Hugging Face's logo

dhritic9
/
vit-base-brain-tumor-detection

Image Classification
Transformers
TensorBoard
Safetensors
vit
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
dhritic9 commited on
Commit
e066afe
•
1 Parent(s): 4e550bc

dhritic9/amodel1

Browse files
Files changed (12) hide show
  1. README.md +110 -0
  2. all_results.json +13 -0
  3. config.json +36 -0
  4. eval_results.json +8 -0
  5. model.safetensors +3 -0
  6. preprocessor_config.json +36 -0
  7. runs/Jul15_10-07-48_ed1e415c8971/events.out.tfevents.1721038070.ed1e415c8971.4691.0 +3 -0
  8. runs/Jul15_10-16-21_ed1e415c8971/events.out.tfevents.1721038594.ed1e415c8971.6989.0 +3 -0
  9. runs/Jul15_10-16-21_ed1e415c8971/events.out.tfevents.1721040783.ed1e415c8971.6989.1 +3 -0
  10. train_results.json +8 -0
  11. trainer_state.json +3992 -0
  12. training_args.bin +3 -0
README.md ADDED
@@ -0,0 +1,110 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: google/vit-base-patch16-224-in21k
4
+ tags:
5
+ - generated_from_trainer
6
+ metrics:
7
+ - accuracy
8
+ model-index:
9
+ - name: vit-base-brain-tumor-detection
10
+ results: []
11
+ ---
12
+
13
+ <!-- This model card has been generated automatically according to the information the Trainer had access to. You
14
+ should probably proofread and complete it, then remove this comment. -->
15
+
16
+ # vit-base-brain-tumor-detection
17
+
18
+ This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
19
+ It achieves the following results on the evaluation set:
20
+ - Loss: 0.5832
21
+ - Accuracy: 0.785
22
+
23
+ ## Model description
24
+
25
+ More information needed
26
+
27
+ ## Intended uses & limitations
28
+
29
+ More information needed
30
+
31
+ ## Training and evaluation data
32
+
33
+ More information needed
34
+
35
+ ## Training procedure
36
+
37
+ ### Training hyperparameters
38
+
39
+ The following hyperparameters were used during training:
40
+ - learning_rate: 0.0002
41
+ - train_batch_size: 16
42
+ - eval_batch_size: 8
43
+ - seed: 42
44
+ - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
+ - lr_scheduler_type: linear
46
+ - num_epochs: 20
47
+ - mixed_precision_training: Native AMP
48
+
49
+ ### Training results
50
+
51
+ | Training Loss | Epoch | Step | Validation Loss | Accuracy |
52
+ |:-------------:|:-----:|:----:|:---------------:|:--------:|
53
+ | 0.9535 | 0.4 | 100 | 0.8966 | 0.618 |
54
+ | 0.862 | 0.8 | 200 | 1.1149 | 0.561 |
55
+ | 0.7373 | 1.2 | 300 | 0.8543 | 0.605 |
56
+ | 0.6476 | 1.6 | 400 | 0.7307 | 0.666 |
57
+ | 0.6712 | 2.0 | 500 | 0.6954 | 0.694 |
58
+ | 0.4892 | 2.4 | 600 | 0.6391 | 0.707 |
59
+ | 0.5801 | 2.8 | 700 | 0.6247 | 0.708 |
60
+ | 0.3505 | 3.2 | 800 | 0.6056 | 0.778 |
61
+ | 0.3503 | 3.6 | 900 | 0.6264 | 0.743 |
62
+ | 0.3416 | 4.0 | 1000 | 0.5832 | 0.785 |
63
+ | 0.1427 | 4.4 | 1100 | 0.7297 | 0.769 |
64
+ | 0.1982 | 4.8 | 1200 | 0.7761 | 0.73 |
65
+ | 0.193 | 5.2 | 1300 | 0.8467 | 0.741 |
66
+ | 0.1831 | 5.6 | 1400 | 0.6975 | 0.774 |
67
+ | 0.2612 | 6.0 | 1500 | 0.8719 | 0.775 |
68
+ | 0.102 | 6.4 | 1600 | 0.9045 | 0.788 |
69
+ | 0.1029 | 6.8 | 1700 | 0.9655 | 0.783 |
70
+ | 0.0735 | 7.2 | 1800 | 0.9906 | 0.78 |
71
+ | 0.0715 | 7.6 | 1900 | 0.8893 | 0.787 |
72
+ | 0.1254 | 8.0 | 2000 | 1.1221 | 0.761 |
73
+ | 0.021 | 8.4 | 2100 | 1.1648 | 0.779 |
74
+ | 0.0133 | 8.8 | 2200 | 0.9857 | 0.806 |
75
+ | 0.0086 | 9.2 | 2300 | 1.0365 | 0.799 |
76
+ | 0.0223 | 9.6 | 2400 | 0.9826 | 0.812 |
77
+ | 0.0023 | 10.0 | 2500 | 1.0697 | 0.795 |
78
+ | 0.0021 | 10.4 | 2600 | 1.0490 | 0.815 |
79
+ | 0.0401 | 10.8 | 2700 | 1.1594 | 0.8 |
80
+ | 0.0012 | 11.2 | 2800 | 1.0811 | 0.817 |
81
+ | 0.0034 | 11.6 | 2900 | 1.0956 | 0.825 |
82
+ | 0.0012 | 12.0 | 3000 | 1.2010 | 0.808 |
83
+ | 0.0011 | 12.4 | 3100 | 1.1712 | 0.81 |
84
+ | 0.0092 | 12.8 | 3200 | 1.1814 | 0.813 |
85
+ | 0.0007 | 13.2 | 3300 | 1.1677 | 0.818 |
86
+ | 0.0007 | 13.6 | 3400 | 1.1723 | 0.818 |
87
+ | 0.0006 | 14.0 | 3500 | 1.1852 | 0.821 |
88
+ | 0.0005 | 14.4 | 3600 | 1.1928 | 0.82 |
89
+ | 0.0005 | 14.8 | 3700 | 1.2030 | 0.819 |
90
+ | 0.0005 | 15.2 | 3800 | 1.2093 | 0.818 |
91
+ | 0.0005 | 15.6 | 3900 | 1.2160 | 0.818 |
92
+ | 0.0004 | 16.0 | 4000 | 1.2232 | 0.819 |
93
+ | 0.0004 | 16.4 | 4100 | 1.2302 | 0.819 |
94
+ | 0.0004 | 16.8 | 4200 | 1.2350 | 0.819 |
95
+ | 0.0004 | 17.2 | 4300 | 1.2400 | 0.82 |
96
+ | 0.0004 | 17.6 | 4400 | 1.2442 | 0.821 |
97
+ | 0.0004 | 18.0 | 4500 | 1.2483 | 0.821 |
98
+ | 0.0004 | 18.4 | 4600 | 1.2518 | 0.821 |
99
+ | 0.0004 | 18.8 | 4700 | 1.2546 | 0.821 |
100
+ | 0.0004 | 19.2 | 4800 | 1.2561 | 0.821 |
101
+ | 0.0004 | 19.6 | 4900 | 1.2574 | 0.82 |
102
+ | 0.0004 | 20.0 | 5000 | 1.2577 | 0.82 |
103
+
104
+
105
+ ### Framework versions
106
+
107
+ - Transformers 4.41.2
108
+ - Pytorch 2.3.0+cu121
109
+ - Datasets 2.20.0
110
+ - Tokenizers 0.19.1
all_results.json ADDED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 0.785,
4
+ "eval_loss": 0.583153247833252,
5
+ "eval_runtime": 6.36,
6
+ "eval_samples_per_second": 157.234,
7
+ "eval_steps_per_second": 19.654,
8
+ "total_flos": 6.19947029495808e+18,
9
+ "train_loss": 0.16214322472065687,
10
+ "train_runtime": 2178.5774,
11
+ "train_samples_per_second": 36.721,
12
+ "train_steps_per_second": 2.295
13
+ }
config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "google/vit-base-patch16-224-in21k",
3
+ "architectures": [
4
+ "ViTForImageClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "encoder_stride": 16,
8
+ "hidden_act": "gelu",
9
+ "hidden_dropout_prob": 0.0,
10
+ "hidden_size": 768,
11
+ "id2label": {
12
+ "0": 0,
13
+ "1": 1,
14
+ "2": 2,
15
+ "3": 3
16
+ },
17
+ "image_size": 224,
18
+ "initializer_range": 0.02,
19
+ "intermediate_size": 3072,
20
+ "label2id": {
21
+ "0": "0",
22
+ "1": "1",
23
+ "2": "2",
24
+ "3": "3"
25
+ },
26
+ "layer_norm_eps": 1e-12,
27
+ "model_type": "vit",
28
+ "num_attention_heads": 12,
29
+ "num_channels": 3,
30
+ "num_hidden_layers": 12,
31
+ "patch_size": 16,
32
+ "problem_type": "single_label_classification",
33
+ "qkv_bias": true,
34
+ "torch_dtype": "float32",
35
+ "transformers_version": "4.41.2"
36
+ }
eval_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "eval_accuracy": 0.785,
4
+ "eval_loss": 0.583153247833252,
5
+ "eval_runtime": 6.36,
6
+ "eval_samples_per_second": 157.234,
7
+ "eval_steps_per_second": 19.654
8
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:177ebc3f9c2c01bd8c0f19c64c4540689aa4b1e2a2554668df1693af173fe0be
3
+ size 343230128
preprocessor_config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_valid_processor_keys": [
3
+ "images",
4
+ "do_resize",
5
+ "size",
6
+ "resample",
7
+ "do_rescale",
8
+ "rescale_factor",
9
+ "do_normalize",
10
+ "image_mean",
11
+ "image_std",
12
+ "return_tensors",
13
+ "data_format",
14
+ "input_data_format"
15
+ ],
16
+ "do_normalize": true,
17
+ "do_rescale": true,
18
+ "do_resize": true,
19
+ "image_mean": [
20
+ 0.5,
21
+ 0.5,
22
+ 0.5
23
+ ],
24
+ "image_processor_type": "ViTImageProcessor",
25
+ "image_std": [
26
+ 0.5,
27
+ 0.5,
28
+ 0.5
29
+ ],
30
+ "resample": 2,
31
+ "rescale_factor": 0.00392156862745098,
32
+ "size": {
33
+ "height": 224,
34
+ "width": 224
35
+ }
36
+ }
runs/Jul15_10-07-48_ed1e415c8971/events.out.tfevents.1721038070.ed1e415c8971.4691.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b76a3267b1fe833f28c82804217f5cfad365ecec9d76732f0e3090d1fc543b81
3
+ size 4184
runs/Jul15_10-16-21_ed1e415c8971/events.out.tfevents.1721038594.ed1e415c8971.6989.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ceddaa98a872f7c2646ca0c6853dd7a4ecf976c85707c27a171f5edf52ecc1a2
3
+ size 126826
runs/Jul15_10-16-21_ed1e415c8971/events.out.tfevents.1721040783.ed1e415c8971.6989.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ac5653db79d5524a688ffaf032bf1ea9d40509841055df6d7cd07231059e770
3
+ size 411
train_results.json ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "epoch": 20.0,
3
+ "total_flos": 6.19947029495808e+18,
4
+ "train_loss": 0.16214322472065687,
5
+ "train_runtime": 2178.5774,
6
+ "train_samples_per_second": 36.721,
7
+ "train_steps_per_second": 2.295
8
+ }
trainer_state.json ADDED
@@ -0,0 +1,3992 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.583153247833252,
3
+ "best_model_checkpoint": "./vit-base-brain-tumor-detection/checkpoint-1000",
4
+ "epoch": 20.0,
5
+ "eval_steps": 100,
6
+ "global_step": 5000,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.04,
13
+ "grad_norm": 1.653470516204834,
14
+ "learning_rate": 0.0001996,
15
+ "loss": 1.3897,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.08,
20
+ "grad_norm": 0.5039440989494324,
21
+ "learning_rate": 0.00019920000000000002,
22
+ "loss": 1.4048,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.12,
27
+ "grad_norm": 0.8407747745513916,
28
+ "learning_rate": 0.0001988,
29
+ "loss": 1.3912,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.16,
34
+ "grad_norm": 0.898512601852417,
35
+ "learning_rate": 0.0001984,
36
+ "loss": 1.3661,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.2,
41
+ "grad_norm": 2.638078212738037,
42
+ "learning_rate": 0.00019800000000000002,
43
+ "loss": 1.2842,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.24,
48
+ "grad_norm": 1.7601313591003418,
49
+ "learning_rate": 0.0001976,
50
+ "loss": 1.239,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.28,
55
+ "grad_norm": 5.79738187789917,
56
+ "learning_rate": 0.0001972,
57
+ "loss": 1.1859,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.32,
62
+ "grad_norm": 3.172907829284668,
63
+ "learning_rate": 0.0001968,
64
+ "loss": 1.1265,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.36,
69
+ "grad_norm": 3.0865306854248047,
70
+ "learning_rate": 0.0001964,
71
+ "loss": 1.1506,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.4,
76
+ "grad_norm": 3.093891143798828,
77
+ "learning_rate": 0.000196,
78
+ "loss": 0.9535,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.4,
83
+ "eval_accuracy": 0.618,
84
+ "eval_loss": 0.8965551853179932,
85
+ "eval_runtime": 12.9233,
86
+ "eval_samples_per_second": 77.38,
87
+ "eval_steps_per_second": 9.672,
88
+ "step": 100
89
+ },
90
+ {
91
+ "epoch": 0.44,
92
+ "grad_norm": 1.729243516921997,
93
+ "learning_rate": 0.0001956,
94
+ "loss": 0.9068,
95
+ "step": 110
96
+ },
97
+ {
98
+ "epoch": 0.48,
99
+ "grad_norm": 4.636998653411865,
100
+ "learning_rate": 0.0001952,
101
+ "loss": 0.8465,
102
+ "step": 120
103
+ },
104
+ {
105
+ "epoch": 0.52,
106
+ "grad_norm": 1.6259952783584595,
107
+ "learning_rate": 0.0001948,
108
+ "loss": 1.0864,
109
+ "step": 130
110
+ },
111
+ {
112
+ "epoch": 0.56,
113
+ "grad_norm": 2.269352436065674,
114
+ "learning_rate": 0.0001944,
115
+ "loss": 1.0165,
116
+ "step": 140
117
+ },
118
+ {
119
+ "epoch": 0.6,
120
+ "grad_norm": 2.516747236251831,
121
+ "learning_rate": 0.000194,
122
+ "loss": 0.8877,
123
+ "step": 150
124
+ },
125
+ {
126
+ "epoch": 0.64,
127
+ "grad_norm": 3.543081760406494,
128
+ "learning_rate": 0.00019360000000000002,
129
+ "loss": 0.9845,
130
+ "step": 160
131
+ },
132
+ {
133
+ "epoch": 0.68,
134
+ "grad_norm": 2.1874799728393555,
135
+ "learning_rate": 0.0001932,
136
+ "loss": 0.8626,
137
+ "step": 170
138
+ },
139
+ {
140
+ "epoch": 0.72,
141
+ "grad_norm": 2.4686241149902344,
142
+ "learning_rate": 0.0001928,
143
+ "loss": 0.8459,
144
+ "step": 180
145
+ },
146
+ {
147
+ "epoch": 0.76,
148
+ "grad_norm": 3.529902696609497,
149
+ "learning_rate": 0.00019240000000000001,
150
+ "loss": 1.0014,
151
+ "step": 190
152
+ },
153
+ {
154
+ "epoch": 0.8,
155
+ "grad_norm": 6.457658767700195,
156
+ "learning_rate": 0.000192,
157
+ "loss": 0.862,
158
+ "step": 200
159
+ },
160
+ {
161
+ "epoch": 0.8,
162
+ "eval_accuracy": 0.561,
163
+ "eval_loss": 1.1149431467056274,
164
+ "eval_runtime": 5.617,
165
+ "eval_samples_per_second": 178.029,
166
+ "eval_steps_per_second": 22.254,
167
+ "step": 200
168
+ },
169
+ {
170
+ "epoch": 0.84,
171
+ "grad_norm": 3.746286153793335,
172
+ "learning_rate": 0.0001916,
173
+ "loss": 0.9211,
174
+ "step": 210
175
+ },
176
+ {
177
+ "epoch": 0.88,
178
+ "grad_norm": 0.9981279373168945,
179
+ "learning_rate": 0.0001912,
180
+ "loss": 0.7935,
181
+ "step": 220
182
+ },
183
+ {
184
+ "epoch": 0.92,
185
+ "grad_norm": 4.057267189025879,
186
+ "learning_rate": 0.0001908,
187
+ "loss": 0.9092,
188
+ "step": 230
189
+ },
190
+ {
191
+ "epoch": 0.96,
192
+ "grad_norm": 1.723970890045166,
193
+ "learning_rate": 0.0001904,
194
+ "loss": 0.8029,
195
+ "step": 240
196
+ },
197
+ {
198
+ "epoch": 1.0,
199
+ "grad_norm": 2.485327959060669,
200
+ "learning_rate": 0.00019,
201
+ "loss": 0.8081,
202
+ "step": 250
203
+ },
204
+ {
205
+ "epoch": 1.04,
206
+ "grad_norm": 1.9523849487304688,
207
+ "learning_rate": 0.0001896,
208
+ "loss": 0.6713,
209
+ "step": 260
210
+ },
211
+ {
212
+ "epoch": 1.08,
213
+ "grad_norm": 1.719793438911438,
214
+ "learning_rate": 0.0001892,
215
+ "loss": 0.6921,
216
+ "step": 270
217
+ },
218
+ {
219
+ "epoch": 1.12,
220
+ "grad_norm": 1.283629059791565,
221
+ "learning_rate": 0.0001888,
222
+ "loss": 0.6989,
223
+ "step": 280
224
+ },
225
+ {
226
+ "epoch": 1.16,
227
+ "grad_norm": 14.220324516296387,
228
+ "learning_rate": 0.0001884,
229
+ "loss": 0.9012,
230
+ "step": 290
231
+ },
232
+ {
233
+ "epoch": 1.2,
234
+ "grad_norm": 1.621273398399353,
235
+ "learning_rate": 0.000188,
236
+ "loss": 0.7373,
237
+ "step": 300
238
+ },
239
+ {
240
+ "epoch": 1.2,
241
+ "eval_accuracy": 0.605,
242
+ "eval_loss": 0.8542758822441101,
243
+ "eval_runtime": 5.4409,
244
+ "eval_samples_per_second": 183.793,
245
+ "eval_steps_per_second": 22.974,
246
+ "step": 300
247
+ },
248
+ {
249
+ "epoch": 1.24,
250
+ "grad_norm": 1.4123197793960571,
251
+ "learning_rate": 0.0001876,
252
+ "loss": 0.8087,
253
+ "step": 310
254
+ },
255
+ {
256
+ "epoch": 1.28,
257
+ "grad_norm": 4.211037635803223,
258
+ "learning_rate": 0.00018720000000000002,
259
+ "loss": 0.7055,
260
+ "step": 320
261
+ },
262
+ {
263
+ "epoch": 1.32,
264
+ "grad_norm": 1.4883663654327393,
265
+ "learning_rate": 0.00018680000000000001,
266
+ "loss": 0.7988,
267
+ "step": 330
268
+ },
269
+ {
270
+ "epoch": 1.3599999999999999,
271
+ "grad_norm": 1.515805959701538,
272
+ "learning_rate": 0.00018640000000000003,
273
+ "loss": 0.8664,
274
+ "step": 340
275
+ },
276
+ {
277
+ "epoch": 1.4,
278
+ "grad_norm": 1.3907535076141357,
279
+ "learning_rate": 0.00018600000000000002,
280
+ "loss": 0.6773,
281
+ "step": 350
282
+ },
283
+ {
284
+ "epoch": 1.44,
285
+ "grad_norm": 1.5013153553009033,
286
+ "learning_rate": 0.0001856,
287
+ "loss": 0.6876,
288
+ "step": 360
289
+ },
290
+ {
291
+ "epoch": 1.48,
292
+ "grad_norm": 2.7530360221862793,
293
+ "learning_rate": 0.00018520000000000003,
294
+ "loss": 0.5832,
295
+ "step": 370
296
+ },
297
+ {
298
+ "epoch": 1.52,
299
+ "grad_norm": 2.470029354095459,
300
+ "learning_rate": 0.00018480000000000002,
301
+ "loss": 0.7209,
302
+ "step": 380
303
+ },
304
+ {
305
+ "epoch": 1.56,
306
+ "grad_norm": 1.3378897905349731,
307
+ "learning_rate": 0.0001844,
308
+ "loss": 0.6202,
309
+ "step": 390
310
+ },
311
+ {
312
+ "epoch": 1.6,
313
+ "grad_norm": 2.541882038116455,
314
+ "learning_rate": 0.00018400000000000003,
315
+ "loss": 0.6476,
316
+ "step": 400
317
+ },
318
+ {
319
+ "epoch": 1.6,
320
+ "eval_accuracy": 0.666,
321
+ "eval_loss": 0.7306695580482483,
322
+ "eval_runtime": 6.1267,
323
+ "eval_samples_per_second": 163.221,
324
+ "eval_steps_per_second": 20.403,
325
+ "step": 400
326
+ },
327
+ {
328
+ "epoch": 1.6400000000000001,
329
+ "grad_norm": 2.5715789794921875,
330
+ "learning_rate": 0.00018360000000000002,
331
+ "loss": 0.5758,
332
+ "step": 410
333
+ },
334
+ {
335
+ "epoch": 1.6800000000000002,
336
+ "grad_norm": 1.6180702447891235,
337
+ "learning_rate": 0.0001832,
338
+ "loss": 0.7008,
339
+ "step": 420
340
+ },
341
+ {
342
+ "epoch": 1.72,
343
+ "grad_norm": 2.0645174980163574,
344
+ "learning_rate": 0.00018280000000000003,
345
+ "loss": 0.7472,
346
+ "step": 430
347
+ },
348
+ {
349
+ "epoch": 1.76,
350
+ "grad_norm": 6.950054168701172,
351
+ "learning_rate": 0.00018240000000000002,
352
+ "loss": 0.9427,
353
+ "step": 440
354
+ },
355
+ {
356
+ "epoch": 1.8,
357
+ "grad_norm": 2.622056722640991,
358
+ "learning_rate": 0.000182,
359
+ "loss": 0.6746,
360
+ "step": 450
361
+ },
362
+ {
363
+ "epoch": 1.8399999999999999,
364
+ "grad_norm": 1.489776611328125,
365
+ "learning_rate": 0.00018160000000000002,
366
+ "loss": 0.7327,
367
+ "step": 460
368
+ },
369
+ {
370
+ "epoch": 1.88,
371
+ "grad_norm": 1.0776207447052002,
372
+ "learning_rate": 0.0001812,
373
+ "loss": 0.5842,
374
+ "step": 470
375
+ },
376
+ {
377
+ "epoch": 1.92,
378
+ "grad_norm": 2.286405086517334,
379
+ "learning_rate": 0.0001808,
380
+ "loss": 0.6962,
381
+ "step": 480
382
+ },
383
+ {
384
+ "epoch": 1.96,
385
+ "grad_norm": 2.593034267425537,
386
+ "learning_rate": 0.00018040000000000002,
387
+ "loss": 0.6214,
388
+ "step": 490
389
+ },
390
+ {
391
+ "epoch": 2.0,
392
+ "grad_norm": 1.7781627178192139,
393
+ "learning_rate": 0.00018,
394
+ "loss": 0.6712,
395
+ "step": 500
396
+ },
397
+ {
398
+ "epoch": 2.0,
399
+ "eval_accuracy": 0.694,
400
+ "eval_loss": 0.6953811049461365,
401
+ "eval_runtime": 5.961,
402
+ "eval_samples_per_second": 167.757,
403
+ "eval_steps_per_second": 20.97,
404
+ "step": 500
405
+ },
406
+ {
407
+ "epoch": 2.04,
408
+ "grad_norm": 1.331834316253662,
409
+ "learning_rate": 0.0001796,
410
+ "loss": 0.3866,
411
+ "step": 510
412
+ },
413
+ {
414
+ "epoch": 2.08,
415
+ "grad_norm": 1.5159146785736084,
416
+ "learning_rate": 0.00017920000000000002,
417
+ "loss": 0.5219,
418
+ "step": 520
419
+ },
420
+ {
421
+ "epoch": 2.12,
422
+ "grad_norm": 3.525988817214966,
423
+ "learning_rate": 0.0001788,
424
+ "loss": 0.4692,
425
+ "step": 530
426
+ },
427
+ {
428
+ "epoch": 2.16,
429
+ "grad_norm": 6.470268249511719,
430
+ "learning_rate": 0.0001784,
431
+ "loss": 0.4269,
432
+ "step": 540
433
+ },
434
+ {
435
+ "epoch": 2.2,
436
+ "grad_norm": 1.4871269464492798,
437
+ "learning_rate": 0.00017800000000000002,
438
+ "loss": 0.3844,
439
+ "step": 550
440
+ },
441
+ {
442
+ "epoch": 2.24,
443
+ "grad_norm": 3.404496669769287,
444
+ "learning_rate": 0.0001776,
445
+ "loss": 0.6704,
446
+ "step": 560
447
+ },
448
+ {
449
+ "epoch": 2.2800000000000002,
450
+ "grad_norm": 1.3112822771072388,
451
+ "learning_rate": 0.0001772,
452
+ "loss": 0.6746,
453
+ "step": 570
454
+ },
455
+ {
456
+ "epoch": 2.32,
457
+ "grad_norm": 5.380651473999023,
458
+ "learning_rate": 0.00017680000000000001,
459
+ "loss": 0.7249,
460
+ "step": 580
461
+ },
462
+ {
463
+ "epoch": 2.36,
464
+ "grad_norm": 1.006377935409546,
465
+ "learning_rate": 0.0001764,
466
+ "loss": 0.4615,
467
+ "step": 590
468
+ },
469
+ {
470
+ "epoch": 2.4,
471
+ "grad_norm": 2.6458024978637695,
472
+ "learning_rate": 0.00017600000000000002,
473
+ "loss": 0.4892,
474
+ "step": 600
475
+ },
476
+ {
477
+ "epoch": 2.4,
478
+ "eval_accuracy": 0.707,
479
+ "eval_loss": 0.6391066908836365,
480
+ "eval_runtime": 5.4367,
481
+ "eval_samples_per_second": 183.934,
482
+ "eval_steps_per_second": 22.992,
483
+ "step": 600
484
+ },
485
+ {
486
+ "epoch": 2.44,
487
+ "grad_norm": 3.180305004119873,
488
+ "learning_rate": 0.0001756,
489
+ "loss": 0.5289,
490
+ "step": 610
491
+ },
492
+ {
493
+ "epoch": 2.48,
494
+ "grad_norm": 3.7770426273345947,
495
+ "learning_rate": 0.0001752,
496
+ "loss": 0.4695,
497
+ "step": 620
498
+ },
499
+ {
500
+ "epoch": 2.52,
501
+ "grad_norm": 3.895836591720581,
502
+ "learning_rate": 0.00017480000000000002,
503
+ "loss": 0.4387,
504
+ "step": 630
505
+ },
506
+ {
507
+ "epoch": 2.56,
508
+ "grad_norm": 3.006532907485962,
509
+ "learning_rate": 0.0001744,
510
+ "loss": 0.5095,
511
+ "step": 640
512
+ },
513
+ {
514
+ "epoch": 2.6,
515
+ "grad_norm": 4.034090995788574,
516
+ "learning_rate": 0.000174,
517
+ "loss": 0.5297,
518
+ "step": 650
519
+ },
520
+ {
521
+ "epoch": 2.64,
522
+ "grad_norm": 2.2521889209747314,
523
+ "learning_rate": 0.00017360000000000002,
524
+ "loss": 0.4649,
525
+ "step": 660
526
+ },
527
+ {
528
+ "epoch": 2.68,
529
+ "grad_norm": 2.8466689586639404,
530
+ "learning_rate": 0.0001732,
531
+ "loss": 0.4819,
532
+ "step": 670
533
+ },
534
+ {
535
+ "epoch": 2.7199999999999998,
536
+ "grad_norm": 8.159728050231934,
537
+ "learning_rate": 0.0001728,
538
+ "loss": 0.5005,
539
+ "step": 680
540
+ },
541
+ {
542
+ "epoch": 2.76,
543
+ "grad_norm": 2.4270811080932617,
544
+ "learning_rate": 0.00017240000000000002,
545
+ "loss": 0.4146,
546
+ "step": 690
547
+ },
548
+ {
549
+ "epoch": 2.8,
550
+ "grad_norm": 4.790659427642822,
551
+ "learning_rate": 0.000172,
552
+ "loss": 0.5801,
553
+ "step": 700
554
+ },
555
+ {
556
+ "epoch": 2.8,
557
+ "eval_accuracy": 0.708,
558
+ "eval_loss": 0.6246920824050903,
559
+ "eval_runtime": 5.4112,
560
+ "eval_samples_per_second": 184.802,
561
+ "eval_steps_per_second": 23.1,
562
+ "step": 700
563
+ },
564
+ {
565
+ "epoch": 2.84,
566
+ "grad_norm": 2.981046676635742,
567
+ "learning_rate": 0.0001716,
568
+ "loss": 0.5311,
569
+ "step": 710
570
+ },
571
+ {
572
+ "epoch": 2.88,
573
+ "grad_norm": 3.6522932052612305,
574
+ "learning_rate": 0.00017120000000000001,
575
+ "loss": 0.4836,
576
+ "step": 720
577
+ },
578
+ {
579
+ "epoch": 2.92,
580
+ "grad_norm": 2.448866367340088,
581
+ "learning_rate": 0.0001708,
582
+ "loss": 0.6052,
583
+ "step": 730
584
+ },
585
+ {
586
+ "epoch": 2.96,
587
+ "grad_norm": 3.3822741508483887,
588
+ "learning_rate": 0.0001704,
589
+ "loss": 0.4695,
590
+ "step": 740
591
+ },
592
+ {
593
+ "epoch": 3.0,
594
+ "grad_norm": 3.214433431625366,
595
+ "learning_rate": 0.00017,
596
+ "loss": 0.6105,
597
+ "step": 750
598
+ },
599
+ {
600
+ "epoch": 3.04,
601
+ "grad_norm": 1.5671595335006714,
602
+ "learning_rate": 0.0001696,
603
+ "loss": 0.4189,
604
+ "step": 760
605
+ },
606
+ {
607
+ "epoch": 3.08,
608
+ "grad_norm": 3.648240327835083,
609
+ "learning_rate": 0.0001692,
610
+ "loss": 0.2745,
611
+ "step": 770
612
+ },
613
+ {
614
+ "epoch": 3.12,
615
+ "grad_norm": 1.588549017906189,
616
+ "learning_rate": 0.0001688,
617
+ "loss": 0.2882,
618
+ "step": 780
619
+ },
620
+ {
621
+ "epoch": 3.16,
622
+ "grad_norm": 2.041724920272827,
623
+ "learning_rate": 0.0001684,
624
+ "loss": 0.2642,
625
+ "step": 790
626
+ },
627
+ {
628
+ "epoch": 3.2,
629
+ "grad_norm": 4.308396816253662,
630
+ "learning_rate": 0.000168,
631
+ "loss": 0.3505,
632
+ "step": 800
633
+ },
634
+ {
635
+ "epoch": 3.2,
636
+ "eval_accuracy": 0.778,
637
+ "eval_loss": 0.6055798530578613,
638
+ "eval_runtime": 6.0689,
639
+ "eval_samples_per_second": 164.773,
640
+ "eval_steps_per_second": 20.597,
641
+ "step": 800
642
+ },
643
+ {
644
+ "epoch": 3.24,
645
+ "grad_norm": 2.717337131500244,
646
+ "learning_rate": 0.0001676,
647
+ "loss": 0.2626,
648
+ "step": 810
649
+ },
650
+ {
651
+ "epoch": 3.2800000000000002,
652
+ "grad_norm": 2.7859904766082764,
653
+ "learning_rate": 0.0001672,
654
+ "loss": 0.3826,
655
+ "step": 820
656
+ },
657
+ {
658
+ "epoch": 3.32,
659
+ "grad_norm": 3.864297389984131,
660
+ "learning_rate": 0.0001668,
661
+ "loss": 0.4313,
662
+ "step": 830
663
+ },
664
+ {
665
+ "epoch": 3.36,
666
+ "grad_norm": 3.782465934753418,
667
+ "learning_rate": 0.0001664,
668
+ "loss": 0.3361,
669
+ "step": 840
670
+ },
671
+ {
672
+ "epoch": 3.4,
673
+ "grad_norm": 3.1425442695617676,
674
+ "learning_rate": 0.000166,
675
+ "loss": 0.2759,
676
+ "step": 850
677
+ },
678
+ {
679
+ "epoch": 3.44,
680
+ "grad_norm": 3.1139609813690186,
681
+ "learning_rate": 0.0001656,
682
+ "loss": 0.3949,
683
+ "step": 860
684
+ },
685
+ {
686
+ "epoch": 3.48,
687
+ "grad_norm": 4.945987224578857,
688
+ "learning_rate": 0.0001652,
689
+ "loss": 0.3701,
690
+ "step": 870
691
+ },
692
+ {
693
+ "epoch": 3.52,
694
+ "grad_norm": 1.7310290336608887,
695
+ "learning_rate": 0.0001648,
696
+ "loss": 0.2879,
697
+ "step": 880
698
+ },
699
+ {
700
+ "epoch": 3.56,
701
+ "grad_norm": 5.650954246520996,
702
+ "learning_rate": 0.0001644,
703
+ "loss": 0.4443,
704
+ "step": 890
705
+ },
706
+ {
707
+ "epoch": 3.6,
708
+ "grad_norm": 2.8908536434173584,
709
+ "learning_rate": 0.000164,
710
+ "loss": 0.3503,
711
+ "step": 900
712
+ },
713
+ {
714
+ "epoch": 3.6,
715
+ "eval_accuracy": 0.743,
716
+ "eval_loss": 0.6263792514801025,
717
+ "eval_runtime": 5.7324,
718
+ "eval_samples_per_second": 174.449,
719
+ "eval_steps_per_second": 21.806,
720
+ "step": 900
721
+ },
722
+ {
723
+ "epoch": 3.64,
724
+ "grad_norm": 1.3399735689163208,
725
+ "learning_rate": 0.0001636,
726
+ "loss": 0.3594,
727
+ "step": 910
728
+ },
729
+ {
730
+ "epoch": 3.68,
731
+ "grad_norm": 1.3512450456619263,
732
+ "learning_rate": 0.0001632,
733
+ "loss": 0.31,
734
+ "step": 920
735
+ },
736
+ {
737
+ "epoch": 3.7199999999999998,
738
+ "grad_norm": 2.5052506923675537,
739
+ "learning_rate": 0.0001628,
740
+ "loss": 0.3234,
741
+ "step": 930
742
+ },
743
+ {
744
+ "epoch": 3.76,
745
+ "grad_norm": 7.230278968811035,
746
+ "learning_rate": 0.00016240000000000002,
747
+ "loss": 0.3559,
748
+ "step": 940
749
+ },
750
+ {
751
+ "epoch": 3.8,
752
+ "grad_norm": 1.464324712753296,
753
+ "learning_rate": 0.000162,
754
+ "loss": 0.2914,
755
+ "step": 950
756
+ },
757
+ {
758
+ "epoch": 3.84,
759
+ "grad_norm": 3.140822649002075,
760
+ "learning_rate": 0.00016160000000000002,
761
+ "loss": 0.2292,
762
+ "step": 960
763
+ },
764
+ {
765
+ "epoch": 3.88,
766
+ "grad_norm": 7.812990665435791,
767
+ "learning_rate": 0.00016120000000000002,
768
+ "loss": 0.401,
769
+ "step": 970
770
+ },
771
+ {
772
+ "epoch": 3.92,
773
+ "grad_norm": 7.805127143859863,
774
+ "learning_rate": 0.0001608,
775
+ "loss": 0.4106,
776
+ "step": 980
777
+ },
778
+ {
779
+ "epoch": 3.96,
780
+ "grad_norm": 5.645721435546875,
781
+ "learning_rate": 0.00016040000000000002,
782
+ "loss": 0.3844,
783
+ "step": 990
784
+ },
785
+ {
786
+ "epoch": 4.0,
787
+ "grad_norm": 2.4538416862487793,
788
+ "learning_rate": 0.00016,
789
+ "loss": 0.3416,
790
+ "step": 1000
791
+ },
792
+ {
793
+ "epoch": 4.0,
794
+ "eval_accuracy": 0.785,
795
+ "eval_loss": 0.583153247833252,
796
+ "eval_runtime": 5.857,
797
+ "eval_samples_per_second": 170.736,
798
+ "eval_steps_per_second": 21.342,
799
+ "step": 1000
800
+ },
801
+ {
802
+ "epoch": 4.04,
803
+ "grad_norm": 3.2518820762634277,
804
+ "learning_rate": 0.0001596,
805
+ "loss": 0.1471,
806
+ "step": 1010
807
+ },
808
+ {
809
+ "epoch": 4.08,
810
+ "grad_norm": 5.489699840545654,
811
+ "learning_rate": 0.00015920000000000002,
812
+ "loss": 0.1807,
813
+ "step": 1020
814
+ },
815
+ {
816
+ "epoch": 4.12,
817
+ "grad_norm": 0.6528148651123047,
818
+ "learning_rate": 0.0001588,
819
+ "loss": 0.189,
820
+ "step": 1030
821
+ },
822
+ {
823
+ "epoch": 4.16,
824
+ "grad_norm": 5.080958843231201,
825
+ "learning_rate": 0.00015840000000000003,
826
+ "loss": 0.246,
827
+ "step": 1040
828
+ },
829
+ {
830
+ "epoch": 4.2,
831
+ "grad_norm": 2.6048026084899902,
832
+ "learning_rate": 0.00015800000000000002,
833
+ "loss": 0.1937,
834
+ "step": 1050
835
+ },
836
+ {
837
+ "epoch": 4.24,
838
+ "grad_norm": 2.226956844329834,
839
+ "learning_rate": 0.0001576,
840
+ "loss": 0.1781,
841
+ "step": 1060
842
+ },
843
+ {
844
+ "epoch": 4.28,
845
+ "grad_norm": 1.7226955890655518,
846
+ "learning_rate": 0.00015720000000000003,
847
+ "loss": 0.2304,
848
+ "step": 1070
849
+ },
850
+ {
851
+ "epoch": 4.32,
852
+ "grad_norm": 4.231192111968994,
853
+ "learning_rate": 0.00015680000000000002,
854
+ "loss": 0.1828,
855
+ "step": 1080
856
+ },
857
+ {
858
+ "epoch": 4.36,
859
+ "grad_norm": 2.015580177307129,
860
+ "learning_rate": 0.0001564,
861
+ "loss": 0.2052,
862
+ "step": 1090
863
+ },
864
+ {
865
+ "epoch": 4.4,
866
+ "grad_norm": 1.2273974418640137,
867
+ "learning_rate": 0.00015600000000000002,
868
+ "loss": 0.1427,
869
+ "step": 1100
870
+ },
871
+ {
872
+ "epoch": 4.4,
873
+ "eval_accuracy": 0.769,
874
+ "eval_loss": 0.7296926975250244,
875
+ "eval_runtime": 5.2856,
876
+ "eval_samples_per_second": 189.193,
877
+ "eval_steps_per_second": 23.649,
878
+ "step": 1100
879
+ },
880
+ {
881
+ "epoch": 4.44,
882
+ "grad_norm": 12.377334594726562,
883
+ "learning_rate": 0.00015560000000000001,
884
+ "loss": 0.2476,
885
+ "step": 1110
886
+ },
887
+ {
888
+ "epoch": 4.48,
889
+ "grad_norm": 5.310695171356201,
890
+ "learning_rate": 0.0001552,
891
+ "loss": 0.2956,
892
+ "step": 1120
893
+ },
894
+ {
895
+ "epoch": 4.52,
896
+ "grad_norm": 1.6223011016845703,
897
+ "learning_rate": 0.00015480000000000002,
898
+ "loss": 0.3322,
899
+ "step": 1130
900
+ },
901
+ {
902
+ "epoch": 4.5600000000000005,
903
+ "grad_norm": 2.9505221843719482,
904
+ "learning_rate": 0.0001544,
905
+ "loss": 0.2848,
906
+ "step": 1140
907
+ },
908
+ {
909
+ "epoch": 4.6,
910
+ "grad_norm": 8.578746795654297,
911
+ "learning_rate": 0.000154,
912
+ "loss": 0.1866,
913
+ "step": 1150
914
+ },
915
+ {
916
+ "epoch": 4.64,
917
+ "grad_norm": 3.874375820159912,
918
+ "learning_rate": 0.00015360000000000002,
919
+ "loss": 0.2218,
920
+ "step": 1160
921
+ },
922
+ {
923
+ "epoch": 4.68,
924
+ "grad_norm": 2.2798233032226562,
925
+ "learning_rate": 0.0001532,
926
+ "loss": 0.2767,
927
+ "step": 1170
928
+ },
929
+ {
930
+ "epoch": 4.72,
931
+ "grad_norm": 1.274636149406433,
932
+ "learning_rate": 0.0001528,
933
+ "loss": 0.2295,
934
+ "step": 1180
935
+ },
936
+ {
937
+ "epoch": 4.76,
938
+ "grad_norm": 4.265562534332275,
939
+ "learning_rate": 0.00015240000000000002,
940
+ "loss": 0.2062,
941
+ "step": 1190
942
+ },
943
+ {
944
+ "epoch": 4.8,
945
+ "grad_norm": 3.9114277362823486,
946
+ "learning_rate": 0.000152,
947
+ "loss": 0.1982,
948
+ "step": 1200
949
+ },
950
+ {
951
+ "epoch": 4.8,
952
+ "eval_accuracy": 0.73,
953
+ "eval_loss": 0.7761247754096985,
954
+ "eval_runtime": 5.3281,
955
+ "eval_samples_per_second": 187.684,
956
+ "eval_steps_per_second": 23.46,
957
+ "step": 1200
958
+ },
959
+ {
960
+ "epoch": 4.84,
961
+ "grad_norm": 3.885077953338623,
962
+ "learning_rate": 0.0001516,
963
+ "loss": 0.2723,
964
+ "step": 1210
965
+ },
966
+ {
967
+ "epoch": 4.88,
968
+ "grad_norm": 3.1466147899627686,
969
+ "learning_rate": 0.00015120000000000002,
970
+ "loss": 0.1816,
971
+ "step": 1220
972
+ },
973
+ {
974
+ "epoch": 4.92,
975
+ "grad_norm": 1.5894757509231567,
976
+ "learning_rate": 0.0001508,
977
+ "loss": 0.2388,
978
+ "step": 1230
979
+ },
980
+ {
981
+ "epoch": 4.96,
982
+ "grad_norm": 2.17746901512146,
983
+ "learning_rate": 0.0001504,
984
+ "loss": 0.2346,
985
+ "step": 1240
986
+ },
987
+ {
988
+ "epoch": 5.0,
989
+ "grad_norm": 4.807671546936035,
990
+ "learning_rate": 0.00015000000000000001,
991
+ "loss": 0.1019,
992
+ "step": 1250
993
+ },
994
+ {
995
+ "epoch": 5.04,
996
+ "grad_norm": 0.5939441919326782,
997
+ "learning_rate": 0.0001496,
998
+ "loss": 0.2222,
999
+ "step": 1260
1000
+ },
1001
+ {
1002
+ "epoch": 5.08,
1003
+ "grad_norm": 6.009125232696533,
1004
+ "learning_rate": 0.0001492,
1005
+ "loss": 0.1772,
1006
+ "step": 1270
1007
+ },
1008
+ {
1009
+ "epoch": 5.12,
1010
+ "grad_norm": 0.48813578486442566,
1011
+ "learning_rate": 0.0001488,
1012
+ "loss": 0.1592,
1013
+ "step": 1280
1014
+ },
1015
+ {
1016
+ "epoch": 5.16,
1017
+ "grad_norm": 2.24546480178833,
1018
+ "learning_rate": 0.0001484,
1019
+ "loss": 0.2064,
1020
+ "step": 1290
1021
+ },
1022
+ {
1023
+ "epoch": 5.2,
1024
+ "grad_norm": 1.3820618391036987,
1025
+ "learning_rate": 0.000148,
1026
+ "loss": 0.193,
1027
+ "step": 1300
1028
+ },
1029
+ {
1030
+ "epoch": 5.2,
1031
+ "eval_accuracy": 0.741,
1032
+ "eval_loss": 0.8467380404472351,
1033
+ "eval_runtime": 6.1473,
1034
+ "eval_samples_per_second": 162.672,
1035
+ "eval_steps_per_second": 20.334,
1036
+ "step": 1300
1037
+ },
1038
+ {
1039
+ "epoch": 5.24,
1040
+ "grad_norm": 1.7147108316421509,
1041
+ "learning_rate": 0.0001476,
1042
+ "loss": 0.1436,
1043
+ "step": 1310
1044
+ },
1045
+ {
1046
+ "epoch": 5.28,
1047
+ "grad_norm": 0.31574681401252747,
1048
+ "learning_rate": 0.0001472,
1049
+ "loss": 0.1127,
1050
+ "step": 1320
1051
+ },
1052
+ {
1053
+ "epoch": 5.32,
1054
+ "grad_norm": 0.05109202861785889,
1055
+ "learning_rate": 0.00014680000000000002,
1056
+ "loss": 0.0844,
1057
+ "step": 1330
1058
+ },
1059
+ {
1060
+ "epoch": 5.36,
1061
+ "grad_norm": 3.080343246459961,
1062
+ "learning_rate": 0.0001464,
1063
+ "loss": 0.1685,
1064
+ "step": 1340
1065
+ },
1066
+ {
1067
+ "epoch": 5.4,
1068
+ "grad_norm": 0.8458659648895264,
1069
+ "learning_rate": 0.000146,
1070
+ "loss": 0.0856,
1071
+ "step": 1350
1072
+ },
1073
+ {
1074
+ "epoch": 5.44,
1075
+ "grad_norm": 2.8861546516418457,
1076
+ "learning_rate": 0.00014560000000000002,
1077
+ "loss": 0.1465,
1078
+ "step": 1360
1079
+ },
1080
+ {
1081
+ "epoch": 5.48,
1082
+ "grad_norm": 0.9648165702819824,
1083
+ "learning_rate": 0.0001452,
1084
+ "loss": 0.1534,
1085
+ "step": 1370
1086
+ },
1087
+ {
1088
+ "epoch": 5.52,
1089
+ "grad_norm": 0.8869705200195312,
1090
+ "learning_rate": 0.0001448,
1091
+ "loss": 0.1499,
1092
+ "step": 1380
1093
+ },
1094
+ {
1095
+ "epoch": 5.5600000000000005,
1096
+ "grad_norm": 5.056468963623047,
1097
+ "learning_rate": 0.0001444,
1098
+ "loss": 0.3252,
1099
+ "step": 1390
1100
+ },
1101
+ {
1102
+ "epoch": 5.6,
1103
+ "grad_norm": 4.4490485191345215,
1104
+ "learning_rate": 0.000144,
1105
+ "loss": 0.1831,
1106
+ "step": 1400
1107
+ },
1108
+ {
1109
+ "epoch": 5.6,
1110
+ "eval_accuracy": 0.774,
1111
+ "eval_loss": 0.6975213885307312,
1112
+ "eval_runtime": 5.5981,
1113
+ "eval_samples_per_second": 178.633,
1114
+ "eval_steps_per_second": 22.329,
1115
+ "step": 1400
1116
+ },
1117
+ {
1118
+ "epoch": 5.64,
1119
+ "grad_norm": 5.352523326873779,
1120
+ "learning_rate": 0.0001436,
1121
+ "loss": 0.1022,
1122
+ "step": 1410
1123
+ },
1124
+ {
1125
+ "epoch": 5.68,
1126
+ "grad_norm": 0.33864977955818176,
1127
+ "learning_rate": 0.0001432,
1128
+ "loss": 0.2627,
1129
+ "step": 1420
1130
+ },
1131
+ {
1132
+ "epoch": 5.72,
1133
+ "grad_norm": 3.7317004203796387,
1134
+ "learning_rate": 0.0001428,
1135
+ "loss": 0.1334,
1136
+ "step": 1430
1137
+ },
1138
+ {
1139
+ "epoch": 5.76,
1140
+ "grad_norm": 3.7961854934692383,
1141
+ "learning_rate": 0.0001424,
1142
+ "loss": 0.1658,
1143
+ "step": 1440
1144
+ },
1145
+ {
1146
+ "epoch": 5.8,
1147
+ "grad_norm": 0.13848626613616943,
1148
+ "learning_rate": 0.000142,
1149
+ "loss": 0.069,
1150
+ "step": 1450
1151
+ },
1152
+ {
1153
+ "epoch": 5.84,
1154
+ "grad_norm": 4.271240234375,
1155
+ "learning_rate": 0.0001416,
1156
+ "loss": 0.1401,
1157
+ "step": 1460
1158
+ },
1159
+ {
1160
+ "epoch": 5.88,
1161
+ "grad_norm": 3.202373504638672,
1162
+ "learning_rate": 0.0001412,
1163
+ "loss": 0.1468,
1164
+ "step": 1470
1165
+ },
1166
+ {
1167
+ "epoch": 5.92,
1168
+ "grad_norm": 0.39670008420944214,
1169
+ "learning_rate": 0.0001408,
1170
+ "loss": 0.0742,
1171
+ "step": 1480
1172
+ },
1173
+ {
1174
+ "epoch": 5.96,
1175
+ "grad_norm": 3.4956166744232178,
1176
+ "learning_rate": 0.0001404,
1177
+ "loss": 0.0929,
1178
+ "step": 1490
1179
+ },
1180
+ {
1181
+ "epoch": 6.0,
1182
+ "grad_norm": 3.5953352451324463,
1183
+ "learning_rate": 0.00014,
1184
+ "loss": 0.2612,
1185
+ "step": 1500
1186
+ },
1187
+ {
1188
+ "epoch": 6.0,
1189
+ "eval_accuracy": 0.775,
1190
+ "eval_loss": 0.8718728423118591,
1191
+ "eval_runtime": 5.287,
1192
+ "eval_samples_per_second": 189.142,
1193
+ "eval_steps_per_second": 23.643,
1194
+ "step": 1500
1195
+ },
1196
+ {
1197
+ "epoch": 6.04,
1198
+ "grad_norm": 0.334338515996933,
1199
+ "learning_rate": 0.0001396,
1200
+ "loss": 0.0812,
1201
+ "step": 1510
1202
+ },
1203
+ {
1204
+ "epoch": 6.08,
1205
+ "grad_norm": 4.4882612228393555,
1206
+ "learning_rate": 0.0001392,
1207
+ "loss": 0.0565,
1208
+ "step": 1520
1209
+ },
1210
+ {
1211
+ "epoch": 6.12,
1212
+ "grad_norm": 6.175742149353027,
1213
+ "learning_rate": 0.00013879999999999999,
1214
+ "loss": 0.0966,
1215
+ "step": 1530
1216
+ },
1217
+ {
1218
+ "epoch": 6.16,
1219
+ "grad_norm": 3.152432680130005,
1220
+ "learning_rate": 0.0001384,
1221
+ "loss": 0.1777,
1222
+ "step": 1540
1223
+ },
1224
+ {
1225
+ "epoch": 6.2,
1226
+ "grad_norm": 0.0891793891787529,
1227
+ "learning_rate": 0.000138,
1228
+ "loss": 0.0406,
1229
+ "step": 1550
1230
+ },
1231
+ {
1232
+ "epoch": 6.24,
1233
+ "grad_norm": 11.810345649719238,
1234
+ "learning_rate": 0.00013759999999999998,
1235
+ "loss": 0.0862,
1236
+ "step": 1560
1237
+ },
1238
+ {
1239
+ "epoch": 6.28,
1240
+ "grad_norm": 0.12118436396121979,
1241
+ "learning_rate": 0.00013720000000000003,
1242
+ "loss": 0.0366,
1243
+ "step": 1570
1244
+ },
1245
+ {
1246
+ "epoch": 6.32,
1247
+ "grad_norm": 12.970296859741211,
1248
+ "learning_rate": 0.00013680000000000002,
1249
+ "loss": 0.0596,
1250
+ "step": 1580
1251
+ },
1252
+ {
1253
+ "epoch": 6.36,
1254
+ "grad_norm": 0.05544751510024071,
1255
+ "learning_rate": 0.0001364,
1256
+ "loss": 0.1192,
1257
+ "step": 1590
1258
+ },
1259
+ {
1260
+ "epoch": 6.4,
1261
+ "grad_norm": 0.16358701884746552,
1262
+ "learning_rate": 0.00013600000000000003,
1263
+ "loss": 0.102,
1264
+ "step": 1600
1265
+ },
1266
+ {
1267
+ "epoch": 6.4,
1268
+ "eval_accuracy": 0.788,
1269
+ "eval_loss": 0.9045199155807495,
1270
+ "eval_runtime": 5.1633,
1271
+ "eval_samples_per_second": 193.676,
1272
+ "eval_steps_per_second": 24.21,
1273
+ "step": 1600
1274
+ },
1275
+ {
1276
+ "epoch": 6.44,
1277
+ "grad_norm": 2.016242742538452,
1278
+ "learning_rate": 0.00013560000000000002,
1279
+ "loss": 0.0633,
1280
+ "step": 1610
1281
+ },
1282
+ {
1283
+ "epoch": 6.48,
1284
+ "grad_norm": 0.19605639576911926,
1285
+ "learning_rate": 0.0001352,
1286
+ "loss": 0.0223,
1287
+ "step": 1620
1288
+ },
1289
+ {
1290
+ "epoch": 6.52,
1291
+ "grad_norm": 0.13813719153404236,
1292
+ "learning_rate": 0.00013480000000000002,
1293
+ "loss": 0.1615,
1294
+ "step": 1630
1295
+ },
1296
+ {
1297
+ "epoch": 6.5600000000000005,
1298
+ "grad_norm": 0.18716655671596527,
1299
+ "learning_rate": 0.00013440000000000001,
1300
+ "loss": 0.0248,
1301
+ "step": 1640
1302
+ },
1303
+ {
1304
+ "epoch": 6.6,
1305
+ "grad_norm": 0.14301587641239166,
1306
+ "learning_rate": 0.000134,
1307
+ "loss": 0.0507,
1308
+ "step": 1650
1309
+ },
1310
+ {
1311
+ "epoch": 6.64,
1312
+ "grad_norm": 0.10028518736362457,
1313
+ "learning_rate": 0.00013360000000000002,
1314
+ "loss": 0.0334,
1315
+ "step": 1660
1316
+ },
1317
+ {
1318
+ "epoch": 6.68,
1319
+ "grad_norm": 6.466274261474609,
1320
+ "learning_rate": 0.0001332,
1321
+ "loss": 0.0881,
1322
+ "step": 1670
1323
+ },
1324
+ {
1325
+ "epoch": 6.72,
1326
+ "grad_norm": 0.15147142112255096,
1327
+ "learning_rate": 0.0001328,
1328
+ "loss": 0.1162,
1329
+ "step": 1680
1330
+ },
1331
+ {
1332
+ "epoch": 6.76,
1333
+ "grad_norm": 0.06184665858745575,
1334
+ "learning_rate": 0.00013240000000000002,
1335
+ "loss": 0.085,
1336
+ "step": 1690
1337
+ },
1338
+ {
1339
+ "epoch": 6.8,
1340
+ "grad_norm": 9.578435897827148,
1341
+ "learning_rate": 0.000132,
1342
+ "loss": 0.1029,
1343
+ "step": 1700
1344
+ },
1345
+ {
1346
+ "epoch": 6.8,
1347
+ "eval_accuracy": 0.783,
1348
+ "eval_loss": 0.9655129313468933,
1349
+ "eval_runtime": 5.9259,
1350
+ "eval_samples_per_second": 168.75,
1351
+ "eval_steps_per_second": 21.094,
1352
+ "step": 1700
1353
+ },
1354
+ {
1355
+ "epoch": 6.84,
1356
+ "grad_norm": 9.377839088439941,
1357
+ "learning_rate": 0.0001316,
1358
+ "loss": 0.0559,
1359
+ "step": 1710
1360
+ },
1361
+ {
1362
+ "epoch": 6.88,
1363
+ "grad_norm": 9.196260452270508,
1364
+ "learning_rate": 0.00013120000000000002,
1365
+ "loss": 0.0919,
1366
+ "step": 1720
1367
+ },
1368
+ {
1369
+ "epoch": 6.92,
1370
+ "grad_norm": 0.6224649548530579,
1371
+ "learning_rate": 0.0001308,
1372
+ "loss": 0.0328,
1373
+ "step": 1730
1374
+ },
1375
+ {
1376
+ "epoch": 6.96,
1377
+ "grad_norm": 0.8349862098693848,
1378
+ "learning_rate": 0.0001304,
1379
+ "loss": 0.0837,
1380
+ "step": 1740
1381
+ },
1382
+ {
1383
+ "epoch": 7.0,
1384
+ "grad_norm": 7.744389057159424,
1385
+ "learning_rate": 0.00013000000000000002,
1386
+ "loss": 0.1493,
1387
+ "step": 1750
1388
+ },
1389
+ {
1390
+ "epoch": 7.04,
1391
+ "grad_norm": 19.332578659057617,
1392
+ "learning_rate": 0.0001296,
1393
+ "loss": 0.1517,
1394
+ "step": 1760
1395
+ },
1396
+ {
1397
+ "epoch": 7.08,
1398
+ "grad_norm": 9.477415084838867,
1399
+ "learning_rate": 0.00012920000000000002,
1400
+ "loss": 0.0612,
1401
+ "step": 1770
1402
+ },
1403
+ {
1404
+ "epoch": 7.12,
1405
+ "grad_norm": 0.04479631036520004,
1406
+ "learning_rate": 0.00012880000000000001,
1407
+ "loss": 0.0629,
1408
+ "step": 1780
1409
+ },
1410
+ {
1411
+ "epoch": 7.16,
1412
+ "grad_norm": 0.03189521282911301,
1413
+ "learning_rate": 0.0001284,
1414
+ "loss": 0.0544,
1415
+ "step": 1790
1416
+ },
1417
+ {
1418
+ "epoch": 7.2,
1419
+ "grad_norm": 0.04392462968826294,
1420
+ "learning_rate": 0.00012800000000000002,
1421
+ "loss": 0.0735,
1422
+ "step": 1800
1423
+ },
1424
+ {
1425
+ "epoch": 7.2,
1426
+ "eval_accuracy": 0.78,
1427
+ "eval_loss": 0.9905818700790405,
1428
+ "eval_runtime": 5.871,
1429
+ "eval_samples_per_second": 170.328,
1430
+ "eval_steps_per_second": 21.291,
1431
+ "step": 1800
1432
+ },
1433
+ {
1434
+ "epoch": 7.24,
1435
+ "grad_norm": 0.06359164416790009,
1436
+ "learning_rate": 0.0001276,
1437
+ "loss": 0.1629,
1438
+ "step": 1810
1439
+ },
1440
+ {
1441
+ "epoch": 7.28,
1442
+ "grad_norm": 1.7004547119140625,
1443
+ "learning_rate": 0.0001272,
1444
+ "loss": 0.1058,
1445
+ "step": 1820
1446
+ },
1447
+ {
1448
+ "epoch": 7.32,
1449
+ "grad_norm": 0.6779342293739319,
1450
+ "learning_rate": 0.00012680000000000002,
1451
+ "loss": 0.0404,
1452
+ "step": 1830
1453
+ },
1454
+ {
1455
+ "epoch": 7.36,
1456
+ "grad_norm": 0.050379179418087006,
1457
+ "learning_rate": 0.0001264,
1458
+ "loss": 0.0565,
1459
+ "step": 1840
1460
+ },
1461
+ {
1462
+ "epoch": 7.4,
1463
+ "grad_norm": 0.23603555560112,
1464
+ "learning_rate": 0.000126,
1465
+ "loss": 0.0316,
1466
+ "step": 1850
1467
+ },
1468
+ {
1469
+ "epoch": 7.44,
1470
+ "grad_norm": 0.030859237536787987,
1471
+ "learning_rate": 0.00012560000000000002,
1472
+ "loss": 0.0781,
1473
+ "step": 1860
1474
+ },
1475
+ {
1476
+ "epoch": 7.48,
1477
+ "grad_norm": 0.14111267030239105,
1478
+ "learning_rate": 0.0001252,
1479
+ "loss": 0.1127,
1480
+ "step": 1870
1481
+ },
1482
+ {
1483
+ "epoch": 7.52,
1484
+ "grad_norm": 5.0431084632873535,
1485
+ "learning_rate": 0.0001248,
1486
+ "loss": 0.0226,
1487
+ "step": 1880
1488
+ },
1489
+ {
1490
+ "epoch": 7.5600000000000005,
1491
+ "grad_norm": 0.1641249656677246,
1492
+ "learning_rate": 0.00012440000000000002,
1493
+ "loss": 0.0875,
1494
+ "step": 1890
1495
+ },
1496
+ {
1497
+ "epoch": 7.6,
1498
+ "grad_norm": 1.8398263454437256,
1499
+ "learning_rate": 0.000124,
1500
+ "loss": 0.0715,
1501
+ "step": 1900
1502
+ },
1503
+ {
1504
+ "epoch": 7.6,
1505
+ "eval_accuracy": 0.787,
1506
+ "eval_loss": 0.8893071413040161,
1507
+ "eval_runtime": 5.6004,
1508
+ "eval_samples_per_second": 178.559,
1509
+ "eval_steps_per_second": 22.32,
1510
+ "step": 1900
1511
+ },
1512
+ {
1513
+ "epoch": 7.64,
1514
+ "grad_norm": 11.241652488708496,
1515
+ "learning_rate": 0.0001236,
1516
+ "loss": 0.0418,
1517
+ "step": 1910
1518
+ },
1519
+ {
1520
+ "epoch": 7.68,
1521
+ "grad_norm": 1.0991768836975098,
1522
+ "learning_rate": 0.0001232,
1523
+ "loss": 0.0755,
1524
+ "step": 1920
1525
+ },
1526
+ {
1527
+ "epoch": 7.72,
1528
+ "grad_norm": 19.279315948486328,
1529
+ "learning_rate": 0.0001228,
1530
+ "loss": 0.2096,
1531
+ "step": 1930
1532
+ },
1533
+ {
1534
+ "epoch": 7.76,
1535
+ "grad_norm": 0.7445815801620483,
1536
+ "learning_rate": 0.0001224,
1537
+ "loss": 0.0395,
1538
+ "step": 1940
1539
+ },
1540
+ {
1541
+ "epoch": 7.8,
1542
+ "grad_norm": 0.07692953944206238,
1543
+ "learning_rate": 0.000122,
1544
+ "loss": 0.0305,
1545
+ "step": 1950
1546
+ },
1547
+ {
1548
+ "epoch": 7.84,
1549
+ "grad_norm": 0.1866171956062317,
1550
+ "learning_rate": 0.0001216,
1551
+ "loss": 0.0379,
1552
+ "step": 1960
1553
+ },
1554
+ {
1555
+ "epoch": 7.88,
1556
+ "grad_norm": 11.213932037353516,
1557
+ "learning_rate": 0.0001212,
1558
+ "loss": 0.0625,
1559
+ "step": 1970
1560
+ },
1561
+ {
1562
+ "epoch": 7.92,
1563
+ "grad_norm": 0.762454092502594,
1564
+ "learning_rate": 0.0001208,
1565
+ "loss": 0.0336,
1566
+ "step": 1980
1567
+ },
1568
+ {
1569
+ "epoch": 7.96,
1570
+ "grad_norm": 9.390318870544434,
1571
+ "learning_rate": 0.0001204,
1572
+ "loss": 0.0813,
1573
+ "step": 1990
1574
+ },
1575
+ {
1576
+ "epoch": 8.0,
1577
+ "grad_norm": 0.019455738365650177,
1578
+ "learning_rate": 0.00012,
1579
+ "loss": 0.1254,
1580
+ "step": 2000
1581
+ },
1582
+ {
1583
+ "epoch": 8.0,
1584
+ "eval_accuracy": 0.761,
1585
+ "eval_loss": 1.1220637559890747,
1586
+ "eval_runtime": 5.5365,
1587
+ "eval_samples_per_second": 180.619,
1588
+ "eval_steps_per_second": 22.577,
1589
+ "step": 2000
1590
+ },
1591
+ {
1592
+ "epoch": 8.04,
1593
+ "grad_norm": 0.1315418779850006,
1594
+ "learning_rate": 0.00011960000000000001,
1595
+ "loss": 0.1103,
1596
+ "step": 2010
1597
+ },
1598
+ {
1599
+ "epoch": 8.08,
1600
+ "grad_norm": 0.4731353521347046,
1601
+ "learning_rate": 0.0001192,
1602
+ "loss": 0.0696,
1603
+ "step": 2020
1604
+ },
1605
+ {
1606
+ "epoch": 8.12,
1607
+ "grad_norm": 2.56032133102417,
1608
+ "learning_rate": 0.0001188,
1609
+ "loss": 0.0488,
1610
+ "step": 2030
1611
+ },
1612
+ {
1613
+ "epoch": 8.16,
1614
+ "grad_norm": 1.2172237634658813,
1615
+ "learning_rate": 0.0001184,
1616
+ "loss": 0.031,
1617
+ "step": 2040
1618
+ },
1619
+ {
1620
+ "epoch": 8.2,
1621
+ "grad_norm": 0.04160616546869278,
1622
+ "learning_rate": 0.000118,
1623
+ "loss": 0.0332,
1624
+ "step": 2050
1625
+ },
1626
+ {
1627
+ "epoch": 8.24,
1628
+ "grad_norm": 0.02498347871005535,
1629
+ "learning_rate": 0.0001176,
1630
+ "loss": 0.0046,
1631
+ "step": 2060
1632
+ },
1633
+ {
1634
+ "epoch": 8.28,
1635
+ "grad_norm": 0.023147333413362503,
1636
+ "learning_rate": 0.0001172,
1637
+ "loss": 0.0354,
1638
+ "step": 2070
1639
+ },
1640
+ {
1641
+ "epoch": 8.32,
1642
+ "grad_norm": 24.763408660888672,
1643
+ "learning_rate": 0.00011679999999999999,
1644
+ "loss": 0.0182,
1645
+ "step": 2080
1646
+ },
1647
+ {
1648
+ "epoch": 8.36,
1649
+ "grad_norm": 0.1869715005159378,
1650
+ "learning_rate": 0.0001164,
1651
+ "loss": 0.0198,
1652
+ "step": 2090
1653
+ },
1654
+ {
1655
+ "epoch": 8.4,
1656
+ "grad_norm": 0.013836881145834923,
1657
+ "learning_rate": 0.000116,
1658
+ "loss": 0.021,
1659
+ "step": 2100
1660
+ },
1661
+ {
1662
+ "epoch": 8.4,
1663
+ "eval_accuracy": 0.779,
1664
+ "eval_loss": 1.1648471355438232,
1665
+ "eval_runtime": 5.2413,
1666
+ "eval_samples_per_second": 190.794,
1667
+ "eval_steps_per_second": 23.849,
1668
+ "step": 2100
1669
+ },
1670
+ {
1671
+ "epoch": 8.44,
1672
+ "grad_norm": 0.022183042019605637,
1673
+ "learning_rate": 0.00011559999999999999,
1674
+ "loss": 0.0551,
1675
+ "step": 2110
1676
+ },
1677
+ {
1678
+ "epoch": 8.48,
1679
+ "grad_norm": 7.832488536834717,
1680
+ "learning_rate": 0.0001152,
1681
+ "loss": 0.1164,
1682
+ "step": 2120
1683
+ },
1684
+ {
1685
+ "epoch": 8.52,
1686
+ "grad_norm": 0.02375981956720352,
1687
+ "learning_rate": 0.0001148,
1688
+ "loss": 0.0164,
1689
+ "step": 2130
1690
+ },
1691
+ {
1692
+ "epoch": 8.56,
1693
+ "grad_norm": 0.018566014245152473,
1694
+ "learning_rate": 0.0001144,
1695
+ "loss": 0.0155,
1696
+ "step": 2140
1697
+ },
1698
+ {
1699
+ "epoch": 8.6,
1700
+ "grad_norm": 2.625162124633789,
1701
+ "learning_rate": 0.00011399999999999999,
1702
+ "loss": 0.0702,
1703
+ "step": 2150
1704
+ },
1705
+ {
1706
+ "epoch": 8.64,
1707
+ "grad_norm": 0.022067397832870483,
1708
+ "learning_rate": 0.0001136,
1709
+ "loss": 0.0441,
1710
+ "step": 2160
1711
+ },
1712
+ {
1713
+ "epoch": 8.68,
1714
+ "grad_norm": 0.030472321435809135,
1715
+ "learning_rate": 0.0001132,
1716
+ "loss": 0.012,
1717
+ "step": 2170
1718
+ },
1719
+ {
1720
+ "epoch": 8.72,
1721
+ "grad_norm": 0.957785964012146,
1722
+ "learning_rate": 0.00011279999999999999,
1723
+ "loss": 0.0151,
1724
+ "step": 2180
1725
+ },
1726
+ {
1727
+ "epoch": 8.76,
1728
+ "grad_norm": 3.766643524169922,
1729
+ "learning_rate": 0.00011240000000000002,
1730
+ "loss": 0.0853,
1731
+ "step": 2190
1732
+ },
1733
+ {
1734
+ "epoch": 8.8,
1735
+ "grad_norm": 0.32392239570617676,
1736
+ "learning_rate": 0.00011200000000000001,
1737
+ "loss": 0.0133,
1738
+ "step": 2200
1739
+ },
1740
+ {
1741
+ "epoch": 8.8,
1742
+ "eval_accuracy": 0.806,
1743
+ "eval_loss": 0.9857252240180969,
1744
+ "eval_runtime": 5.2301,
1745
+ "eval_samples_per_second": 191.201,
1746
+ "eval_steps_per_second": 23.9,
1747
+ "step": 2200
1748
+ },
1749
+ {
1750
+ "epoch": 8.84,
1751
+ "grad_norm": 0.021277396008372307,
1752
+ "learning_rate": 0.00011160000000000002,
1753
+ "loss": 0.0314,
1754
+ "step": 2210
1755
+ },
1756
+ {
1757
+ "epoch": 8.88,
1758
+ "grad_norm": 13.449085235595703,
1759
+ "learning_rate": 0.00011120000000000002,
1760
+ "loss": 0.0629,
1761
+ "step": 2220
1762
+ },
1763
+ {
1764
+ "epoch": 8.92,
1765
+ "grad_norm": 0.02039525657892227,
1766
+ "learning_rate": 0.00011080000000000001,
1767
+ "loss": 0.0358,
1768
+ "step": 2230
1769
+ },
1770
+ {
1771
+ "epoch": 8.96,
1772
+ "grad_norm": 0.10957217961549759,
1773
+ "learning_rate": 0.00011040000000000001,
1774
+ "loss": 0.0379,
1775
+ "step": 2240
1776
+ },
1777
+ {
1778
+ "epoch": 9.0,
1779
+ "grad_norm": 0.016647523269057274,
1780
+ "learning_rate": 0.00011000000000000002,
1781
+ "loss": 0.0125,
1782
+ "step": 2250
1783
+ },
1784
+ {
1785
+ "epoch": 9.04,
1786
+ "grad_norm": 13.669293403625488,
1787
+ "learning_rate": 0.00010960000000000001,
1788
+ "loss": 0.035,
1789
+ "step": 2260
1790
+ },
1791
+ {
1792
+ "epoch": 9.08,
1793
+ "grad_norm": 0.012543587014079094,
1794
+ "learning_rate": 0.00010920000000000001,
1795
+ "loss": 0.031,
1796
+ "step": 2270
1797
+ },
1798
+ {
1799
+ "epoch": 9.12,
1800
+ "grad_norm": 0.011763609014451504,
1801
+ "learning_rate": 0.00010880000000000002,
1802
+ "loss": 0.0299,
1803
+ "step": 2280
1804
+ },
1805
+ {
1806
+ "epoch": 9.16,
1807
+ "grad_norm": 0.0155358100309968,
1808
+ "learning_rate": 0.00010840000000000002,
1809
+ "loss": 0.0339,
1810
+ "step": 2290
1811
+ },
1812
+ {
1813
+ "epoch": 9.2,
1814
+ "grad_norm": 0.2667885720729828,
1815
+ "learning_rate": 0.00010800000000000001,
1816
+ "loss": 0.0086,
1817
+ "step": 2300
1818
+ },
1819
+ {
1820
+ "epoch": 9.2,
1821
+ "eval_accuracy": 0.799,
1822
+ "eval_loss": 1.036458134651184,
1823
+ "eval_runtime": 6.0166,
1824
+ "eval_samples_per_second": 166.207,
1825
+ "eval_steps_per_second": 20.776,
1826
+ "step": 2300
1827
+ },
1828
+ {
1829
+ "epoch": 9.24,
1830
+ "grad_norm": 0.008514340966939926,
1831
+ "learning_rate": 0.00010760000000000001,
1832
+ "loss": 0.0177,
1833
+ "step": 2310
1834
+ },
1835
+ {
1836
+ "epoch": 9.28,
1837
+ "grad_norm": 0.033006951212882996,
1838
+ "learning_rate": 0.00010720000000000002,
1839
+ "loss": 0.0041,
1840
+ "step": 2320
1841
+ },
1842
+ {
1843
+ "epoch": 9.32,
1844
+ "grad_norm": 17.495553970336914,
1845
+ "learning_rate": 0.00010680000000000001,
1846
+ "loss": 0.0506,
1847
+ "step": 2330
1848
+ },
1849
+ {
1850
+ "epoch": 9.36,
1851
+ "grad_norm": 29.127471923828125,
1852
+ "learning_rate": 0.00010640000000000001,
1853
+ "loss": 0.0216,
1854
+ "step": 2340
1855
+ },
1856
+ {
1857
+ "epoch": 9.4,
1858
+ "grad_norm": 0.03006252832710743,
1859
+ "learning_rate": 0.00010600000000000002,
1860
+ "loss": 0.0273,
1861
+ "step": 2350
1862
+ },
1863
+ {
1864
+ "epoch": 9.44,
1865
+ "grad_norm": 0.021872328594326973,
1866
+ "learning_rate": 0.0001056,
1867
+ "loss": 0.0027,
1868
+ "step": 2360
1869
+ },
1870
+ {
1871
+ "epoch": 9.48,
1872
+ "grad_norm": 0.01040344312787056,
1873
+ "learning_rate": 0.00010520000000000001,
1874
+ "loss": 0.018,
1875
+ "step": 2370
1876
+ },
1877
+ {
1878
+ "epoch": 9.52,
1879
+ "grad_norm": 9.743282318115234,
1880
+ "learning_rate": 0.00010480000000000001,
1881
+ "loss": 0.1394,
1882
+ "step": 2380
1883
+ },
1884
+ {
1885
+ "epoch": 9.56,
1886
+ "grad_norm": 0.012420209124684334,
1887
+ "learning_rate": 0.0001044,
1888
+ "loss": 0.0257,
1889
+ "step": 2390
1890
+ },
1891
+ {
1892
+ "epoch": 9.6,
1893
+ "grad_norm": 0.21533189713954926,
1894
+ "learning_rate": 0.00010400000000000001,
1895
+ "loss": 0.0223,
1896
+ "step": 2400
1897
+ },
1898
+ {
1899
+ "epoch": 9.6,
1900
+ "eval_accuracy": 0.812,
1901
+ "eval_loss": 0.982612669467926,
1902
+ "eval_runtime": 5.176,
1903
+ "eval_samples_per_second": 193.201,
1904
+ "eval_steps_per_second": 24.15,
1905
+ "step": 2400
1906
+ },
1907
+ {
1908
+ "epoch": 9.64,
1909
+ "grad_norm": 0.02012082375586033,
1910
+ "learning_rate": 0.00010360000000000001,
1911
+ "loss": 0.0371,
1912
+ "step": 2410
1913
+ },
1914
+ {
1915
+ "epoch": 9.68,
1916
+ "grad_norm": 0.03367308899760246,
1917
+ "learning_rate": 0.0001032,
1918
+ "loss": 0.0053,
1919
+ "step": 2420
1920
+ },
1921
+ {
1922
+ "epoch": 9.72,
1923
+ "grad_norm": 0.9374167919158936,
1924
+ "learning_rate": 0.0001028,
1925
+ "loss": 0.058,
1926
+ "step": 2430
1927
+ },
1928
+ {
1929
+ "epoch": 9.76,
1930
+ "grad_norm": 0.010282558389008045,
1931
+ "learning_rate": 0.00010240000000000001,
1932
+ "loss": 0.0438,
1933
+ "step": 2440
1934
+ },
1935
+ {
1936
+ "epoch": 9.8,
1937
+ "grad_norm": 0.009324116632342339,
1938
+ "learning_rate": 0.00010200000000000001,
1939
+ "loss": 0.0028,
1940
+ "step": 2450
1941
+ },
1942
+ {
1943
+ "epoch": 9.84,
1944
+ "grad_norm": 0.009123372845351696,
1945
+ "learning_rate": 0.0001016,
1946
+ "loss": 0.1,
1947
+ "step": 2460
1948
+ },
1949
+ {
1950
+ "epoch": 9.88,
1951
+ "grad_norm": 0.010366716422140598,
1952
+ "learning_rate": 0.00010120000000000001,
1953
+ "loss": 0.1172,
1954
+ "step": 2470
1955
+ },
1956
+ {
1957
+ "epoch": 9.92,
1958
+ "grad_norm": 3.760880708694458,
1959
+ "learning_rate": 0.00010080000000000001,
1960
+ "loss": 0.005,
1961
+ "step": 2480
1962
+ },
1963
+ {
1964
+ "epoch": 9.96,
1965
+ "grad_norm": 0.033989980816841125,
1966
+ "learning_rate": 0.0001004,
1967
+ "loss": 0.0441,
1968
+ "step": 2490
1969
+ },
1970
+ {
1971
+ "epoch": 10.0,
1972
+ "grad_norm": 0.008832584135234356,
1973
+ "learning_rate": 0.0001,
1974
+ "loss": 0.0023,
1975
+ "step": 2500
1976
+ },
1977
+ {
1978
+ "epoch": 10.0,
1979
+ "eval_accuracy": 0.795,
1980
+ "eval_loss": 1.0696669816970825,
1981
+ "eval_runtime": 5.262,
1982
+ "eval_samples_per_second": 190.041,
1983
+ "eval_steps_per_second": 23.755,
1984
+ "step": 2500
1985
+ },
1986
+ {
1987
+ "epoch": 10.04,
1988
+ "grad_norm": 0.06487290561199188,
1989
+ "learning_rate": 9.960000000000001e-05,
1990
+ "loss": 0.0359,
1991
+ "step": 2510
1992
+ },
1993
+ {
1994
+ "epoch": 10.08,
1995
+ "grad_norm": 0.06183789297938347,
1996
+ "learning_rate": 9.92e-05,
1997
+ "loss": 0.0019,
1998
+ "step": 2520
1999
+ },
2000
+ {
2001
+ "epoch": 10.12,
2002
+ "grad_norm": 0.007254609372466803,
2003
+ "learning_rate": 9.88e-05,
2004
+ "loss": 0.0443,
2005
+ "step": 2530
2006
+ },
2007
+ {
2008
+ "epoch": 10.16,
2009
+ "grad_norm": 27.718936920166016,
2010
+ "learning_rate": 9.84e-05,
2011
+ "loss": 0.0429,
2012
+ "step": 2540
2013
+ },
2014
+ {
2015
+ "epoch": 10.2,
2016
+ "grad_norm": 0.007617979310452938,
2017
+ "learning_rate": 9.8e-05,
2018
+ "loss": 0.0034,
2019
+ "step": 2550
2020
+ },
2021
+ {
2022
+ "epoch": 10.24,
2023
+ "grad_norm": 0.42735064029693604,
2024
+ "learning_rate": 9.76e-05,
2025
+ "loss": 0.028,
2026
+ "step": 2560
2027
+ },
2028
+ {
2029
+ "epoch": 10.28,
2030
+ "grad_norm": 0.00830338429659605,
2031
+ "learning_rate": 9.72e-05,
2032
+ "loss": 0.0049,
2033
+ "step": 2570
2034
+ },
2035
+ {
2036
+ "epoch": 10.32,
2037
+ "grad_norm": 0.007815328426659107,
2038
+ "learning_rate": 9.680000000000001e-05,
2039
+ "loss": 0.0015,
2040
+ "step": 2580
2041
+ },
2042
+ {
2043
+ "epoch": 10.36,
2044
+ "grad_norm": 0.023354820907115936,
2045
+ "learning_rate": 9.64e-05,
2046
+ "loss": 0.0017,
2047
+ "step": 2590
2048
+ },
2049
+ {
2050
+ "epoch": 10.4,
2051
+ "grad_norm": 0.02857878804206848,
2052
+ "learning_rate": 9.6e-05,
2053
+ "loss": 0.0021,
2054
+ "step": 2600
2055
+ },
2056
+ {
2057
+ "epoch": 10.4,
2058
+ "eval_accuracy": 0.815,
2059
+ "eval_loss": 1.0489550828933716,
2060
+ "eval_runtime": 5.2508,
2061
+ "eval_samples_per_second": 190.447,
2062
+ "eval_steps_per_second": 23.806,
2063
+ "step": 2600
2064
+ },
2065
+ {
2066
+ "epoch": 10.44,
2067
+ "grad_norm": 0.020630279555916786,
2068
+ "learning_rate": 9.56e-05,
2069
+ "loss": 0.0337,
2070
+ "step": 2610
2071
+ },
2072
+ {
2073
+ "epoch": 10.48,
2074
+ "grad_norm": 0.2516959607601166,
2075
+ "learning_rate": 9.52e-05,
2076
+ "loss": 0.0099,
2077
+ "step": 2620
2078
+ },
2079
+ {
2080
+ "epoch": 10.52,
2081
+ "grad_norm": 0.09867813438177109,
2082
+ "learning_rate": 9.48e-05,
2083
+ "loss": 0.0258,
2084
+ "step": 2630
2085
+ },
2086
+ {
2087
+ "epoch": 10.56,
2088
+ "grad_norm": 0.005374481901526451,
2089
+ "learning_rate": 9.44e-05,
2090
+ "loss": 0.0014,
2091
+ "step": 2640
2092
+ },
2093
+ {
2094
+ "epoch": 10.6,
2095
+ "grad_norm": 0.012418796308338642,
2096
+ "learning_rate": 9.4e-05,
2097
+ "loss": 0.0052,
2098
+ "step": 2650
2099
+ },
2100
+ {
2101
+ "epoch": 10.64,
2102
+ "grad_norm": 0.006738652009516954,
2103
+ "learning_rate": 9.360000000000001e-05,
2104
+ "loss": 0.0015,
2105
+ "step": 2660
2106
+ },
2107
+ {
2108
+ "epoch": 10.68,
2109
+ "grad_norm": 3.031865358352661,
2110
+ "learning_rate": 9.320000000000002e-05,
2111
+ "loss": 0.0023,
2112
+ "step": 2670
2113
+ },
2114
+ {
2115
+ "epoch": 10.72,
2116
+ "grad_norm": 0.007892146706581116,
2117
+ "learning_rate": 9.28e-05,
2118
+ "loss": 0.0014,
2119
+ "step": 2680
2120
+ },
2121
+ {
2122
+ "epoch": 10.76,
2123
+ "grad_norm": 0.008096442557871342,
2124
+ "learning_rate": 9.240000000000001e-05,
2125
+ "loss": 0.0321,
2126
+ "step": 2690
2127
+ },
2128
+ {
2129
+ "epoch": 10.8,
2130
+ "grad_norm": 10.601773262023926,
2131
+ "learning_rate": 9.200000000000001e-05,
2132
+ "loss": 0.0401,
2133
+ "step": 2700
2134
+ },
2135
+ {
2136
+ "epoch": 10.8,
2137
+ "eval_accuracy": 0.8,
2138
+ "eval_loss": 1.1593725681304932,
2139
+ "eval_runtime": 5.7948,
2140
+ "eval_samples_per_second": 172.569,
2141
+ "eval_steps_per_second": 21.571,
2142
+ "step": 2700
2143
+ },
2144
+ {
2145
+ "epoch": 10.84,
2146
+ "grad_norm": 0.00845256820321083,
2147
+ "learning_rate": 9.16e-05,
2148
+ "loss": 0.0016,
2149
+ "step": 2710
2150
+ },
2151
+ {
2152
+ "epoch": 10.88,
2153
+ "grad_norm": 0.016398636624217033,
2154
+ "learning_rate": 9.120000000000001e-05,
2155
+ "loss": 0.0108,
2156
+ "step": 2720
2157
+ },
2158
+ {
2159
+ "epoch": 10.92,
2160
+ "grad_norm": 0.015316460281610489,
2161
+ "learning_rate": 9.080000000000001e-05,
2162
+ "loss": 0.0394,
2163
+ "step": 2730
2164
+ },
2165
+ {
2166
+ "epoch": 10.96,
2167
+ "grad_norm": 0.35788559913635254,
2168
+ "learning_rate": 9.04e-05,
2169
+ "loss": 0.0017,
2170
+ "step": 2740
2171
+ },
2172
+ {
2173
+ "epoch": 11.0,
2174
+ "grad_norm": 0.011767162941396236,
2175
+ "learning_rate": 9e-05,
2176
+ "loss": 0.0047,
2177
+ "step": 2750
2178
+ },
2179
+ {
2180
+ "epoch": 11.04,
2181
+ "grad_norm": 0.007368568331003189,
2182
+ "learning_rate": 8.960000000000001e-05,
2183
+ "loss": 0.0012,
2184
+ "step": 2760
2185
+ },
2186
+ {
2187
+ "epoch": 11.08,
2188
+ "grad_norm": 0.011666523292660713,
2189
+ "learning_rate": 8.92e-05,
2190
+ "loss": 0.0014,
2191
+ "step": 2770
2192
+ },
2193
+ {
2194
+ "epoch": 11.12,
2195
+ "grad_norm": 0.00521031254902482,
2196
+ "learning_rate": 8.88e-05,
2197
+ "loss": 0.0012,
2198
+ "step": 2780
2199
+ },
2200
+ {
2201
+ "epoch": 11.16,
2202
+ "grad_norm": 0.025624191388487816,
2203
+ "learning_rate": 8.840000000000001e-05,
2204
+ "loss": 0.0014,
2205
+ "step": 2790
2206
+ },
2207
+ {
2208
+ "epoch": 11.2,
2209
+ "grad_norm": 0.008872034959495068,
2210
+ "learning_rate": 8.800000000000001e-05,
2211
+ "loss": 0.0012,
2212
+ "step": 2800
2213
+ },
2214
+ {
2215
+ "epoch": 11.2,
2216
+ "eval_accuracy": 0.817,
2217
+ "eval_loss": 1.0810606479644775,
2218
+ "eval_runtime": 5.7202,
2219
+ "eval_samples_per_second": 174.82,
2220
+ "eval_steps_per_second": 21.853,
2221
+ "step": 2800
2222
+ },
2223
+ {
2224
+ "epoch": 11.24,
2225
+ "grad_norm": 0.006208410952240229,
2226
+ "learning_rate": 8.76e-05,
2227
+ "loss": 0.0014,
2228
+ "step": 2810
2229
+ },
2230
+ {
2231
+ "epoch": 11.28,
2232
+ "grad_norm": 0.006170503329485655,
2233
+ "learning_rate": 8.72e-05,
2234
+ "loss": 0.0011,
2235
+ "step": 2820
2236
+ },
2237
+ {
2238
+ "epoch": 11.32,
2239
+ "grad_norm": 0.005829779896885157,
2240
+ "learning_rate": 8.680000000000001e-05,
2241
+ "loss": 0.0015,
2242
+ "step": 2830
2243
+ },
2244
+ {
2245
+ "epoch": 11.36,
2246
+ "grad_norm": 0.0047014919109642506,
2247
+ "learning_rate": 8.64e-05,
2248
+ "loss": 0.001,
2249
+ "step": 2840
2250
+ },
2251
+ {
2252
+ "epoch": 11.4,
2253
+ "grad_norm": 0.01678207516670227,
2254
+ "learning_rate": 8.6e-05,
2255
+ "loss": 0.0019,
2256
+ "step": 2850
2257
+ },
2258
+ {
2259
+ "epoch": 11.44,
2260
+ "grad_norm": 0.004566165152937174,
2261
+ "learning_rate": 8.560000000000001e-05,
2262
+ "loss": 0.0014,
2263
+ "step": 2860
2264
+ },
2265
+ {
2266
+ "epoch": 11.48,
2267
+ "grad_norm": 0.005536119919270277,
2268
+ "learning_rate": 8.52e-05,
2269
+ "loss": 0.001,
2270
+ "step": 2870
2271
+ },
2272
+ {
2273
+ "epoch": 11.52,
2274
+ "grad_norm": 0.005134043283760548,
2275
+ "learning_rate": 8.48e-05,
2276
+ "loss": 0.001,
2277
+ "step": 2880
2278
+ },
2279
+ {
2280
+ "epoch": 11.56,
2281
+ "grad_norm": 0.005173493642359972,
2282
+ "learning_rate": 8.44e-05,
2283
+ "loss": 0.001,
2284
+ "step": 2890
2285
+ },
2286
+ {
2287
+ "epoch": 11.6,
2288
+ "grad_norm": 0.0057559204287827015,
2289
+ "learning_rate": 8.4e-05,
2290
+ "loss": 0.0034,
2291
+ "step": 2900
2292
+ },
2293
+ {
2294
+ "epoch": 11.6,
2295
+ "eval_accuracy": 0.825,
2296
+ "eval_loss": 1.0955668687820435,
2297
+ "eval_runtime": 5.4405,
2298
+ "eval_samples_per_second": 183.805,
2299
+ "eval_steps_per_second": 22.976,
2300
+ "step": 2900
2301
+ },
2302
+ {
2303
+ "epoch": 11.64,
2304
+ "grad_norm": 0.004013891331851482,
2305
+ "learning_rate": 8.36e-05,
2306
+ "loss": 0.0009,
2307
+ "step": 2910
2308
+ },
2309
+ {
2310
+ "epoch": 11.68,
2311
+ "grad_norm": 4.541440486907959,
2312
+ "learning_rate": 8.32e-05,
2313
+ "loss": 0.0744,
2314
+ "step": 2920
2315
+ },
2316
+ {
2317
+ "epoch": 11.72,
2318
+ "grad_norm": 0.004702790640294552,
2319
+ "learning_rate": 8.28e-05,
2320
+ "loss": 0.001,
2321
+ "step": 2930
2322
+ },
2323
+ {
2324
+ "epoch": 11.76,
2325
+ "grad_norm": 0.004574855323880911,
2326
+ "learning_rate": 8.24e-05,
2327
+ "loss": 0.0028,
2328
+ "step": 2940
2329
+ },
2330
+ {
2331
+ "epoch": 11.8,
2332
+ "grad_norm": 0.006892206612974405,
2333
+ "learning_rate": 8.2e-05,
2334
+ "loss": 0.0031,
2335
+ "step": 2950
2336
+ },
2337
+ {
2338
+ "epoch": 11.84,
2339
+ "grad_norm": 0.28937533497810364,
2340
+ "learning_rate": 8.16e-05,
2341
+ "loss": 0.002,
2342
+ "step": 2960
2343
+ },
2344
+ {
2345
+ "epoch": 11.88,
2346
+ "grad_norm": 0.004892553202807903,
2347
+ "learning_rate": 8.120000000000001e-05,
2348
+ "loss": 0.0009,
2349
+ "step": 2970
2350
+ },
2351
+ {
2352
+ "epoch": 11.92,
2353
+ "grad_norm": 0.00817884597927332,
2354
+ "learning_rate": 8.080000000000001e-05,
2355
+ "loss": 0.0012,
2356
+ "step": 2980
2357
+ },
2358
+ {
2359
+ "epoch": 11.96,
2360
+ "grad_norm": 0.007845502346754074,
2361
+ "learning_rate": 8.04e-05,
2362
+ "loss": 0.0169,
2363
+ "step": 2990
2364
+ },
2365
+ {
2366
+ "epoch": 12.0,
2367
+ "grad_norm": 0.004683198872953653,
2368
+ "learning_rate": 8e-05,
2369
+ "loss": 0.0012,
2370
+ "step": 3000
2371
+ },
2372
+ {
2373
+ "epoch": 12.0,
2374
+ "eval_accuracy": 0.808,
2375
+ "eval_loss": 1.200992226600647,
2376
+ "eval_runtime": 5.2017,
2377
+ "eval_samples_per_second": 192.245,
2378
+ "eval_steps_per_second": 24.031,
2379
+ "step": 3000
2380
+ },
2381
+ {
2382
+ "epoch": 12.04,
2383
+ "grad_norm": 0.004240355920046568,
2384
+ "learning_rate": 7.960000000000001e-05,
2385
+ "loss": 0.001,
2386
+ "step": 3010
2387
+ },
2388
+ {
2389
+ "epoch": 12.08,
2390
+ "grad_norm": 0.04097750037908554,
2391
+ "learning_rate": 7.920000000000001e-05,
2392
+ "loss": 0.001,
2393
+ "step": 3020
2394
+ },
2395
+ {
2396
+ "epoch": 12.12,
2397
+ "grad_norm": 0.036695607006549835,
2398
+ "learning_rate": 7.88e-05,
2399
+ "loss": 0.0009,
2400
+ "step": 3030
2401
+ },
2402
+ {
2403
+ "epoch": 12.16,
2404
+ "grad_norm": 0.003702098038047552,
2405
+ "learning_rate": 7.840000000000001e-05,
2406
+ "loss": 0.0009,
2407
+ "step": 3040
2408
+ },
2409
+ {
2410
+ "epoch": 12.2,
2411
+ "grad_norm": 10.573301315307617,
2412
+ "learning_rate": 7.800000000000001e-05,
2413
+ "loss": 0.0321,
2414
+ "step": 3050
2415
+ },
2416
+ {
2417
+ "epoch": 12.24,
2418
+ "grad_norm": 0.004409339744597673,
2419
+ "learning_rate": 7.76e-05,
2420
+ "loss": 0.0008,
2421
+ "step": 3060
2422
+ },
2423
+ {
2424
+ "epoch": 12.28,
2425
+ "grad_norm": 0.004956026095896959,
2426
+ "learning_rate": 7.72e-05,
2427
+ "loss": 0.0008,
2428
+ "step": 3070
2429
+ },
2430
+ {
2431
+ "epoch": 12.32,
2432
+ "grad_norm": 0.006153183523565531,
2433
+ "learning_rate": 7.680000000000001e-05,
2434
+ "loss": 0.0009,
2435
+ "step": 3080
2436
+ },
2437
+ {
2438
+ "epoch": 12.36,
2439
+ "grad_norm": 0.005817431956529617,
2440
+ "learning_rate": 7.64e-05,
2441
+ "loss": 0.0009,
2442
+ "step": 3090
2443
+ },
2444
+ {
2445
+ "epoch": 12.4,
2446
+ "grad_norm": 0.004047242924571037,
2447
+ "learning_rate": 7.6e-05,
2448
+ "loss": 0.0011,
2449
+ "step": 3100
2450
+ },
2451
+ {
2452
+ "epoch": 12.4,
2453
+ "eval_accuracy": 0.81,
2454
+ "eval_loss": 1.1712357997894287,
2455
+ "eval_runtime": 5.9454,
2456
+ "eval_samples_per_second": 168.198,
2457
+ "eval_steps_per_second": 21.025,
2458
+ "step": 3100
2459
+ },
2460
+ {
2461
+ "epoch": 12.44,
2462
+ "grad_norm": 0.05110397934913635,
2463
+ "learning_rate": 7.560000000000001e-05,
2464
+ "loss": 0.0009,
2465
+ "step": 3110
2466
+ },
2467
+ {
2468
+ "epoch": 12.48,
2469
+ "grad_norm": 0.005057331640273333,
2470
+ "learning_rate": 7.52e-05,
2471
+ "loss": 0.0008,
2472
+ "step": 3120
2473
+ },
2474
+ {
2475
+ "epoch": 12.52,
2476
+ "grad_norm": 0.00844608899205923,
2477
+ "learning_rate": 7.48e-05,
2478
+ "loss": 0.0008,
2479
+ "step": 3130
2480
+ },
2481
+ {
2482
+ "epoch": 12.56,
2483
+ "grad_norm": 0.004293734673410654,
2484
+ "learning_rate": 7.44e-05,
2485
+ "loss": 0.0011,
2486
+ "step": 3140
2487
+ },
2488
+ {
2489
+ "epoch": 12.6,
2490
+ "grad_norm": 0.004264953080564737,
2491
+ "learning_rate": 7.4e-05,
2492
+ "loss": 0.0008,
2493
+ "step": 3150
2494
+ },
2495
+ {
2496
+ "epoch": 12.64,
2497
+ "grad_norm": 0.021664146333932877,
2498
+ "learning_rate": 7.36e-05,
2499
+ "loss": 0.0432,
2500
+ "step": 3160
2501
+ },
2502
+ {
2503
+ "epoch": 12.68,
2504
+ "grad_norm": 0.004172308370471001,
2505
+ "learning_rate": 7.32e-05,
2506
+ "loss": 0.0008,
2507
+ "step": 3170
2508
+ },
2509
+ {
2510
+ "epoch": 12.72,
2511
+ "grad_norm": 0.006536141969263554,
2512
+ "learning_rate": 7.280000000000001e-05,
2513
+ "loss": 0.0008,
2514
+ "step": 3180
2515
+ },
2516
+ {
2517
+ "epoch": 12.76,
2518
+ "grad_norm": 0.0048880744725465775,
2519
+ "learning_rate": 7.24e-05,
2520
+ "loss": 0.0008,
2521
+ "step": 3190
2522
+ },
2523
+ {
2524
+ "epoch": 12.8,
2525
+ "grad_norm": 0.004492601379752159,
2526
+ "learning_rate": 7.2e-05,
2527
+ "loss": 0.0092,
2528
+ "step": 3200
2529
+ },
2530
+ {
2531
+ "epoch": 12.8,
2532
+ "eval_accuracy": 0.813,
2533
+ "eval_loss": 1.1813948154449463,
2534
+ "eval_runtime": 5.2063,
2535
+ "eval_samples_per_second": 192.074,
2536
+ "eval_steps_per_second": 24.009,
2537
+ "step": 3200
2538
+ },
2539
+ {
2540
+ "epoch": 12.84,
2541
+ "grad_norm": 0.003316229674965143,
2542
+ "learning_rate": 7.16e-05,
2543
+ "loss": 0.0008,
2544
+ "step": 3210
2545
+ },
2546
+ {
2547
+ "epoch": 12.88,
2548
+ "grad_norm": 0.004760683514177799,
2549
+ "learning_rate": 7.12e-05,
2550
+ "loss": 0.0007,
2551
+ "step": 3220
2552
+ },
2553
+ {
2554
+ "epoch": 12.92,
2555
+ "grad_norm": 0.004011666867882013,
2556
+ "learning_rate": 7.08e-05,
2557
+ "loss": 0.0007,
2558
+ "step": 3230
2559
+ },
2560
+ {
2561
+ "epoch": 12.96,
2562
+ "grad_norm": 0.004005063325166702,
2563
+ "learning_rate": 7.04e-05,
2564
+ "loss": 0.0362,
2565
+ "step": 3240
2566
+ },
2567
+ {
2568
+ "epoch": 13.0,
2569
+ "grad_norm": 0.0050388905219733715,
2570
+ "learning_rate": 7e-05,
2571
+ "loss": 0.0008,
2572
+ "step": 3250
2573
+ },
2574
+ {
2575
+ "epoch": 13.04,
2576
+ "grad_norm": 0.01386964414268732,
2577
+ "learning_rate": 6.96e-05,
2578
+ "loss": 0.0009,
2579
+ "step": 3260
2580
+ },
2581
+ {
2582
+ "epoch": 13.08,
2583
+ "grad_norm": 0.004433898255228996,
2584
+ "learning_rate": 6.92e-05,
2585
+ "loss": 0.0008,
2586
+ "step": 3270
2587
+ },
2588
+ {
2589
+ "epoch": 13.12,
2590
+ "grad_norm": 0.004408930893987417,
2591
+ "learning_rate": 6.879999999999999e-05,
2592
+ "loss": 0.0008,
2593
+ "step": 3280
2594
+ },
2595
+ {
2596
+ "epoch": 13.16,
2597
+ "grad_norm": 0.004361864645034075,
2598
+ "learning_rate": 6.840000000000001e-05,
2599
+ "loss": 0.0007,
2600
+ "step": 3290
2601
+ },
2602
+ {
2603
+ "epoch": 13.2,
2604
+ "grad_norm": 0.010670163668692112,
2605
+ "learning_rate": 6.800000000000001e-05,
2606
+ "loss": 0.0007,
2607
+ "step": 3300
2608
+ },
2609
+ {
2610
+ "epoch": 13.2,
2611
+ "eval_accuracy": 0.818,
2612
+ "eval_loss": 1.1677281856536865,
2613
+ "eval_runtime": 5.3663,
2614
+ "eval_samples_per_second": 186.347,
2615
+ "eval_steps_per_second": 23.293,
2616
+ "step": 3300
2617
+ },
2618
+ {
2619
+ "epoch": 13.24,
2620
+ "grad_norm": 0.004130377899855375,
2621
+ "learning_rate": 6.76e-05,
2622
+ "loss": 0.0007,
2623
+ "step": 3310
2624
+ },
2625
+ {
2626
+ "epoch": 13.28,
2627
+ "grad_norm": 0.0031264375429600477,
2628
+ "learning_rate": 6.720000000000001e-05,
2629
+ "loss": 0.0007,
2630
+ "step": 3320
2631
+ },
2632
+ {
2633
+ "epoch": 13.32,
2634
+ "grad_norm": 0.004546809010207653,
2635
+ "learning_rate": 6.680000000000001e-05,
2636
+ "loss": 0.0007,
2637
+ "step": 3330
2638
+ },
2639
+ {
2640
+ "epoch": 13.36,
2641
+ "grad_norm": 0.0035698998253792524,
2642
+ "learning_rate": 6.64e-05,
2643
+ "loss": 0.0007,
2644
+ "step": 3340
2645
+ },
2646
+ {
2647
+ "epoch": 13.4,
2648
+ "grad_norm": 0.004544632509350777,
2649
+ "learning_rate": 6.6e-05,
2650
+ "loss": 0.0007,
2651
+ "step": 3350
2652
+ },
2653
+ {
2654
+ "epoch": 13.44,
2655
+ "grad_norm": 0.0035377303138375282,
2656
+ "learning_rate": 6.560000000000001e-05,
2657
+ "loss": 0.0006,
2658
+ "step": 3360
2659
+ },
2660
+ {
2661
+ "epoch": 13.48,
2662
+ "grad_norm": 0.003488726681098342,
2663
+ "learning_rate": 6.52e-05,
2664
+ "loss": 0.0007,
2665
+ "step": 3370
2666
+ },
2667
+ {
2668
+ "epoch": 13.52,
2669
+ "grad_norm": 0.016187245026230812,
2670
+ "learning_rate": 6.48e-05,
2671
+ "loss": 0.0007,
2672
+ "step": 3380
2673
+ },
2674
+ {
2675
+ "epoch": 13.56,
2676
+ "grad_norm": 0.0037824937608093023,
2677
+ "learning_rate": 6.440000000000001e-05,
2678
+ "loss": 0.0006,
2679
+ "step": 3390
2680
+ },
2681
+ {
2682
+ "epoch": 13.6,
2683
+ "grad_norm": 0.0033098619896918535,
2684
+ "learning_rate": 6.400000000000001e-05,
2685
+ "loss": 0.0007,
2686
+ "step": 3400
2687
+ },
2688
+ {
2689
+ "epoch": 13.6,
2690
+ "eval_accuracy": 0.818,
2691
+ "eval_loss": 1.172300934791565,
2692
+ "eval_runtime": 5.2344,
2693
+ "eval_samples_per_second": 191.045,
2694
+ "eval_steps_per_second": 23.881,
2695
+ "step": 3400
2696
+ },
2697
+ {
2698
+ "epoch": 13.64,
2699
+ "grad_norm": 0.0036126759368926287,
2700
+ "learning_rate": 6.36e-05,
2701
+ "loss": 0.0007,
2702
+ "step": 3410
2703
+ },
2704
+ {
2705
+ "epoch": 13.68,
2706
+ "grad_norm": 0.0037607806734740734,
2707
+ "learning_rate": 6.32e-05,
2708
+ "loss": 0.0006,
2709
+ "step": 3420
2710
+ },
2711
+ {
2712
+ "epoch": 13.72,
2713
+ "grad_norm": 0.002861644374206662,
2714
+ "learning_rate": 6.280000000000001e-05,
2715
+ "loss": 0.0006,
2716
+ "step": 3430
2717
+ },
2718
+ {
2719
+ "epoch": 13.76,
2720
+ "grad_norm": 0.0034740378614515066,
2721
+ "learning_rate": 6.24e-05,
2722
+ "loss": 0.0006,
2723
+ "step": 3440
2724
+ },
2725
+ {
2726
+ "epoch": 13.8,
2727
+ "grad_norm": 0.0031694734934717417,
2728
+ "learning_rate": 6.2e-05,
2729
+ "loss": 0.0006,
2730
+ "step": 3450
2731
+ },
2732
+ {
2733
+ "epoch": 13.84,
2734
+ "grad_norm": 0.0038870801217854023,
2735
+ "learning_rate": 6.16e-05,
2736
+ "loss": 0.0006,
2737
+ "step": 3460
2738
+ },
2739
+ {
2740
+ "epoch": 13.88,
2741
+ "grad_norm": 0.0030017481185495853,
2742
+ "learning_rate": 6.12e-05,
2743
+ "loss": 0.0006,
2744
+ "step": 3470
2745
+ },
2746
+ {
2747
+ "epoch": 13.92,
2748
+ "grad_norm": 0.004931191448122263,
2749
+ "learning_rate": 6.08e-05,
2750
+ "loss": 0.0006,
2751
+ "step": 3480
2752
+ },
2753
+ {
2754
+ "epoch": 13.96,
2755
+ "grad_norm": 0.0034845368936657906,
2756
+ "learning_rate": 6.04e-05,
2757
+ "loss": 0.0008,
2758
+ "step": 3490
2759
+ },
2760
+ {
2761
+ "epoch": 14.0,
2762
+ "grad_norm": 0.003087324323132634,
2763
+ "learning_rate": 6e-05,
2764
+ "loss": 0.0006,
2765
+ "step": 3500
2766
+ },
2767
+ {
2768
+ "epoch": 14.0,
2769
+ "eval_accuracy": 0.821,
2770
+ "eval_loss": 1.1852285861968994,
2771
+ "eval_runtime": 5.8798,
2772
+ "eval_samples_per_second": 170.074,
2773
+ "eval_steps_per_second": 21.259,
2774
+ "step": 3500
2775
+ },
2776
+ {
2777
+ "epoch": 14.04,
2778
+ "grad_norm": 0.003496804041787982,
2779
+ "learning_rate": 5.96e-05,
2780
+ "loss": 0.0006,
2781
+ "step": 3510
2782
+ },
2783
+ {
2784
+ "epoch": 14.08,
2785
+ "grad_norm": 0.0027559709269553423,
2786
+ "learning_rate": 5.92e-05,
2787
+ "loss": 0.0006,
2788
+ "step": 3520
2789
+ },
2790
+ {
2791
+ "epoch": 14.12,
2792
+ "grad_norm": 0.0026562130078673363,
2793
+ "learning_rate": 5.88e-05,
2794
+ "loss": 0.0006,
2795
+ "step": 3530
2796
+ },
2797
+ {
2798
+ "epoch": 14.16,
2799
+ "grad_norm": 0.002727353246882558,
2800
+ "learning_rate": 5.8399999999999997e-05,
2801
+ "loss": 0.0006,
2802
+ "step": 3540
2803
+ },
2804
+ {
2805
+ "epoch": 14.2,
2806
+ "grad_norm": 0.02825564332306385,
2807
+ "learning_rate": 5.8e-05,
2808
+ "loss": 0.0006,
2809
+ "step": 3550
2810
+ },
2811
+ {
2812
+ "epoch": 14.24,
2813
+ "grad_norm": 0.0026983567513525486,
2814
+ "learning_rate": 5.76e-05,
2815
+ "loss": 0.0007,
2816
+ "step": 3560
2817
+ },
2818
+ {
2819
+ "epoch": 14.28,
2820
+ "grad_norm": 0.005061229225248098,
2821
+ "learning_rate": 5.72e-05,
2822
+ "loss": 0.0006,
2823
+ "step": 3570
2824
+ },
2825
+ {
2826
+ "epoch": 14.32,
2827
+ "grad_norm": 0.002966745523735881,
2828
+ "learning_rate": 5.68e-05,
2829
+ "loss": 0.0005,
2830
+ "step": 3580
2831
+ },
2832
+ {
2833
+ "epoch": 14.36,
2834
+ "grad_norm": 0.002947271103039384,
2835
+ "learning_rate": 5.6399999999999995e-05,
2836
+ "loss": 0.0006,
2837
+ "step": 3590
2838
+ },
2839
+ {
2840
+ "epoch": 14.4,
2841
+ "grad_norm": 0.0029176445677876472,
2842
+ "learning_rate": 5.6000000000000006e-05,
2843
+ "loss": 0.0005,
2844
+ "step": 3600
2845
+ },
2846
+ {
2847
+ "epoch": 14.4,
2848
+ "eval_accuracy": 0.82,
2849
+ "eval_loss": 1.1927635669708252,
2850
+ "eval_runtime": 5.2928,
2851
+ "eval_samples_per_second": 188.936,
2852
+ "eval_steps_per_second": 23.617,
2853
+ "step": 3600
2854
+ },
2855
+ {
2856
+ "epoch": 14.44,
2857
+ "grad_norm": 0.0033958940766751766,
2858
+ "learning_rate": 5.560000000000001e-05,
2859
+ "loss": 0.0005,
2860
+ "step": 3610
2861
+ },
2862
+ {
2863
+ "epoch": 14.48,
2864
+ "grad_norm": 0.003239579265937209,
2865
+ "learning_rate": 5.520000000000001e-05,
2866
+ "loss": 0.0006,
2867
+ "step": 3620
2868
+ },
2869
+ {
2870
+ "epoch": 14.52,
2871
+ "grad_norm": 0.002811953192576766,
2872
+ "learning_rate": 5.4800000000000004e-05,
2873
+ "loss": 0.0005,
2874
+ "step": 3630
2875
+ },
2876
+ {
2877
+ "epoch": 14.56,
2878
+ "grad_norm": 0.0023668641224503517,
2879
+ "learning_rate": 5.440000000000001e-05,
2880
+ "loss": 0.0005,
2881
+ "step": 3640
2882
+ },
2883
+ {
2884
+ "epoch": 14.6,
2885
+ "grad_norm": 0.0036032095085829496,
2886
+ "learning_rate": 5.4000000000000005e-05,
2887
+ "loss": 0.0005,
2888
+ "step": 3650
2889
+ },
2890
+ {
2891
+ "epoch": 14.64,
2892
+ "grad_norm": 0.0026333522982895374,
2893
+ "learning_rate": 5.360000000000001e-05,
2894
+ "loss": 0.0005,
2895
+ "step": 3660
2896
+ },
2897
+ {
2898
+ "epoch": 14.68,
2899
+ "grad_norm": 0.0027461478020995855,
2900
+ "learning_rate": 5.3200000000000006e-05,
2901
+ "loss": 0.0005,
2902
+ "step": 3670
2903
+ },
2904
+ {
2905
+ "epoch": 14.72,
2906
+ "grad_norm": 0.002880874089896679,
2907
+ "learning_rate": 5.28e-05,
2908
+ "loss": 0.0005,
2909
+ "step": 3680
2910
+ },
2911
+ {
2912
+ "epoch": 14.76,
2913
+ "grad_norm": 0.003313728841021657,
2914
+ "learning_rate": 5.2400000000000007e-05,
2915
+ "loss": 0.0005,
2916
+ "step": 3690
2917
+ },
2918
+ {
2919
+ "epoch": 14.8,
2920
+ "grad_norm": 0.002891551936045289,
2921
+ "learning_rate": 5.2000000000000004e-05,
2922
+ "loss": 0.0005,
2923
+ "step": 3700
2924
+ },
2925
+ {
2926
+ "epoch": 14.8,
2927
+ "eval_accuracy": 0.819,
2928
+ "eval_loss": 1.203009009361267,
2929
+ "eval_runtime": 5.2248,
2930
+ "eval_samples_per_second": 191.394,
2931
+ "eval_steps_per_second": 23.924,
2932
+ "step": 3700
2933
+ },
2934
+ {
2935
+ "epoch": 14.84,
2936
+ "grad_norm": 0.002585276961326599,
2937
+ "learning_rate": 5.16e-05,
2938
+ "loss": 0.0005,
2939
+ "step": 3710
2940
+ },
2941
+ {
2942
+ "epoch": 14.88,
2943
+ "grad_norm": 0.002555152401328087,
2944
+ "learning_rate": 5.1200000000000004e-05,
2945
+ "loss": 0.0005,
2946
+ "step": 3720
2947
+ },
2948
+ {
2949
+ "epoch": 14.92,
2950
+ "grad_norm": 0.0025339240673929453,
2951
+ "learning_rate": 5.08e-05,
2952
+ "loss": 0.0005,
2953
+ "step": 3730
2954
+ },
2955
+ {
2956
+ "epoch": 14.96,
2957
+ "grad_norm": 0.0028862394392490387,
2958
+ "learning_rate": 5.0400000000000005e-05,
2959
+ "loss": 0.0005,
2960
+ "step": 3740
2961
+ },
2962
+ {
2963
+ "epoch": 15.0,
2964
+ "grad_norm": 0.0030768855940550566,
2965
+ "learning_rate": 5e-05,
2966
+ "loss": 0.0005,
2967
+ "step": 3750
2968
+ },
2969
+ {
2970
+ "epoch": 15.04,
2971
+ "grad_norm": 0.0024443198926746845,
2972
+ "learning_rate": 4.96e-05,
2973
+ "loss": 0.0005,
2974
+ "step": 3760
2975
+ },
2976
+ {
2977
+ "epoch": 15.08,
2978
+ "grad_norm": 0.0026144583243876696,
2979
+ "learning_rate": 4.92e-05,
2980
+ "loss": 0.0005,
2981
+ "step": 3770
2982
+ },
2983
+ {
2984
+ "epoch": 15.12,
2985
+ "grad_norm": 0.002447100356221199,
2986
+ "learning_rate": 4.88e-05,
2987
+ "loss": 0.0005,
2988
+ "step": 3780
2989
+ },
2990
+ {
2991
+ "epoch": 15.16,
2992
+ "grad_norm": 0.003170330310240388,
2993
+ "learning_rate": 4.8400000000000004e-05,
2994
+ "loss": 0.0005,
2995
+ "step": 3790
2996
+ },
2997
+ {
2998
+ "epoch": 15.2,
2999
+ "grad_norm": 0.0029565359000116587,
3000
+ "learning_rate": 4.8e-05,
3001
+ "loss": 0.0005,
3002
+ "step": 3800
3003
+ },
3004
+ {
3005
+ "epoch": 15.2,
3006
+ "eval_accuracy": 0.818,
3007
+ "eval_loss": 1.2092506885528564,
3008
+ "eval_runtime": 5.9208,
3009
+ "eval_samples_per_second": 168.895,
3010
+ "eval_steps_per_second": 21.112,
3011
+ "step": 3800
3012
+ },
3013
+ {
3014
+ "epoch": 15.24,
3015
+ "grad_norm": 0.0024090250954031944,
3016
+ "learning_rate": 4.76e-05,
3017
+ "loss": 0.0005,
3018
+ "step": 3810
3019
+ },
3020
+ {
3021
+ "epoch": 15.28,
3022
+ "grad_norm": 0.0021153378766030073,
3023
+ "learning_rate": 4.72e-05,
3024
+ "loss": 0.0005,
3025
+ "step": 3820
3026
+ },
3027
+ {
3028
+ "epoch": 15.32,
3029
+ "grad_norm": 0.0027159815654158592,
3030
+ "learning_rate": 4.6800000000000006e-05,
3031
+ "loss": 0.0005,
3032
+ "step": 3830
3033
+ },
3034
+ {
3035
+ "epoch": 15.36,
3036
+ "grad_norm": 0.0028497877065092325,
3037
+ "learning_rate": 4.64e-05,
3038
+ "loss": 0.0005,
3039
+ "step": 3840
3040
+ },
3041
+ {
3042
+ "epoch": 15.4,
3043
+ "grad_norm": 0.0024237327743321657,
3044
+ "learning_rate": 4.600000000000001e-05,
3045
+ "loss": 0.0005,
3046
+ "step": 3850
3047
+ },
3048
+ {
3049
+ "epoch": 15.44,
3050
+ "grad_norm": 0.0031945749651640654,
3051
+ "learning_rate": 4.5600000000000004e-05,
3052
+ "loss": 0.0005,
3053
+ "step": 3860
3054
+ },
3055
+ {
3056
+ "epoch": 15.48,
3057
+ "grad_norm": 0.0026720704045146704,
3058
+ "learning_rate": 4.52e-05,
3059
+ "loss": 0.0005,
3060
+ "step": 3870
3061
+ },
3062
+ {
3063
+ "epoch": 15.52,
3064
+ "grad_norm": 0.0030961642041802406,
3065
+ "learning_rate": 4.4800000000000005e-05,
3066
+ "loss": 0.0005,
3067
+ "step": 3880
3068
+ },
3069
+ {
3070
+ "epoch": 15.56,
3071
+ "grad_norm": 0.002183083677664399,
3072
+ "learning_rate": 4.44e-05,
3073
+ "loss": 0.0005,
3074
+ "step": 3890
3075
+ },
3076
+ {
3077
+ "epoch": 15.6,
3078
+ "grad_norm": 0.0029423057567328215,
3079
+ "learning_rate": 4.4000000000000006e-05,
3080
+ "loss": 0.0005,
3081
+ "step": 3900
3082
+ },
3083
+ {
3084
+ "epoch": 15.6,
3085
+ "eval_accuracy": 0.818,
3086
+ "eval_loss": 1.2159565687179565,
3087
+ "eval_runtime": 5.3023,
3088
+ "eval_samples_per_second": 188.596,
3089
+ "eval_steps_per_second": 23.575,
3090
+ "step": 3900
3091
+ },
3092
+ {
3093
+ "epoch": 15.64,
3094
+ "grad_norm": 0.002488673897460103,
3095
+ "learning_rate": 4.36e-05,
3096
+ "loss": 0.0005,
3097
+ "step": 3910
3098
+ },
3099
+ {
3100
+ "epoch": 15.68,
3101
+ "grad_norm": 0.0022629755549132824,
3102
+ "learning_rate": 4.32e-05,
3103
+ "loss": 0.0005,
3104
+ "step": 3920
3105
+ },
3106
+ {
3107
+ "epoch": 15.72,
3108
+ "grad_norm": 0.0022780627477914095,
3109
+ "learning_rate": 4.2800000000000004e-05,
3110
+ "loss": 0.0004,
3111
+ "step": 3930
3112
+ },
3113
+ {
3114
+ "epoch": 15.76,
3115
+ "grad_norm": 0.002429110463708639,
3116
+ "learning_rate": 4.24e-05,
3117
+ "loss": 0.0005,
3118
+ "step": 3940
3119
+ },
3120
+ {
3121
+ "epoch": 15.8,
3122
+ "grad_norm": 0.003042061347514391,
3123
+ "learning_rate": 4.2e-05,
3124
+ "loss": 0.0004,
3125
+ "step": 3950
3126
+ },
3127
+ {
3128
+ "epoch": 15.84,
3129
+ "grad_norm": 0.0023479724768549204,
3130
+ "learning_rate": 4.16e-05,
3131
+ "loss": 0.0005,
3132
+ "step": 3960
3133
+ },
3134
+ {
3135
+ "epoch": 15.88,
3136
+ "grad_norm": 0.0032163311261683702,
3137
+ "learning_rate": 4.12e-05,
3138
+ "loss": 0.0005,
3139
+ "step": 3970
3140
+ },
3141
+ {
3142
+ "epoch": 15.92,
3143
+ "grad_norm": 0.0020656727720052004,
3144
+ "learning_rate": 4.08e-05,
3145
+ "loss": 0.0004,
3146
+ "step": 3980
3147
+ },
3148
+ {
3149
+ "epoch": 15.96,
3150
+ "grad_norm": 0.002548342105001211,
3151
+ "learning_rate": 4.0400000000000006e-05,
3152
+ "loss": 0.0005,
3153
+ "step": 3990
3154
+ },
3155
+ {
3156
+ "epoch": 16.0,
3157
+ "grad_norm": 0.001894648070447147,
3158
+ "learning_rate": 4e-05,
3159
+ "loss": 0.0004,
3160
+ "step": 4000
3161
+ },
3162
+ {
3163
+ "epoch": 16.0,
3164
+ "eval_accuracy": 0.819,
3165
+ "eval_loss": 1.2232067584991455,
3166
+ "eval_runtime": 5.3459,
3167
+ "eval_samples_per_second": 187.058,
3168
+ "eval_steps_per_second": 23.382,
3169
+ "step": 4000
3170
+ },
3171
+ {
3172
+ "epoch": 16.04,
3173
+ "grad_norm": 0.002135420450940728,
3174
+ "learning_rate": 3.960000000000001e-05,
3175
+ "loss": 0.0004,
3176
+ "step": 4010
3177
+ },
3178
+ {
3179
+ "epoch": 16.08,
3180
+ "grad_norm": 0.002859125379472971,
3181
+ "learning_rate": 3.9200000000000004e-05,
3182
+ "loss": 0.0004,
3183
+ "step": 4020
3184
+ },
3185
+ {
3186
+ "epoch": 16.12,
3187
+ "grad_norm": 0.002459143288433552,
3188
+ "learning_rate": 3.88e-05,
3189
+ "loss": 0.0004,
3190
+ "step": 4030
3191
+ },
3192
+ {
3193
+ "epoch": 16.16,
3194
+ "grad_norm": 0.0026586554013192654,
3195
+ "learning_rate": 3.8400000000000005e-05,
3196
+ "loss": 0.0004,
3197
+ "step": 4040
3198
+ },
3199
+ {
3200
+ "epoch": 16.2,
3201
+ "grad_norm": 0.0031422963365912437,
3202
+ "learning_rate": 3.8e-05,
3203
+ "loss": 0.0004,
3204
+ "step": 4050
3205
+ },
3206
+ {
3207
+ "epoch": 16.24,
3208
+ "grad_norm": 0.0020834060851484537,
3209
+ "learning_rate": 3.76e-05,
3210
+ "loss": 0.0004,
3211
+ "step": 4060
3212
+ },
3213
+ {
3214
+ "epoch": 16.28,
3215
+ "grad_norm": 0.00235546356998384,
3216
+ "learning_rate": 3.72e-05,
3217
+ "loss": 0.0004,
3218
+ "step": 4070
3219
+ },
3220
+ {
3221
+ "epoch": 16.32,
3222
+ "grad_norm": 0.0019886652007699013,
3223
+ "learning_rate": 3.68e-05,
3224
+ "loss": 0.0004,
3225
+ "step": 4080
3226
+ },
3227
+ {
3228
+ "epoch": 16.36,
3229
+ "grad_norm": 0.0024375116918236017,
3230
+ "learning_rate": 3.6400000000000004e-05,
3231
+ "loss": 0.0004,
3232
+ "step": 4090
3233
+ },
3234
+ {
3235
+ "epoch": 16.4,
3236
+ "grad_norm": 0.002063624793663621,
3237
+ "learning_rate": 3.6e-05,
3238
+ "loss": 0.0004,
3239
+ "step": 4100
3240
+ },
3241
+ {
3242
+ "epoch": 16.4,
3243
+ "eval_accuracy": 0.819,
3244
+ "eval_loss": 1.2302378416061401,
3245
+ "eval_runtime": 5.3045,
3246
+ "eval_samples_per_second": 188.52,
3247
+ "eval_steps_per_second": 23.565,
3248
+ "step": 4100
3249
+ },
3250
+ {
3251
+ "epoch": 16.44,
3252
+ "grad_norm": 0.0025726554449647665,
3253
+ "learning_rate": 3.56e-05,
3254
+ "loss": 0.0004,
3255
+ "step": 4110
3256
+ },
3257
+ {
3258
+ "epoch": 16.48,
3259
+ "grad_norm": 0.0025369401555508375,
3260
+ "learning_rate": 3.52e-05,
3261
+ "loss": 0.0004,
3262
+ "step": 4120
3263
+ },
3264
+ {
3265
+ "epoch": 16.52,
3266
+ "grad_norm": 0.0020231506787240505,
3267
+ "learning_rate": 3.48e-05,
3268
+ "loss": 0.0004,
3269
+ "step": 4130
3270
+ },
3271
+ {
3272
+ "epoch": 16.56,
3273
+ "grad_norm": 0.0021831688936799765,
3274
+ "learning_rate": 3.4399999999999996e-05,
3275
+ "loss": 0.0004,
3276
+ "step": 4140
3277
+ },
3278
+ {
3279
+ "epoch": 16.6,
3280
+ "grad_norm": 0.0023912801407277584,
3281
+ "learning_rate": 3.4000000000000007e-05,
3282
+ "loss": 0.0004,
3283
+ "step": 4150
3284
+ },
3285
+ {
3286
+ "epoch": 16.64,
3287
+ "grad_norm": 0.0023613148368895054,
3288
+ "learning_rate": 3.3600000000000004e-05,
3289
+ "loss": 0.0004,
3290
+ "step": 4160
3291
+ },
3292
+ {
3293
+ "epoch": 16.68,
3294
+ "grad_norm": 0.0024576077703386545,
3295
+ "learning_rate": 3.32e-05,
3296
+ "loss": 0.0004,
3297
+ "step": 4170
3298
+ },
3299
+ {
3300
+ "epoch": 16.72,
3301
+ "grad_norm": 0.002091441536322236,
3302
+ "learning_rate": 3.2800000000000004e-05,
3303
+ "loss": 0.0004,
3304
+ "step": 4180
3305
+ },
3306
+ {
3307
+ "epoch": 16.76,
3308
+ "grad_norm": 0.0024098132271319628,
3309
+ "learning_rate": 3.24e-05,
3310
+ "loss": 0.0004,
3311
+ "step": 4190
3312
+ },
3313
+ {
3314
+ "epoch": 16.8,
3315
+ "grad_norm": 0.0022078922484070063,
3316
+ "learning_rate": 3.2000000000000005e-05,
3317
+ "loss": 0.0004,
3318
+ "step": 4200
3319
+ },
3320
+ {
3321
+ "epoch": 16.8,
3322
+ "eval_accuracy": 0.819,
3323
+ "eval_loss": 1.2350249290466309,
3324
+ "eval_runtime": 5.4057,
3325
+ "eval_samples_per_second": 184.99,
3326
+ "eval_steps_per_second": 23.124,
3327
+ "step": 4200
3328
+ },
3329
+ {
3330
+ "epoch": 16.84,
3331
+ "grad_norm": 0.0018520988523960114,
3332
+ "learning_rate": 3.16e-05,
3333
+ "loss": 0.0004,
3334
+ "step": 4210
3335
+ },
3336
+ {
3337
+ "epoch": 16.88,
3338
+ "grad_norm": 0.00181812874507159,
3339
+ "learning_rate": 3.12e-05,
3340
+ "loss": 0.0004,
3341
+ "step": 4220
3342
+ },
3343
+ {
3344
+ "epoch": 16.92,
3345
+ "grad_norm": 0.0023901008535176516,
3346
+ "learning_rate": 3.08e-05,
3347
+ "loss": 0.0004,
3348
+ "step": 4230
3349
+ },
3350
+ {
3351
+ "epoch": 16.96,
3352
+ "grad_norm": 0.002264365553855896,
3353
+ "learning_rate": 3.04e-05,
3354
+ "loss": 0.0004,
3355
+ "step": 4240
3356
+ },
3357
+ {
3358
+ "epoch": 17.0,
3359
+ "grad_norm": 0.002284231362864375,
3360
+ "learning_rate": 3e-05,
3361
+ "loss": 0.0004,
3362
+ "step": 4250
3363
+ },
3364
+ {
3365
+ "epoch": 17.04,
3366
+ "grad_norm": 0.0025141683872789145,
3367
+ "learning_rate": 2.96e-05,
3368
+ "loss": 0.0004,
3369
+ "step": 4260
3370
+ },
3371
+ {
3372
+ "epoch": 17.08,
3373
+ "grad_norm": 0.0023226479534059763,
3374
+ "learning_rate": 2.9199999999999998e-05,
3375
+ "loss": 0.0004,
3376
+ "step": 4270
3377
+ },
3378
+ {
3379
+ "epoch": 17.12,
3380
+ "grad_norm": 0.002532349433749914,
3381
+ "learning_rate": 2.88e-05,
3382
+ "loss": 0.0004,
3383
+ "step": 4280
3384
+ },
3385
+ {
3386
+ "epoch": 17.16,
3387
+ "grad_norm": 0.0018973592668771744,
3388
+ "learning_rate": 2.84e-05,
3389
+ "loss": 0.0004,
3390
+ "step": 4290
3391
+ },
3392
+ {
3393
+ "epoch": 17.2,
3394
+ "grad_norm": 0.0026193023659288883,
3395
+ "learning_rate": 2.8000000000000003e-05,
3396
+ "loss": 0.0004,
3397
+ "step": 4300
3398
+ },
3399
+ {
3400
+ "epoch": 17.2,
3401
+ "eval_accuracy": 0.82,
3402
+ "eval_loss": 1.2400437593460083,
3403
+ "eval_runtime": 5.314,
3404
+ "eval_samples_per_second": 188.181,
3405
+ "eval_steps_per_second": 23.523,
3406
+ "step": 4300
3407
+ },
3408
+ {
3409
+ "epoch": 17.24,
3410
+ "grad_norm": 0.001868457649834454,
3411
+ "learning_rate": 2.7600000000000003e-05,
3412
+ "loss": 0.0004,
3413
+ "step": 4310
3414
+ },
3415
+ {
3416
+ "epoch": 17.28,
3417
+ "grad_norm": 0.0027018135879188776,
3418
+ "learning_rate": 2.7200000000000004e-05,
3419
+ "loss": 0.0004,
3420
+ "step": 4320
3421
+ },
3422
+ {
3423
+ "epoch": 17.32,
3424
+ "grad_norm": 0.0023467419669032097,
3425
+ "learning_rate": 2.6800000000000004e-05,
3426
+ "loss": 0.0004,
3427
+ "step": 4330
3428
+ },
3429
+ {
3430
+ "epoch": 17.36,
3431
+ "grad_norm": 0.002205600030720234,
3432
+ "learning_rate": 2.64e-05,
3433
+ "loss": 0.0004,
3434
+ "step": 4340
3435
+ },
3436
+ {
3437
+ "epoch": 17.4,
3438
+ "grad_norm": 0.0021359582897275686,
3439
+ "learning_rate": 2.6000000000000002e-05,
3440
+ "loss": 0.0004,
3441
+ "step": 4350
3442
+ },
3443
+ {
3444
+ "epoch": 17.44,
3445
+ "grad_norm": 0.0024898636620491743,
3446
+ "learning_rate": 2.5600000000000002e-05,
3447
+ "loss": 0.0004,
3448
+ "step": 4360
3449
+ },
3450
+ {
3451
+ "epoch": 17.48,
3452
+ "grad_norm": 0.0017528502503409982,
3453
+ "learning_rate": 2.5200000000000003e-05,
3454
+ "loss": 0.0004,
3455
+ "step": 4370
3456
+ },
3457
+ {
3458
+ "epoch": 17.52,
3459
+ "grad_norm": 0.0022504818625748158,
3460
+ "learning_rate": 2.48e-05,
3461
+ "loss": 0.0004,
3462
+ "step": 4380
3463
+ },
3464
+ {
3465
+ "epoch": 17.56,
3466
+ "grad_norm": 0.00209965487010777,
3467
+ "learning_rate": 2.44e-05,
3468
+ "loss": 0.0004,
3469
+ "step": 4390
3470
+ },
3471
+ {
3472
+ "epoch": 17.6,
3473
+ "grad_norm": 0.0020423857495188713,
3474
+ "learning_rate": 2.4e-05,
3475
+ "loss": 0.0004,
3476
+ "step": 4400
3477
+ },
3478
+ {
3479
+ "epoch": 17.6,
3480
+ "eval_accuracy": 0.821,
3481
+ "eval_loss": 1.2442196607589722,
3482
+ "eval_runtime": 5.4679,
3483
+ "eval_samples_per_second": 182.887,
3484
+ "eval_steps_per_second": 22.861,
3485
+ "step": 4400
3486
+ },
3487
+ {
3488
+ "epoch": 17.64,
3489
+ "grad_norm": 0.0022337392438203096,
3490
+ "learning_rate": 2.36e-05,
3491
+ "loss": 0.0004,
3492
+ "step": 4410
3493
+ },
3494
+ {
3495
+ "epoch": 17.68,
3496
+ "grad_norm": 0.0018173023127019405,
3497
+ "learning_rate": 2.32e-05,
3498
+ "loss": 0.0004,
3499
+ "step": 4420
3500
+ },
3501
+ {
3502
+ "epoch": 17.72,
3503
+ "grad_norm": 0.002295331796631217,
3504
+ "learning_rate": 2.2800000000000002e-05,
3505
+ "loss": 0.0004,
3506
+ "step": 4430
3507
+ },
3508
+ {
3509
+ "epoch": 17.76,
3510
+ "grad_norm": 0.001931387116201222,
3511
+ "learning_rate": 2.2400000000000002e-05,
3512
+ "loss": 0.0004,
3513
+ "step": 4440
3514
+ },
3515
+ {
3516
+ "epoch": 17.8,
3517
+ "grad_norm": 0.0019796870183199644,
3518
+ "learning_rate": 2.2000000000000003e-05,
3519
+ "loss": 0.0004,
3520
+ "step": 4450
3521
+ },
3522
+ {
3523
+ "epoch": 17.84,
3524
+ "grad_norm": 0.0016401337925344706,
3525
+ "learning_rate": 2.16e-05,
3526
+ "loss": 0.0004,
3527
+ "step": 4460
3528
+ },
3529
+ {
3530
+ "epoch": 17.88,
3531
+ "grad_norm": 0.001987120369449258,
3532
+ "learning_rate": 2.12e-05,
3533
+ "loss": 0.0004,
3534
+ "step": 4470
3535
+ },
3536
+ {
3537
+ "epoch": 17.92,
3538
+ "grad_norm": 0.002337078796699643,
3539
+ "learning_rate": 2.08e-05,
3540
+ "loss": 0.0004,
3541
+ "step": 4480
3542
+ },
3543
+ {
3544
+ "epoch": 17.96,
3545
+ "grad_norm": 0.0024157376028597355,
3546
+ "learning_rate": 2.04e-05,
3547
+ "loss": 0.0004,
3548
+ "step": 4490
3549
+ },
3550
+ {
3551
+ "epoch": 18.0,
3552
+ "grad_norm": 0.0021758731454610825,
3553
+ "learning_rate": 2e-05,
3554
+ "loss": 0.0004,
3555
+ "step": 4500
3556
+ },
3557
+ {
3558
+ "epoch": 18.0,
3559
+ "eval_accuracy": 0.821,
3560
+ "eval_loss": 1.2483272552490234,
3561
+ "eval_runtime": 5.8485,
3562
+ "eval_samples_per_second": 170.983,
3563
+ "eval_steps_per_second": 21.373,
3564
+ "step": 4500
3565
+ },
3566
+ {
3567
+ "epoch": 18.04,
3568
+ "grad_norm": 0.002017402555793524,
3569
+ "learning_rate": 1.9600000000000002e-05,
3570
+ "loss": 0.0004,
3571
+ "step": 4510
3572
+ },
3573
+ {
3574
+ "epoch": 18.08,
3575
+ "grad_norm": 0.0020504500716924667,
3576
+ "learning_rate": 1.9200000000000003e-05,
3577
+ "loss": 0.0004,
3578
+ "step": 4520
3579
+ },
3580
+ {
3581
+ "epoch": 18.12,
3582
+ "grad_norm": 0.0019253026694059372,
3583
+ "learning_rate": 1.88e-05,
3584
+ "loss": 0.0004,
3585
+ "step": 4530
3586
+ },
3587
+ {
3588
+ "epoch": 18.16,
3589
+ "grad_norm": 0.0020694080740213394,
3590
+ "learning_rate": 1.84e-05,
3591
+ "loss": 0.0004,
3592
+ "step": 4540
3593
+ },
3594
+ {
3595
+ "epoch": 18.2,
3596
+ "grad_norm": 0.00169233581982553,
3597
+ "learning_rate": 1.8e-05,
3598
+ "loss": 0.0004,
3599
+ "step": 4550
3600
+ },
3601
+ {
3602
+ "epoch": 18.24,
3603
+ "grad_norm": 0.0019015150610357523,
3604
+ "learning_rate": 1.76e-05,
3605
+ "loss": 0.0004,
3606
+ "step": 4560
3607
+ },
3608
+ {
3609
+ "epoch": 18.28,
3610
+ "grad_norm": 0.002108811168000102,
3611
+ "learning_rate": 1.7199999999999998e-05,
3612
+ "loss": 0.0004,
3613
+ "step": 4570
3614
+ },
3615
+ {
3616
+ "epoch": 18.32,
3617
+ "grad_norm": 0.0021513872779905796,
3618
+ "learning_rate": 1.6800000000000002e-05,
3619
+ "loss": 0.0004,
3620
+ "step": 4580
3621
+ },
3622
+ {
3623
+ "epoch": 18.36,
3624
+ "grad_norm": 0.0019856479484587908,
3625
+ "learning_rate": 1.6400000000000002e-05,
3626
+ "loss": 0.0004,
3627
+ "step": 4590
3628
+ },
3629
+ {
3630
+ "epoch": 18.4,
3631
+ "grad_norm": 0.0020635281689465046,
3632
+ "learning_rate": 1.6000000000000003e-05,
3633
+ "loss": 0.0004,
3634
+ "step": 4600
3635
+ },
3636
+ {
3637
+ "epoch": 18.4,
3638
+ "eval_accuracy": 0.821,
3639
+ "eval_loss": 1.25178062915802,
3640
+ "eval_runtime": 5.3158,
3641
+ "eval_samples_per_second": 188.119,
3642
+ "eval_steps_per_second": 23.515,
3643
+ "step": 4600
3644
+ },
3645
+ {
3646
+ "epoch": 18.44,
3647
+ "grad_norm": 0.0020957705564796925,
3648
+ "learning_rate": 1.56e-05,
3649
+ "loss": 0.0004,
3650
+ "step": 4610
3651
+ },
3652
+ {
3653
+ "epoch": 18.48,
3654
+ "grad_norm": 0.002100448589771986,
3655
+ "learning_rate": 1.52e-05,
3656
+ "loss": 0.0004,
3657
+ "step": 4620
3658
+ },
3659
+ {
3660
+ "epoch": 18.52,
3661
+ "grad_norm": 0.0017688070656731725,
3662
+ "learning_rate": 1.48e-05,
3663
+ "loss": 0.0004,
3664
+ "step": 4630
3665
+ },
3666
+ {
3667
+ "epoch": 18.56,
3668
+ "grad_norm": 0.0018529172521084547,
3669
+ "learning_rate": 1.44e-05,
3670
+ "loss": 0.0004,
3671
+ "step": 4640
3672
+ },
3673
+ {
3674
+ "epoch": 18.6,
3675
+ "grad_norm": 0.0016175085911527276,
3676
+ "learning_rate": 1.4000000000000001e-05,
3677
+ "loss": 0.0004,
3678
+ "step": 4650
3679
+ },
3680
+ {
3681
+ "epoch": 18.64,
3682
+ "grad_norm": 0.001927876495756209,
3683
+ "learning_rate": 1.3600000000000002e-05,
3684
+ "loss": 0.0004,
3685
+ "step": 4660
3686
+ },
3687
+ {
3688
+ "epoch": 18.68,
3689
+ "grad_norm": 0.00210273964330554,
3690
+ "learning_rate": 1.32e-05,
3691
+ "loss": 0.0004,
3692
+ "step": 4670
3693
+ },
3694
+ {
3695
+ "epoch": 18.72,
3696
+ "grad_norm": 0.002222416689619422,
3697
+ "learning_rate": 1.2800000000000001e-05,
3698
+ "loss": 0.0004,
3699
+ "step": 4680
3700
+ },
3701
+ {
3702
+ "epoch": 18.76,
3703
+ "grad_norm": 0.0018316478235647082,
3704
+ "learning_rate": 1.24e-05,
3705
+ "loss": 0.0004,
3706
+ "step": 4690
3707
+ },
3708
+ {
3709
+ "epoch": 18.8,
3710
+ "grad_norm": 0.0019819007720798254,
3711
+ "learning_rate": 1.2e-05,
3712
+ "loss": 0.0004,
3713
+ "step": 4700
3714
+ },
3715
+ {
3716
+ "epoch": 18.8,
3717
+ "eval_accuracy": 0.821,
3718
+ "eval_loss": 1.254600167274475,
3719
+ "eval_runtime": 5.2417,
3720
+ "eval_samples_per_second": 190.777,
3721
+ "eval_steps_per_second": 23.847,
3722
+ "step": 4700
3723
+ },
3724
+ {
3725
+ "epoch": 18.84,
3726
+ "grad_norm": 0.0016140459338203073,
3727
+ "learning_rate": 1.16e-05,
3728
+ "loss": 0.0003,
3729
+ "step": 4710
3730
+ },
3731
+ {
3732
+ "epoch": 18.88,
3733
+ "grad_norm": 0.0018461669096723199,
3734
+ "learning_rate": 1.1200000000000001e-05,
3735
+ "loss": 0.0003,
3736
+ "step": 4720
3737
+ },
3738
+ {
3739
+ "epoch": 18.92,
3740
+ "grad_norm": 0.0018047132762148976,
3741
+ "learning_rate": 1.08e-05,
3742
+ "loss": 0.0003,
3743
+ "step": 4730
3744
+ },
3745
+ {
3746
+ "epoch": 18.96,
3747
+ "grad_norm": 0.0021749059669673443,
3748
+ "learning_rate": 1.04e-05,
3749
+ "loss": 0.0004,
3750
+ "step": 4740
3751
+ },
3752
+ {
3753
+ "epoch": 19.0,
3754
+ "grad_norm": 0.002118439180776477,
3755
+ "learning_rate": 1e-05,
3756
+ "loss": 0.0004,
3757
+ "step": 4750
3758
+ },
3759
+ {
3760
+ "epoch": 19.04,
3761
+ "grad_norm": 0.0019508595578372478,
3762
+ "learning_rate": 9.600000000000001e-06,
3763
+ "loss": 0.0004,
3764
+ "step": 4760
3765
+ },
3766
+ {
3767
+ "epoch": 19.08,
3768
+ "grad_norm": 0.002842509187757969,
3769
+ "learning_rate": 9.2e-06,
3770
+ "loss": 0.0004,
3771
+ "step": 4770
3772
+ },
3773
+ {
3774
+ "epoch": 19.12,
3775
+ "grad_norm": 0.0016302353469654918,
3776
+ "learning_rate": 8.8e-06,
3777
+ "loss": 0.0003,
3778
+ "step": 4780
3779
+ },
3780
+ {
3781
+ "epoch": 19.16,
3782
+ "grad_norm": 0.001962054753676057,
3783
+ "learning_rate": 8.400000000000001e-06,
3784
+ "loss": 0.0003,
3785
+ "step": 4790
3786
+ },
3787
+ {
3788
+ "epoch": 19.2,
3789
+ "grad_norm": 0.0021732356399297714,
3790
+ "learning_rate": 8.000000000000001e-06,
3791
+ "loss": 0.0004,
3792
+ "step": 4800
3793
+ },
3794
+ {
3795
+ "epoch": 19.2,
3796
+ "eval_accuracy": 0.821,
3797
+ "eval_loss": 1.2561498880386353,
3798
+ "eval_runtime": 10.3614,
3799
+ "eval_samples_per_second": 96.512,
3800
+ "eval_steps_per_second": 12.064,
3801
+ "step": 4800
3802
+ },
3803
+ {
3804
+ "epoch": 19.24,
3805
+ "grad_norm": 0.0018920012516900897,
3806
+ "learning_rate": 7.6e-06,
3807
+ "loss": 0.0004,
3808
+ "step": 4810
3809
+ },
3810
+ {
3811
+ "epoch": 19.28,
3812
+ "grad_norm": 0.002033686963841319,
3813
+ "learning_rate": 7.2e-06,
3814
+ "loss": 0.0003,
3815
+ "step": 4820
3816
+ },
3817
+ {
3818
+ "epoch": 19.32,
3819
+ "grad_norm": 0.0019939634948968887,
3820
+ "learning_rate": 6.800000000000001e-06,
3821
+ "loss": 0.0004,
3822
+ "step": 4830
3823
+ },
3824
+ {
3825
+ "epoch": 19.36,
3826
+ "grad_norm": 0.0024624329525977373,
3827
+ "learning_rate": 6.4000000000000006e-06,
3828
+ "loss": 0.0004,
3829
+ "step": 4840
3830
+ },
3831
+ {
3832
+ "epoch": 19.4,
3833
+ "grad_norm": 0.002102550817653537,
3834
+ "learning_rate": 6e-06,
3835
+ "loss": 0.0004,
3836
+ "step": 4850
3837
+ },
3838
+ {
3839
+ "epoch": 19.44,
3840
+ "grad_norm": 0.0017401843797415495,
3841
+ "learning_rate": 5.600000000000001e-06,
3842
+ "loss": 0.0004,
3843
+ "step": 4860
3844
+ },
3845
+ {
3846
+ "epoch": 19.48,
3847
+ "grad_norm": 0.0016316601540893316,
3848
+ "learning_rate": 5.2e-06,
3849
+ "loss": 0.0003,
3850
+ "step": 4870
3851
+ },
3852
+ {
3853
+ "epoch": 19.52,
3854
+ "grad_norm": 0.0024664264637976885,
3855
+ "learning_rate": 4.800000000000001e-06,
3856
+ "loss": 0.0003,
3857
+ "step": 4880
3858
+ },
3859
+ {
3860
+ "epoch": 19.56,
3861
+ "grad_norm": 0.0017682453617453575,
3862
+ "learning_rate": 4.4e-06,
3863
+ "loss": 0.0004,
3864
+ "step": 4890
3865
+ },
3866
+ {
3867
+ "epoch": 19.6,
3868
+ "grad_norm": 0.0019737447146326303,
3869
+ "learning_rate": 4.000000000000001e-06,
3870
+ "loss": 0.0004,
3871
+ "step": 4900
3872
+ },
3873
+ {
3874
+ "epoch": 19.6,
3875
+ "eval_accuracy": 0.82,
3876
+ "eval_loss": 1.2573806047439575,
3877
+ "eval_runtime": 5.3728,
3878
+ "eval_samples_per_second": 186.124,
3879
+ "eval_steps_per_second": 23.265,
3880
+ "step": 4900
3881
+ },
3882
+ {
3883
+ "epoch": 19.64,
3884
+ "grad_norm": 0.002168442588299513,
3885
+ "learning_rate": 3.6e-06,
3886
+ "loss": 0.0003,
3887
+ "step": 4910
3888
+ },
3889
+ {
3890
+ "epoch": 19.68,
3891
+ "grad_norm": 0.0016441222978755832,
3892
+ "learning_rate": 3.2000000000000003e-06,
3893
+ "loss": 0.0003,
3894
+ "step": 4920
3895
+ },
3896
+ {
3897
+ "epoch": 19.72,
3898
+ "grad_norm": 0.0016924934461712837,
3899
+ "learning_rate": 2.8000000000000003e-06,
3900
+ "loss": 0.0003,
3901
+ "step": 4930
3902
+ },
3903
+ {
3904
+ "epoch": 19.76,
3905
+ "grad_norm": 0.0026684573385864496,
3906
+ "learning_rate": 2.4000000000000003e-06,
3907
+ "loss": 0.0004,
3908
+ "step": 4940
3909
+ },
3910
+ {
3911
+ "epoch": 19.8,
3912
+ "grad_norm": 0.002037678612396121,
3913
+ "learning_rate": 2.0000000000000003e-06,
3914
+ "loss": 0.0004,
3915
+ "step": 4950
3916
+ },
3917
+ {
3918
+ "epoch": 19.84,
3919
+ "grad_norm": 0.0019053890137001872,
3920
+ "learning_rate": 1.6000000000000001e-06,
3921
+ "loss": 0.0004,
3922
+ "step": 4960
3923
+ },
3924
+ {
3925
+ "epoch": 19.88,
3926
+ "grad_norm": 0.002319651423022151,
3927
+ "learning_rate": 1.2000000000000002e-06,
3928
+ "loss": 0.0003,
3929
+ "step": 4970
3930
+ },
3931
+ {
3932
+ "epoch": 19.92,
3933
+ "grad_norm": 0.0018768333829939365,
3934
+ "learning_rate": 8.000000000000001e-07,
3935
+ "loss": 0.0004,
3936
+ "step": 4980
3937
+ },
3938
+ {
3939
+ "epoch": 19.96,
3940
+ "grad_norm": 0.0020151715725660324,
3941
+ "learning_rate": 4.0000000000000003e-07,
3942
+ "loss": 0.0004,
3943
+ "step": 4990
3944
+ },
3945
+ {
3946
+ "epoch": 20.0,
3947
+ "grad_norm": 0.0020873937755823135,
3948
+ "learning_rate": 0.0,
3949
+ "loss": 0.0004,
3950
+ "step": 5000
3951
+ },
3952
+ {
3953
+ "epoch": 20.0,
3954
+ "eval_accuracy": 0.82,
3955
+ "eval_loss": 1.257745385169983,
3956
+ "eval_runtime": 5.3256,
3957
+ "eval_samples_per_second": 187.773,
3958
+ "eval_steps_per_second": 23.472,
3959
+ "step": 5000
3960
+ },
3961
+ {
3962
+ "epoch": 20.0,
3963
+ "step": 5000,
3964
+ "total_flos": 6.19947029495808e+18,
3965
+ "train_loss": 0.16214322472065687,
3966
+ "train_runtime": 2178.5774,
3967
+ "train_samples_per_second": 36.721,
3968
+ "train_steps_per_second": 2.295
3969
+ }
3970
+ ],
3971
+ "logging_steps": 10,
3972
+ "max_steps": 5000,
3973
+ "num_input_tokens_seen": 0,
3974
+ "num_train_epochs": 20,
3975
+ "save_steps": 100,
3976
+ "stateful_callbacks": {
3977
+ "TrainerControl": {
3978
+ "args": {
3979
+ "should_epoch_stop": false,
3980
+ "should_evaluate": false,
3981
+ "should_log": false,
3982
+ "should_save": true,
3983
+ "should_training_stop": true
3984
+ },
3985
+ "attributes": {}
3986
+ }
3987
+ },
3988
+ "total_flos": 6.19947029495808e+18,
3989
+ "train_batch_size": 16,
3990
+ "trial_name": null,
3991
+ "trial_params": null
3992
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b7d0d95ca4cafd9a1f4fe4b75ffc2554e3cdebbeab547f2279846c3b2eae1d52
3
+ size 5112