loss": 0.4286,
931
+ "step": 1150
932
+ },
933
+ {
934
+ "epoch": 13.850746268656717,
935
+ "grad_norm": 18.54501724243164,
936
+ "learning_rate": 4.002677376171352e-05,
937
+ "loss": 0.465,
938
+ "step": 1160
939
+ },
940
+ {
941
+ "epoch": 13.970149253731343,
942
+ "grad_norm": 14.206164360046387,
943
+ "learning_rate": 3.989290495314592e-05,
944
+ "loss": 0.3862,
945
+ "step": 1170
946
+ },
947
+ {
948
+ "epoch": 13.99402985074627,
949
+ "eval_accuracy": 0.8772486772486773,
950
+ "eval_loss": 0.4498014450073242,
951
+ "eval_runtime": 21.3493,
952
+ "eval_samples_per_second": 88.528,
953
+ "eval_steps_per_second": 2.81,
954
+ "step": 1172
955
+ },
956
+ {
957
+ "epoch": 14.08955223880597,
958
+ "grad_norm": 16.436885833740234,
959
+ "learning_rate": 3.9759036144578314e-05,
960
+ "loss": 0.4198,
961
+ "step": 1180
962
+ },
963
+ {
964
+ "epoch": 14.208955223880597,
965
+ "grad_norm": 18.03758430480957,
966
+ "learning_rate": 3.962516733601071e-05,
967
+ "loss": 0.3928,
968
+ "step": 1190
969
+ },
970
+ {
971
+ "epoch": 14.328358208955224,
972
+ "grad_norm": 18.285844802856445,
973
+ "learning_rate": 3.949129852744311e-05,
974
+ "loss": 0.41,
975
+ "step": 1200
976
+ },
977
+ {
978
+ "epoch": 14.447761194029852,
979
+ "grad_norm": 25.42701530456543,
980
+ "learning_rate": 3.93574297188755e-05,
981
+ "loss": 0.369,
982
+ "step": 1210
983
+ },
984
+ {
985
+ "epoch": 14.567164179104477,
986
+ "grad_norm": 37.38109588623047,
987
+ "learning_rate": 3.92235609103079e-05,
988
+ "loss": 0.3675,
989
+ "step": 1220
990
+ },
991
+ {
992
+ "epoch": 14.686567164179104,
993
+ "grad_norm": 18.741134643554688,
994
+ "learning_rate": 3.9089692101740296e-05,
995
+ "loss": 0.399,
996
+ "step": 1230
997
+ },
998
+ {
999
+ "epoch": 14.805970149253731,
1000
+ "grad_norm": 13.165870666503906,
1001
+ "learning_rate": 3.895582329317269e-05,
1002
+ "loss": 0.4279,
1003
+ "step": 1240
1004
+ },
1005
+ {
1006
+ "epoch": 14.925373134328359,
1007
+ "grad_norm": 13.853320121765137,
1008
+ "learning_rate": 3.882195448460509e-05,
1009
+ "loss": 0.351,
1010
+ "step": 1250
1011
+ },
1012
+ {
1013
+ "epoch": 14.997014925373135,
1014
+ "eval_accuracy": 0.8677248677248677,
1015
+ "eval_loss": 0.48586639761924744,
1016
+ "eval_runtime": 23.1342,
1017
+ "eval_samples_per_second": 81.697,
1018
+ "eval_steps_per_second": 2.594,
1019
+ "step": 1256
1020
+ },
1021
+ {
1022
+ "epoch": 15.044776119402986,
1023
+ "grad_norm": 17.546924591064453,
1024
+ "learning_rate": 3.8688085676037485e-05,
1025
+ "loss": 0.3519,
1026
+ "step": 1260
1027
+ },
1028
+ {
1029
+ "epoch": 15.164179104477611,
1030
+ "grad_norm": 15.299020767211914,
1031
+ "learning_rate": 3.855421686746988e-05,
1032
+ "loss": 0.3644,
1033
+ "step": 1270
1034
+ },
1035
+ {
1036
+ "epoch": 15.283582089552239,
1037
+ "grad_norm": 27.95186424255371,
1038
+ "learning_rate": 3.842034805890228e-05,
1039
+ "loss": 0.3707,
1040
+ "step": 1280
1041
+ },
1042
+ {
1043
+ "epoch": 15.402985074626866,
1044
+ "grad_norm": 24.1954288482666,
1045
+ "learning_rate": 3.8286479250334675e-05,
1046
+ "loss": 0.3937,
1047
+ "step": 1290
1048
+ },
1049
+ {
1050
+ "epoch": 15.522388059701493,
1051
+ "grad_norm": 15.104287147521973,
1052
+ "learning_rate": 3.815261044176707e-05,
1053
+ "loss": 0.3693,
1054
+ "step": 1300
1055
+ },
1056
+ {
1057
+ "epoch": 15.64179104477612,
1058
+ "grad_norm": 20.633682250976562,
1059
+ "learning_rate": 3.801874163319947e-05,
1060
+ "loss": 0.3565,
1061
+ "step": 1310
1062
+ },
1063
+ {
1064
+ "epoch": 15.761194029850746,
1065
+ "grad_norm": 25.37423324584961,
1066
+ "learning_rate": 3.7884872824631864e-05,
1067
+ "loss": 0.3697,
1068
+ "step": 1320
1069
+ },
1070
+ {
1071
+ "epoch": 15.880597014925373,
1072
+ "grad_norm": 15.787097930908203,
1073
+ "learning_rate": 3.7751004016064253e-05,
1074
+ "loss": 0.3586,
1075
+ "step": 1330
1076
+ },
1077
+ {
1078
+ "epoch": 16.0,
1079
+ "grad_norm": 23.120027542114258,
1080
+ "learning_rate": 3.761713520749666e-05,
1081
+ "loss": 0.3807,
1082
+ "step": 1340
1083
+ },
1084
+ {
1085
+ "epoch": 16.0,
1086
+ "eval_accuracy": 0.8555555555555555,
1087
+ "eval_loss": 0.5188712477684021,
1088
+ "eval_runtime": 25.4894,
1089
+ "eval_samples_per_second": 74.148,
1090
+ "eval_steps_per_second": 2.354,
1091
+ "step": 1340
1092
+ },
1093
+ {
1094
+ "epoch": 16.119402985074625,
1095
+ "grad_norm": 17.341602325439453,
1096
+ "learning_rate": 3.748326639892905e-05,
1097
+ "loss": 0.4134,
1098
+ "step": 1350
1099
+ },
1100
+ {
1101
+ "epoch": 16.238805970149254,
1102
+ "grad_norm": 25.106281280517578,
1103
+ "learning_rate": 3.734939759036144e-05,
1104
+ "loss": 0.3528,
1105
+ "step": 1360
1106
+ },
1107
+ {
1108
+ "epoch": 16.35820895522388,
1109
+ "grad_norm": 12.993470191955566,
1110
+ "learning_rate": 3.7215528781793846e-05,
1111
+ "loss": 0.3435,
1112
+ "step": 1370
1113
+ },
1114
+ {
1115
+ "epoch": 16.47761194029851,
1116
+ "grad_norm": 14.296526908874512,
1117
+ "learning_rate": 3.708165997322624e-05,
1118
+ "loss": 0.3641,
1119
+ "step": 1380
1120
+ },
1121
+ {
1122
+ "epoch": 16.597014925373134,
1123
+ "grad_norm": 17.266536712646484,
1124
+ "learning_rate": 3.694779116465863e-05,
1125
+ "loss": 0.4076,
1126
+ "step": 1390
1127
+ },
1128
+ {
1129
+ "epoch": 16.71641791044776,
1130
+ "grad_norm": 18.806644439697266,
1131
+ "learning_rate": 3.6813922356091035e-05,
1132
+ "loss": 0.312,
1133
+ "step": 1400
1134
+ },
1135
+ {
1136
+ "epoch": 16.83582089552239,
1137
+ "grad_norm": 20.539464950561523,
1138
+ "learning_rate": 3.668005354752343e-05,
1139
+ "loss": 0.3232,
1140
+ "step": 1410
1141
+ },
1142
+ {
1143
+ "epoch": 16.955223880597014,
1144
+ "grad_norm": 22.800634384155273,
1145
+ "learning_rate": 3.654618473895582e-05,
1146
+ "loss": 0.3538,
1147
+ "step": 1420
1148
+ },
1149
+ {
1150
+ "epoch": 16.991044776119402,
1151
+ "eval_accuracy": 0.8645502645502645,
1152
+ "eval_loss": 0.4959017336368561,
1153
+ "eval_runtime": 29.6901,
1154
+ "eval_samples_per_second": 63.658,
1155
+ "eval_steps_per_second": 2.021,
1156
+ "step": 1423
1157
+ },
1158
+ {
1159
+ "epoch": 17.074626865671643,
1160
+ "grad_norm": 28.333608627319336,
1161
+ "learning_rate": 3.641231593038822e-05,
1162
+ "loss": 0.3491,
1163
+ "step": 1430
1164
+ },
1165
+ {
1166
+ "epoch": 17.19402985074627,
1167
+ "grad_norm": 15.66159439086914,
1168
+ "learning_rate": 3.627844712182062e-05,
1169
+ "loss": 0.2808,
1170
+ "step": 1440
1171
+ },
1172
+ {
1173
+ "epoch": 17.313432835820894,
1174
+ "grad_norm": 17.449249267578125,
1175
+ "learning_rate": 3.614457831325301e-05,
1176
+ "loss": 0.3763,
1177
+ "step": 1450
1178
+ },
1179
+ {
1180
+ "epoch": 17.432835820895523,
1181
+ "grad_norm": 16.44668197631836,
1182
+ "learning_rate": 3.601070950468541e-05,
1183
+ "loss": 0.3173,
1184
+ "step": 1460
1185
+ },
1186
+ {
1187
+ "epoch": 17.55223880597015,
1188
+ "grad_norm": 12.903731346130371,
1189
+ "learning_rate": 3.587684069611781e-05,
1190
+ "loss": 0.3423,
1191
+ "step": 1470
1192
+ },
1193
+ {
1194
+ "epoch": 17.671641791044777,
1195
+ "grad_norm": 12.897427558898926,
1196
+ "learning_rate": 3.57429718875502e-05,
1197
+ "loss": 0.3042,
1198
+ "step": 1480
1199
+ },
1200
+ {
1201
+ "epoch": 17.791044776119403,
1202
+ "grad_norm": 21.114595413208008,
1203
+ "learning_rate": 3.5609103078982596e-05,
1204
+ "loss": 0.3684,
1205
+ "step": 1490
1206
+ },
1207
+ {
1208
+ "epoch": 17.91044776119403,
1209
+ "grad_norm": 27.15323829650879,
1210
+ "learning_rate": 3.5475234270415e-05,
1211
+ "loss": 0.3181,
1212
+ "step": 1500
1213
+ },
1214
+ {
1215
+ "epoch": 17.99402985074627,
1216
+ "eval_accuracy": 0.8698412698412699,
1217
+ "eval_loss": 0.48311787843704224,
1218
+ "eval_runtime": 20.9604,
1219
+ "eval_samples_per_second": 90.17,
1220
+ "eval_steps_per_second": 2.863,
1221
+ "step": 1507
1222
+ },
1223
+ {
1224
+ "epoch": 18.029850746268657,
1225
+ "grad_norm": 20.24980926513672,
1226
+ "learning_rate": 3.534136546184739e-05,
1227
+ "loss": 0.3785,
1228
+ "step": 1510
1229
+ },
1230
+ {
1231
+ "epoch": 18.149253731343283,
1232
+ "grad_norm": 20.919986724853516,
1233
+ "learning_rate": 3.5207496653279785e-05,
1234
+ "loss": 0.3465,
1235
+ "step": 1520
1236
+ },
1237
+ {
1238
+ "epoch": 18.26865671641791,
1239
+ "grad_norm": 18.047338485717773,
1240
+ "learning_rate": 3.507362784471218e-05,
1241
+ "loss": 0.3131,
1242
+ "step": 1530
1243
+ },
1244
+ {
1245
+ "epoch": 18.388059701492537,
1246
+ "grad_norm": 18.01116943359375,
1247
+ "learning_rate": 3.4939759036144585e-05,
1248
+ "loss": 0.3251,
1249
+ "step": 1540
1250
+ },
1251
+ {
1252
+ "epoch": 18.507462686567163,
1253
+ "grad_norm": 17.661109924316406,
1254
+ "learning_rate": 3.4805890227576974e-05,
1255
+ "loss": 0.3534,
1256
+ "step": 1550
1257
+ },
1258
+ {
1259
+ "epoch": 18.62686567164179,
1260
+ "grad_norm": 13.190411567687988,
1261
+ "learning_rate": 3.467202141900937e-05,
1262
+ "loss": 0.3443,
1263
+ "step": 1560
1264
+ },
1265
+ {
1266
+ "epoch": 18.746268656716417,
1267
+ "grad_norm": 11.302985191345215,
1268
+ "learning_rate": 3.4538152610441774e-05,
1269
+ "loss": 0.294,
1270
+ "step": 1570
1271
+ },
1272
+ {
1273
+ "epoch": 18.865671641791046,
1274
+ "grad_norm": 17.95700454711914,
1275
+ "learning_rate": 3.4404283801874164e-05,
1276
+ "loss": 0.3354,
1277
+ "step": 1580
1278
+ },
1279
+ {
1280
+ "epoch": 18.98507462686567,
1281
+ "grad_norm": 21.134395599365234,
1282
+ "learning_rate": 3.427041499330656e-05,
1283
+ "loss": 0.3225,
1284
+ "step": 1590
1285
+ },
1286
+ {
1287
+ "epoch": 18.997014925373133,
1288
+ "eval_accuracy": 0.8804232804232804,
1289
+ "eval_loss": 0.4890158772468567,
1290
+ "eval_runtime": 21.2829,
1291
+ "eval_samples_per_second": 88.804,
1292
+ "eval_steps_per_second": 2.819,
1293
+ "step": 1591
1294
+ },
1295
+ {
1296
+ "epoch": 19.104477611940297,
1297
+ "grad_norm": 20.105192184448242,
1298
+ "learning_rate": 3.413654618473896e-05,
1299
+ "loss": 0.2791,
1300
+ "step": 1600
1301
+ },
1302
+ {
1303
+ "epoch": 19.223880597014926,
1304
+ "grad_norm": 19.9129638671875,
1305
+ "learning_rate": 3.400267737617135e-05,
1306
+ "loss": 0.3031,
1307
+ "step": 1610
1308
+ },
1309
+ {
1310
+ "epoch": 19.34328358208955,
1311
+ "grad_norm": 18.273677825927734,
1312
+ "learning_rate": 3.386880856760375e-05,
1313
+ "loss": 0.3341,
1314
+ "step": 1620
1315
+ },
1316
+ {
1317
+ "epoch": 19.46268656716418,
1318
+ "grad_norm": 22.22028350830078,
1319
+ "learning_rate": 3.3734939759036146e-05,
1320
+ "loss": 0.2613,
1321
+ "step": 1630
1322
+ },
1323
+ {
1324
+ "epoch": 19.582089552238806,
1325
+ "grad_norm": 17.404489517211914,
1326
+ "learning_rate": 3.360107095046854e-05,
1327
+ "loss": 0.3275,
1328
+ "step": 1640
1329
+ },
1330
+ {
1331
+ "epoch": 19.701492537313435,
1332
+ "grad_norm": 11.650886535644531,
1333
+ "learning_rate": 3.346720214190094e-05,
1334
+ "loss": 0.3002,
1335
+ "step": 1650
1336
+ },
1337
+ {
1338
+ "epoch": 19.82089552238806,
1339
+ "grad_norm": 21.051326751708984,
1340
+ "learning_rate": 3.3333333333333335e-05,
1341
+ "loss": 0.3075,
1342
+ "step": 1660
1343
+ },
1344
+ {
1345
+ "epoch": 19.940298507462686,
1346
+ "grad_norm": 11.516342163085938,
1347
+ "learning_rate": 3.319946452476573e-05,
1348
+ "loss": 0.3257,
1349
+ "step": 1670
1350
+ },
1351
+ {
1352
+ "epoch": 20.0,
1353
+ "eval_accuracy": 0.8735449735449735,
1354
+ "eval_loss": 0.48166623711586,
1355
+ "eval_runtime": 35.9635,
1356
+ "eval_samples_per_second": 52.553,
1357
+ "eval_steps_per_second": 1.668,
1358
+ "step": 1675
1359
+ },
1360
+ {
1361
+ "epoch": 20.059701492537314,
1362
+ "grad_norm": 21.99034881591797,
1363
+ "learning_rate": 3.306559571619813e-05,
1364
+ "loss": 0.2885,
1365
+ "step": 1680
1366
+ },
1367
+ {
1368
+ "epoch": 20.17910447761194,
1369
+ "grad_norm": 15.522448539733887,
1370
+ "learning_rate": 3.2931726907630524e-05,
1371
+ "loss": 0.2835,
1372
+ "step": 1690
1373
+ },
1374
+ {
1375
+ "epoch": 20.298507462686565,
1376
+ "grad_norm": 16.3590030670166,
1377
+ "learning_rate": 3.279785809906292e-05,
1378
+ "loss": 0.2958,
1379
+ "step": 1700
1380
+ },
1381
+ {
1382
+ "epoch": 20.417910447761194,
1383
+ "grad_norm": 15.87469482421875,
1384
+ "learning_rate": 3.266398929049532e-05,
1385
+ "loss": 0.2831,
1386
+ "step": 1710
1387
+ },
1388
+ {
1389
+ "epoch": 20.53731343283582,
1390
+ "grad_norm": 16.18557357788086,
1391
+ "learning_rate": 3.253012048192771e-05,
1392
+ "loss": 0.2622,
1393
+ "step": 1720
1394
+ },
1395
+ {
1396
+ "epoch": 20.65671641791045,
1397
+ "grad_norm": 15.619478225708008,
1398
+ "learning_rate": 3.23962516733601e-05,
1399
+ "loss": 0.3094,
1400
+ "step": 1730
1401
+ },
1402
+ {
1403
+ "epoch": 20.776119402985074,
1404
+ "grad_norm": 16.444242477416992,
1405
+ "learning_rate": 3.2262382864792506e-05,
1406
+ "loss": 0.3312,
1407
+ "step": 1740
1408
+ },
1409
+ {
1410
+ "epoch": 20.895522388059703,
1411
+ "grad_norm": 14.049259185791016,
1412
+ "learning_rate": 3.21285140562249e-05,
1413
+ "loss": 0.2667,
1414
+ "step": 1750
1415
+ },
1416
+ {
1417
+ "epoch": 20.991044776119402,
1418
+ "eval_accuracy": 0.8682539682539683,
1419
+ "eval_loss": 0.5198962688446045,
1420
+ "eval_runtime": 21.6457,
1421
+ "eval_samples_per_second": 87.315,
1422
+ "eval_steps_per_second": 2.772,
1423
+ "step": 1758
1424
+ },
1425
+ {
1426
+ "epoch": 21.01492537313433,
1427
+ "grad_norm": 25.527427673339844,
1428
+ "learning_rate": 3.199464524765729e-05,
1429
+ "loss": 0.2932,
1430
+ "step": 1760
1431
+ },
1432
+ {
1433
+ "epoch": 21.134328358208954,
1434
+ "grad_norm": 20.6369686126709,
1435
+ "learning_rate": 3.1860776439089695e-05,
1436
+ "loss": 0.3097,
1437
+ "step": 1770
1438
+ },
1439
+ {
1440
+ "epoch": 21.253731343283583,
1441
+ "grad_norm": 17.76145362854004,
1442
+ "learning_rate": 3.172690763052209e-05,
1443
+ "loss": 0.3055,
1444
+ "step": 1780
1445
+ },
1446
+ {
1447
+ "epoch": 21.37313432835821,
1448
+ "grad_norm": 17.357999801635742,
1449
+ "learning_rate": 3.159303882195448e-05,
1450
+ "loss": 0.3102,
1451
+ "step": 1790
1452
+ },
1453
+ {
1454
+ "epoch": 21.492537313432837,
1455
+ "grad_norm": 16.11004638671875,
1456
+ "learning_rate": 3.1459170013386885e-05,
1457
+ "loss": 0.2703,
1458
+ "step": 1800
1459
+ },
1460
+ {
1461
+ "epoch": 21.611940298507463,
1462
+ "grad_norm": 14.16020679473877,
1463
+ "learning_rate": 3.132530120481928e-05,
1464
+ "loss": 0.3295,
1465
+ "step": 1810
1466
+ },
1467
+ {
1468
+ "epoch": 21.73134328358209,
1469
+ "grad_norm": 19.754079818725586,
1470
+ "learning_rate": 3.119143239625168e-05,
1471
+ "loss": 0.2677,
1472
+ "step": 1820
1473
+ },
1474
+ {
1475
+ "epoch": 21.850746268656717,
1476
+ "grad_norm": 15.820847511291504,
1477
+ "learning_rate": 3.105756358768407e-05,
1478
+ "loss": 0.2737,
1479
+ "step": 1830
1480
+ },
1481
+ {
1482
+ "epoch": 21.970149253731343,
1483
+ "grad_norm": 11.689366340637207,
1484
+ "learning_rate": 3.092369477911647e-05,
1485
+ "loss": 0.2863,
1486
+ "step": 1840
1487
+ },
1488
+ {
1489
+ "epoch": 21.99402985074627,
1490
+ "eval_accuracy": 0.8682539682539683,
1491
+ "eval_loss": 0.48346272110939026,
1492
+ "eval_runtime": 21.1384,
1493
+ "eval_samples_per_second": 89.411,
1494
+ "eval_steps_per_second": 2.838,
1495
+ "step": 1842
1496
+ },
1497
+ {
1498
+ "epoch": 22.08955223880597,
1499
+ "grad_norm": 9.886032104492188,
1500
+ "learning_rate": 3.078982597054887e-05,
1501
+ "loss": 0.2935,
1502
+ "step": 1850
1503
+ },
1504
+ {
1505
+ "epoch": 22.208955223880597,
1506
+ "grad_norm": 26.104557037353516,
1507
+ "learning_rate": 3.0655957161981256e-05,
1508
+ "loss": 0.2708,
1509
+ "step": 1860
1510
+ },
1511
+ {
1512
+ "epoch": 22.328358208955223,
1513
+ "grad_norm": 19.244731903076172,
1514
+ "learning_rate": 3.052208835341366e-05,
1515
+ "loss": 0.3206,
1516
+ "step": 1870
1517
+ },
1518
+ {
1519
+ "epoch": 22.44776119402985,
1520
+ "grad_norm": 12.097124099731445,
1521
+ "learning_rate": 3.0388219544846053e-05,
1522
+ "loss": 0.296,
1523
+ "step": 1880
1524
+ },
1525
+ {
1526
+ "epoch": 22.567164179104477,
1527
+ "grad_norm": 22.592735290527344,
1528
+ "learning_rate": 3.025435073627845e-05,
1529
+ "loss": 0.2638,
1530
+ "step": 1890
1531
+ },
1532
+ {
1533
+ "epoch": 22.686567164179106,
1534
+ "grad_norm": 9.944023132324219,
1535
+ "learning_rate": 3.012048192771085e-05,
1536
+ "loss": 0.3055,
1537
+ "step": 1900
1538
+ },
1539
+ {
1540
+ "epoch": 22.80597014925373,
1541
+ "grad_norm": 28.136188507080078,
1542
+ "learning_rate": 2.9986613119143242e-05,
1543
+ "loss": 0.3577,
1544
+ "step": 1910
1545
+ },
1546
+ {
1547
+ "epoch": 22.925373134328357,
1548
+ "grad_norm": 14.673059463500977,
1549
+ "learning_rate": 2.9852744310575638e-05,
1550
+ "loss": 0.2527,
1551
+ "step": 1920
1552
+ },
1553
+ {
1554
+ "epoch": 22.997014925373133,
1555
+ "eval_accuracy": 0.8772486772486773,
1556
+ "eval_loss": 0.4763975739479065,
1557
+ "eval_runtime": 21.7063,
1558
+ "eval_samples_per_second": 87.071,
1559
+ "eval_steps_per_second": 2.764,
1560
+ "step": 1926
1561
+ },
1562
+ {
1563
+ "epoch": 23.044776119402986,
1564
+ "grad_norm": 16.39044189453125,
1565
+ "learning_rate": 2.971887550200803e-05,
1566
+ "loss": 0.2727,
1567
+ "step": 1930
1568
+ },
1569
+ {
1570
+ "epoch": 23.16417910447761,
1571
+ "grad_norm": 15.90278148651123,
1572
+ "learning_rate": 2.958500669344043e-05,
1573
+ "loss": 0.3085,
1574
+ "step": 1940
1575
+ },
1576
+ {
1577
+ "epoch": 23.28358208955224,
1578
+ "grad_norm": 17.350341796875,
1579
+ "learning_rate": 2.9451137884872827e-05,
1580
+ "loss": 0.2484,
1581
+ "step": 1950
1582
+ },
1583
+ {
1584
+ "epoch": 23.402985074626866,
1585
+ "grad_norm": 25.032800674438477,
1586
+ "learning_rate": 2.931726907630522e-05,
1587
+ "loss": 0.2594,
1588
+ "step": 1960
1589
+ },
1590
+ {
1591
+ "epoch": 23.52238805970149,
1592
+ "grad_norm": 16.619842529296875,
1593
+ "learning_rate": 2.918340026773762e-05,
1594
+ "loss": 0.2444,
1595
+ "step": 1970
1596
+ },
1597
+ {
1598
+ "epoch": 23.64179104477612,
1599
+ "grad_norm": 16.149154663085938,
1600
+ "learning_rate": 2.9049531459170017e-05,
1601
+ "loss": 0.2535,
1602
+ "step": 1980
1603
+ },
1604
+ {
1605
+ "epoch": 23.761194029850746,
1606
+ "grad_norm": 14.45046615600586,
1607
+ "learning_rate": 2.891566265060241e-05,
1608
+ "loss": 0.2122,
1609
+ "step": 1990
1610
+ },
1611
+ {
1612
+ "epoch": 23.880597014925375,
1613
+ "grad_norm": 12.895037651062012,
1614
+ "learning_rate": 2.878179384203481e-05,
1615
+ "loss": 0.2739,
1616
+ "step": 2000
1617
+ },
1618
+ {
1619
+ "epoch": 24.0,
1620
+ "grad_norm": 10.731025695800781,
1621
+ "learning_rate": 2.8647925033467206e-05,
1622
+ "loss": 0.2657,
1623
+ "step": 2010
1624
+ },
1625
+ {
1626
+ "epoch": 24.0,
1627
+ "eval_accuracy": 0.8767195767195767,
1628
+ "eval_loss": 0.4650660455226898,
1629
+ "eval_runtime": 21.5221,
1630
+ "eval_samples_per_second": 87.817,
1631
+ "eval_steps_per_second": 2.788,
1632
+ "step": 2010
1633
+ },
1634
+ {
1635
+ "epoch": 24.119402985074625,
1636
+ "grad_norm": 15.251605987548828,
1637
+ "learning_rate": 2.85140562248996e-05,
1638
+ "loss": 0.2214,
1639
+ "step": 2020
1640
+ },
1641
+ {
1642
+ "epoch": 24.238805970149254,
1643
+ "grad_norm": 22.7297420501709,
1644
+ "learning_rate": 2.8380187416331995e-05,
1645
+ "loss": 0.2274,
1646
+ "step": 2030
1647
+ },
1648
+ {
1649
+ "epoch": 24.35820895522388,
1650
+ "grad_norm": 15.054763793945312,
1651
+ "learning_rate": 2.8246318607764395e-05,
1652
+ "loss": 0.2559,
1653
+ "step": 2040
1654
+ },
1655
+ {
1656
+ "epoch": 24.47761194029851,
1657
+ "grad_norm": 11.38253402709961,
1658
+ "learning_rate": 2.8112449799196788e-05,
1659
+ "loss": 0.2383,
1660
+ "step": 2050
1661
+ },
1662
+ {
1663
+ "epoch": 24.597014925373134,
1664
+ "grad_norm": 10.91185188293457,
1665
+ "learning_rate": 2.7978580990629184e-05,
1666
+ "loss": 0.2419,
1667
+ "step": 2060
1668
+ },
1669
+ {
1670
+ "epoch": 24.71641791044776,
1671
+ "grad_norm": 12.653990745544434,
1672
+ "learning_rate": 2.7844712182061584e-05,
1673
+ "loss": 0.2794,
1674
+ "step": 2070
1675
+ },
1676
+ {
1677
+ "epoch": 24.83582089552239,
1678
+ "grad_norm": 16.15665626525879,
1679
+ "learning_rate": 2.7710843373493977e-05,
1680
+ "loss": 0.2554,
1681
+ "step": 2080
1682
+ },
1683
+ {
1684
+ "epoch": 24.955223880597014,
1685
+ "grad_norm": 18.719255447387695,
1686
+ "learning_rate": 2.7576974564926374e-05,
1687
+ "loss": 0.1995,
1688
+ "step": 2090
1689
+ },
1690
+ {
1691
+ "epoch": 24.991044776119402,
1692
+ "eval_accuracy": 0.8693121693121693,
1693
+ "eval_loss": 0.5079193711280823,
1694
+ "eval_runtime": 22.8272,
1695
+ "eval_samples_per_second": 82.796,
1696
+ "eval_steps_per_second": 2.628,
1697
+ "step": 2093
1698
+ },
1699
+ {
1700
+ "epoch": 25.074626865671643,
1701
+ "grad_norm": 18.27308464050293,
1702
+ "learning_rate": 2.7443105756358774e-05,
1703
+ "loss": 0.2431,
1704
+ "step": 2100
1705
+ },
1706
+ {
1707
+ "epoch": 25.19402985074627,
1708
+ "grad_norm": 13.238642692565918,
1709
+ "learning_rate": 2.7309236947791167e-05,
1710
+ "loss": 0.2215,
1711
+ "step": 2110
1712
+ },
1713
+ {
1714
+ "epoch": 25.313432835820894,
1715
+ "grad_norm": 19.07526397705078,
1716
+ "learning_rate": 2.7175368139223563e-05,
1717
+ "loss": 0.2356,
1718
+ "step": 2120
1719
+ },
1720
+ {
1721
+ "epoch": 25.432835820895523,
1722
+ "grad_norm": 23.106143951416016,
1723
+ "learning_rate": 2.7041499330655956e-05,
1724
+ "loss": 0.2428,
1725
+ "step": 2130
1726
+ },
1727
+ {
1728
+ "epoch": 25.55223880597015,
1729
+ "grad_norm": 16.487586975097656,
1730
+ "learning_rate": 2.6907630522088356e-05,
1731
+ "loss": 0.2199,
1732
+ "step": 2140
1733
+ },
1734
+ {
1735
+ "epoch": 25.671641791044777,
1736
+ "grad_norm": 18.651390075683594,
1737
+ "learning_rate": 2.6773761713520752e-05,
1738
+ "loss": 0.2386,
1739
+ "step": 2150
1740
+ },
1741
+ {
1742
+ "epoch": 25.791044776119403,
1743
+ "grad_norm": 15.018671989440918,
1744
+ "learning_rate": 2.6639892904953145e-05,
1745
+ "loss": 0.2338,
1746
+ "step": 2160
1747
+ },
1748
+ {
1749
+ "epoch": 25.91044776119403,
1750
+ "grad_norm": 15.358821868896484,
1751
+ "learning_rate": 2.6506024096385545e-05,
1752
+ "loss": 0.2481,
1753
+ "step": 2170
1754
+ },
1755
+ {
1756
+ "epoch": 25.99402985074627,
1757
+ "eval_accuracy": 0.8698412698412699,
1758
+ "eval_loss": 0.5112302899360657,
1759
+ "eval_runtime": 21.5553,
1760
+ "eval_samples_per_second": 87.682,
1761
+ "eval_steps_per_second": 2.784,
1762
+ "step": 2177
1763
+ },
1764
+ {
1765
+ "epoch": 26.029850746268657,
1766
+ "grad_norm": 15.894144058227539,
1767
+ "learning_rate": 2.637215528781794e-05,
1768
+ "loss": 0.2625,
1769
+ "step": 2180
1770
+ },
1771
+ {
1772
+ "epoch": 26.149253731343283,
1773
+ "grad_norm": 16.53984260559082,
1774
+ "learning_rate": 2.6238286479250334e-05,
1775
+ "loss": 0.2675,
1776
+ "step": 2190
1777
+ },
1778
+ {
1779
+ "epoch": 26.26865671641791,
1780
+ "grad_norm": 16.194046020507812,
1781
+ "learning_rate": 2.6104417670682734e-05,
1782
+ "loss": 0.2586,
1783
+ "step": 2200
1784
+ },
1785
+ {
1786
+ "epoch": 26.388059701492537,
1787
+ "grad_norm": 11.86839771270752,
1788
+ "learning_rate": 2.597054886211513e-05,
1789
+ "loss": 0.237,
1790
+ "step": 2210
1791
+ },
1792
+ {
1793
+ "epoch": 26.507462686567163,
1794
+ "grad_norm": 10.908110618591309,
1795
+ "learning_rate": 2.5836680053547524e-05,
1796
+ "loss": 0.2136,
1797
+ "step": 2220
1798
+ },
1799
+ {
1800
+ "epoch": 26.62686567164179,
1801
+ "grad_norm": 13.298651695251465,
1802
+ "learning_rate": 2.570281124497992e-05,
1803
+ "loss": 0.2459,
1804
+ "step": 2230
1805
+ },
1806
+ {
1807
+ "epoch": 26.746268656716417,
1808
+ "grad_norm": 13.250383377075195,
1809
+ "learning_rate": 2.556894243641232e-05,
1810
+ "loss": 0.245,
1811
+ "step": 2240
1812
+ },
1813
+ {
1814
+ "epoch": 26.865671641791046,
1815
+ "grad_norm": 11.171323776245117,
1816
+ "learning_rate": 2.5435073627844713e-05,
1817
+ "loss": 0.2943,
1818
+ "step": 2250
1819
+ },
1820
+ {
1821
+ "epoch": 26.98507462686567,
1822
+ "grad_norm": 19.93524932861328,
1823
+ "learning_rate": 2.530120481927711e-05,
1824
+ "loss": 0.2072,
1825
+ "step": 2260
1826
+ },
1827
+ {
1828
+ "epoch": 26.997014925373133,
1829
+ "eval_accuracy": 0.8830687830687831,
1830
+ "eval_loss": 0.5082184076309204,
1831
+ "eval_runtime": 21.764,
1832
+ "eval_samples_per_second": 86.841,
1833
+ "eval_steps_per_second": 2.757,
1834
+ "step": 2261
1835
+ },
1836
+ {
1837
+ "epoch": 27.104477611940297,
1838
+ "grad_norm": 9.708877563476562,
1839
+ "learning_rate": 2.516733601070951e-05,
1840
+ "loss": 0.2333,
1841
+ "step": 2270
1842
+ },
1843
+ {
1844
+ "epoch": 27.223880597014926,
1845
+ "grad_norm": 17.46422004699707,
1846
+ "learning_rate": 2.5033467202141902e-05,
1847
+ "loss": 0.2399,
1848
+ "step": 2280
1849
+ },
1850
+ {
1851
+ "epoch": 27.34328358208955,
1852
+ "grad_norm": 15.301572799682617,
1853
+ "learning_rate": 2.48995983935743e-05,
1854
+ "loss": 0.2569,
1855
+ "step": 2290
1856
+ },
1857
+ {
1858
+ "epoch": 27.46268656716418,
1859
+ "grad_norm": 10.849388122558594,
1860
+ "learning_rate": 2.4765729585006695e-05,
1861
+ "loss": 0.2426,
1862
+ "step": 2300
1863
+ },
1864
+ {
1865
+ "epoch": 27.582089552238806,
1866
+ "grad_norm": 9.713058471679688,
1867
+ "learning_rate": 2.463186077643909e-05,
1868
+ "loss": 0.2205,
1869
+ "step": 2310
1870
+ },
1871
+ {
1872
+ "epoch": 27.701492537313435,
1873
+ "grad_norm": 11.311200141906738,
1874
+ "learning_rate": 2.4497991967871488e-05,
1875
+ "loss": 0.2402,
1876
+ "step": 2320
1877
+ },
1878
+ {
1879
+ "epoch": 27.82089552238806,
1880
+ "grad_norm": 16.44293785095215,
1881
+ "learning_rate": 2.4364123159303884e-05,
1882
+ "loss": 0.2145,
1883
+ "step": 2330
1884
+ },
1885
+ {
1886
+ "epoch": 27.940298507462686,
1887
+ "grad_norm": 12.815253257751465,
1888
+ "learning_rate": 2.4230254350736277e-05,
1889
+ "loss": 0.2164,
1890
+ "step": 2340
1891
+ },
1892
+ {
1893
+ "epoch": 28.0,
1894
+ "eval_accuracy": 0.873015873015873,
1895
+ "eval_loss": 0.5001527070999146,
1896
+ "eval_runtime": 24.2145,
1897
+ "eval_samples_per_second": 78.052,
1898
+ "eval_steps_per_second": 2.478,
1899
+ "step": 2345
1900
+ },
1901
+ {
1902
+ "epoch": 28.059701492537314,
1903
+ "grad_norm": 16.615896224975586,
1904
+ "learning_rate": 2.4096385542168677e-05,
1905
+ "loss": 0.2217,
1906
+ "step": 2350
1907
+ },
1908
+ {
1909
+ "epoch": 28.17910447761194,
1910
+ "grad_norm": 11.069056510925293,
1911
+ "learning_rate": 2.3962516733601073e-05,
1912
+ "loss": 0.1978,
1913
+ "step": 2360
1914
+ },
1915
+ {
1916
+ "epoch": 28.298507462686565,
1917
+ "grad_norm": 17.405532836914062,
1918
+ "learning_rate": 2.3828647925033466e-05,
1919
+ "loss": 0.2542,
1920
+ "step": 2370
1921
+ },
1922
+ {
1923
+ "epoch": 28.417910447761194,
1924
+ "grad_norm": 13.685500144958496,
1925
+ "learning_rate": 2.3694779116465866e-05,
1926
+ "loss": 0.2102,
1927
+ "step": 2380
1928
+ },
1929
+ {
1930
+ "epoch": 28.53731343283582,
1931
+ "grad_norm": 14.143411636352539,
1932
+ "learning_rate": 2.356091030789826e-05,
1933
+ "loss": 0.2151,
1934
+ "step": 2390
1935
+ },
1936
+ {
1937
+ "epoch": 28.65671641791045,
1938
+ "grad_norm": 11.186328887939453,
1939
+ "learning_rate": 2.3427041499330656e-05,
1940
+ "loss": 0.2109,
1941
+ "step": 2400
1942
+ },
1943
+ {
1944
+ "epoch": 28.776119402985074,
1945
+ "grad_norm": 11.518386840820312,
1946
+ "learning_rate": 2.3293172690763055e-05,
1947
+ "loss": 0.2041,
1948
+ "step": 2410
1949
+ },
1950
+ {
1951
+ "epoch": 28.895522388059703,
1952
+ "grad_norm": 14.913932800292969,
1953
+ "learning_rate": 2.315930388219545e-05,
1954
+ "loss": 0.2198,
1955
+ "step": 2420
1956
+ },
1957
+ {
1958
+ "epoch": 28.991044776119402,
1959
+ "eval_accuracy": 0.8777777777777778,
1960
+ "eval_loss": 0.47848638892173767,
1961
+ "eval_runtime": 23.896,
1962
+ "eval_samples_per_second": 79.093,
1963
+ "eval_steps_per_second": 2.511,
1964
+ "step": 2428
1965
+ },
1966
+ {
1967
+ "epoch": 29.01492537313433,
1968
+ "grad_norm": 13.153636932373047,
1969
+ "learning_rate": 2.3025435073627845e-05,
1970
+ "loss": 0.2036,
1971
+ "step": 2430
1972
+ },
1973
+ {
1974
+ "epoch": 29.134328358208954,
1975
+ "grad_norm": 11.16688346862793,
1976
+ "learning_rate": 2.289156626506024e-05,
1977
+ "loss": 0.1935,
1978
+ "step": 2440
1979
+ },
1980
+ {
1981
+ "epoch": 29.253731343283583,
1982
+ "grad_norm": 14.297533988952637,
1983
+ "learning_rate": 2.2757697456492638e-05,
1984
+ "loss": 0.2095,
1985
+ "step": 2450
1986
+ },
1987
+ {
1988
+ "epoch": 29.37313432835821,
1989
+ "grad_norm": 15.996660232543945,
1990
+ "learning_rate": 2.2623828647925034e-05,
1991
+ "loss": 0.1946,
1992
+ "step": 2460
1993
+ },
1994
+ {
1995
+ "epoch": 29.492537313432837,
1996
+ "grad_norm": 22.007244110107422,
1997
+ "learning_rate": 2.248995983935743e-05,
1998
+ "loss": 0.2171,
1999
+ "step": 2470
2000
+ },
2001
+ {
2002
+ "epoch": 29.611940298507463,
2003
+ "grad_norm": 19.894161224365234,
2004
+ "learning_rate": 2.2356091030789827e-05,
2005
+ "loss": 0.2542,
2006
+ "step": 2480
2007
+ },
2008
+ {
2009
+ "epoch": 29.73134328358209,
2010
+ "grad_norm": 13.56891918182373,
2011
+ "learning_rate": 2.2222222222222223e-05,
2012
+ "loss": 0.2392,
2013
+ "step": 2490
2014
+ },
2015
+ {
2016
+ "epoch": 29.850746268656717,
2017
+ "grad_norm": 15.41037368774414,
2018
+ "learning_rate": 2.208835341365462e-05,
2019
+ "loss": 0.1737,
2020
+ "step": 2500
2021
+ },
2022
+ {
2023
+ "epoch": 29.970149253731343,
2024
+ "grad_norm": 12.46347713470459,
2025
+ "learning_rate": 2.1954484605087016e-05,
2026
+ "loss": 0.2137,
2027
+ "step": 2510
2028
+ },
2029
+ {
2030
+ "epoch": 29.99402985074627,
2031
+ "eval_accuracy": 0.8888888888888888,
2032
+ "eval_loss": 0.5012401342391968,
2033
+ "eval_runtime": 24.7826,
2034
+ "eval_samples_per_second": 76.263,
2035
+ "eval_steps_per_second": 2.421,
2036
+ "step": 2512
2037
+ },
2038
+ {
2039
+ "epoch": 30.08955223880597,
2040
+ "grad_norm": 12.577794075012207,
2041
+ "learning_rate": 2.1820615796519413e-05,
2042
+ "loss": 0.1764,
2043
+ "step": 2520
2044
+ },
2045
+ {
2046
+ "epoch": 30.208955223880597,
2047
+ "grad_norm": 10.100226402282715,
2048
+ "learning_rate": 2.168674698795181e-05,
2049
+ "loss": 0.2085,
2050
+ "step": 2530
2051
+ },
2052
+ {
2053
+ "epoch": 30.328358208955223,
2054
+ "grad_norm": 7.882339000701904,
2055
+ "learning_rate": 2.1552878179384202e-05,
2056
+ "loss": 0.217,
2057
+ "step": 2540
2058
+ },
2059
+ {
2060
+ "epoch": 30.44776119402985,
2061
+ "grad_norm": 9.06692886352539,
2062
+ "learning_rate": 2.1419009370816602e-05,
2063
+ "loss": 0.2082,
2064
+ "step": 2550
2065
+ },
2066
+ {
2067
+ "epoch": 30.567164179104477,
2068
+ "grad_norm": 24.43864631652832,
2069
+ "learning_rate": 2.1285140562248998e-05,
2070
+ "loss": 0.2089,
2071
+ "step": 2560
2072
+ },
2073
+ {
2074
+ "epoch": 30.686567164179106,
2075
+ "grad_norm": 15.762516021728516,
2076
+ "learning_rate": 2.115127175368139e-05,
2077
+ "loss": 0.176,
2078
+ "step": 2570
2079
+ },
2080
+ {
2081
+ "epoch": 30.80597014925373,
2082
+ "grad_norm": 13.477516174316406,
2083
+ "learning_rate": 2.101740294511379e-05,
2084
+ "loss": 0.2034,
2085
+ "step": 2580
2086
+ },
2087
+ {
2088
+ "epoch": 30.925373134328357,
2089
+ "grad_norm": 8.844980239868164,
2090
+ "learning_rate": 2.0883534136546184e-05,
2091
+ "loss": 0.1936,
2092
+ "step": 2590
2093
+ },
2094
+ {
2095
+ "epoch": 30.997014925373133,
2096
+ "eval_accuracy": 0.8756613756613757,
2097
+ "eval_loss": 0.4961310923099518,
2098
+ "eval_runtime": 23.3753,
2099
+ "eval_samples_per_second": 80.855,
2100
+ "eval_steps_per_second": 2.567,
2101
+ "step": 2596
2102
+ },
2103
+ {
2104
+ "epoch": 31.044776119402986,
2105
+ "grad_norm": 9.131314277648926,
2106
+ "learning_rate": 2.074966532797858e-05,
2107
+ "loss": 0.1943,
2108
+ "step": 2600
2109
+ },
2110
+ {
2111
+ "epoch": 31.16417910447761,
2112
+ "grad_norm": 15.152297019958496,
2113
+ "learning_rate": 2.061579651941098e-05,
2114
+ "loss": 0.23,
2115
+ "step": 2610
2116
+ },
2117
+ {
2118
+ "epoch": 31.28358208955224,
2119
+ "grad_norm": 13.151653289794922,
2120
+ "learning_rate": 2.0481927710843373e-05,
2121
+ "loss": 0.1769,
2122
+ "step": 2620
2123
+ },
2124
+ {
2125
+ "epoch": 31.402985074626866,
2126
+ "grad_norm": 12.334269523620605,
2127
+ "learning_rate": 2.034805890227577e-05,
2128
+ "loss": 0.1562,
2129
+ "step": 2630
2130
+ },
2131
+ {
2132
+ "epoch": 31.52238805970149,
2133
+ "grad_norm": 26.206056594848633,
2134
+ "learning_rate": 2.0214190093708166e-05,
2135
+ "loss": 0.1864,
2136
+ "step": 2640
2137
+ },
2138
+ {
2139
+ "epoch": 31.64179104477612,
2140
+ "grad_norm": 13.51135540008545,
2141
+ "learning_rate": 2.0080321285140562e-05,
2142
+ "loss": 0.1935,
2143
+ "step": 2650
2144
+ },
2145
+ {
2146
+ "epoch": 31.761194029850746,
2147
+ "grad_norm": 18.67013931274414,
2148
+ "learning_rate": 1.994645247657296e-05,
2149
+ "loss": 0.1884,
2150
+ "step": 2660
2151
+ },
2152
+ {
2153
+ "epoch": 31.880597014925375,
2154
+ "grad_norm": 14.84253978729248,
2155
+ "learning_rate": 1.9812583668005355e-05,
2156
+ "loss": 0.1825,
2157
+ "step": 2670
2158
+ },
2159
+ {
2160
+ "epoch": 32.0,
2161
+ "grad_norm": 11.984930038452148,
2162
+ "learning_rate": 1.967871485943775e-05,
2163
+ "loss": 0.2255,
2164
+ "step": 2680
2165
+ },
2166
+ {
2167
+ "epoch": 32.0,
2168
+ "eval_accuracy": 0.8788359788359789,
2169
+ "eval_loss": 0.4986724257469177,
2170
+ "eval_runtime": 21.5574,
2171
+ "eval_samples_per_second": 87.673,
2172
+ "eval_steps_per_second": 2.783,
2173
+ "step": 2680
2174
+ },
2175
+ {
2176
+ "epoch": 32.11940298507463,
2177
+ "grad_norm": 8.543514251708984,
2178
+ "learning_rate": 1.9544846050870148e-05,
2179
+ "loss": 0.1831,
2180
+ "step": 2690
2181
+ },
2182
+ {
2183
+ "epoch": 32.23880597014925,
2184
+ "grad_norm": 16.21164894104004,
2185
+ "learning_rate": 1.9410977242302544e-05,
2186
+ "loss": 0.1642,
2187
+ "step": 2700
2188
+ },
2189
+ {
2190
+ "epoch": 32.35820895522388,
2191
+ "grad_norm": 32.741031646728516,
2192
+ "learning_rate": 1.927710843373494e-05,
2193
+ "loss": 0.2028,
2194
+ "step": 2710
2195
+ },
2196
+ {
2197
+ "epoch": 32.47761194029851,
2198
+ "grad_norm": 11.744430541992188,
2199
+ "learning_rate": 1.9143239625167337e-05,
2200
+ "loss": 0.182,
2201
+ "step": 2720
2202
+ },
2203
+ {
2204
+ "epoch": 32.59701492537313,
2205
+ "grad_norm": 16.407419204711914,
2206
+ "learning_rate": 1.9009370816599734e-05,
2207
+ "loss": 0.2066,
2208
+ "step": 2730
2209
+ },
2210
+ {
2211
+ "epoch": 32.71641791044776,
2212
+ "grad_norm": 13.327704429626465,
2213
+ "learning_rate": 1.8875502008032127e-05,
2214
+ "loss": 0.1831,
2215
+ "step": 2740
2216
+ },
2217
+ {
2218
+ "epoch": 32.83582089552239,
2219
+ "grad_norm": 14.911913871765137,
2220
+ "learning_rate": 1.8741633199464527e-05,
2221
+ "loss": 0.1994,
2222
+ "step": 2750
2223
+ },
2224
+ {
2225
+ "epoch": 32.95522388059702,
2226
+ "grad_norm": 11.125310897827148,
2227
+ "learning_rate": 1.8607764390896923e-05,
2228
+ "loss": 0.1818,
2229
+ "step": 2760
2230
+ },
2231
+ {
2232
+ "epoch": 32.9910447761194,
2233
+ "eval_accuracy": 0.8851851851851852,
2234
+ "eval_loss": 0.48399239778518677,
2235
+ "eval_runtime": 21.334,
2236
+ "eval_samples_per_second": 88.591,
2237
+ "eval_steps_per_second": 2.812,
2238
+ "step": 2763
2239
+ },
2240
+ {
2241
+ "epoch": 33.07462686567164,
2242
+ "grad_norm": 12.510457038879395,
2243
+ "learning_rate": 1.8473895582329316e-05,
2244
+ "loss": 0.1833,
2245
+ "step": 2770
2246
+ },
2247
+ {
2248
+ "epoch": 33.19402985074627,
2249
+ "grad_norm": 15.175230979919434,
2250
+ "learning_rate": 1.8340026773761716e-05,
2251
+ "loss": 0.1788,
2252
+ "step": 2780
2253
+ },
2254
+ {
2255
+ "epoch": 33.3134328358209,
2256
+ "grad_norm": 8.436702728271484,
2257
+ "learning_rate": 1.820615796519411e-05,
2258
+ "loss": 0.1814,
2259
+ "step": 2790
2260
+ },
2261
+ {
2262
+ "epoch": 33.43283582089552,
2263
+ "grad_norm": 11.283780097961426,
2264
+ "learning_rate": 1.8072289156626505e-05,
2265
+ "loss": 0.1644,
2266
+ "step": 2800
2267
+ },
2268
+ {
2269
+ "epoch": 33.55223880597015,
2270
+ "grad_norm": 8.91975212097168,
2271
+ "learning_rate": 1.7938420348058905e-05,
2272
+ "loss": 0.1648,
2273
+ "step": 2810
2274
+ },
2275
+ {
2276
+ "epoch": 33.67164179104478,
2277
+ "grad_norm": 9.153727531433105,
2278
+ "learning_rate": 1.7804551539491298e-05,
2279
+ "loss": 0.1749,
2280
+ "step": 2820
2281
+ },
2282
+ {
2283
+ "epoch": 33.791044776119406,
2284
+ "grad_norm": 11.47763442993164,
2285
+ "learning_rate": 1.7670682730923694e-05,
2286
+ "loss": 0.1978,
2287
+ "step": 2830
2288
+ },
2289
+ {
2290
+ "epoch": 33.91044776119403,
2291
+ "grad_norm": 15.642449378967285,
2292
+ "learning_rate": 1.753681392235609e-05,
2293
+ "loss": 0.1644,
2294
+ "step": 2840
2295
+ },
2296
+ {
2297
+ "epoch": 33.994029850746266,
2298
+ "eval_accuracy": 0.8862433862433863,
2299
+ "eval_loss": 0.46936145424842834,
2300
+ "eval_runtime": 22.5625,
2301
+ "eval_samples_per_second": 83.767,
2302
+ "eval_steps_per_second": 2.659,
2303
+ "step": 2847
2304
+ },
2305
+ {
2306
+ "epoch": 34.02985074626866,
2307
+ "grad_norm": 11.902774810791016,
2308
+ "learning_rate": 1.7402945113788487e-05,
2309
+ "loss": 0.1615,
2310
+ "step": 2850
2311
+ },
2312
+ {
2313
+ "epoch": 34.149253731343286,
2314
+ "grad_norm": 35.593807220458984,
2315
+ "learning_rate": 1.7269076305220887e-05,
2316
+ "loss": 0.1855,
2317
+ "step": 2860
2318
+ },
2319
+ {
2320
+ "epoch": 34.26865671641791,
2321
+ "grad_norm": 15.615717887878418,
2322
+ "learning_rate": 1.713520749665328e-05,
2323
+ "loss": 0.182,
2324
+ "step": 2870
2325
+ },
2326
+ {
2327
+ "epoch": 34.38805970149254,
2328
+ "grad_norm": 13.55231761932373,
2329
+ "learning_rate": 1.7001338688085676e-05,
2330
+ "loss": 0.2104,
2331
+ "step": 2880
2332
+ },
2333
+ {
2334
+ "epoch": 34.507462686567166,
2335
+ "grad_norm": 5.884557723999023,
2336
+ "learning_rate": 1.6867469879518073e-05,
2337
+ "loss": 0.1501,
2338
+ "step": 2890
2339
+ },
2340
+ {
2341
+ "epoch": 34.62686567164179,
2342
+ "grad_norm": 15.07630729675293,
2343
+ "learning_rate": 1.673360107095047e-05,
2344
+ "loss": 0.1585,
2345
+ "step": 2900
2346
+ },
2347
+ {
2348
+ "epoch": 34.74626865671642,
2349
+ "grad_norm": 9.989094734191895,
2350
+ "learning_rate": 1.6599732262382866e-05,
2351
+ "loss": 0.144,
2352
+ "step": 2910
2353
+ },
2354
+ {
2355
+ "epoch": 34.865671641791046,
2356
+ "grad_norm": 11.948739051818848,
2357
+ "learning_rate": 1.6465863453815262e-05,
2358
+ "loss": 0.1658,
2359
+ "step": 2920
2360
+ },
2361
+ {
2362
+ "epoch": 34.985074626865675,
2363
+ "grad_norm": 13.029720306396484,
2364
+ "learning_rate": 1.633199464524766e-05,
2365
+ "loss": 0.1799,
2366
+ "step": 2930
2367
+ },
2368
+ {
2369
+ "epoch": 34.997014925373136,
2370
+ "eval_accuracy": 0.8915343915343915,
2371
+ "eval_loss": 0.4599432647228241,
2372
+ "eval_runtime": 21.3881,
2373
+ "eval_samples_per_second": 88.367,
2374
+ "eval_steps_per_second": 2.805,
2375
+ "step": 2931
2376
+ },
2377
+ {
2378
+ "epoch": 35.1044776119403,
2379
+ "grad_norm": 12.619296073913574,
2380
+ "learning_rate": 1.619812583668005e-05,
2381
+ "loss": 0.1677,
2382
+ "step": 2940
2383
+ },
2384
+ {
2385
+ "epoch": 35.223880597014926,
2386
+ "grad_norm": 13.858939170837402,
2387
+ "learning_rate": 1.606425702811245e-05,
2388
+ "loss": 0.175,
2389
+ "step": 2950
2390
+ },
2391
+ {
2392
+ "epoch": 35.343283582089555,
2393
+ "grad_norm": 11.862871170043945,
2394
+ "learning_rate": 1.5930388219544848e-05,
2395
+ "loss": 0.1736,
2396
+ "step": 2960
2397
+ },
2398
+ {
2399
+ "epoch": 35.46268656716418,
2400
+ "grad_norm": 7.9211859703063965,
2401
+ "learning_rate": 1.579651941097724e-05,
2402
+ "loss": 0.1585,
2403
+ "step": 2970
2404
+ },
2405
+ {
2406
+ "epoch": 35.582089552238806,
2407
+ "grad_norm": 14.684250831604004,
2408
+ "learning_rate": 1.566265060240964e-05,
2409
+ "loss": 0.1576,
2410
+ "step": 2980
2411
+ },
2412
+ {
2413
+ "epoch": 35.701492537313435,
2414
+ "grad_norm": 8.843223571777344,
2415
+ "learning_rate": 1.5528781793842034e-05,
2416
+ "loss": 0.1594,
2417
+ "step": 2990
2418
+ },
2419
+ {
2420
+ "epoch": 35.82089552238806,
2421
+ "grad_norm": 9.077608108520508,
2422
+ "learning_rate": 1.5394912985274433e-05,
2423
+ "loss": 0.1745,
2424
+ "step": 3000
2425
+ },
2426
+ {
2427
+ "epoch": 35.940298507462686,
2428
+ "grad_norm": 14.813329696655273,
2429
+ "learning_rate": 1.526104417670683e-05,
2430
+ "loss": 0.1624,
2431
+ "step": 3010
2432
+ },
2433
+ {
2434
+ "epoch": 36.0,
2435
+ "eval_accuracy": 0.8851851851851852,
2436
+ "eval_loss": 0.5122030377388,
2437
+ "eval_runtime": 22.4119,
2438
+ "eval_samples_per_second": 84.33,
2439
+ "eval_steps_per_second": 2.677,
2440
+ "step": 3015
2441
+ },
2442
+ {
2443
+ "epoch": 36.059701492537314,
2444
+ "grad_norm": 17.36454200744629,
2445
+ "learning_rate": 1.5127175368139224e-05,
2446
+ "loss": 0.1414,
2447
+ "step": 3020
2448
+ },
2449
+ {
2450
+ "epoch": 36.17910447761194,
2451
+ "grad_norm": 9.699509620666504,
2452
+ "learning_rate": 1.4993306559571621e-05,
2453
+ "loss": 0.15,
2454
+ "step": 3030
2455
+ },
2456
+ {
2457
+ "epoch": 36.298507462686565,
2458
+ "grad_norm": 8.813672065734863,
2459
+ "learning_rate": 1.4859437751004016e-05,
2460
+ "loss": 0.1924,
2461
+ "step": 3040
2462
+ },
2463
+ {
2464
+ "epoch": 36.417910447761194,
2465
+ "grad_norm": 12.653448104858398,
2466
+ "learning_rate": 1.4725568942436414e-05,
2467
+ "loss": 0.1925,
2468
+ "step": 3050
2469
+ },
2470
+ {
2471
+ "epoch": 36.53731343283582,
2472
+ "grad_norm": 9.532696723937988,
2473
+ "learning_rate": 1.459170013386881e-05,
2474
+ "loss": 0.1822,
2475
+ "step": 3060
2476
+ },
2477
+ {
2478
+ "epoch": 36.656716417910445,
2479
+ "grad_norm": 9.864669799804688,
2480
+ "learning_rate": 1.4457831325301205e-05,
2481
+ "loss": 0.1688,
2482
+ "step": 3070
2483
+ },
2484
+ {
2485
+ "epoch": 36.776119402985074,
2486
+ "grad_norm": 13.674315452575684,
2487
+ "learning_rate": 1.4323962516733603e-05,
2488
+ "loss": 0.1568,
2489
+ "step": 3080
2490
+ },
2491
+ {
2492
+ "epoch": 36.8955223880597,
2493
+ "grad_norm": 8.822213172912598,
2494
+ "learning_rate": 1.4190093708165998e-05,
2495
+ "loss": 0.157,
2496
+ "step": 3090
2497
+ },
2498
+ {
2499
+ "epoch": 36.9910447761194,
2500
+ "eval_accuracy": 0.8899470899470899,
2501
+ "eval_loss": 0.4546496570110321,
2502
+ "eval_runtime": 23.0569,
2503
+ "eval_samples_per_second": 81.971,
2504
+ "eval_steps_per_second": 2.602,
2505
+ "step": 3098
2506
+ },
2507
+ {
2508
+ "epoch": 37.014925373134325,
2509
+ "grad_norm": 13.846744537353516,
2510
+ "learning_rate": 1.4056224899598394e-05,
2511
+ "loss": 0.2011,
2512
+ "step": 3100
2513
+ },
2514
+ {
2515
+ "epoch": 37.134328358208954,
2516
+ "grad_norm": 12.753867149353027,
2517
+ "learning_rate": 1.3922356091030792e-05,
2518
+ "loss": 0.1594,
2519
+ "step": 3110
2520
+ },
2521
+ {
2522
+ "epoch": 37.25373134328358,
2523
+ "grad_norm": 9.448025703430176,
2524
+ "learning_rate": 1.3788487282463187e-05,
2525
+ "loss": 0.1583,
2526
+ "step": 3120
2527
+ },
2528
+ {
2529
+ "epoch": 37.37313432835821,
2530
+ "grad_norm": 26.652421951293945,
2531
+ "learning_rate": 1.3654618473895583e-05,
2532
+ "loss": 0.1584,
2533
+ "step": 3130
2534
+ },
2535
+ {
2536
+ "epoch": 37.492537313432834,
2537
+ "grad_norm": 13.557148933410645,
2538
+ "learning_rate": 1.3520749665327978e-05,
2539
+ "loss": 0.1753,
2540
+ "step": 3140
2541
+ },
2542
+ {
2543
+ "epoch": 37.61194029850746,
2544
+ "grad_norm": 12.321662902832031,
2545
+ "learning_rate": 1.3386880856760376e-05,
2546
+ "loss": 0.179,
2547
+ "step": 3150
2548
+ },
2549
+ {
2550
+ "epoch": 37.73134328358209,
2551
+ "grad_norm": 14.634367942810059,
2552
+ "learning_rate": 1.3253012048192772e-05,
2553
+ "loss": 0.1761,
2554
+ "step": 3160
2555
+ },
2556
+ {
2557
+ "epoch": 37.850746268656714,
2558
+ "grad_norm": 10.881927490234375,
2559
+ "learning_rate": 1.3119143239625167e-05,
2560
+ "loss": 0.1583,
2561
+ "step": 3170
2562
+ },
2563
+ {
2564
+ "epoch": 37.97014925373134,
2565
+ "grad_norm": 13.106158256530762,
2566
+ "learning_rate": 1.2985274431057565e-05,
2567
+ "loss": 0.2165,
2568
+ "step": 3180
2569
+ },
2570
+ {
2571
+ "epoch": 37.994029850746266,
2572
+ "eval_accuracy": 0.8835978835978836,
2573
+ "eval_loss": 0.5097155570983887,
2574
+ "eval_runtime": 21.5365,
2575
+ "eval_samples_per_second": 87.758,
2576
+ "eval_steps_per_second": 2.786,
2577
+ "step": 3182
2578
+ },
2579
+ {
2580
+ "epoch": 38.08955223880597,
2581
+ "grad_norm": 7.807220935821533,
2582
+ "learning_rate": 1.285140562248996e-05,
2583
+ "loss": 0.1416,
2584
+ "step": 3190
2585
+ },
2586
+ {
2587
+ "epoch": 38.208955223880594,
2588
+ "grad_norm": 9.47774887084961,
2589
+ "learning_rate": 1.2717536813922356e-05,
2590
+ "loss": 0.1422,
2591
+ "step": 3200
2592
+ },
2593
+ {
2594
+ "epoch": 38.32835820895522,
2595
+ "grad_norm": 12.726134300231934,
2596
+ "learning_rate": 1.2583668005354755e-05,
2597
+ "loss": 0.1493,
2598
+ "step": 3210
2599
+ },
2600
+ {
2601
+ "epoch": 38.44776119402985,
2602
+ "grad_norm": 10.369739532470703,
2603
+ "learning_rate": 1.244979919678715e-05,
2604
+ "loss": 0.1441,
2605
+ "step": 3220
2606
+ },
2607
+ {
2608
+ "epoch": 38.56716417910448,
2609
+ "grad_norm": 9.966578483581543,
2610
+ "learning_rate": 1.2315930388219546e-05,
2611
+ "loss": 0.1566,
2612
+ "step": 3230
2613
+ },
2614
+ {
2615
+ "epoch": 38.6865671641791,
2616
+ "grad_norm": 7.988857746124268,
2617
+ "learning_rate": 1.2182061579651942e-05,
2618
+ "loss": 0.13,
2619
+ "step": 3240
2620
+ },
2621
+ {
2622
+ "epoch": 38.80597014925373,
2623
+ "grad_norm": 6.763943195343018,
2624
+ "learning_rate": 1.2048192771084338e-05,
2625
+ "loss": 0.1633,
2626
+ "step": 3250
2627
+ },
2628
+ {
2629
+ "epoch": 38.92537313432836,
2630
+ "grad_norm": 22.221744537353516,
2631
+ "learning_rate": 1.1914323962516733e-05,
2632
+ "loss": 0.1565,
2633
+ "step": 3260
2634
+ },
2635
+ {
2636
+ "epoch": 38.997014925373136,
2637
+ "eval_accuracy": 0.8952380952380953,
2638
+ "eval_loss": 0.4565820097923279,
2639
+ "eval_runtime": 21.5334,
2640
+ "eval_samples_per_second": 87.771,
2641
+ "eval_steps_per_second": 2.786,
2642
+ "step": 3266
2643
+ },
2644
+ {
2645
+ "epoch": 39.04477611940298,
2646
+ "grad_norm": 7.209897518157959,
2647
+ "learning_rate": 1.178045515394913e-05,
2648
+ "loss": 0.1382,
2649
+ "step": 3270
2650
+ },
2651
+ {
2652
+ "epoch": 39.16417910447761,
2653
+ "grad_norm": 8.185746192932129,
2654
+ "learning_rate": 1.1646586345381528e-05,
2655
+ "loss": 0.1426,
2656
+ "step": 3280
2657
+ },
2658
+ {
2659
+ "epoch": 39.28358208955224,
2660
+ "grad_norm": 14.66981315612793,
2661
+ "learning_rate": 1.1512717536813922e-05,
2662
+ "loss": 0.1291,
2663
+ "step": 3290
2664
+ },
2665
+ {
2666
+ "epoch": 39.40298507462686,
2667
+ "grad_norm": 7.977197647094727,
2668
+ "learning_rate": 1.1378848728246319e-05,
2669
+ "loss": 0.1293,
2670
+ "step": 3300
2671
+ },
2672
+ {
2673
+ "epoch": 39.52238805970149,
2674
+ "grad_norm": 9.289730072021484,
2675
+ "learning_rate": 1.1244979919678715e-05,
2676
+ "loss": 0.1353,
2677
+ "step": 3310
2678
+ },
2679
+ {
2680
+ "epoch": 39.64179104477612,
2681
+ "grad_norm": 8.643695831298828,
2682
+ "learning_rate": 1.1111111111111112e-05,
2683
+ "loss": 0.1558,
2684
+ "step": 3320
2685
+ },
2686
+ {
2687
+ "epoch": 39.76119402985075,
2688
+ "grad_norm": 10.772273063659668,
2689
+ "learning_rate": 1.0977242302543508e-05,
2690
+ "loss": 0.18,
2691
+ "step": 3330
2692
+ },
2693
+ {
2694
+ "epoch": 39.88059701492537,
2695
+ "grad_norm": 16.730188369750977,
2696
+ "learning_rate": 1.0843373493975904e-05,
2697
+ "loss": 0.1323,
2698
+ "step": 3340
2699
+ },
2700
+ {
2701
+ "epoch": 40.0,
2702
+ "grad_norm": 16.59427261352539,
2703
+ "learning_rate": 1.0709504685408301e-05,
2704
+ "loss": 0.1476,
2705
+ "step": 3350
2706
+ },
2707
+ {
2708
+ "epoch": 40.0,
2709
+ "eval_accuracy": 0.8915343915343915,
2710
+ "eval_loss": 0.45787477493286133,
2711
+ "eval_runtime": 20.9113,
2712
+ "eval_samples_per_second": 90.382,
2713
+ "eval_steps_per_second": 2.869,
2714
+ "step": 3350
2715
+ },
2716
+ {
2717
+ "epoch": 40.11940298507463,
2718
+ "grad_norm": 13.793547630310059,
2719
+ "learning_rate": 1.0575635876840696e-05,
2720
+ "loss": 0.1335,
2721
+ "step": 3360
2722
+ },
2723
+ {
2724
+ "epoch": 40.23880597014925,
2725
+ "grad_norm": 12.108443260192871,
2726
+ "learning_rate": 1.0441767068273092e-05,
2727
+ "loss": 0.1571,
2728
+ "step": 3370
2729
+ },
2730
+ {
2731
+ "epoch": 40.35820895522388,
2732
+ "grad_norm": 6.3094024658203125,
2733
+ "learning_rate": 1.030789825970549e-05,
2734
+ "loss": 0.0985,
2735
+ "step": 3380
2736
+ },
2737
+ {
2738
+ "epoch": 40.47761194029851,
2739
+ "grad_norm": 12.814749717712402,
2740
+ "learning_rate": 1.0174029451137885e-05,
2741
+ "loss": 0.1461,
2742
+ "step": 3390
2743
+ },
2744
+ {
2745
+ "epoch": 40.59701492537313,
2746
+ "grad_norm": 4.608739852905273,
2747
+ "learning_rate": 1.0040160642570281e-05,
2748
+ "loss": 0.1116,
2749
+ "step": 3400
2750
+ },
2751
+ {
2752
+ "epoch": 40.71641791044776,
2753
+ "grad_norm": 11.582805633544922,
2754
+ "learning_rate": 9.906291834002678e-06,
2755
+ "loss": 0.1358,
2756
+ "step": 3410
2757
+ },
2758
+ {
2759
+ "epoch": 40.83582089552239,
2760
+ "grad_norm": 11.274500846862793,
2761
+ "learning_rate": 9.772423025435074e-06,
2762
+ "loss": 0.1246,
2763
+ "step": 3420
2764
+ },
2765
+ {
2766
+ "epoch": 40.95522388059702,
2767
+ "grad_norm": 10.500862121582031,
2768
+ "learning_rate": 9.63855421686747e-06,
2769
+ "loss": 0.1296,
2770
+ "step": 3430
2771
+ },
2772
+ {
2773
+ "epoch": 40.9910447761194,
2774
+ "eval_accuracy": 0.8931216931216931,
2775
+ "eval_loss": 0.4594806134700775,
2776
+ "eval_runtime": 21.5399,
2777
+ "eval_samples_per_second": 87.744,
2778
+ "eval_steps_per_second": 2.786,
2779
+ "step": 3433
2780
+ },
2781
+ {
2782
+ "epoch": 41.07462686567164,
2783
+ "grad_norm": 13.058525085449219,
2784
+ "learning_rate": 9.504685408299867e-06,
2785
+ "loss": 0.1284,
2786
+ "step": 3440
2787
+ },
2788
+ {
2789
+ "epoch": 41.19402985074627,
2790
+ "grad_norm": 11.443678855895996,
2791
+ "learning_rate": 9.370816599732263e-06,
2792
+ "loss": 0.1214,
2793
+ "step": 3450
2794
+ },
2795
+ {
2796
+ "epoch": 41.3134328358209,
2797
+ "grad_norm": 7.747429370880127,
2798
+ "learning_rate": 9.236947791164658e-06,
2799
+ "loss": 0.1001,
2800
+ "step": 3460
2801
+ },
2802
+ {
2803
+ "epoch": 41.43283582089552,
2804
+ "grad_norm": 13.356925010681152,
2805
+ "learning_rate": 9.103078982597054e-06,
2806
+ "loss": 0.1258,
2807
+ "step": 3470
2808
+ },
2809
+ {
2810
+ "epoch": 41.55223880597015,
2811
+ "grad_norm": 7.233834743499756,
2812
+ "learning_rate": 8.969210174029452e-06,
2813
+ "loss": 0.1238,
2814
+ "step": 3480
2815
+ },
2816
+ {
2817
+ "epoch": 41.67164179104478,
2818
+ "grad_norm": 9.122505187988281,
2819
+ "learning_rate": 8.835341365461847e-06,
2820
+ "loss": 0.1224,
2821
+ "step": 3490
2822
+ },
2823
+ {
2824
+ "epoch": 41.791044776119406,
2825
+ "grad_norm": 11.804553031921387,
2826
+ "learning_rate": 8.701472556894244e-06,
2827
+ "loss": 0.1173,
2828
+ "step": 3500
2829
+ },
2830
+ {
2831
+ "epoch": 41.91044776119403,
2832
+ "grad_norm": 16.831106185913086,
2833
+ "learning_rate": 8.56760374832664e-06,
2834
+ "loss": 0.1159,
2835
+ "step": 3510
2836
+ },
2837
+ {
2838
+ "epoch": 41.994029850746266,
2839
+ "eval_accuracy": 0.8883597883597883,
2840
+ "eval_loss": 0.4840790629386902,
2841
+ "eval_runtime": 21.8549,
2842
+ "eval_samples_per_second": 86.48,
2843
+ "eval_steps_per_second": 2.745,
2844
+ "step": 3517
2845
+ },
2846
+ {
2847
+ "epoch": 42.02985074626866,
2848
+ "grad_norm": 7.735755443572998,
2849
+ "learning_rate": 8.433734939759036e-06,
2850
+ "loss": 0.1211,
2851
+ "step": 3520
2852
+ },
2853
+ {
2854
+ "epoch": 42.149253731343286,
2855
+ "grad_norm": 9.165584564208984,
2856
+ "learning_rate": 8.299866131191433e-06,
2857
+ "loss": 0.1441,
2858
+ "step": 3530
2859
+ },
2860
+ {
2861
+ "epoch": 42.26865671641791,
2862
+ "grad_norm": 11.504267692565918,
2863
+ "learning_rate": 8.16599732262383e-06,
2864
+ "loss": 0.0981,
2865
+ "step": 3540
2866
+ },
2867
+ {
2868
+ "epoch": 42.38805970149254,
2869
+ "grad_norm": 13.387564659118652,
2870
+ "learning_rate": 8.032128514056226e-06,
2871
+ "loss": 0.1017,
2872
+ "step": 3550
2873
+ },
2874
+ {
2875
+ "epoch": 42.507462686567166,
2876
+ "grad_norm": 13.6415376663208,
2877
+ "learning_rate": 7.89825970548862e-06,
2878
+ "loss": 0.145,
2879
+ "step": 3560
2880
+ },
2881
+ {
2882
+ "epoch": 42.62686567164179,
2883
+ "grad_norm": 15.573339462280273,
2884
+ "learning_rate": 7.764390896921017e-06,
2885
+ "loss": 0.1396,
2886
+ "step": 3570
2887
+ },
2888
+ {
2889
+ "epoch": 42.74626865671642,
2890
+ "grad_norm": 8.934632301330566,
2891
+ "learning_rate": 7.630522088353415e-06,
2892
+ "loss": 0.129,
2893
+ "step": 3580
2894
+ },
2895
+ {
2896
+ "epoch": 42.865671641791046,
2897
+ "grad_norm": 6.53822135925293,
2898
+ "learning_rate": 7.4966532797858104e-06,
2899
+ "loss": 0.1384,
2900
+ "step": 3590
2901
+ },
2902
+ {
2903
+ "epoch": 42.985074626865675,
2904
+ "grad_norm": 14.102238655090332,
2905
+ "learning_rate": 7.362784471218207e-06,
2906
+ "loss": 0.1071,
2907
+ "step": 3600
2908
+ },
2909
+ {
2910
+ "epoch": 42.997014925373136,
2911
+ "eval_accuracy": 0.882010582010582,
2912
+ "eval_loss": 0.4729897379875183,
2913
+ "eval_runtime": 23.368,
2914
+ "eval_samples_per_second": 80.88,
2915
+ "eval_steps_per_second": 2.568,
2916
+ "step": 3601
2917
+ },
2918
+ {
2919
+ "epoch": 43.1044776119403,
2920
+ "grad_norm": 10.040472030639648,
2921
+ "learning_rate": 7.228915662650602e-06,
2922
+ "loss": 0.1247,
2923
+ "step": 3610
2924
+ },
2925
+ {
2926
+ "epoch": 43.223880597014926,
2927
+ "grad_norm": 9.124853134155273,
2928
+ "learning_rate": 7.095046854082999e-06,
2929
+ "loss": 0.1059,
2930
+ "step": 3620
2931
+ },
2932
+ {
2933
+ "epoch": 43.343283582089555,
2934
+ "grad_norm": 9.409137725830078,
2935
+ "learning_rate": 6.961178045515396e-06,
2936
+ "loss": 0.1062,
2937
+ "step": 3630
2938
+ },
2939
+ {
2940
+ "epoch": 43.46268656716418,
2941
+ "grad_norm": 12.034777641296387,
2942
+ "learning_rate": 6.827309236947792e-06,
2943
+ "loss": 0.1295,
2944
+ "step": 3640
2945
+ },
2946
+ {
2947
+ "epoch": 43.582089552238806,
2948
+ "grad_norm": 11.372798919677734,
2949
+ "learning_rate": 6.693440428380188e-06,
2950
+ "loss": 0.0976,
2951
+ "step": 3650
2952
+ },
2953
+ {
2954
+ "epoch": 43.701492537313435,
2955
+ "grad_norm": 19.926820755004883,
2956
+ "learning_rate": 6.559571619812584e-06,
2957
+ "loss": 0.1318,
2958
+ "step": 3660
2959
+ },
2960
+ {
2961
+ "epoch": 43.82089552238806,
2962
+ "grad_norm": 10.970476150512695,
2963
+ "learning_rate": 6.42570281124498e-06,
2964
+ "loss": 0.1251,
2965
+ "step": 3670
2966
+ },
2967
+ {
2968
+ "epoch": 43.940298507462686,
2969
+ "grad_norm": 12.275060653686523,
2970
+ "learning_rate": 6.291834002677377e-06,
2971
+ "loss": 0.1017,
2972
+ "step": 3680
2973
+ },
2974
+ {
2975
+ "epoch": 44.0,
2976
+ "eval_accuracy": 0.8931216931216931,
2977
+ "eval_loss": 0.4470039904117584,
2978
+ "eval_runtime": 23.3521,
2979
+ "eval_samples_per_second": 80.935,
2980
+ "eval_steps_per_second": 2.569,
2981
+ "step": 3685
2982
+ },
2983
+ {
2984
+ "epoch": 44.059701492537314,
2985
+ "grad_norm": 8.43974494934082,
2986
+ "learning_rate": 6.157965194109773e-06,
2987
+ "loss": 0.1239,
2988
+ "step": 3690
2989
+ },
2990
+ {
2991
+ "epoch": 44.17910447761194,
2992
+ "grad_norm": 14.687185287475586,
2993
+ "learning_rate": 6.024096385542169e-06,
2994
+ "loss": 0.1189,
2995
+ "step": 3700
2996
+ },
2997
+ {
2998
+ "epoch": 44.298507462686565,
2999
+ "grad_norm": 9.467026710510254,
3000
+ "learning_rate": 5.890227576974565e-06,
3001
+ "loss": 0.1491,
3002
+ "step": 3710
3003
+ },
3004
+ {
3005
+ "epoch": 44.417910447761194,
3006
+ "grad_norm": 5.684982776641846,
3007
+ "learning_rate": 5.756358768406961e-06,
3008
+ "loss": 0.0982,
3009
+ "step": 3720
3010
+ },
3011
+ {
3012
+ "epoch": 44.53731343283582,
3013
+ "grad_norm": 7.686473369598389,
3014
+ "learning_rate": 5.622489959839358e-06,
3015
+ "loss": 0.1147,
3016
+ "step": 3730
3017
+ },
3018
+ {
3019
+ "epoch": 44.656716417910445,
3020
+ "grad_norm": 11.38039779663086,
3021
+ "learning_rate": 5.488621151271754e-06,
3022
+ "loss": 0.1179,
3023
+ "step": 3740
3024
+ },
3025
+ {
3026
+ "epoch": 44.776119402985074,
3027
+ "grad_norm": 6.498850345611572,
3028
+ "learning_rate": 5.3547523427041504e-06,
3029
+ "loss": 0.12,
3030
+ "step": 3750
3031
+ },
3032
+ {
3033
+ "epoch": 44.8955223880597,
3034
+ "grad_norm": 9.847461700439453,
3035
+ "learning_rate": 5.220883534136546e-06,
3036
+ "loss": 0.11,
3037
+ "step": 3760
3038
+ },
3039
+ {
3040
+ "epoch": 44.9910447761194,
3041
+ "eval_accuracy": 0.891005291005291,
3042
+ "eval_loss": 0.4557030200958252,
3043
+ "eval_runtime": 24.0952,
3044
+ "eval_samples_per_second": 78.439,
3045
+ "eval_steps_per_second": 2.49,
3046
+ "step": 3768
3047
+ },
3048
+ {
3049
+ "epoch": 45.014925373134325,
3050
+ "grad_norm": 9.032334327697754,
3051
+ "learning_rate": 5.087014725568942e-06,
3052
+ "loss": 0.1352,
3053
+ "step": 3770
3054
+ },
3055
+ {
3056
+ "epoch": 45.134328358208954,
3057
+ "grad_norm": 9.14494514465332,
3058
+ "learning_rate": 4.953145917001339e-06,
3059
+ "loss": 0.1014,
3060
+ "step": 3780
3061
+ },
3062
+ {
3063
+ "epoch": 45.25373134328358,
3064
+ "grad_norm": 8.348182678222656,
3065
+ "learning_rate": 4.819277108433735e-06,
3066
+ "loss": 0.1052,
3067
+ "step": 3790
3068
+ },
3069
+ {
3070
+ "epoch": 45.37313432835821,
3071
+ "grad_norm": 5.649065971374512,
3072
+ "learning_rate": 4.685408299866132e-06,
3073
+ "loss": 0.1138,
3074
+ "step": 3800
3075
+ },
3076
+ {
3077
+ "epoch": 45.492537313432834,
3078
+ "grad_norm": 7.422122955322266,
3079
+ "learning_rate": 4.551539491298527e-06,
3080
+ "loss": 0.1091,
3081
+ "step": 3810
3082
+ },
3083
+ {
3084
+ "epoch": 45.61194029850746,
3085
+ "grad_norm": 9.711326599121094,
3086
+ "learning_rate": 4.417670682730924e-06,
3087
+ "loss": 0.1068,
3088
+ "step": 3820
3089
+ },
3090
+ {
3091
+ "epoch": 45.73134328358209,
3092
+ "grad_norm": 6.854335784912109,
3093
+ "learning_rate": 4.28380187416332e-06,
3094
+ "loss": 0.1331,
3095
+ "step": 3830
3096
+ },
3097
+ {
3098
+ "epoch": 45.850746268656714,
3099
+ "grad_norm": 7.027961254119873,
3100
+ "learning_rate": 4.149933065595716e-06,
3101
+ "loss": 0.0979,
3102
+ "step": 3840
3103
+ },
3104
+ {
3105
+ "epoch": 45.97014925373134,
3106
+ "grad_norm": 8.980549812316895,
3107
+ "learning_rate": 4.016064257028113e-06,
3108
+ "loss": 0.126,
3109
+ "step": 3850
3110
+ },
3111
+ {
3112
+ "epoch": 45.994029850746266,
3113
+ "eval_accuracy": 0.8925925925925926,
3114
+ "eval_loss": 0.4585404098033905,
3115
+ "eval_runtime": 23.6784,
3116
+ "eval_samples_per_second": 79.82,
3117
+ "eval_steps_per_second": 2.534,
3118
+ "step": 3852
3119
+ },
3120
+ {
3121
+ "epoch": 46.08955223880597,
3122
+ "grad_norm": 6.251053810119629,
3123
+ "learning_rate": 3.882195448460508e-06,
3124
+ "loss": 0.1199,
3125
+ "step": 3860
3126
+ },
3127
+ {
3128
+ "epoch": 46.208955223880594,
3129
+ "grad_norm": 9.849762916564941,
3130
+ "learning_rate": 3.7483266398929052e-06,
3131
+ "loss": 0.1068,
3132
+ "step": 3870
3133
+ },
3134
+ {
3135
+ "epoch": 46.32835820895522,
3136
+ "grad_norm": 4.6796722412109375,
3137
+ "learning_rate": 3.614457831325301e-06,
3138
+ "loss": 0.0909,
3139
+ "step": 3880
3140
+ },
3141
+ {
3142
+ "epoch": 46.44776119402985,
3143
+ "grad_norm": 7.510217666625977,
3144
+ "learning_rate": 3.480589022757698e-06,
3145
+ "loss": 0.1058,
3146
+ "step": 3890
3147
+ },
3148
+ {
3149
+ "epoch": 46.56716417910448,
3150
+ "grad_norm": 10.298948287963867,
3151
+ "learning_rate": 3.346720214190094e-06,
3152
+ "loss": 0.1045,
3153
+ "step": 3900
3154
+ },
3155
+ {
3156
+ "epoch": 46.6865671641791,
3157
+ "grad_norm": 10.80692195892334,
3158
+ "learning_rate": 3.21285140562249e-06,
3159
+ "loss": 0.1051,
3160
+ "step": 3910
3161
+ },
3162
+ {
3163
+ "epoch": 46.80597014925373,
3164
+ "grad_norm": 7.66585111618042,
3165
+ "learning_rate": 3.0789825970548864e-06,
3166
+ "loss": 0.1152,
3167
+ "step": 3920
3168
+ },
3169
+ {
3170
+ "epoch": 46.92537313432836,
3171
+ "grad_norm": 12.45518970489502,
3172
+ "learning_rate": 2.9451137884872824e-06,
3173
+ "loss": 0.1079,
3174
+ "step": 3930
3175
+ },
3176
+ {
3177
+ "epoch": 46.997014925373136,
3178
+ "eval_accuracy": 0.8904761904761904,
3179
+ "eval_loss": 0.4550909698009491,
3180
+ "eval_runtime": 22.5814,
3181
+ "eval_samples_per_second": 83.697,
3182
+ "eval_steps_per_second": 2.657,
3183
+ "step": 3936
3184
+ },
3185
+ {
3186
+ "epoch": 47.04477611940298,
3187
+ "grad_norm": 8.93836498260498,
3188
+ "learning_rate": 2.811244979919679e-06,
3189
+ "loss": 0.0922,
3190
+ "step": 3940
3191
+ },
3192
+ {
3193
+ "epoch": 47.16417910447761,
3194
+ "grad_norm": 9.170904159545898,
3195
+ "learning_rate": 2.6773761713520752e-06,
3196
+ "loss": 0.1062,
3197
+ "step": 3950
3198
+ },
3199
+ {
3200
+ "epoch": 47.28358208955224,
3201
+ "grad_norm": 8.85617446899414,
3202
+ "learning_rate": 2.543507362784471e-06,
3203
+ "loss": 0.1051,
3204
+ "step": 3960
3205
+ },
3206
+ {
3207
+ "epoch": 47.40298507462686,
3208
+ "grad_norm": 12.655815124511719,
3209
+ "learning_rate": 2.4096385542168676e-06,
3210
+ "loss": 0.1178,
3211
+ "step": 3970
3212
+ },
3213
+ {
3214
+ "epoch": 47.52238805970149,
3215
+ "grad_norm": 7.482461929321289,
3216
+ "learning_rate": 2.2757697456492636e-06,
3217
+ "loss": 0.1011,
3218
+ "step": 3980
3219
+ },
3220
+ {
3221
+ "epoch": 47.64179104477612,
3222
+ "grad_norm": 9.295207023620605,
3223
+ "learning_rate": 2.14190093708166e-06,
3224
+ "loss": 0.1175,
3225
+ "step": 3990
3226
+ },
3227
+ {
3228
+ "epoch": 47.76119402985075,
3229
+ "grad_norm": 8.872883796691895,
3230
+ "learning_rate": 2.0080321285140564e-06,
3231
+ "loss": 0.1046,
3232
+ "step": 4000
3233
+ },
3234
+ {
3235
+ "epoch": 47.88059701492537,
3236
+ "grad_norm": 11.744367599487305,
3237
+ "learning_rate": 1.8741633199464526e-06,
3238
+ "loss": 0.0891,
3239
+ "step": 4010
3240
+ },
3241
+ {
3242
+ "epoch": 48.0,
3243
+ "grad_norm": 14.95531177520752,
3244
+ "learning_rate": 1.740294511378849e-06,
3245
+ "loss": 0.1194,
3246
+ "step": 4020
3247
+ },
3248
+ {
3249
+ "epoch": 48.0,
3250
+ "eval_accuracy": 0.8947089947089947,
3251
+ "eval_loss": 0.44007420539855957,
3252
+ "eval_runtime": 21.2726,
3253
+ "eval_samples_per_second": 88.847,
3254
+ "eval_steps_per_second": 2.821,
3255
+ "step": 4020
3256
+ },
3257
+ {
3258
+ "epoch": 48.11940298507463,
3259
+ "grad_norm": 12.020395278930664,
3260
+ "learning_rate": 1.606425702811245e-06,
3261
+ "loss": 0.1122,
3262
+ "step": 4030
3263
+ },
3264
+ {
3265
+ "epoch": 48.23880597014925,
3266
+ "grad_norm": 10.32168197631836,
3267
+ "learning_rate": 1.4725568942436412e-06,
3268
+ "loss": 0.1076,
3269
+ "step": 4040
3270
+ },
3271
+ {
3272
+ "epoch": 48.35820895522388,
3273
+ "grad_norm": 12.109696388244629,
3274
+ "learning_rate": 1.3386880856760376e-06,
3275
+ "loss": 0.1145,
3276
+ "step": 4050
3277
+ },
3278
+ {
3279
+ "epoch": 48.47761194029851,
3280
+ "grad_norm": 5.61909294128418,
3281
+ "learning_rate": 1.2048192771084338e-06,
3282
+ "loss": 0.0886,
3283
+ "step": 4060
3284
+ },
3285
+ {
3286
+ "epoch": 48.59701492537313,
3287
+ "grad_norm": 5.84828519821167,
3288
+ "learning_rate": 1.07095046854083e-06,
3289
+ "loss": 0.088,
3290
+ "step": 4070
3291
+ },
3292
+ {
3293
+ "epoch": 48.71641791044776,
3294
+ "grad_norm": 7.48928165435791,
3295
+ "learning_rate": 9.370816599732263e-07,
3296
+ "loss": 0.0943,
3297
+ "step": 4080
3298
+ },
3299
+ {
3300
+ "epoch": 48.83582089552239,
3301
+ "grad_norm": 6.499556541442871,
3302
+ "learning_rate": 8.032128514056225e-07,
3303
+ "loss": 0.104,
3304
+ "step": 4090
3305
+ },
3306
+ {
3307
+ "epoch": 48.95522388059702,
3308
+ "grad_norm": 8.104216575622559,
3309
+ "learning_rate": 6.693440428380188e-07,
3310
+ "loss": 0.11,
3311
+ "step": 4100
3312
+ },
3313
+ {
3314
+ "epoch": 48.9910447761194,
3315
+ "eval_accuracy": 0.8968253968253969,
3316
+ "eval_loss": 0.44243764877319336,
3317
+ "eval_runtime": 21.6947,
3318
+ "eval_samples_per_second": 87.118,
3319
+ "eval_steps_per_second": 2.766,
3320
+ "step": 4103
3321
+ },
3322
+ {
3323
+ "epoch": 49.07462686567164,
3324
+ "grad_norm": 5.112276077270508,
3325
+ "learning_rate": 5.35475234270415e-07,
3326
+ "loss": 0.0958,
3327
+ "step": 4110
3328
+ },
3329
+ {
3330
+ "epoch": 49.19402985074627,
3331
+ "grad_norm": 9.789267539978027,
3332
+ "learning_rate": 4.0160642570281125e-07,
3333
+ "loss": 0.1036,
3334
+ "step": 4120
3335
+ },
3336
+ {
3337
+ "epoch": 49.3134328358209,
3338
+ "grad_norm": 7.143287658691406,
3339
+ "learning_rate": 2.677376171352075e-07,
3340
+ "loss": 0.1019,
3341
+ "step": 4130
3342
+ },
3343
+ {
3344
+ "epoch": 49.43283582089552,
3345
+ "grad_norm": 6.7176384925842285,
3346
+ "learning_rate": 1.3386880856760375e-07,
3347
+ "loss": 0.1089,
3348
+ "step": 4140
3349
+ },
3350
+ {
3351
+ "epoch": 49.55223880597015,
3352
+ "grad_norm": 11.27464771270752,
3353
+ "learning_rate": 0.0,
3354
+ "loss": 0.1104,
3355
+ "step": 4150
3356
+ },
3357
+ {
3358
+ "epoch": 49.55223880597015,
3359
+ "eval_accuracy": 0.8957671957671958,
3360
+ "eval_loss": 0.44136127829551697,
3361
+ "eval_runtime": 22.143,
3362
+ "eval_samples_per_second": 85.354,
3363
+ "eval_steps_per_second": 2.71,
3364
+ "step": 4150
3365
+ },
3366
+ {
3367
+ "epoch": 49.55223880597015,
3368
+ "step": 4150,
3369
+ "total_flos": 5.421822883117793e+19,
3370
+ "train_loss": 0.33880508939903903,
3371
+ "train_runtime": 14225.4065,
3372
+ "train_samples_per_second": 37.644,
3373
+ "train_steps_per_second": 0.292
3374
  }
3375
  ],
3376
  "logging_steps": 10,
3377
+ "max_steps": 4150,
3378
  "num_input_tokens_seen": 0,
3379
+ "num_train_epochs": 50,
3380
  "save_steps": 500,
3381
  "stateful_callbacks": {
3382
  "TrainerControl": {
 
3390
  "attributes": {}
3391
  }
3392
  },
3393
+ "total_flos": 5.421822883117793e+19,
3394
+ "train_batch_size": 32,
3395
  "trial_name": null,
3396
  "trial_params": null
3397
  }