Hugging Face's logo

haywoodsloan
/
ai-image-detector-deploy

Image Classification
Transformers
TensorBoard
Safetensors
swinv2
Trained with AutoTrain
Model card Files Metrics Community
1
haywoodsloan commited on
Commit
0e1cf08
·
0 Parent(s):

Collapse history

Browse files
Files changed (16) hide show
  1. .gitattributes +35 -0
  2. README.md +32 -0
  3. checkpoint-11947/config.json +68 -0
  4. checkpoint-11947/model.safetensors +3 -0
  5. checkpoint-11947/optimizer.pt +3 -0
  6. checkpoint-11947/rng_state.pth +3 -0
  7. checkpoint-11947/scheduler.pt +3 -0
  8. checkpoint-11947/trainer_state.json +3394 -0
  9. checkpoint-11947/training_args.bin +3 -0
  10. config.json +68 -0
  11. model.safetensors +3 -0
  12. preprocessor_config.json +23 -0
  13. runs/Apr18_01-52-02_2b9b531dbda5/events.out.tfevents.1744941124.2b9b531dbda5.218.0 +3 -0
  14. runs/Apr18_01-52-02_2b9b531dbda5/events.out.tfevents.1744985308.2b9b531dbda5.218.1 +3 -0
  15. training_args.bin +3 -0
  16. training_params.json +29 -0
.gitattributes ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ *.7z filter=lfs diff=lfs merge=lfs -text
2
+ *.arrow filter=lfs diff=lfs merge=lfs -text
3
+ *.bin filter=lfs diff=lfs merge=lfs -text
4
+ *.bz2 filter=lfs diff=lfs merge=lfs -text
5
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
6
+ *.ftz filter=lfs diff=lfs merge=lfs -text
7
+ *.gz filter=lfs diff=lfs merge=lfs -text
8
+ *.h5 filter=lfs diff=lfs merge=lfs -text
9
+ *.joblib filter=lfs diff=lfs merge=lfs -text
10
+ *.lfs.* filter=lfs diff=lfs merge=lfs -text
11
+ *.mlmodel filter=lfs diff=lfs merge=lfs -text
12
+ *.model filter=lfs diff=lfs merge=lfs -text
13
+ *.msgpack filter=lfs diff=lfs merge=lfs -text
14
+ *.npy filter=lfs diff=lfs merge=lfs -text
15
+ *.npz filter=lfs diff=lfs merge=lfs -text
16
+ *.onnx filter=lfs diff=lfs merge=lfs -text
17
+ *.ot filter=lfs diff=lfs merge=lfs -text
18
+ *.parquet filter=lfs diff=lfs merge=lfs -text
19
+ *.pb filter=lfs diff=lfs merge=lfs -text
20
+ *.pickle filter=lfs diff=lfs merge=lfs -text
21
+ *.pkl filter=lfs diff=lfs merge=lfs -text
22
+ *.pt filter=lfs diff=lfs merge=lfs -text
23
+ *.pth filter=lfs diff=lfs merge=lfs -text
24
+ *.rar filter=lfs diff=lfs merge=lfs -text
25
+ *.safetensors filter=lfs diff=lfs merge=lfs -text
26
+ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
+ *.tar.* filter=lfs diff=lfs merge=lfs -text
28
+ *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.tflite filter=lfs diff=lfs merge=lfs -text
30
+ *.tgz filter=lfs diff=lfs merge=lfs -text
31
+ *.wasm filter=lfs diff=lfs merge=lfs -text
32
+ *.xz filter=lfs diff=lfs merge=lfs -text
33
+ *.zip filter=lfs diff=lfs merge=lfs -text
34
+ *.zst filter=lfs diff=lfs merge=lfs -text
35
+ *tfevents* filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ ---
3
+ tags:
4
+ - autotrain
5
+ - transformers
6
+ - image-classification
7
+ base_model: haywoodsloan/ai-image-detector-deploy
8
+ widget:
9
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
10
+ example_title: Tiger
11
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
12
+ example_title: Teapot
13
+ - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
14
+ example_title: Palace
15
+ ---
16
+
17
+ # Model Trained Using AutoTrain
18
+
19
+ - Problem type: Image Classification
20
+
21
+ ## Validation Metrics
22
+ loss: 0.09541802108287811
23
+
24
+ f1: 0.9853826686447335
25
+
26
+ precision: 0.9808886765408504
27
+
28
+ recall: 0.9899180291938807
29
+
30
+ auc: 0.9957081876919603
31
+
32
+ accuracy: 0.9794339738473816
checkpoint-11947/config.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "haywoodsloan/ai-image-detector-deploy",
3
+ "_num_labels": 2,
4
+ "architectures": [
5
+ "Swinv2ForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "depths": [
9
+ 2,
10
+ 2,
11
+ 18,
12
+ 2
13
+ ],
14
+ "drop_path_rate": 0.1,
15
+ "embed_dim": 192,
16
+ "encoder_stride": 32,
17
+ "hidden_act": "gelu",
18
+ "hidden_dropout_prob": 0.0,
19
+ "hidden_size": 1536,
20
+ "id2label": {
21
+ "0": "artificial",
22
+ "1": "real"
23
+ },
24
+ "image_size": 256,
25
+ "initializer_range": 0.02,
26
+ "label2id": {
27
+ "artificial": 0,
28
+ "real": 1
29
+ },
30
+ "layer_norm_eps": 1e-05,
31
+ "mlp_ratio": 4.0,
32
+ "model_type": "swinv2",
33
+ "num_channels": 3,
34
+ "num_heads": [
35
+ 6,
36
+ 12,
37
+ 24,
38
+ 48
39
+ ],
40
+ "num_layers": 4,
41
+ "out_features": [
42
+ "stage4"
43
+ ],
44
+ "out_indices": [
45
+ 4
46
+ ],
47
+ "patch_size": 4,
48
+ "path_norm": true,
49
+ "pretrained_window_sizes": [
50
+ 12,
51
+ 12,
52
+ 12,
53
+ 6
54
+ ],
55
+ "problem_type": "single_label_classification",
56
+ "qkv_bias": true,
57
+ "stage_names": [
58
+ "stem",
59
+ "stage1",
60
+ "stage2",
61
+ "stage3",
62
+ "stage4"
63
+ ],
64
+ "torch_dtype": "float32",
65
+ "transformers_version": "4.48.0",
66
+ "use_absolute_embeddings": false,
67
+ "window_size": 16
68
+ }
checkpoint-11947/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:802ba4e65566633ea393c79a8f7a54ab37e312bd6055442274592086c1a8e8f9
3
+ size 780886280
checkpoint-11947/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:df7e49a707d39efb20f7a4ebe23766e927bc4a9493bb16853ac8a9a3a8c6d4b5
3
+ size 1562042973
checkpoint-11947/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:63b91129ba9e9f0776a474177e74224169c0672b4a88f78f9bf3cf0327d5d8fa
3
+ size 14244
checkpoint-11947/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9e6d3afd36e75f1b911970670ed62e1e81c9ab90d0f6218dff05fe39b3c6d669
3
+ size 1064
checkpoint-11947/trainer_state.json ADDED
@@ -0,0 +1,3394 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.09541802108287811,
3
+ "best_model_checkpoint": "autotrain-ai-image-detect-20250418-0133/checkpoint-11947",
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 11947,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.0020925755419770654,
13
+ "grad_norm": 0.03698592260479927,
14
+ "learning_rate": 3.9999567824173485e-05,
15
+ "loss": 0.0423,
16
+ "step": 25
17
+ },
18
+ {
19
+ "epoch": 0.004185151083954131,
20
+ "grad_norm": 8.344589233398438,
21
+ "learning_rate": 3.999827131537151e-05,
22
+ "loss": 0.0367,
23
+ "step": 50
24
+ },
25
+ {
26
+ "epoch": 0.006277726625931196,
27
+ "grad_norm": 0.0733426958322525,
28
+ "learning_rate": 3.999611052962606e-05,
29
+ "loss": 0.0461,
30
+ "step": 75
31
+ },
32
+ {
33
+ "epoch": 0.008370302167908262,
34
+ "grad_norm": 0.00953048188239336,
35
+ "learning_rate": 3.999308556032107e-05,
36
+ "loss": 0.0482,
37
+ "step": 100
38
+ },
39
+ {
40
+ "epoch": 0.010462877709885327,
41
+ "grad_norm": 4.583376884460449,
42
+ "learning_rate": 3.9989196538188406e-05,
43
+ "loss": 0.0225,
44
+ "step": 125
45
+ },
46
+ {
47
+ "epoch": 0.012555453251862391,
48
+ "grad_norm": 0.005306304432451725,
49
+ "learning_rate": 3.9984443631302194e-05,
50
+ "loss": 0.0391,
51
+ "step": 150
52
+ },
53
+ {
54
+ "epoch": 0.014648028793839458,
55
+ "grad_norm": 13.573140144348145,
56
+ "learning_rate": 3.99788270450716e-05,
57
+ "loss": 0.108,
58
+ "step": 175
59
+ },
60
+ {
61
+ "epoch": 0.016740604335816523,
62
+ "grad_norm": 4.633761405944824,
63
+ "learning_rate": 3.997234702223188e-05,
64
+ "loss": 0.1012,
65
+ "step": 200
66
+ },
67
+ {
68
+ "epoch": 0.01883317987779359,
69
+ "grad_norm": 0.4037593603134155,
70
+ "learning_rate": 3.996500384283397e-05,
71
+ "loss": 0.0122,
72
+ "step": 225
73
+ },
74
+ {
75
+ "epoch": 0.020925755419770653,
76
+ "grad_norm": 0.5655341148376465,
77
+ "learning_rate": 3.995679782423232e-05,
78
+ "loss": 0.0216,
79
+ "step": 250
80
+ },
81
+ {
82
+ "epoch": 0.02301833096174772,
83
+ "grad_norm": 0.48408448696136475,
84
+ "learning_rate": 3.9947729321071243e-05,
85
+ "loss": 0.0208,
86
+ "step": 275
87
+ },
88
+ {
89
+ "epoch": 0.025110906503724783,
90
+ "grad_norm": 0.5579467415809631,
91
+ "learning_rate": 3.99377987252695e-05,
92
+ "loss": 0.0292,
93
+ "step": 300
94
+ },
95
+ {
96
+ "epoch": 0.02720348204570185,
97
+ "grad_norm": 6.891158103942871,
98
+ "learning_rate": 3.992700646600344e-05,
99
+ "loss": 0.0899,
100
+ "step": 325
101
+ },
102
+ {
103
+ "epoch": 0.029296057587678916,
104
+ "grad_norm": 9.514092445373535,
105
+ "learning_rate": 3.991535300968842e-05,
106
+ "loss": 0.0542,
107
+ "step": 350
108
+ },
109
+ {
110
+ "epoch": 0.03138863312965598,
111
+ "grad_norm": 5.77216911315918,
112
+ "learning_rate": 3.990283885995866e-05,
113
+ "loss": 0.0797,
114
+ "step": 375
115
+ },
116
+ {
117
+ "epoch": 0.033481208671633046,
118
+ "grad_norm": 0.1123751625418663,
119
+ "learning_rate": 3.988946455764545e-05,
120
+ "loss": 0.0527,
121
+ "step": 400
122
+ },
123
+ {
124
+ "epoch": 0.03557378421361011,
125
+ "grad_norm": 12.38884449005127,
126
+ "learning_rate": 3.9875230680753816e-05,
127
+ "loss": 0.0204,
128
+ "step": 425
129
+ },
130
+ {
131
+ "epoch": 0.03766635975558718,
132
+ "grad_norm": 0.028740063309669495,
133
+ "learning_rate": 3.98601378444375e-05,
134
+ "loss": 0.0279,
135
+ "step": 450
136
+ },
137
+ {
138
+ "epoch": 0.03975893529756424,
139
+ "grad_norm": 0.9096512794494629,
140
+ "learning_rate": 3.984418670097241e-05,
141
+ "loss": 0.0575,
142
+ "step": 475
143
+ },
144
+ {
145
+ "epoch": 0.041851510839541306,
146
+ "grad_norm": 0.3465224802494049,
147
+ "learning_rate": 3.98273779397284e-05,
148
+ "loss": 0.0248,
149
+ "step": 500
150
+ },
151
+ {
152
+ "epoch": 0.04394408638151837,
153
+ "grad_norm": 0.26555371284484863,
154
+ "learning_rate": 3.98097122871395e-05,
155
+ "loss": 0.0073,
156
+ "step": 525
157
+ },
158
+ {
159
+ "epoch": 0.04603666192349544,
160
+ "grad_norm": 0.39954063296318054,
161
+ "learning_rate": 3.9791190506672516e-05,
162
+ "loss": 0.0297,
163
+ "step": 550
164
+ },
165
+ {
166
+ "epoch": 0.048129237465472506,
167
+ "grad_norm": 0.08370428532361984,
168
+ "learning_rate": 3.977181339879402e-05,
169
+ "loss": 0.0357,
170
+ "step": 575
171
+ },
172
+ {
173
+ "epoch": 0.050221813007449566,
174
+ "grad_norm": 0.010254177264869213,
175
+ "learning_rate": 3.975158180093578e-05,
176
+ "loss": 0.0209,
177
+ "step": 600
178
+ },
179
+ {
180
+ "epoch": 0.05231438854942663,
181
+ "grad_norm": 0.012340884655714035,
182
+ "learning_rate": 3.973049658745854e-05,
183
+ "loss": 0.0755,
184
+ "step": 625
185
+ },
186
+ {
187
+ "epoch": 0.0544069640914037,
188
+ "grad_norm": 0.42715218663215637,
189
+ "learning_rate": 3.9708558669614266e-05,
190
+ "loss": 0.0609,
191
+ "step": 650
192
+ },
193
+ {
194
+ "epoch": 0.056499539633380766,
195
+ "grad_norm": 0.14249233901500702,
196
+ "learning_rate": 3.968576899550673e-05,
197
+ "loss": 0.0424,
198
+ "step": 675
199
+ },
200
+ {
201
+ "epoch": 0.05859211517535783,
202
+ "grad_norm": 0.015084848739206791,
203
+ "learning_rate": 3.966212855005055e-05,
204
+ "loss": 0.0463,
205
+ "step": 700
206
+ },
207
+ {
208
+ "epoch": 0.06068469071733489,
209
+ "grad_norm": 0.2930925488471985,
210
+ "learning_rate": 3.963763835492865e-05,
211
+ "loss": 0.0465,
212
+ "step": 725
213
+ },
214
+ {
215
+ "epoch": 0.06277726625931196,
216
+ "grad_norm": 0.27187275886535645,
217
+ "learning_rate": 3.961229946854804e-05,
218
+ "loss": 0.0441,
219
+ "step": 750
220
+ },
221
+ {
222
+ "epoch": 0.06486984180128903,
223
+ "grad_norm": 14.745759963989258,
224
+ "learning_rate": 3.9586112985994154e-05,
225
+ "loss": 0.0369,
226
+ "step": 775
227
+ },
228
+ {
229
+ "epoch": 0.06696241734326609,
230
+ "grad_norm": 0.0668809786438942,
231
+ "learning_rate": 3.955908003898345e-05,
232
+ "loss": 0.0223,
233
+ "step": 800
234
+ },
235
+ {
236
+ "epoch": 0.06905499288524315,
237
+ "grad_norm": 0.01884956657886505,
238
+ "learning_rate": 3.953120179581457e-05,
239
+ "loss": 0.0709,
240
+ "step": 825
241
+ },
242
+ {
243
+ "epoch": 0.07114756842722023,
244
+ "grad_norm": 0.06137616187334061,
245
+ "learning_rate": 3.950247946131777e-05,
246
+ "loss": 0.0746,
247
+ "step": 850
248
+ },
249
+ {
250
+ "epoch": 0.07324014396919729,
251
+ "grad_norm": 6.313259601593018,
252
+ "learning_rate": 3.947291427680293e-05,
253
+ "loss": 0.0722,
254
+ "step": 875
255
+ },
256
+ {
257
+ "epoch": 0.07533271951117436,
258
+ "grad_norm": 0.09329789131879807,
259
+ "learning_rate": 3.944250752000585e-05,
260
+ "loss": 0.0475,
261
+ "step": 900
262
+ },
263
+ {
264
+ "epoch": 0.07742529505315142,
265
+ "grad_norm": 0.4573472738265991,
266
+ "learning_rate": 3.9411260505033065e-05,
267
+ "loss": 0.0454,
268
+ "step": 925
269
+ },
270
+ {
271
+ "epoch": 0.07951787059512848,
272
+ "grad_norm": 43.10087585449219,
273
+ "learning_rate": 3.937917458230501e-05,
274
+ "loss": 0.0217,
275
+ "step": 950
276
+ },
277
+ {
278
+ "epoch": 0.08161044613710555,
279
+ "grad_norm": 0.033992718905210495,
280
+ "learning_rate": 3.934625113849771e-05,
281
+ "loss": 0.0336,
282
+ "step": 975
283
+ },
284
+ {
285
+ "epoch": 0.08370302167908261,
286
+ "grad_norm": 18.495559692382812,
287
+ "learning_rate": 3.9312491596482826e-05,
288
+ "loss": 0.0366,
289
+ "step": 1000
290
+ },
291
+ {
292
+ "epoch": 0.08579559722105969,
293
+ "grad_norm": 0.8050076961517334,
294
+ "learning_rate": 3.9277897415266146e-05,
295
+ "loss": 0.0519,
296
+ "step": 1025
297
+ },
298
+ {
299
+ "epoch": 0.08788817276303675,
300
+ "grad_norm": 7.1265106201171875,
301
+ "learning_rate": 3.924247008992456e-05,
302
+ "loss": 0.0687,
303
+ "step": 1050
304
+ },
305
+ {
306
+ "epoch": 0.0899807483050138,
307
+ "grad_norm": 0.0298299677670002,
308
+ "learning_rate": 3.920621115154143e-05,
309
+ "loss": 0.0591,
310
+ "step": 1075
311
+ },
312
+ {
313
+ "epoch": 0.09207332384699088,
314
+ "grad_norm": 17.500030517578125,
315
+ "learning_rate": 3.916912216714043e-05,
316
+ "loss": 0.0513,
317
+ "step": 1100
318
+ },
319
+ {
320
+ "epoch": 0.09416589938896794,
321
+ "grad_norm": 8.270520210266113,
322
+ "learning_rate": 3.9131204739617785e-05,
323
+ "loss": 0.0614,
324
+ "step": 1125
325
+ },
326
+ {
327
+ "epoch": 0.09625847493094501,
328
+ "grad_norm": 0.8669015765190125,
329
+ "learning_rate": 3.909246050767307e-05,
330
+ "loss": 0.0556,
331
+ "step": 1150
332
+ },
333
+ {
334
+ "epoch": 0.09835105047292207,
335
+ "grad_norm": 1.3969131708145142,
336
+ "learning_rate": 3.905289114573833e-05,
337
+ "loss": 0.0511,
338
+ "step": 1175
339
+ },
340
+ {
341
+ "epoch": 0.10044362601489913,
342
+ "grad_norm": 0.07863520085811615,
343
+ "learning_rate": 3.901249836390574e-05,
344
+ "loss": 0.0445,
345
+ "step": 1200
346
+ },
347
+ {
348
+ "epoch": 0.1025362015568762,
349
+ "grad_norm": 10.990960121154785,
350
+ "learning_rate": 3.897128390785368e-05,
351
+ "loss": 0.0525,
352
+ "step": 1225
353
+ },
354
+ {
355
+ "epoch": 0.10462877709885327,
356
+ "grad_norm": 0.1187528669834137,
357
+ "learning_rate": 3.892924955877131e-05,
358
+ "loss": 0.0538,
359
+ "step": 1250
360
+ },
361
+ {
362
+ "epoch": 0.10672135264083034,
363
+ "grad_norm": 0.329466313123703,
364
+ "learning_rate": 3.888639713328159e-05,
365
+ "loss": 0.0606,
366
+ "step": 1275
367
+ },
368
+ {
369
+ "epoch": 0.1088139281828074,
370
+ "grad_norm": 0.12424540519714355,
371
+ "learning_rate": 3.8842728483362756e-05,
372
+ "loss": 0.0476,
373
+ "step": 1300
374
+ },
375
+ {
376
+ "epoch": 0.11090650372478446,
377
+ "grad_norm": 0.8104080557823181,
378
+ "learning_rate": 3.8798245496268297e-05,
379
+ "loss": 0.0388,
380
+ "step": 1325
381
+ },
382
+ {
383
+ "epoch": 0.11299907926676153,
384
+ "grad_norm": 0.26356926560401917,
385
+ "learning_rate": 3.875295009444539e-05,
386
+ "loss": 0.0435,
387
+ "step": 1350
388
+ },
389
+ {
390
+ "epoch": 0.11509165480873859,
391
+ "grad_norm": 0.3834102749824524,
392
+ "learning_rate": 3.8706844235451795e-05,
393
+ "loss": 0.0295,
394
+ "step": 1375
395
+ },
396
+ {
397
+ "epoch": 0.11718423035071567,
398
+ "grad_norm": 10.899662971496582,
399
+ "learning_rate": 3.865992991187129e-05,
400
+ "loss": 0.064,
401
+ "step": 1400
402
+ },
403
+ {
404
+ "epoch": 0.11927680589269272,
405
+ "grad_norm": 0.03342663124203682,
406
+ "learning_rate": 3.861220915122755e-05,
407
+ "loss": 0.0311,
408
+ "step": 1425
409
+ },
410
+ {
411
+ "epoch": 0.12136938143466978,
412
+ "grad_norm": 7.466312885284424,
413
+ "learning_rate": 3.856368401589646e-05,
414
+ "loss": 0.0135,
415
+ "step": 1450
416
+ },
417
+ {
418
+ "epoch": 0.12346195697664686,
419
+ "grad_norm": 0.12502658367156982,
420
+ "learning_rate": 3.8514356603017085e-05,
421
+ "loss": 0.0299,
422
+ "step": 1475
423
+ },
424
+ {
425
+ "epoch": 0.12555453251862392,
426
+ "grad_norm": 0.0441652350127697,
427
+ "learning_rate": 3.846422904440096e-05,
428
+ "loss": 0.0305,
429
+ "step": 1500
430
+ },
431
+ {
432
+ "epoch": 0.12764710806060098,
433
+ "grad_norm": 8.360950469970703,
434
+ "learning_rate": 3.8413303506440005e-05,
435
+ "loss": 0.0612,
436
+ "step": 1525
437
+ },
438
+ {
439
+ "epoch": 0.12973968360257807,
440
+ "grad_norm": 30.80620765686035,
441
+ "learning_rate": 3.836158219001286e-05,
442
+ "loss": 0.0324,
443
+ "step": 1550
444
+ },
445
+ {
446
+ "epoch": 0.13183225914455513,
447
+ "grad_norm": 5.32338285446167,
448
+ "learning_rate": 3.830906733038979e-05,
449
+ "loss": 0.1051,
450
+ "step": 1575
451
+ },
452
+ {
453
+ "epoch": 0.13392483468653218,
454
+ "grad_norm": 0.44424283504486084,
455
+ "learning_rate": 3.8255761197136084e-05,
456
+ "loss": 0.0862,
457
+ "step": 1600
458
+ },
459
+ {
460
+ "epoch": 0.13601741022850924,
461
+ "grad_norm": 0.15658941864967346,
462
+ "learning_rate": 3.820166609401396e-05,
463
+ "loss": 0.0393,
464
+ "step": 1625
465
+ },
466
+ {
467
+ "epoch": 0.1381099857704863,
468
+ "grad_norm": 0.022251861169934273,
469
+ "learning_rate": 3.8146784358883e-05,
470
+ "loss": 0.0592,
471
+ "step": 1650
472
+ },
473
+ {
474
+ "epoch": 0.1402025613124634,
475
+ "grad_norm": 0.01628238894045353,
476
+ "learning_rate": 3.8091118363599146e-05,
477
+ "loss": 0.0539,
478
+ "step": 1675
479
+ },
480
+ {
481
+ "epoch": 0.14229513685444045,
482
+ "grad_norm": 0.7614484429359436,
483
+ "learning_rate": 3.8034670513912136e-05,
484
+ "loss": 0.0852,
485
+ "step": 1700
486
+ },
487
+ {
488
+ "epoch": 0.1443877123964175,
489
+ "grad_norm": 0.04743648320436478,
490
+ "learning_rate": 3.797744324936159e-05,
491
+ "loss": 0.0344,
492
+ "step": 1725
493
+ },
494
+ {
495
+ "epoch": 0.14648028793839457,
496
+ "grad_norm": 10.843757629394531,
497
+ "learning_rate": 3.7919439043171536e-05,
498
+ "loss": 0.0516,
499
+ "step": 1750
500
+ },
501
+ {
502
+ "epoch": 0.14857286348037163,
503
+ "grad_norm": 11.178771018981934,
504
+ "learning_rate": 3.786066040214355e-05,
505
+ "loss": 0.0657,
506
+ "step": 1775
507
+ },
508
+ {
509
+ "epoch": 0.15066543902234872,
510
+ "grad_norm": 0.1333361119031906,
511
+ "learning_rate": 3.780110986654841e-05,
512
+ "loss": 0.0321,
513
+ "step": 1800
514
+ },
515
+ {
516
+ "epoch": 0.15275801456432578,
517
+ "grad_norm": 8.613201141357422,
518
+ "learning_rate": 3.774079001001631e-05,
519
+ "loss": 0.0344,
520
+ "step": 1825
521
+ },
522
+ {
523
+ "epoch": 0.15485059010630284,
524
+ "grad_norm": 17.0685977935791,
525
+ "learning_rate": 3.7679703439425635e-05,
526
+ "loss": 0.0392,
527
+ "step": 1850
528
+ },
529
+ {
530
+ "epoch": 0.1569431656482799,
531
+ "grad_norm": 0.14994333684444427,
532
+ "learning_rate": 3.76178527947903e-05,
533
+ "loss": 0.0831,
534
+ "step": 1875
535
+ },
536
+ {
537
+ "epoch": 0.15903574119025696,
538
+ "grad_norm": 0.13761335611343384,
539
+ "learning_rate": 3.7555240749145644e-05,
540
+ "loss": 0.0669,
541
+ "step": 1900
542
+ },
543
+ {
544
+ "epoch": 0.16112831673223404,
545
+ "grad_norm": 0.015248261392116547,
546
+ "learning_rate": 3.7491870008432935e-05,
547
+ "loss": 0.038,
548
+ "step": 1925
549
+ },
550
+ {
551
+ "epoch": 0.1632208922742111,
552
+ "grad_norm": 0.13382545113563538,
553
+ "learning_rate": 3.7427743311382395e-05,
554
+ "loss": 0.0298,
555
+ "step": 1950
556
+ },
557
+ {
558
+ "epoch": 0.16531346781618816,
559
+ "grad_norm": 7.744543552398682,
560
+ "learning_rate": 3.7362863429394855e-05,
561
+ "loss": 0.0405,
562
+ "step": 1975
563
+ },
564
+ {
565
+ "epoch": 0.16740604335816522,
566
+ "grad_norm": 0.06504332274198532,
567
+ "learning_rate": 3.7297233166421974e-05,
568
+ "loss": 0.0549,
569
+ "step": 2000
570
+ },
571
+ {
572
+ "epoch": 0.16949861890014228,
573
+ "grad_norm": 0.023874608799815178,
574
+ "learning_rate": 3.723085535884507e-05,
575
+ "loss": 0.0328,
576
+ "step": 2025
577
+ },
578
+ {
579
+ "epoch": 0.17159119444211937,
580
+ "grad_norm": 0.1030099168419838,
581
+ "learning_rate": 3.716373287535252e-05,
582
+ "loss": 0.0523,
583
+ "step": 2050
584
+ },
585
+ {
586
+ "epoch": 0.17368376998409643,
587
+ "grad_norm": 12.099047660827637,
588
+ "learning_rate": 3.709586861681581e-05,
589
+ "loss": 0.0452,
590
+ "step": 2075
591
+ },
592
+ {
593
+ "epoch": 0.1757763455260735,
594
+ "grad_norm": 1.0645301342010498,
595
+ "learning_rate": 3.702726551616413e-05,
596
+ "loss": 0.0376,
597
+ "step": 2100
598
+ },
599
+ {
600
+ "epoch": 0.17786892106805055,
601
+ "grad_norm": 0.7780424356460571,
602
+ "learning_rate": 3.6957926538257676e-05,
603
+ "loss": 0.08,
604
+ "step": 2125
605
+ },
606
+ {
607
+ "epoch": 0.1799614966100276,
608
+ "grad_norm": 0.07100946456193924,
609
+ "learning_rate": 3.688785467975944e-05,
610
+ "loss": 0.0382,
611
+ "step": 2150
612
+ },
613
+ {
614
+ "epoch": 0.1820540721520047,
615
+ "grad_norm": 0.06628437340259552,
616
+ "learning_rate": 3.681705296900576e-05,
617
+ "loss": 0.0544,
618
+ "step": 2175
619
+ },
620
+ {
621
+ "epoch": 0.18414664769398176,
622
+ "grad_norm": 13.79120922088623,
623
+ "learning_rate": 3.674552446587543e-05,
624
+ "loss": 0.08,
625
+ "step": 2200
626
+ },
627
+ {
628
+ "epoch": 0.18623922323595882,
629
+ "grad_norm": 4.065523147583008,
630
+ "learning_rate": 3.667327226165743e-05,
631
+ "loss": 0.0284,
632
+ "step": 2225
633
+ },
634
+ {
635
+ "epoch": 0.18833179877793588,
636
+ "grad_norm": 20.047752380371094,
637
+ "learning_rate": 3.660029947891738e-05,
638
+ "loss": 0.0808,
639
+ "step": 2250
640
+ },
641
+ {
642
+ "epoch": 0.19042437431991294,
643
+ "grad_norm": 0.29820454120635986,
644
+ "learning_rate": 3.652660927136256e-05,
645
+ "loss": 0.0445,
646
+ "step": 2275
647
+ },
648
+ {
649
+ "epoch": 0.19251694986189002,
650
+ "grad_norm": 0.5741401314735413,
651
+ "learning_rate": 3.645220482370558e-05,
652
+ "loss": 0.0782,
653
+ "step": 2300
654
+ },
655
+ {
656
+ "epoch": 0.19460952540386708,
657
+ "grad_norm": 0.03809121996164322,
658
+ "learning_rate": 3.637708935152683e-05,
659
+ "loss": 0.0324,
660
+ "step": 2325
661
+ },
662
+ {
663
+ "epoch": 0.19670210094584414,
664
+ "grad_norm": 0.1253562569618225,
665
+ "learning_rate": 3.630126610113541e-05,
666
+ "loss": 0.0573,
667
+ "step": 2350
668
+ },
669
+ {
670
+ "epoch": 0.1987946764878212,
671
+ "grad_norm": 10.205113410949707,
672
+ "learning_rate": 3.622473834942894e-05,
673
+ "loss": 0.0449,
674
+ "step": 2375
675
+ },
676
+ {
677
+ "epoch": 0.20088725202979826,
678
+ "grad_norm": 0.18986789882183075,
679
+ "learning_rate": 3.6147509403751844e-05,
680
+ "loss": 0.061,
681
+ "step": 2400
682
+ },
683
+ {
684
+ "epoch": 0.20297982757177535,
685
+ "grad_norm": 0.2179853469133377,
686
+ "learning_rate": 3.606958260175245e-05,
687
+ "loss": 0.0501,
688
+ "step": 2425
689
+ },
690
+ {
691
+ "epoch": 0.2050724031137524,
692
+ "grad_norm": 8.819538116455078,
693
+ "learning_rate": 3.599096131123879e-05,
694
+ "loss": 0.0484,
695
+ "step": 2450
696
+ },
697
+ {
698
+ "epoch": 0.20716497865572947,
699
+ "grad_norm": 11.121920585632324,
700
+ "learning_rate": 3.591164893003296e-05,
701
+ "loss": 0.0516,
702
+ "step": 2475
703
+ },
704
+ {
705
+ "epoch": 0.20925755419770653,
706
+ "grad_norm": 6.540257453918457,
707
+ "learning_rate": 3.5831648885824375e-05,
708
+ "loss": 0.0473,
709
+ "step": 2500
710
+ },
711
+ {
712
+ "epoch": 0.2113501297396836,
713
+ "grad_norm": 0.03221260383725166,
714
+ "learning_rate": 3.5750964636021535e-05,
715
+ "loss": 0.0161,
716
+ "step": 2525
717
+ },
718
+ {
719
+ "epoch": 0.21344270528166068,
720
+ "grad_norm": 0.006673712749034166,
721
+ "learning_rate": 3.566959966760269e-05,
722
+ "loss": 0.0543,
723
+ "step": 2550
724
+ },
725
+ {
726
+ "epoch": 0.21553528082363774,
727
+ "grad_norm": 0.0166669562458992,
728
+ "learning_rate": 3.5587557496965086e-05,
729
+ "loss": 0.05,
730
+ "step": 2575
731
+ },
732
+ {
733
+ "epoch": 0.2176278563656148,
734
+ "grad_norm": 0.056203585118055344,
735
+ "learning_rate": 3.5504841669773e-05,
736
+ "loss": 0.0318,
737
+ "step": 2600
738
+ },
739
+ {
740
+ "epoch": 0.21972043190759186,
741
+ "grad_norm": 0.15732666850090027,
742
+ "learning_rate": 3.542145576080454e-05,
743
+ "loss": 0.0285,
744
+ "step": 2625
745
+ },
746
+ {
747
+ "epoch": 0.22181300744956892,
748
+ "grad_norm": 0.6591875553131104,
749
+ "learning_rate": 3.533740337379713e-05,
750
+ "loss": 0.0366,
751
+ "step": 2650
752
+ },
753
+ {
754
+ "epoch": 0.223905582991546,
755
+ "grad_norm": 110.1420669555664,
756
+ "learning_rate": 3.5252688141291725e-05,
757
+ "loss": 0.0378,
758
+ "step": 2675
759
+ },
760
+ {
761
+ "epoch": 0.22599815853352306,
762
+ "grad_norm": 2.0431766510009766,
763
+ "learning_rate": 3.516731372447592e-05,
764
+ "loss": 0.0819,
765
+ "step": 2700
766
+ },
767
+ {
768
+ "epoch": 0.22809073407550012,
769
+ "grad_norm": 25.620094299316406,
770
+ "learning_rate": 3.5081283813025604e-05,
771
+ "loss": 0.0576,
772
+ "step": 2725
773
+ },
774
+ {
775
+ "epoch": 0.23018330961747718,
776
+ "grad_norm": 21.387821197509766,
777
+ "learning_rate": 3.4994602124945604e-05,
778
+ "loss": 0.08,
779
+ "step": 2750
780
+ },
781
+ {
782
+ "epoch": 0.23227588515945424,
783
+ "grad_norm": 3.099421262741089,
784
+ "learning_rate": 3.490727240640893e-05,
785
+ "loss": 0.0561,
786
+ "step": 2775
787
+ },
788
+ {
789
+ "epoch": 0.23436846070143133,
790
+ "grad_norm": 8.681183815002441,
791
+ "learning_rate": 3.481929843159491e-05,
792
+ "loss": 0.062,
793
+ "step": 2800
794
+ },
795
+ {
796
+ "epoch": 0.2364610362434084,
797
+ "grad_norm": 0.09572477638721466,
798
+ "learning_rate": 3.473068400252607e-05,
799
+ "loss": 0.0503,
800
+ "step": 2825
801
+ },
802
+ {
803
+ "epoch": 0.23855361178538545,
804
+ "grad_norm": 7.184039115905762,
805
+ "learning_rate": 3.464143294890383e-05,
806
+ "loss": 0.0848,
807
+ "step": 2850
808
+ },
809
+ {
810
+ "epoch": 0.2406461873273625,
811
+ "grad_norm": 5.73546838760376,
812
+ "learning_rate": 3.455154912794297e-05,
813
+ "loss": 0.0455,
814
+ "step": 2875
815
+ },
816
+ {
817
+ "epoch": 0.24273876286933957,
818
+ "grad_norm": 0.046288516372442245,
819
+ "learning_rate": 3.4461036424204955e-05,
820
+ "loss": 0.0306,
821
+ "step": 2900
822
+ },
823
+ {
824
+ "epoch": 0.24483133841131666,
825
+ "grad_norm": 0.867224395275116,
826
+ "learning_rate": 3.4369898749430044e-05,
827
+ "loss": 0.0382,
828
+ "step": 2925
829
+ },
830
+ {
831
+ "epoch": 0.24692391395329372,
832
+ "grad_norm": 0.01112967636436224,
833
+ "learning_rate": 3.427814004236822e-05,
834
+ "loss": 0.0927,
835
+ "step": 2950
836
+ },
837
+ {
838
+ "epoch": 0.24901648949527078,
839
+ "grad_norm": 0.26373812556266785,
840
+ "learning_rate": 3.4185764268608996e-05,
841
+ "loss": 0.0338,
842
+ "step": 2975
843
+ },
844
+ {
845
+ "epoch": 0.25110906503724784,
846
+ "grad_norm": 9.9061279296875,
847
+ "learning_rate": 3.409277542041001e-05,
848
+ "loss": 0.0344,
849
+ "step": 3000
850
+ },
851
+ {
852
+ "epoch": 0.2532016405792249,
853
+ "grad_norm": 0.27759966254234314,
854
+ "learning_rate": 3.399917751652448e-05,
855
+ "loss": 0.0332,
856
+ "step": 3025
857
+ },
858
+ {
859
+ "epoch": 0.25529421612120196,
860
+ "grad_norm": 0.04884539917111397,
861
+ "learning_rate": 3.390497460202758e-05,
862
+ "loss": 0.0848,
863
+ "step": 3050
864
+ },
865
+ {
866
+ "epoch": 0.257386791663179,
867
+ "grad_norm": 4.758861541748047,
868
+ "learning_rate": 3.3810170748141534e-05,
869
+ "loss": 0.0408,
870
+ "step": 3075
871
+ },
872
+ {
873
+ "epoch": 0.25947936720515613,
874
+ "grad_norm": 0.06590425223112106,
875
+ "learning_rate": 3.371477005205974e-05,
876
+ "loss": 0.0334,
877
+ "step": 3100
878
+ },
879
+ {
880
+ "epoch": 0.2615719427471332,
881
+ "grad_norm": 0.018905460834503174,
882
+ "learning_rate": 3.361877663676966e-05,
883
+ "loss": 0.0278,
884
+ "step": 3125
885
+ },
886
+ {
887
+ "epoch": 0.26366451828911025,
888
+ "grad_norm": 0.08253883570432663,
889
+ "learning_rate": 3.352219465087466e-05,
890
+ "loss": 0.0406,
891
+ "step": 3150
892
+ },
893
+ {
894
+ "epoch": 0.2657570938310873,
895
+ "grad_norm": 23.68890380859375,
896
+ "learning_rate": 3.34250282684147e-05,
897
+ "loss": 0.0697,
898
+ "step": 3175
899
+ },
900
+ {
901
+ "epoch": 0.26784966937306437,
902
+ "grad_norm": 0.23610574007034302,
903
+ "learning_rate": 3.332728168868593e-05,
904
+ "loss": 0.0315,
905
+ "step": 3200
906
+ },
907
+ {
908
+ "epoch": 0.26994224491504143,
909
+ "grad_norm": 1.0449060201644897,
910
+ "learning_rate": 3.322895913605925e-05,
911
+ "loss": 0.0252,
912
+ "step": 3225
913
+ },
914
+ {
915
+ "epoch": 0.2720348204570185,
916
+ "grad_norm": 26.833650588989258,
917
+ "learning_rate": 3.313006485979772e-05,
918
+ "loss": 0.0832,
919
+ "step": 3250
920
+ },
921
+ {
922
+ "epoch": 0.27412739599899555,
923
+ "grad_norm": 0.09942333400249481,
924
+ "learning_rate": 3.303060313387286e-05,
925
+ "loss": 0.0174,
926
+ "step": 3275
927
+ },
928
+ {
929
+ "epoch": 0.2762199715409726,
930
+ "grad_norm": 0.023540310561656952,
931
+ "learning_rate": 3.293057825678007e-05,
932
+ "loss": 0.0957,
933
+ "step": 3300
934
+ },
935
+ {
936
+ "epoch": 0.27831254708294967,
937
+ "grad_norm": 0.07260198891162872,
938
+ "learning_rate": 3.282999455135272e-05,
939
+ "loss": 0.0699,
940
+ "step": 3325
941
+ },
942
+ {
943
+ "epoch": 0.2804051226249268,
944
+ "grad_norm": 20.466096878051758,
945
+ "learning_rate": 3.272885636457541e-05,
946
+ "loss": 0.069,
947
+ "step": 3350
948
+ },
949
+ {
950
+ "epoch": 0.28249769816690384,
951
+ "grad_norm": 0.5947694778442383,
952
+ "learning_rate": 3.2627168067396105e-05,
953
+ "loss": 0.0471,
954
+ "step": 3375
955
+ },
956
+ {
957
+ "epoch": 0.2845902737088809,
958
+ "grad_norm": 0.021415280178189278,
959
+ "learning_rate": 3.252493405453718e-05,
960
+ "loss": 0.0423,
961
+ "step": 3400
962
+ },
963
+ {
964
+ "epoch": 0.28668284925085796,
965
+ "grad_norm": 0.10097396373748779,
966
+ "learning_rate": 3.242215874430554e-05,
967
+ "loss": 0.0285,
968
+ "step": 3425
969
+ },
970
+ {
971
+ "epoch": 0.288775424792835,
972
+ "grad_norm": 16.974971771240234,
973
+ "learning_rate": 3.231884657840165e-05,
974
+ "loss": 0.0559,
975
+ "step": 3450
976
+ },
977
+ {
978
+ "epoch": 0.2908680003348121,
979
+ "grad_norm": 7.112433433532715,
980
+ "learning_rate": 3.221500202172757e-05,
981
+ "loss": 0.0499,
982
+ "step": 3475
983
+ },
984
+ {
985
+ "epoch": 0.29296057587678914,
986
+ "grad_norm": 0.02784404344856739,
987
+ "learning_rate": 3.2110629562194026e-05,
988
+ "loss": 0.0308,
989
+ "step": 3500
990
+ },
991
+ {
992
+ "epoch": 0.2950531514187662,
993
+ "grad_norm": 0.3016861379146576,
994
+ "learning_rate": 3.20057337105264e-05,
995
+ "loss": 0.0636,
996
+ "step": 3525
997
+ },
998
+ {
999
+ "epoch": 0.29714572696074326,
1000
+ "grad_norm": 16.54718780517578,
1001
+ "learning_rate": 3.190031900006984e-05,
1002
+ "loss": 0.0791,
1003
+ "step": 3550
1004
+ },
1005
+ {
1006
+ "epoch": 0.2992383025027203,
1007
+ "grad_norm": 4.460518836975098,
1008
+ "learning_rate": 3.1794389986593306e-05,
1009
+ "loss": 0.0682,
1010
+ "step": 3575
1011
+ },
1012
+ {
1013
+ "epoch": 0.30133087804469744,
1014
+ "grad_norm": 0.12846969068050385,
1015
+ "learning_rate": 3.168795124809269e-05,
1016
+ "loss": 0.0428,
1017
+ "step": 3600
1018
+ },
1019
+ {
1020
+ "epoch": 0.3034234535866745,
1021
+ "grad_norm": 32.92333221435547,
1022
+ "learning_rate": 3.158100738459298e-05,
1023
+ "loss": 0.0472,
1024
+ "step": 3625
1025
+ },
1026
+ {
1027
+ "epoch": 0.30551602912865156,
1028
+ "grad_norm": 7.472573757171631,
1029
+ "learning_rate": 3.147356301794943e-05,
1030
+ "loss": 0.0663,
1031
+ "step": 3650
1032
+ },
1033
+ {
1034
+ "epoch": 0.3076086046706286,
1035
+ "grad_norm": 0.17467910051345825,
1036
+ "learning_rate": 3.136562279164784e-05,
1037
+ "loss": 0.0449,
1038
+ "step": 3675
1039
+ },
1040
+ {
1041
+ "epoch": 0.3097011802126057,
1042
+ "grad_norm": 0.05833444744348526,
1043
+ "learning_rate": 3.125719137060384e-05,
1044
+ "loss": 0.0311,
1045
+ "step": 3700
1046
+ },
1047
+ {
1048
+ "epoch": 0.31179375575458274,
1049
+ "grad_norm": 0.027352772653102875,
1050
+ "learning_rate": 3.114827344096136e-05,
1051
+ "loss": 0.0474,
1052
+ "step": 3725
1053
+ },
1054
+ {
1055
+ "epoch": 0.3138863312965598,
1056
+ "grad_norm": 29.073902130126953,
1057
+ "learning_rate": 3.103887370989002e-05,
1058
+ "loss": 0.0631,
1059
+ "step": 3750
1060
+ },
1061
+ {
1062
+ "epoch": 0.31597890683853685,
1063
+ "grad_norm": 6.369044303894043,
1064
+ "learning_rate": 3.092899690538173e-05,
1065
+ "loss": 0.0718,
1066
+ "step": 3775
1067
+ },
1068
+ {
1069
+ "epoch": 0.3180714823805139,
1070
+ "grad_norm": 0.3031097948551178,
1071
+ "learning_rate": 3.081864777604637e-05,
1072
+ "loss": 0.0338,
1073
+ "step": 3800
1074
+ },
1075
+ {
1076
+ "epoch": 0.320164057922491,
1077
+ "grad_norm": 0.14146162569522858,
1078
+ "learning_rate": 3.070783109090657e-05,
1079
+ "loss": 0.0462,
1080
+ "step": 3825
1081
+ },
1082
+ {
1083
+ "epoch": 0.3222566334644681,
1084
+ "grad_norm": 7.264471054077148,
1085
+ "learning_rate": 3.0596551639191564e-05,
1086
+ "loss": 0.0507,
1087
+ "step": 3850
1088
+ },
1089
+ {
1090
+ "epoch": 0.32434920900644515,
1091
+ "grad_norm": 5.5302205085754395,
1092
+ "learning_rate": 3.0484814230130264e-05,
1093
+ "loss": 0.0594,
1094
+ "step": 3875
1095
+ },
1096
+ {
1097
+ "epoch": 0.3264417845484222,
1098
+ "grad_norm": 0.14829619228839874,
1099
+ "learning_rate": 3.0372623692743386e-05,
1100
+ "loss": 0.0154,
1101
+ "step": 3900
1102
+ },
1103
+ {
1104
+ "epoch": 0.32853436009039927,
1105
+ "grad_norm": 0.028743617236614227,
1106
+ "learning_rate": 3.0259984875634737e-05,
1107
+ "loss": 0.0758,
1108
+ "step": 3925
1109
+ },
1110
+ {
1111
+ "epoch": 0.33062693563237633,
1112
+ "grad_norm": 33.016883850097656,
1113
+ "learning_rate": 3.014690264678172e-05,
1114
+ "loss": 0.0608,
1115
+ "step": 3950
1116
+ },
1117
+ {
1118
+ "epoch": 0.3327195111743534,
1119
+ "grad_norm": 8.965993881225586,
1120
+ "learning_rate": 3.00333818933249e-05,
1121
+ "loss": 0.036,
1122
+ "step": 3975
1123
+ },
1124
+ {
1125
+ "epoch": 0.33481208671633045,
1126
+ "grad_norm": 0.054518092423677444,
1127
+ "learning_rate": 2.9919427521356827e-05,
1128
+ "loss": 0.0458,
1129
+ "step": 4000
1130
+ },
1131
+ {
1132
+ "epoch": 0.3369046622583075,
1133
+ "grad_norm": 0.04685276001691818,
1134
+ "learning_rate": 2.980504445570998e-05,
1135
+ "loss": 0.0587,
1136
+ "step": 4025
1137
+ },
1138
+ {
1139
+ "epoch": 0.33899723780028457,
1140
+ "grad_norm": 22.658573150634766,
1141
+ "learning_rate": 2.9690237639743965e-05,
1142
+ "loss": 0.0122,
1143
+ "step": 4050
1144
+ },
1145
+ {
1146
+ "epoch": 0.3410898133422617,
1147
+ "grad_norm": 0.01873711682856083,
1148
+ "learning_rate": 2.9575012035131823e-05,
1149
+ "loss": 0.047,
1150
+ "step": 4075
1151
+ },
1152
+ {
1153
+ "epoch": 0.34318238888423874,
1154
+ "grad_norm": 0.04110574722290039,
1155
+ "learning_rate": 2.9459372621645667e-05,
1156
+ "loss": 0.0196,
1157
+ "step": 4100
1158
+ },
1159
+ {
1160
+ "epoch": 0.3452749644262158,
1161
+ "grad_norm": 0.44694823026657104,
1162
+ "learning_rate": 2.934332439694139e-05,
1163
+ "loss": 0.0772,
1164
+ "step": 4125
1165
+ },
1166
+ {
1167
+ "epoch": 0.34736753996819286,
1168
+ "grad_norm": 5.95962381362915,
1169
+ "learning_rate": 2.9226872376342745e-05,
1170
+ "loss": 0.0203,
1171
+ "step": 4150
1172
+ },
1173
+ {
1174
+ "epoch": 0.3494601155101699,
1175
+ "grad_norm": 0.111228346824646,
1176
+ "learning_rate": 2.9110021592624553e-05,
1177
+ "loss": 0.068,
1178
+ "step": 4175
1179
+ },
1180
+ {
1181
+ "epoch": 0.351552691052147,
1182
+ "grad_norm": 0.08670425415039062,
1183
+ "learning_rate": 2.899277709579522e-05,
1184
+ "loss": 0.0349,
1185
+ "step": 4200
1186
+ },
1187
+ {
1188
+ "epoch": 0.35364526659412404,
1189
+ "grad_norm": 7.51801872253418,
1190
+ "learning_rate": 2.8875143952878474e-05,
1191
+ "loss": 0.0656,
1192
+ "step": 4225
1193
+ },
1194
+ {
1195
+ "epoch": 0.3557378421361011,
1196
+ "grad_norm": 0.4301106333732605,
1197
+ "learning_rate": 2.875712724769439e-05,
1198
+ "loss": 0.0337,
1199
+ "step": 4250
1200
+ },
1201
+ {
1202
+ "epoch": 0.35783041767807816,
1203
+ "grad_norm": 0.04058774933218956,
1204
+ "learning_rate": 2.863873208063969e-05,
1205
+ "loss": 0.0633,
1206
+ "step": 4275
1207
+ },
1208
+ {
1209
+ "epoch": 0.3599229932200552,
1210
+ "grad_norm": 0.29651474952697754,
1211
+ "learning_rate": 2.8519963568467277e-05,
1212
+ "loss": 0.0156,
1213
+ "step": 4300
1214
+ },
1215
+ {
1216
+ "epoch": 0.36201556876203234,
1217
+ "grad_norm": 0.7200632691383362,
1218
+ "learning_rate": 2.8400826844065146e-05,
1219
+ "loss": 0.0601,
1220
+ "step": 4325
1221
+ },
1222
+ {
1223
+ "epoch": 0.3641081443040094,
1224
+ "grad_norm": 0.13928024470806122,
1225
+ "learning_rate": 2.828132705623454e-05,
1226
+ "loss": 0.0505,
1227
+ "step": 4350
1228
+ },
1229
+ {
1230
+ "epoch": 0.36620071984598646,
1231
+ "grad_norm": 2.0252275466918945,
1232
+ "learning_rate": 2.8161469369467413e-05,
1233
+ "loss": 0.0213,
1234
+ "step": 4375
1235
+ },
1236
+ {
1237
+ "epoch": 0.3682932953879635,
1238
+ "grad_norm": 0.16088655591011047,
1239
+ "learning_rate": 2.8041258963723246e-05,
1240
+ "loss": 0.0167,
1241
+ "step": 4400
1242
+ },
1243
+ {
1244
+ "epoch": 0.3703858709299406,
1245
+ "grad_norm": 15.17540454864502,
1246
+ "learning_rate": 2.7920701034205184e-05,
1247
+ "loss": 0.0874,
1248
+ "step": 4425
1249
+ },
1250
+ {
1251
+ "epoch": 0.37247844647191763,
1252
+ "grad_norm": 6.356795787811279,
1253
+ "learning_rate": 2.7799800791135513e-05,
1254
+ "loss": 0.0839,
1255
+ "step": 4450
1256
+ },
1257
+ {
1258
+ "epoch": 0.3745710220138947,
1259
+ "grad_norm": 0.16258519887924194,
1260
+ "learning_rate": 2.767856345953048e-05,
1261
+ "loss": 0.0287,
1262
+ "step": 4475
1263
+ },
1264
+ {
1265
+ "epoch": 0.37666359755587175,
1266
+ "grad_norm": 0.0377529114484787,
1267
+ "learning_rate": 2.7556994278974483e-05,
1268
+ "loss": 0.0285,
1269
+ "step": 4500
1270
+ },
1271
+ {
1272
+ "epoch": 0.3787561730978488,
1273
+ "grad_norm": 0.036431267857551575,
1274
+ "learning_rate": 2.743509850339363e-05,
1275
+ "loss": 0.0326,
1276
+ "step": 4525
1277
+ },
1278
+ {
1279
+ "epoch": 0.3808487486398259,
1280
+ "grad_norm": 2.534477710723877,
1281
+ "learning_rate": 2.7312881400828678e-05,
1282
+ "loss": 0.0373,
1283
+ "step": 4550
1284
+ },
1285
+ {
1286
+ "epoch": 0.382941324181803,
1287
+ "grad_norm": 0.5988636612892151,
1288
+ "learning_rate": 2.719034825320736e-05,
1289
+ "loss": 0.0544,
1290
+ "step": 4575
1291
+ },
1292
+ {
1293
+ "epoch": 0.38503389972378005,
1294
+ "grad_norm": 15.286279678344727,
1295
+ "learning_rate": 2.706750435611611e-05,
1296
+ "loss": 0.0494,
1297
+ "step": 4600
1298
+ },
1299
+ {
1300
+ "epoch": 0.3871264752657571,
1301
+ "grad_norm": 0.05748358368873596,
1302
+ "learning_rate": 2.694435501857121e-05,
1303
+ "loss": 0.0355,
1304
+ "step": 4625
1305
+ },
1306
+ {
1307
+ "epoch": 0.38921905080773417,
1308
+ "grad_norm": 0.20348195731639862,
1309
+ "learning_rate": 2.682090556278932e-05,
1310
+ "loss": 0.0311,
1311
+ "step": 4650
1312
+ },
1313
+ {
1314
+ "epoch": 0.39131162634971123,
1315
+ "grad_norm": 0.05684299394488335,
1316
+ "learning_rate": 2.669716132395751e-05,
1317
+ "loss": 0.0366,
1318
+ "step": 4675
1319
+ },
1320
+ {
1321
+ "epoch": 0.3934042018916883,
1322
+ "grad_norm": 0.6135233044624329,
1323
+ "learning_rate": 2.6573127650002647e-05,
1324
+ "loss": 0.0439,
1325
+ "step": 4700
1326
+ },
1327
+ {
1328
+ "epoch": 0.39549677743366535,
1329
+ "grad_norm": 0.07451584190130234,
1330
+ "learning_rate": 2.644880990136029e-05,
1331
+ "loss": 0.0221,
1332
+ "step": 4725
1333
+ },
1334
+ {
1335
+ "epoch": 0.3975893529756424,
1336
+ "grad_norm": 0.18352004885673523,
1337
+ "learning_rate": 2.632421345074301e-05,
1338
+ "loss": 0.014,
1339
+ "step": 4750
1340
+ },
1341
+ {
1342
+ "epoch": 0.39968192851761947,
1343
+ "grad_norm": 15.234163284301758,
1344
+ "learning_rate": 2.619934368290821e-05,
1345
+ "loss": 0.0506,
1346
+ "step": 4775
1347
+ },
1348
+ {
1349
+ "epoch": 0.4017745040595965,
1350
+ "grad_norm": 0.15898753702640533,
1351
+ "learning_rate": 2.60742059944254e-05,
1352
+ "loss": 0.0339,
1353
+ "step": 4800
1354
+ },
1355
+ {
1356
+ "epoch": 0.40386707960157364,
1357
+ "grad_norm": 10.560185432434082,
1358
+ "learning_rate": 2.594880579344298e-05,
1359
+ "loss": 0.0681,
1360
+ "step": 4825
1361
+ },
1362
+ {
1363
+ "epoch": 0.4059596551435507,
1364
+ "grad_norm": 0.027397403493523598,
1365
+ "learning_rate": 2.5823148499454503e-05,
1366
+ "loss": 0.0387,
1367
+ "step": 4850
1368
+ },
1369
+ {
1370
+ "epoch": 0.40805223068552776,
1371
+ "grad_norm": 0.042963869869709015,
1372
+ "learning_rate": 2.5697239543064453e-05,
1373
+ "loss": 0.0575,
1374
+ "step": 4875
1375
+ },
1376
+ {
1377
+ "epoch": 0.4101448062275048,
1378
+ "grad_norm": 0.041683413088321686,
1379
+ "learning_rate": 2.557108436575356e-05,
1380
+ "loss": 0.0586,
1381
+ "step": 4900
1382
+ },
1383
+ {
1384
+ "epoch": 0.4122373817694819,
1385
+ "grad_norm": 0.016998037695884705,
1386
+ "learning_rate": 2.544468841964363e-05,
1387
+ "loss": 0.0532,
1388
+ "step": 4925
1389
+ },
1390
+ {
1391
+ "epoch": 0.41432995731145894,
1392
+ "grad_norm": 0.28552761673927307,
1393
+ "learning_rate": 2.531805716726191e-05,
1394
+ "loss": 0.0724,
1395
+ "step": 4950
1396
+ },
1397
+ {
1398
+ "epoch": 0.416422532853436,
1399
+ "grad_norm": 11.638047218322754,
1400
+ "learning_rate": 2.5191196081305008e-05,
1401
+ "loss": 0.0705,
1402
+ "step": 4975
1403
+ },
1404
+ {
1405
+ "epoch": 0.41851510839541306,
1406
+ "grad_norm": 0.08826717734336853,
1407
+ "learning_rate": 2.50641106444024e-05,
1408
+ "loss": 0.0216,
1409
+ "step": 5000
1410
+ },
1411
+ {
1412
+ "epoch": 0.4206076839373901,
1413
+ "grad_norm": 0.15376928448677063,
1414
+ "learning_rate": 2.493680634887946e-05,
1415
+ "loss": 0.0349,
1416
+ "step": 5025
1417
+ },
1418
+ {
1419
+ "epoch": 0.4227002594793672,
1420
+ "grad_norm": 0.09139838069677353,
1421
+ "learning_rate": 2.48092886965201e-05,
1422
+ "loss": 0.0422,
1423
+ "step": 5050
1424
+ },
1425
+ {
1426
+ "epoch": 0.4247928350213443,
1427
+ "grad_norm": 0.6814232468605042,
1428
+ "learning_rate": 2.4681563198328993e-05,
1429
+ "loss": 0.0522,
1430
+ "step": 5075
1431
+ },
1432
+ {
1433
+ "epoch": 0.42688541056332135,
1434
+ "grad_norm": 0.7719263434410095,
1435
+ "learning_rate": 2.4553635374293428e-05,
1436
+ "loss": 0.0218,
1437
+ "step": 5100
1438
+ },
1439
+ {
1440
+ "epoch": 0.4289779861052984,
1441
+ "grad_norm": 4.256403923034668,
1442
+ "learning_rate": 2.4425510753144697e-05,
1443
+ "loss": 0.0255,
1444
+ "step": 5125
1445
+ },
1446
+ {
1447
+ "epoch": 0.4310705616472755,
1448
+ "grad_norm": 16.040773391723633,
1449
+ "learning_rate": 2.429719487211922e-05,
1450
+ "loss": 0.0389,
1451
+ "step": 5150
1452
+ },
1453
+ {
1454
+ "epoch": 0.43316313718925253,
1455
+ "grad_norm": 0.01401565782725811,
1456
+ "learning_rate": 2.4168693276719184e-05,
1457
+ "loss": 0.0672,
1458
+ "step": 5175
1459
+ },
1460
+ {
1461
+ "epoch": 0.4352557127312296,
1462
+ "grad_norm": 3.7595765590667725,
1463
+ "learning_rate": 2.404001152047291e-05,
1464
+ "loss": 0.0497,
1465
+ "step": 5200
1466
+ },
1467
+ {
1468
+ "epoch": 0.43734828827320665,
1469
+ "grad_norm": 0.5674498081207275,
1470
+ "learning_rate": 2.391115516469483e-05,
1471
+ "loss": 0.025,
1472
+ "step": 5225
1473
+ },
1474
+ {
1475
+ "epoch": 0.4394408638151837,
1476
+ "grad_norm": 3.231126308441162,
1477
+ "learning_rate": 2.3782129778245158e-05,
1478
+ "loss": 0.0363,
1479
+ "step": 5250
1480
+ },
1481
+ {
1482
+ "epoch": 0.4415334393571608,
1483
+ "grad_norm": 0.01875438168644905,
1484
+ "learning_rate": 2.365294093728919e-05,
1485
+ "loss": 0.0503,
1486
+ "step": 5275
1487
+ },
1488
+ {
1489
+ "epoch": 0.44362601489913783,
1490
+ "grad_norm": 0.7183122634887695,
1491
+ "learning_rate": 2.3523594225056342e-05,
1492
+ "loss": 0.0562,
1493
+ "step": 5300
1494
+ },
1495
+ {
1496
+ "epoch": 0.44571859044111495,
1497
+ "grad_norm": 25.533050537109375,
1498
+ "learning_rate": 2.3394095231598835e-05,
1499
+ "loss": 0.0704,
1500
+ "step": 5325
1501
+ },
1502
+ {
1503
+ "epoch": 0.447811165983092,
1504
+ "grad_norm": 0.07413307577371597,
1505
+ "learning_rate": 2.326444955355013e-05,
1506
+ "loss": 0.0144,
1507
+ "step": 5350
1508
+ },
1509
+ {
1510
+ "epoch": 0.44990374152506907,
1511
+ "grad_norm": 0.1055075004696846,
1512
+ "learning_rate": 2.3134662793883032e-05,
1513
+ "loss": 0.1008,
1514
+ "step": 5375
1515
+ },
1516
+ {
1517
+ "epoch": 0.4519963170670461,
1518
+ "grad_norm": 0.017579292878508568,
1519
+ "learning_rate": 2.3004740561667554e-05,
1520
+ "loss": 0.0628,
1521
+ "step": 5400
1522
+ },
1523
+ {
1524
+ "epoch": 0.4540888926090232,
1525
+ "grad_norm": 8.102964401245117,
1526
+ "learning_rate": 2.28746884718285e-05,
1527
+ "loss": 0.0642,
1528
+ "step": 5425
1529
+ },
1530
+ {
1531
+ "epoch": 0.45618146815100025,
1532
+ "grad_norm": 10.838987350463867,
1533
+ "learning_rate": 2.2744512144902816e-05,
1534
+ "loss": 0.0271,
1535
+ "step": 5450
1536
+ },
1537
+ {
1538
+ "epoch": 0.4582740436929773,
1539
+ "grad_norm": 13.289299964904785,
1540
+ "learning_rate": 2.2614217206796668e-05,
1541
+ "loss": 0.0474,
1542
+ "step": 5475
1543
+ },
1544
+ {
1545
+ "epoch": 0.46036661923495437,
1546
+ "grad_norm": 49.52151107788086,
1547
+ "learning_rate": 2.2483809288542317e-05,
1548
+ "loss": 0.0635,
1549
+ "step": 5500
1550
+ },
1551
+ {
1552
+ "epoch": 0.4624591947769314,
1553
+ "grad_norm": 0.045649878680706024,
1554
+ "learning_rate": 2.2353294026054745e-05,
1555
+ "loss": 0.0723,
1556
+ "step": 5525
1557
+ },
1558
+ {
1559
+ "epoch": 0.4645517703189085,
1560
+ "grad_norm": 19.275131225585938,
1561
+ "learning_rate": 2.22226770598881e-05,
1562
+ "loss": 0.0483,
1563
+ "step": 5550
1564
+ },
1565
+ {
1566
+ "epoch": 0.4666443458608856,
1567
+ "grad_norm": 0.4754731357097626,
1568
+ "learning_rate": 2.2091964034991903e-05,
1569
+ "loss": 0.0602,
1570
+ "step": 5575
1571
+ },
1572
+ {
1573
+ "epoch": 0.46873692140286266,
1574
+ "grad_norm": 0.023405829444527626,
1575
+ "learning_rate": 2.196116060046712e-05,
1576
+ "loss": 0.0669,
1577
+ "step": 5600
1578
+ },
1579
+ {
1580
+ "epoch": 0.4708294969448397,
1581
+ "grad_norm": 0.06915944814682007,
1582
+ "learning_rate": 2.1830272409321997e-05,
1583
+ "loss": 0.0604,
1584
+ "step": 5625
1585
+ },
1586
+ {
1587
+ "epoch": 0.4729220724868168,
1588
+ "grad_norm": 0.10562706738710403,
1589
+ "learning_rate": 2.1699305118227745e-05,
1590
+ "loss": 0.0737,
1591
+ "step": 5650
1592
+ },
1593
+ {
1594
+ "epoch": 0.47501464802879384,
1595
+ "grad_norm": 0.24281857907772064,
1596
+ "learning_rate": 2.1568264387274095e-05,
1597
+ "loss": 0.0368,
1598
+ "step": 5675
1599
+ },
1600
+ {
1601
+ "epoch": 0.4771072235707709,
1602
+ "grad_norm": 5.6087517738342285,
1603
+ "learning_rate": 2.1437155879724664e-05,
1604
+ "loss": 0.111,
1605
+ "step": 5700
1606
+ },
1607
+ {
1608
+ "epoch": 0.47919979911274796,
1609
+ "grad_norm": 7.008552551269531,
1610
+ "learning_rate": 2.130598526177222e-05,
1611
+ "loss": 0.0397,
1612
+ "step": 5725
1613
+ },
1614
+ {
1615
+ "epoch": 0.481292374654725,
1616
+ "grad_norm": 0.5012567043304443,
1617
+ "learning_rate": 2.1174758202293783e-05,
1618
+ "loss": 0.0257,
1619
+ "step": 5750
1620
+ },
1621
+ {
1622
+ "epoch": 0.4833849501967021,
1623
+ "grad_norm": 0.17054887115955353,
1624
+ "learning_rate": 2.104348037260564e-05,
1625
+ "loss": 0.0635,
1626
+ "step": 5775
1627
+ },
1628
+ {
1629
+ "epoch": 0.48547752573867914,
1630
+ "grad_norm": 18.033811569213867,
1631
+ "learning_rate": 2.091215744621825e-05,
1632
+ "loss": 0.075,
1633
+ "step": 5800
1634
+ },
1635
+ {
1636
+ "epoch": 0.48757010128065625,
1637
+ "grad_norm": 6.623977184295654,
1638
+ "learning_rate": 2.078079509859103e-05,
1639
+ "loss": 0.0448,
1640
+ "step": 5825
1641
+ },
1642
+ {
1643
+ "epoch": 0.4896626768226333,
1644
+ "grad_norm": 12.964604377746582,
1645
+ "learning_rate": 2.0649399006887106e-05,
1646
+ "loss": 0.0385,
1647
+ "step": 5850
1648
+ },
1649
+ {
1650
+ "epoch": 0.4917552523646104,
1651
+ "grad_norm": 19.55506706237793,
1652
+ "learning_rate": 2.051797484972793e-05,
1653
+ "loss": 0.0614,
1654
+ "step": 5875
1655
+ },
1656
+ {
1657
+ "epoch": 0.49384782790658743,
1658
+ "grad_norm": 0.6688134670257568,
1659
+ "learning_rate": 2.0386528306947873e-05,
1660
+ "loss": 0.0567,
1661
+ "step": 5900
1662
+ },
1663
+ {
1664
+ "epoch": 0.4959404034485645,
1665
+ "grad_norm": 0.5797390341758728,
1666
+ "learning_rate": 2.025506505934876e-05,
1667
+ "loss": 0.0378,
1668
+ "step": 5925
1669
+ },
1670
+ {
1671
+ "epoch": 0.49803297899054155,
1672
+ "grad_norm": 5.049008369445801,
1673
+ "learning_rate": 2.0123590788454365e-05,
1674
+ "loss": 0.0649,
1675
+ "step": 5950
1676
+ },
1677
+ {
1678
+ "epoch": 0.5001255545325186,
1679
+ "grad_norm": 0.4083123803138733,
1680
+ "learning_rate": 1.9992111176264857e-05,
1681
+ "loss": 0.0221,
1682
+ "step": 5975
1683
+ },
1684
+ {
1685
+ "epoch": 0.5022181300744957,
1686
+ "grad_norm": 0.05658812075853348,
1687
+ "learning_rate": 1.986063190501124e-05,
1688
+ "loss": 0.0804,
1689
+ "step": 6000
1690
+ },
1691
+ {
1692
+ "epoch": 0.5043107056164727,
1693
+ "grad_norm": 0.026742545887827873,
1694
+ "learning_rate": 1.9729158656909784e-05,
1695
+ "loss": 0.042,
1696
+ "step": 6025
1697
+ },
1698
+ {
1699
+ "epoch": 0.5064032811584498,
1700
+ "grad_norm": 1.0423567295074463,
1701
+ "learning_rate": 1.9597697113916464e-05,
1702
+ "loss": 0.0527,
1703
+ "step": 6050
1704
+ },
1705
+ {
1706
+ "epoch": 0.5084958567004269,
1707
+ "grad_norm": 8.540645599365234,
1708
+ "learning_rate": 1.946625295748137e-05,
1709
+ "loss": 0.0589,
1710
+ "step": 6075
1711
+ },
1712
+ {
1713
+ "epoch": 0.5105884322424039,
1714
+ "grad_norm": 0.3264181315898895,
1715
+ "learning_rate": 1.9334831868303206e-05,
1716
+ "loss": 0.0438,
1717
+ "step": 6100
1718
+ },
1719
+ {
1720
+ "epoch": 0.512681007784381,
1721
+ "grad_norm": 0.10669530928134918,
1722
+ "learning_rate": 1.9203439526083747e-05,
1723
+ "loss": 0.0269,
1724
+ "step": 6125
1725
+ },
1726
+ {
1727
+ "epoch": 0.514773583326358,
1728
+ "grad_norm": 0.03823905065655708,
1729
+ "learning_rate": 1.9072081609282408e-05,
1730
+ "loss": 0.0704,
1731
+ "step": 6150
1732
+ },
1733
+ {
1734
+ "epoch": 0.5168661588683352,
1735
+ "grad_norm": 0.04673992842435837,
1736
+ "learning_rate": 1.894076379487081e-05,
1737
+ "loss": 0.0345,
1738
+ "step": 6175
1739
+ },
1740
+ {
1741
+ "epoch": 0.5189587344103123,
1742
+ "grad_norm": 0.10730260610580444,
1743
+ "learning_rate": 1.8809491758087463e-05,
1744
+ "loss": 0.0485,
1745
+ "step": 6200
1746
+ },
1747
+ {
1748
+ "epoch": 0.5210513099522893,
1749
+ "grad_norm": 0.4243800640106201,
1750
+ "learning_rate": 1.8678271172192458e-05,
1751
+ "loss": 0.0482,
1752
+ "step": 6225
1753
+ },
1754
+ {
1755
+ "epoch": 0.5231438854942664,
1756
+ "grad_norm": 0.14486907422542572,
1757
+ "learning_rate": 1.854710770822231e-05,
1758
+ "loss": 0.0314,
1759
+ "step": 6250
1760
+ },
1761
+ {
1762
+ "epoch": 0.5252364610362434,
1763
+ "grad_norm": 0.09311558306217194,
1764
+ "learning_rate": 1.841600703474487e-05,
1765
+ "loss": 0.0412,
1766
+ "step": 6275
1767
+ },
1768
+ {
1769
+ "epoch": 0.5273290365782205,
1770
+ "grad_norm": 0.009483098983764648,
1771
+ "learning_rate": 1.828497481761432e-05,
1772
+ "loss": 0.0321,
1773
+ "step": 6300
1774
+ },
1775
+ {
1776
+ "epoch": 0.5294216121201976,
1777
+ "grad_norm": 0.03347919136285782,
1778
+ "learning_rate": 1.8154016719726345e-05,
1779
+ "loss": 0.0594,
1780
+ "step": 6325
1781
+ },
1782
+ {
1783
+ "epoch": 0.5315141876621746,
1784
+ "grad_norm": 7.508219242095947,
1785
+ "learning_rate": 1.8023138400773362e-05,
1786
+ "loss": 0.057,
1787
+ "step": 6350
1788
+ },
1789
+ {
1790
+ "epoch": 0.5336067632041517,
1791
+ "grad_norm": 0.15084685385227203,
1792
+ "learning_rate": 1.7892345516999934e-05,
1793
+ "loss": 0.05,
1794
+ "step": 6375
1795
+ },
1796
+ {
1797
+ "epoch": 0.5356993387461287,
1798
+ "grad_norm": 0.13771960139274597,
1799
+ "learning_rate": 1.7761643720958324e-05,
1800
+ "loss": 0.029,
1801
+ "step": 6400
1802
+ },
1803
+ {
1804
+ "epoch": 0.5377919142881058,
1805
+ "grad_norm": 0.024205004796385765,
1806
+ "learning_rate": 1.7631038661264214e-05,
1807
+ "loss": 0.0555,
1808
+ "step": 6425
1809
+ },
1810
+ {
1811
+ "epoch": 0.5398844898300829,
1812
+ "grad_norm": 0.03557073697447777,
1813
+ "learning_rate": 1.750053598235256e-05,
1814
+ "loss": 0.0293,
1815
+ "step": 6450
1816
+ },
1817
+ {
1818
+ "epoch": 0.5419770653720599,
1819
+ "grad_norm": 0.07501186430454254,
1820
+ "learning_rate": 1.7370141324233672e-05,
1821
+ "loss": 0.0283,
1822
+ "step": 6475
1823
+ },
1824
+ {
1825
+ "epoch": 0.544069640914037,
1826
+ "grad_norm": 0.033643439412117004,
1827
+ "learning_rate": 1.723986032224947e-05,
1828
+ "loss": 0.0087,
1829
+ "step": 6500
1830
+ },
1831
+ {
1832
+ "epoch": 0.546162216456014,
1833
+ "grad_norm": 0.08923541009426117,
1834
+ "learning_rate": 1.7109698606829926e-05,
1835
+ "loss": 0.0417,
1836
+ "step": 6525
1837
+ },
1838
+ {
1839
+ "epoch": 0.5482547919979911,
1840
+ "grad_norm": 0.1683102399110794,
1841
+ "learning_rate": 1.6979661803249726e-05,
1842
+ "loss": 0.0081,
1843
+ "step": 6550
1844
+ },
1845
+ {
1846
+ "epoch": 0.5503473675399682,
1847
+ "grad_norm": 0.03645162284374237,
1848
+ "learning_rate": 1.6849755531385184e-05,
1849
+ "loss": 0.0286,
1850
+ "step": 6575
1851
+ },
1852
+ {
1853
+ "epoch": 0.5524399430819452,
1854
+ "grad_norm": 6.731907844543457,
1855
+ "learning_rate": 1.671998540547134e-05,
1856
+ "loss": 0.0511,
1857
+ "step": 6600
1858
+ },
1859
+ {
1860
+ "epoch": 0.5545325186239223,
1861
+ "grad_norm": 0.34839117527008057,
1862
+ "learning_rate": 1.6590357033859334e-05,
1863
+ "loss": 0.0529,
1864
+ "step": 6625
1865
+ },
1866
+ {
1867
+ "epoch": 0.5566250941658993,
1868
+ "grad_norm": 0.04825979471206665,
1869
+ "learning_rate": 1.6460876018774033e-05,
1870
+ "loss": 0.0478,
1871
+ "step": 6650
1872
+ },
1873
+ {
1874
+ "epoch": 0.5587176697078765,
1875
+ "grad_norm": 0.15074151754379272,
1876
+ "learning_rate": 1.63315479560719e-05,
1877
+ "loss": 0.0403,
1878
+ "step": 6675
1879
+ },
1880
+ {
1881
+ "epoch": 0.5608102452498536,
1882
+ "grad_norm": 0.8657410144805908,
1883
+ "learning_rate": 1.6202378434999194e-05,
1884
+ "loss": 0.0297,
1885
+ "step": 6700
1886
+ },
1887
+ {
1888
+ "epoch": 0.5629028207918306,
1889
+ "grad_norm": 25.035709381103516,
1890
+ "learning_rate": 1.607337303795035e-05,
1891
+ "loss": 0.0258,
1892
+ "step": 6725
1893
+ },
1894
+ {
1895
+ "epoch": 0.5649953963338077,
1896
+ "grad_norm": 5.161378860473633,
1897
+ "learning_rate": 1.594453734022678e-05,
1898
+ "loss": 0.0357,
1899
+ "step": 6750
1900
+ },
1901
+ {
1902
+ "epoch": 0.5670879718757847,
1903
+ "grad_norm": 15.704187393188477,
1904
+ "learning_rate": 1.5815876909795907e-05,
1905
+ "loss": 0.0589,
1906
+ "step": 6775
1907
+ },
1908
+ {
1909
+ "epoch": 0.5691805474177618,
1910
+ "grad_norm": 0.023194925859570503,
1911
+ "learning_rate": 1.568739730705051e-05,
1912
+ "loss": 0.0335,
1913
+ "step": 6800
1914
+ },
1915
+ {
1916
+ "epoch": 0.5712731229597389,
1917
+ "grad_norm": 54.290287017822266,
1918
+ "learning_rate": 1.5559104084568443e-05,
1919
+ "loss": 0.0282,
1920
+ "step": 6825
1921
+ },
1922
+ {
1923
+ "epoch": 0.5733656985017159,
1924
+ "grad_norm": 0.055701714009046555,
1925
+ "learning_rate": 1.543100278687265e-05,
1926
+ "loss": 0.0264,
1927
+ "step": 6850
1928
+ },
1929
+ {
1930
+ "epoch": 0.575458274043693,
1931
+ "grad_norm": 5.854345321655273,
1932
+ "learning_rate": 1.5303098950191556e-05,
1933
+ "loss": 0.0544,
1934
+ "step": 6875
1935
+ },
1936
+ {
1937
+ "epoch": 0.57755084958567,
1938
+ "grad_norm": 0.003441957989707589,
1939
+ "learning_rate": 1.5175398102219787e-05,
1940
+ "loss": 0.0356,
1941
+ "step": 6900
1942
+ },
1943
+ {
1944
+ "epoch": 0.5796434251276471,
1945
+ "grad_norm": 0.08189363032579422,
1946
+ "learning_rate": 1.5047905761879303e-05,
1947
+ "loss": 0.0494,
1948
+ "step": 6925
1949
+ },
1950
+ {
1951
+ "epoch": 0.5817360006696242,
1952
+ "grad_norm": 7.302794456481934,
1953
+ "learning_rate": 1.4920627439080858e-05,
1954
+ "loss": 0.0255,
1955
+ "step": 6950
1956
+ },
1957
+ {
1958
+ "epoch": 0.5838285762116012,
1959
+ "grad_norm": 14.938457489013672,
1960
+ "learning_rate": 1.4793568634485885e-05,
1961
+ "loss": 0.0662,
1962
+ "step": 6975
1963
+ },
1964
+ {
1965
+ "epoch": 0.5859211517535783,
1966
+ "grad_norm": 0.264533668756485,
1967
+ "learning_rate": 1.4666734839268773e-05,
1968
+ "loss": 0.0719,
1969
+ "step": 7000
1970
+ },
1971
+ {
1972
+ "epoch": 0.5880137272955553,
1973
+ "grad_norm": 1.2785422801971436,
1974
+ "learning_rate": 1.454013153487955e-05,
1975
+ "loss": 0.0459,
1976
+ "step": 7025
1977
+ },
1978
+ {
1979
+ "epoch": 0.5901063028375324,
1980
+ "grad_norm": 0.13779093325138092,
1981
+ "learning_rate": 1.4413764192806993e-05,
1982
+ "loss": 0.0313,
1983
+ "step": 7050
1984
+ },
1985
+ {
1986
+ "epoch": 0.5921988783795095,
1987
+ "grad_norm": 1.123705506324768,
1988
+ "learning_rate": 1.4287638274342148e-05,
1989
+ "loss": 0.0703,
1990
+ "step": 7075
1991
+ },
1992
+ {
1993
+ "epoch": 0.5942914539214865,
1994
+ "grad_norm": 3.7024033069610596,
1995
+ "learning_rate": 1.4161759230342321e-05,
1996
+ "loss": 0.0347,
1997
+ "step": 7100
1998
+ },
1999
+ {
2000
+ "epoch": 0.5963840294634636,
2001
+ "grad_norm": 4.496721267700195,
2002
+ "learning_rate": 1.40361325009955e-05,
2003
+ "loss": 0.0591,
2004
+ "step": 7125
2005
+ },
2006
+ {
2007
+ "epoch": 0.5984766050054406,
2008
+ "grad_norm": 10.648604393005371,
2009
+ "learning_rate": 1.3910763515585244e-05,
2010
+ "loss": 0.0486,
2011
+ "step": 7150
2012
+ },
2013
+ {
2014
+ "epoch": 0.6005691805474178,
2015
+ "grad_norm": 5.158627033233643,
2016
+ "learning_rate": 1.3785657692256033e-05,
2017
+ "loss": 0.0699,
2018
+ "step": 7175
2019
+ },
2020
+ {
2021
+ "epoch": 0.6026617560893949,
2022
+ "grad_norm": 0.5422956347465515,
2023
+ "learning_rate": 1.3660820437779148e-05,
2024
+ "loss": 0.0637,
2025
+ "step": 7200
2026
+ },
2027
+ {
2028
+ "epoch": 0.6047543316313719,
2029
+ "grad_norm": 0.2532649040222168,
2030
+ "learning_rate": 1.3536257147318938e-05,
2031
+ "loss": 0.0426,
2032
+ "step": 7225
2033
+ },
2034
+ {
2035
+ "epoch": 0.606846907173349,
2036
+ "grad_norm": 5.536031246185303,
2037
+ "learning_rate": 1.341197320419971e-05,
2038
+ "loss": 0.0545,
2039
+ "step": 7250
2040
+ },
2041
+ {
2042
+ "epoch": 0.608939482715326,
2043
+ "grad_norm": 0.14337138831615448,
2044
+ "learning_rate": 1.3287973979673043e-05,
2045
+ "loss": 0.0293,
2046
+ "step": 7275
2047
+ },
2048
+ {
2049
+ "epoch": 0.6110320582573031,
2050
+ "grad_norm": 0.029332490637898445,
2051
+ "learning_rate": 1.3164264832685673e-05,
2052
+ "loss": 0.0628,
2053
+ "step": 7300
2054
+ },
2055
+ {
2056
+ "epoch": 0.6131246337992802,
2057
+ "grad_norm": 0.3548976480960846,
2058
+ "learning_rate": 1.3040851109647892e-05,
2059
+ "loss": 0.0532,
2060
+ "step": 7325
2061
+ },
2062
+ {
2063
+ "epoch": 0.6152172093412572,
2064
+ "grad_norm": 0.19546709954738617,
2065
+ "learning_rate": 1.2917738144202466e-05,
2066
+ "loss": 0.0821,
2067
+ "step": 7350
2068
+ },
2069
+ {
2070
+ "epoch": 0.6173097848832343,
2071
+ "grad_norm": 1.9515585899353027,
2072
+ "learning_rate": 1.2794931256994155e-05,
2073
+ "loss": 0.05,
2074
+ "step": 7375
2075
+ },
2076
+ {
2077
+ "epoch": 0.6194023604252114,
2078
+ "grad_norm": 0.4851491451263428,
2079
+ "learning_rate": 1.2672435755439761e-05,
2080
+ "loss": 0.0314,
2081
+ "step": 7400
2082
+ },
2083
+ {
2084
+ "epoch": 0.6214949359671884,
2085
+ "grad_norm": 1.249005913734436,
2086
+ "learning_rate": 1.2550256933498745e-05,
2087
+ "loss": 0.0552,
2088
+ "step": 7425
2089
+ },
2090
+ {
2091
+ "epoch": 0.6235875115091655,
2092
+ "grad_norm": 7.1881890296936035,
2093
+ "learning_rate": 1.2428400071444447e-05,
2094
+ "loss": 0.036,
2095
+ "step": 7450
2096
+ },
2097
+ {
2098
+ "epoch": 0.6256800870511425,
2099
+ "grad_norm": 0.07499822229146957,
2100
+ "learning_rate": 1.2306870435635867e-05,
2101
+ "loss": 0.0338,
2102
+ "step": 7475
2103
+ },
2104
+ {
2105
+ "epoch": 0.6277726625931196,
2106
+ "grad_norm": 0.27988529205322266,
2107
+ "learning_rate": 1.2185673278290093e-05,
2108
+ "loss": 0.0243,
2109
+ "step": 7500
2110
+ },
2111
+ {
2112
+ "epoch": 0.6298652381350967,
2113
+ "grad_norm": 21.607757568359375,
2114
+ "learning_rate": 1.2064813837255284e-05,
2115
+ "loss": 0.0693,
2116
+ "step": 7525
2117
+ },
2118
+ {
2119
+ "epoch": 0.6319578136770737,
2120
+ "grad_norm": 0.4316962957382202,
2121
+ "learning_rate": 1.1944297335784327e-05,
2122
+ "loss": 0.0602,
2123
+ "step": 7550
2124
+ },
2125
+ {
2126
+ "epoch": 0.6340503892190508,
2127
+ "grad_norm": 5.633823871612549,
2128
+ "learning_rate": 1.1824128982309085e-05,
2129
+ "loss": 0.0552,
2130
+ "step": 7575
2131
+ },
2132
+ {
2133
+ "epoch": 0.6361429647610278,
2134
+ "grad_norm": 0.07708461582660675,
2135
+ "learning_rate": 1.1704313970215302e-05,
2136
+ "loss": 0.0323,
2137
+ "step": 7600
2138
+ },
2139
+ {
2140
+ "epoch": 0.6382355403030049,
2141
+ "grad_norm": 0.06260835379362106,
2142
+ "learning_rate": 1.1584857477618168e-05,
2143
+ "loss": 0.0312,
2144
+ "step": 7625
2145
+ },
2146
+ {
2147
+ "epoch": 0.640328115844982,
2148
+ "grad_norm": 0.11298274248838425,
2149
+ "learning_rate": 1.146576466713852e-05,
2150
+ "loss": 0.0551,
2151
+ "step": 7650
2152
+ },
2153
+ {
2154
+ "epoch": 0.6424206913869591,
2155
+ "grad_norm": 5.66866397857666,
2156
+ "learning_rate": 1.1347040685679747e-05,
2157
+ "loss": 0.0382,
2158
+ "step": 7675
2159
+ },
2160
+ {
2161
+ "epoch": 0.6445132669289362,
2162
+ "grad_norm": 1.05096435546875,
2163
+ "learning_rate": 1.1228690664205335e-05,
2164
+ "loss": 0.0213,
2165
+ "step": 7700
2166
+ },
2167
+ {
2168
+ "epoch": 0.6466058424709132,
2169
+ "grad_norm": 0.7647340893745422,
2170
+ "learning_rate": 1.1110719717517102e-05,
2171
+ "loss": 0.0467,
2172
+ "step": 7725
2173
+ },
2174
+ {
2175
+ "epoch": 0.6486984180128903,
2176
+ "grad_norm": 15.440308570861816,
2177
+ "learning_rate": 1.0993132944034206e-05,
2178
+ "loss": 0.0436,
2179
+ "step": 7750
2180
+ },
2181
+ {
2182
+ "epoch": 0.6507909935548674,
2183
+ "grad_norm": 16.530546188354492,
2184
+ "learning_rate": 1.087593542557273e-05,
2185
+ "loss": 0.0268,
2186
+ "step": 7775
2187
+ },
2188
+ {
2189
+ "epoch": 0.6528835690968444,
2190
+ "grad_norm": 0.1774834841489792,
2191
+ "learning_rate": 1.0759132227126133e-05,
2192
+ "loss": 0.0293,
2193
+ "step": 7800
2194
+ },
2195
+ {
2196
+ "epoch": 0.6549761446388215,
2197
+ "grad_norm": 0.04517792537808418,
2198
+ "learning_rate": 1.064272839664628e-05,
2199
+ "loss": 0.0616,
2200
+ "step": 7825
2201
+ },
2202
+ {
2203
+ "epoch": 0.6570687201807985,
2204
+ "grad_norm": 1.20844566822052,
2205
+ "learning_rate": 1.0526728964825357e-05,
2206
+ "loss": 0.0482,
2207
+ "step": 7850
2208
+ },
2209
+ {
2210
+ "epoch": 0.6591612957227756,
2211
+ "grad_norm": 0.033426906913518906,
2212
+ "learning_rate": 1.0411138944878371e-05,
2213
+ "loss": 0.0274,
2214
+ "step": 7875
2215
+ },
2216
+ {
2217
+ "epoch": 0.6612538712647527,
2218
+ "grad_norm": 0.012136283330619335,
2219
+ "learning_rate": 1.0295963332326587e-05,
2220
+ "loss": 0.0725,
2221
+ "step": 7900
2222
+ },
2223
+ {
2224
+ "epoch": 0.6633464468067297,
2225
+ "grad_norm": 0.18010033667087555,
2226
+ "learning_rate": 1.0181207104781539e-05,
2227
+ "loss": 0.0485,
2228
+ "step": 7925
2229
+ },
2230
+ {
2231
+ "epoch": 0.6654390223487068,
2232
+ "grad_norm": 0.27462857961654663,
2233
+ "learning_rate": 1.0066875221729995e-05,
2234
+ "loss": 0.0314,
2235
+ "step": 7950
2236
+ },
2237
+ {
2238
+ "epoch": 0.6675315978906838,
2239
+ "grad_norm": 7.735607624053955,
2240
+ "learning_rate": 9.952972624319542e-06,
2241
+ "loss": 0.0468,
2242
+ "step": 7975
2243
+ },
2244
+ {
2245
+ "epoch": 0.6696241734326609,
2246
+ "grad_norm": 0.2843437194824219,
2247
+ "learning_rate": 9.839504235145117e-06,
2248
+ "loss": 0.0419,
2249
+ "step": 8000
2250
+ },
2251
+ {
2252
+ "epoch": 0.671716748974638,
2253
+ "grad_norm": 0.6654235124588013,
2254
+ "learning_rate": 9.726474958036188e-06,
2255
+ "loss": 0.0894,
2256
+ "step": 8025
2257
+ },
2258
+ {
2259
+ "epoch": 0.673809324516615,
2260
+ "grad_norm": 6.332837104797363,
2261
+ "learning_rate": 9.613889677844899e-06,
2262
+ "loss": 0.0371,
2263
+ "step": 8050
2264
+ },
2265
+ {
2266
+ "epoch": 0.6759019000585921,
2267
+ "grad_norm": 0.2247881293296814,
2268
+ "learning_rate": 9.501753260234885e-06,
2269
+ "loss": 0.0333,
2270
+ "step": 8075
2271
+ },
2272
+ {
2273
+ "epoch": 0.6779944756005691,
2274
+ "grad_norm": 46.26710891723633,
2275
+ "learning_rate": 9.39007055147106e-06,
2276
+ "loss": 0.0916,
2277
+ "step": 8100
2278
+ },
2279
+ {
2280
+ "epoch": 0.6800870511425462,
2281
+ "grad_norm": 10.392670631408691,
2282
+ "learning_rate": 9.278846378210105e-06,
2283
+ "loss": 0.0376,
2284
+ "step": 8125
2285
+ },
2286
+ {
2287
+ "epoch": 0.6821796266845234,
2288
+ "grad_norm": 0.12523211538791656,
2289
+ "learning_rate": 9.168085547291925e-06,
2290
+ "loss": 0.0604,
2291
+ "step": 8150
2292
+ },
2293
+ {
2294
+ "epoch": 0.6842722022265004,
2295
+ "grad_norm": 0.7956030964851379,
2296
+ "learning_rate": 9.057792845531896e-06,
2297
+ "loss": 0.0348,
2298
+ "step": 8175
2299
+ },
2300
+ {
2301
+ "epoch": 0.6863647777684775,
2302
+ "grad_norm": 0.120228111743927,
2303
+ "learning_rate": 8.947973039513956e-06,
2304
+ "loss": 0.0626,
2305
+ "step": 8200
2306
+ },
2307
+ {
2308
+ "epoch": 0.6884573533104545,
2309
+ "grad_norm": 0.10342678427696228,
2310
+ "learning_rate": 8.838630875384664e-06,
2311
+ "loss": 0.0423,
2312
+ "step": 8225
2313
+ },
2314
+ {
2315
+ "epoch": 0.6905499288524316,
2316
+ "grad_norm": 0.045878130942583084,
2317
+ "learning_rate": 8.72977107864802e-06,
2318
+ "loss": 0.0484,
2319
+ "step": 8250
2320
+ },
2321
+ {
2322
+ "epoch": 0.6926425043944087,
2323
+ "grad_norm": 0.035083040595054626,
2324
+ "learning_rate": 8.6213983539613e-06,
2325
+ "loss": 0.0255,
2326
+ "step": 8275
2327
+ },
2328
+ {
2329
+ "epoch": 0.6947350799363857,
2330
+ "grad_norm": 6.96301794052124,
2331
+ "learning_rate": 8.51351738493168e-06,
2332
+ "loss": 0.0561,
2333
+ "step": 8300
2334
+ },
2335
+ {
2336
+ "epoch": 0.6968276554783628,
2337
+ "grad_norm": 0.13092511892318726,
2338
+ "learning_rate": 8.406132833913869e-06,
2339
+ "loss": 0.0622,
2340
+ "step": 8325
2341
+ },
2342
+ {
2343
+ "epoch": 0.6989202310203398,
2344
+ "grad_norm": 0.3119213879108429,
2345
+ "learning_rate": 8.29924934180856e-06,
2346
+ "loss": 0.0426,
2347
+ "step": 8350
2348
+ },
2349
+ {
2350
+ "epoch": 0.7010128065623169,
2351
+ "grad_norm": 0.057294104248285294,
2352
+ "learning_rate": 8.19287152786192e-06,
2353
+ "loss": 0.0652,
2354
+ "step": 8375
2355
+ },
2356
+ {
2357
+ "epoch": 0.703105382104294,
2358
+ "grad_norm": 0.6926544308662415,
2359
+ "learning_rate": 8.087003989465905e-06,
2360
+ "loss": 0.0447,
2361
+ "step": 8400
2362
+ },
2363
+ {
2364
+ "epoch": 0.705197957646271,
2365
+ "grad_norm": 0.12673336267471313,
2366
+ "learning_rate": 7.981651301959618e-06,
2367
+ "loss": 0.0633,
2368
+ "step": 8425
2369
+ },
2370
+ {
2371
+ "epoch": 0.7072905331882481,
2372
+ "grad_norm": 0.28860193490982056,
2373
+ "learning_rate": 7.876818018431526e-06,
2374
+ "loss": 0.0527,
2375
+ "step": 8450
2376
+ },
2377
+ {
2378
+ "epoch": 0.7093831087302251,
2379
+ "grad_norm": 7.478037357330322,
2380
+ "learning_rate": 7.77250866952274e-06,
2381
+ "loss": 0.0452,
2382
+ "step": 8475
2383
+ },
2384
+ {
2385
+ "epoch": 0.7114756842722022,
2386
+ "grad_norm": 16.230470657348633,
2387
+ "learning_rate": 7.668727763231152e-06,
2388
+ "loss": 0.03,
2389
+ "step": 8500
2390
+ },
2391
+ {
2392
+ "epoch": 0.7135682598141793,
2393
+ "grad_norm": 12.577932357788086,
2394
+ "learning_rate": 7.56547978471667e-06,
2395
+ "loss": 0.0294,
2396
+ "step": 8525
2397
+ },
2398
+ {
2399
+ "epoch": 0.7156608353561563,
2400
+ "grad_norm": 0.06792646646499634,
2401
+ "learning_rate": 7.462769196107325e-06,
2402
+ "loss": 0.0672,
2403
+ "step": 8550
2404
+ },
2405
+ {
2406
+ "epoch": 0.7177534108981334,
2407
+ "grad_norm": 0.8921266794204712,
2408
+ "learning_rate": 7.360600436306487e-06,
2409
+ "loss": 0.0218,
2410
+ "step": 8575
2411
+ },
2412
+ {
2413
+ "epoch": 0.7198459864401104,
2414
+ "grad_norm": 0.06442731618881226,
2415
+ "learning_rate": 7.25897792080096e-06,
2416
+ "loss": 0.0086,
2417
+ "step": 8600
2418
+ },
2419
+ {
2420
+ "epoch": 0.7219385619820875,
2421
+ "grad_norm": 0.8217195272445679,
2422
+ "learning_rate": 7.157906041470222e-06,
2423
+ "loss": 0.0563,
2424
+ "step": 8625
2425
+ },
2426
+ {
2427
+ "epoch": 0.7240311375240647,
2428
+ "grad_norm": 20.53885269165039,
2429
+ "learning_rate": 7.057389166396553e-06,
2430
+ "loss": 0.0598,
2431
+ "step": 8650
2432
+ },
2433
+ {
2434
+ "epoch": 0.7261237130660417,
2435
+ "grad_norm": 0.01888943649828434,
2436
+ "learning_rate": 6.957431639676335e-06,
2437
+ "loss": 0.028,
2438
+ "step": 8675
2439
+ },
2440
+ {
2441
+ "epoch": 0.7282162886080188,
2442
+ "grad_norm": 5.670931816101074,
2443
+ "learning_rate": 6.858037781232218e-06,
2444
+ "loss": 0.0184,
2445
+ "step": 8700
2446
+ },
2447
+ {
2448
+ "epoch": 0.7303088641499959,
2449
+ "grad_norm": 10.210001945495605,
2450
+ "learning_rate": 6.7592118866265136e-06,
2451
+ "loss": 0.065,
2452
+ "step": 8725
2453
+ },
2454
+ {
2455
+ "epoch": 0.7324014396919729,
2456
+ "grad_norm": 7.783575057983398,
2457
+ "learning_rate": 6.660958226875474e-06,
2458
+ "loss": 0.0598,
2459
+ "step": 8750
2460
+ },
2461
+ {
2462
+ "epoch": 0.73449401523395,
2463
+ "grad_norm": 12.363920211791992,
2464
+ "learning_rate": 6.563281048264772e-06,
2465
+ "loss": 0.0893,
2466
+ "step": 8775
2467
+ },
2468
+ {
2469
+ "epoch": 0.736586590775927,
2470
+ "grad_norm": 0.0548861064016819,
2471
+ "learning_rate": 6.466184572165941e-06,
2472
+ "loss": 0.0334,
2473
+ "step": 8800
2474
+ },
2475
+ {
2476
+ "epoch": 0.7386791663179041,
2477
+ "grad_norm": 20.708911895751953,
2478
+ "learning_rate": 6.369672994853957e-06,
2479
+ "loss": 0.0836,
2480
+ "step": 8825
2481
+ },
2482
+ {
2483
+ "epoch": 0.7407717418598811,
2484
+ "grad_norm": 0.15611207485198975,
2485
+ "learning_rate": 6.273750487325902e-06,
2486
+ "loss": 0.0737,
2487
+ "step": 8850
2488
+ },
2489
+ {
2490
+ "epoch": 0.7428643174018582,
2491
+ "grad_norm": 0.1677076667547226,
2492
+ "learning_rate": 6.178421195120661e-06,
2493
+ "loss": 0.0408,
2494
+ "step": 8875
2495
+ },
2496
+ {
2497
+ "epoch": 0.7449568929438353,
2498
+ "grad_norm": 0.5552803874015808,
2499
+ "learning_rate": 6.083689238139811e-06,
2500
+ "loss": 0.0557,
2501
+ "step": 8900
2502
+ },
2503
+ {
2504
+ "epoch": 0.7470494684858123,
2505
+ "grad_norm": 16.739295959472656,
2506
+ "learning_rate": 5.989558710469521e-06,
2507
+ "loss": 0.049,
2508
+ "step": 8925
2509
+ },
2510
+ {
2511
+ "epoch": 0.7491420440277894,
2512
+ "grad_norm": 9.993467330932617,
2513
+ "learning_rate": 5.896033680203663e-06,
2514
+ "loss": 0.0682,
2515
+ "step": 8950
2516
+ },
2517
+ {
2518
+ "epoch": 0.7512346195697664,
2519
+ "grad_norm": 3.3099191188812256,
2520
+ "learning_rate": 5.803118189267949e-06,
2521
+ "loss": 0.0681,
2522
+ "step": 8975
2523
+ },
2524
+ {
2525
+ "epoch": 0.7533271951117435,
2526
+ "grad_norm": 0.18636733293533325,
2527
+ "learning_rate": 5.710816253245299e-06,
2528
+ "loss": 0.0673,
2529
+ "step": 9000
2530
+ },
2531
+ {
2532
+ "epoch": 0.7554197706537206,
2533
+ "grad_norm": 7.4644670486450195,
2534
+ "learning_rate": 5.619131861202254e-06,
2535
+ "loss": 0.0821,
2536
+ "step": 9025
2537
+ },
2538
+ {
2539
+ "epoch": 0.7575123461956976,
2540
+ "grad_norm": 0.2912033200263977,
2541
+ "learning_rate": 5.528068975516616e-06,
2542
+ "loss": 0.0413,
2543
+ "step": 9050
2544
+ },
2545
+ {
2546
+ "epoch": 0.7596049217376747,
2547
+ "grad_norm": 0.06365963816642761,
2548
+ "learning_rate": 5.437631531706158e-06,
2549
+ "loss": 0.059,
2550
+ "step": 9075
2551
+ },
2552
+ {
2553
+ "epoch": 0.7616974972796517,
2554
+ "grad_norm": 0.11069577187299728,
2555
+ "learning_rate": 5.3478234382585995e-06,
2556
+ "loss": 0.0739,
2557
+ "step": 9100
2558
+ },
2559
+ {
2560
+ "epoch": 0.7637900728216288,
2561
+ "grad_norm": 10.11336612701416,
2562
+ "learning_rate": 5.258648576462628e-06,
2563
+ "loss": 0.0497,
2564
+ "step": 9125
2565
+ },
2566
+ {
2567
+ "epoch": 0.765882648363606,
2568
+ "grad_norm": 4.044532299041748,
2569
+ "learning_rate": 5.170110800240209e-06,
2570
+ "loss": 0.0353,
2571
+ "step": 9150
2572
+ },
2573
+ {
2574
+ "epoch": 0.767975223905583,
2575
+ "grad_norm": 1.726006031036377,
2576
+ "learning_rate": 5.082213935980005e-06,
2577
+ "loss": 0.0572,
2578
+ "step": 9175
2579
+ },
2580
+ {
2581
+ "epoch": 0.7700677994475601,
2582
+ "grad_norm": 11.652609825134277,
2583
+ "learning_rate": 4.994961782372016e-06,
2584
+ "loss": 0.0697,
2585
+ "step": 9200
2586
+ },
2587
+ {
2588
+ "epoch": 0.7721603749895372,
2589
+ "grad_norm": 6.057770252227783,
2590
+ "learning_rate": 4.908358110243394e-06,
2591
+ "loss": 0.0572,
2592
+ "step": 9225
2593
+ },
2594
+ {
2595
+ "epoch": 0.7742529505315142,
2596
+ "grad_norm": 0.0476180762052536,
2597
+ "learning_rate": 4.822406662395509e-06,
2598
+ "loss": 0.0812,
2599
+ "step": 9250
2600
+ },
2601
+ {
2602
+ "epoch": 0.7763455260734913,
2603
+ "grad_norm": 0.6360920071601868,
2604
+ "learning_rate": 4.737111153442146e-06,
2605
+ "loss": 0.0308,
2606
+ "step": 9275
2607
+ },
2608
+ {
2609
+ "epoch": 0.7784381016154683,
2610
+ "grad_norm": 0.23660625517368317,
2611
+ "learning_rate": 4.652475269649028e-06,
2612
+ "loss": 0.0298,
2613
+ "step": 9300
2614
+ },
2615
+ {
2616
+ "epoch": 0.7805306771574454,
2617
+ "grad_norm": 1.2915230989456177,
2618
+ "learning_rate": 4.568502668774446e-06,
2619
+ "loss": 0.0447,
2620
+ "step": 9325
2621
+ },
2622
+ {
2623
+ "epoch": 0.7826232526994225,
2624
+ "grad_norm": 0.3824012279510498,
2625
+ "learning_rate": 4.4851969799112304e-06,
2626
+ "loss": 0.0556,
2627
+ "step": 9350
2628
+ },
2629
+ {
2630
+ "epoch": 0.7847158282413995,
2631
+ "grad_norm": 1.5623775720596313,
2632
+ "learning_rate": 4.402561803329866e-06,
2633
+ "loss": 0.041,
2634
+ "step": 9375
2635
+ },
2636
+ {
2637
+ "epoch": 0.7868084037833766,
2638
+ "grad_norm": 4.520791053771973,
2639
+ "learning_rate": 4.320600710322933e-06,
2640
+ "loss": 0.0549,
2641
+ "step": 9400
2642
+ },
2643
+ {
2644
+ "epoch": 0.7889009793253536,
2645
+ "grad_norm": 0.22007611393928528,
2646
+ "learning_rate": 4.239317243050733e-06,
2647
+ "loss": 0.0376,
2648
+ "step": 9425
2649
+ },
2650
+ {
2651
+ "epoch": 0.7909935548673307,
2652
+ "grad_norm": 0.25925832986831665,
2653
+ "learning_rate": 4.158714914388246e-06,
2654
+ "loss": 0.0681,
2655
+ "step": 9450
2656
+ },
2657
+ {
2658
+ "epoch": 0.7930861304093078,
2659
+ "grad_norm": 0.07315188646316528,
2660
+ "learning_rate": 4.078797207773264e-06,
2661
+ "loss": 0.0424,
2662
+ "step": 9475
2663
+ },
2664
+ {
2665
+ "epoch": 0.7951787059512848,
2666
+ "grad_norm": 0.12042330950498581,
2667
+ "learning_rate": 3.999567577055874e-06,
2668
+ "loss": 0.0545,
2669
+ "step": 9500
2670
+ },
2671
+ {
2672
+ "epoch": 0.7972712814932619,
2673
+ "grad_norm": 0.18977761268615723,
2674
+ "learning_rate": 3.921029446349201e-06,
2675
+ "loss": 0.0643,
2676
+ "step": 9525
2677
+ },
2678
+ {
2679
+ "epoch": 0.7993638570352389,
2680
+ "grad_norm": 0.10226922482252121,
2681
+ "learning_rate": 3.843186209881386e-06,
2682
+ "loss": 0.0234,
2683
+ "step": 9550
2684
+ },
2685
+ {
2686
+ "epoch": 0.801456432577216,
2687
+ "grad_norm": 0.10693290829658508,
2688
+ "learning_rate": 3.766041231848949e-06,
2689
+ "loss": 0.0228,
2690
+ "step": 9575
2691
+ },
2692
+ {
2693
+ "epoch": 0.803549008119193,
2694
+ "grad_norm": 0.027522044256329536,
2695
+ "learning_rate": 3.6895978462713443e-06,
2696
+ "loss": 0.0549,
2697
+ "step": 9600
2698
+ },
2699
+ {
2700
+ "epoch": 0.8056415836611701,
2701
+ "grad_norm": 0.1407710462808609,
2702
+ "learning_rate": 3.613859356846916e-06,
2703
+ "loss": 0.0522,
2704
+ "step": 9625
2705
+ },
2706
+ {
2707
+ "epoch": 0.8077341592031473,
2708
+ "grad_norm": 0.6555646061897278,
2709
+ "learning_rate": 3.538829036810074e-06,
2710
+ "loss": 0.0292,
2711
+ "step": 9650
2712
+ },
2713
+ {
2714
+ "epoch": 0.8098267347451243,
2715
+ "grad_norm": 42.79521179199219,
2716
+ "learning_rate": 3.4645101287898975e-06,
2717
+ "loss": 0.0751,
2718
+ "step": 9675
2719
+ },
2720
+ {
2721
+ "epoch": 0.8119193102871014,
2722
+ "grad_norm": 0.2245454490184784,
2723
+ "learning_rate": 3.3909058446699205e-06,
2724
+ "loss": 0.0207,
2725
+ "step": 9700
2726
+ },
2727
+ {
2728
+ "epoch": 0.8140118858290785,
2729
+ "grad_norm": 0.31118181347846985,
2730
+ "learning_rate": 3.3180193654493854e-06,
2731
+ "loss": 0.0474,
2732
+ "step": 9725
2733
+ },
2734
+ {
2735
+ "epoch": 0.8161044613710555,
2736
+ "grad_norm": 0.10519175231456757,
2737
+ "learning_rate": 3.245853841105724e-06,
2738
+ "loss": 0.0384,
2739
+ "step": 9750
2740
+ },
2741
+ {
2742
+ "epoch": 0.8181970369130326,
2743
+ "grad_norm": 0.06231565773487091,
2744
+ "learning_rate": 3.174412390458457e-06,
2745
+ "loss": 0.0373,
2746
+ "step": 9775
2747
+ },
2748
+ {
2749
+ "epoch": 0.8202896124550096,
2750
+ "grad_norm": 1.0349115133285522,
2751
+ "learning_rate": 3.1036981010343737e-06,
2752
+ "loss": 0.0367,
2753
+ "step": 9800
2754
+ },
2755
+ {
2756
+ "epoch": 0.8223821879969867,
2757
+ "grad_norm": 0.1348392218351364,
2758
+ "learning_rate": 3.0337140289341337e-06,
2759
+ "loss": 0.0506,
2760
+ "step": 9825
2761
+ },
2762
+ {
2763
+ "epoch": 0.8244747635389638,
2764
+ "grad_norm": 0.06425004452466965,
2765
+ "learning_rate": 2.9644631987001425e-06,
2766
+ "loss": 0.0427,
2767
+ "step": 9850
2768
+ },
2769
+ {
2770
+ "epoch": 0.8265673390809408,
2771
+ "grad_norm": 4.405979633331299,
2772
+ "learning_rate": 2.895948603185894e-06,
2773
+ "loss": 0.0655,
2774
+ "step": 9875
2775
+ },
2776
+ {
2777
+ "epoch": 0.8286599146229179,
2778
+ "grad_norm": 1.3921059370040894,
2779
+ "learning_rate": 2.828173203426572e-06,
2780
+ "loss": 0.1037,
2781
+ "step": 9900
2782
+ },
2783
+ {
2784
+ "epoch": 0.8307524901648949,
2785
+ "grad_norm": 7.713419437408447,
2786
+ "learning_rate": 2.7611399285111252e-06,
2787
+ "loss": 0.0571,
2788
+ "step": 9925
2789
+ },
2790
+ {
2791
+ "epoch": 0.832845065706872,
2792
+ "grad_norm": 0.28889116644859314,
2793
+ "learning_rate": 2.6948516754556452e-06,
2794
+ "loss": 0.0992,
2795
+ "step": 9950
2796
+ },
2797
+ {
2798
+ "epoch": 0.8349376412488491,
2799
+ "grad_norm": 0.10869034379720688,
2800
+ "learning_rate": 2.6293113090781953e-06,
2801
+ "loss": 0.0404,
2802
+ "step": 9975
2803
+ },
2804
+ {
2805
+ "epoch": 0.8370302167908261,
2806
+ "grad_norm": 6.920375347137451,
2807
+ "learning_rate": 2.564521661874968e-06,
2808
+ "loss": 0.0712,
2809
+ "step": 10000
2810
+ },
2811
+ {
2812
+ "epoch": 0.8391227923328032,
2813
+ "grad_norm": 0.0789741575717926,
2814
+ "learning_rate": 2.5004855338979074e-06,
2815
+ "loss": 0.0727,
2816
+ "step": 10025
2817
+ },
2818
+ {
2819
+ "epoch": 0.8412153678747802,
2820
+ "grad_norm": 0.7709121108055115,
2821
+ "learning_rate": 2.437205692633653e-06,
2822
+ "loss": 0.0676,
2823
+ "step": 10050
2824
+ },
2825
+ {
2826
+ "epoch": 0.8433079434167573,
2827
+ "grad_norm": 0.08355044573545456,
2828
+ "learning_rate": 2.374684872883988e-06,
2829
+ "loss": 0.0442,
2830
+ "step": 10075
2831
+ },
2832
+ {
2833
+ "epoch": 0.8454005189587344,
2834
+ "grad_norm": 7.519137859344482,
2835
+ "learning_rate": 2.3129257766475967e-06,
2836
+ "loss": 0.0689,
2837
+ "step": 10100
2838
+ },
2839
+ {
2840
+ "epoch": 0.8474930945007115,
2841
+ "grad_norm": 0.05053620785474777,
2842
+ "learning_rate": 2.2519310730033305e-06,
2843
+ "loss": 0.039,
2844
+ "step": 10125
2845
+ },
2846
+ {
2847
+ "epoch": 0.8495856700426886,
2848
+ "grad_norm": 0.4234563708305359,
2849
+ "learning_rate": 2.1917033979948377e-06,
2850
+ "loss": 0.042,
2851
+ "step": 10150
2852
+ },
2853
+ {
2854
+ "epoch": 0.8516782455846656,
2855
+ "grad_norm": 0.931300163269043,
2856
+ "learning_rate": 2.1322453545166377e-06,
2857
+ "loss": 0.0359,
2858
+ "step": 10175
2859
+ },
2860
+ {
2861
+ "epoch": 0.8537708211266427,
2862
+ "grad_norm": 0.06214135140180588,
2863
+ "learning_rate": 2.0735595122016417e-06,
2864
+ "loss": 0.0401,
2865
+ "step": 10200
2866
+ },
2867
+ {
2868
+ "epoch": 0.8558633966686198,
2869
+ "grad_norm": 4.354775905609131,
2870
+ "learning_rate": 2.0156484073100845e-06,
2871
+ "loss": 0.1028,
2872
+ "step": 10225
2873
+ },
2874
+ {
2875
+ "epoch": 0.8579559722105968,
2876
+ "grad_norm": 0.26386329531669617,
2877
+ "learning_rate": 1.9585145426199358e-06,
2878
+ "loss": 0.0478,
2879
+ "step": 10250
2880
+ },
2881
+ {
2882
+ "epoch": 0.8600485477525739,
2883
+ "grad_norm": 0.310926228761673,
2884
+ "learning_rate": 1.9021603873187077e-06,
2885
+ "loss": 0.0589,
2886
+ "step": 10275
2887
+ },
2888
+ {
2889
+ "epoch": 0.862141123294551,
2890
+ "grad_norm": 29.527650833129883,
2891
+ "learning_rate": 1.8465883768967718e-06,
2892
+ "loss": 0.0169,
2893
+ "step": 10300
2894
+ },
2895
+ {
2896
+ "epoch": 0.864233698836528,
2897
+ "grad_norm": 0.030782422050833702,
2898
+ "learning_rate": 1.7918009130420744e-06,
2899
+ "loss": 0.0451,
2900
+ "step": 10325
2901
+ },
2902
+ {
2903
+ "epoch": 0.8663262743785051,
2904
+ "grad_norm": 0.1177581399679184,
2905
+ "learning_rate": 1.7378003635363727e-06,
2906
+ "loss": 0.0532,
2907
+ "step": 10350
2908
+ },
2909
+ {
2910
+ "epoch": 0.8684188499204821,
2911
+ "grad_norm": 0.15917450189590454,
2912
+ "learning_rate": 1.684589062152866e-06,
2913
+ "loss": 0.0461,
2914
+ "step": 10375
2915
+ },
2916
+ {
2917
+ "epoch": 0.8705114254624592,
2918
+ "grad_norm": 0.17231865227222443,
2919
+ "learning_rate": 1.6321693085553803e-06,
2920
+ "loss": 0.0732,
2921
+ "step": 10400
2922
+ },
2923
+ {
2924
+ "epoch": 0.8726040010044362,
2925
+ "grad_norm": 0.4389756917953491,
2926
+ "learning_rate": 1.5805433681989434e-06,
2927
+ "loss": 0.058,
2928
+ "step": 10425
2929
+ },
2930
+ {
2931
+ "epoch": 0.8746965765464133,
2932
+ "grad_norm": 0.16644658148288727,
2933
+ "learning_rate": 1.529713472231904e-06,
2934
+ "loss": 0.0241,
2935
+ "step": 10450
2936
+ },
2937
+ {
2938
+ "epoch": 0.8767891520883904,
2939
+ "grad_norm": 0.12109541147947311,
2940
+ "learning_rate": 1.4796818173994853e-06,
2941
+ "loss": 0.0538,
2942
+ "step": 10475
2943
+ },
2944
+ {
2945
+ "epoch": 0.8788817276303674,
2946
+ "grad_norm": 0.07239028066396713,
2947
+ "learning_rate": 1.4304505659488754e-06,
2948
+ "loss": 0.0274,
2949
+ "step": 10500
2950
+ },
2951
+ {
2952
+ "epoch": 0.8809743031723445,
2953
+ "grad_norm": 0.200308695435524,
2954
+ "learning_rate": 1.3820218455357416e-06,
2955
+ "loss": 0.038,
2956
+ "step": 10525
2957
+ },
2958
+ {
2959
+ "epoch": 0.8830668787143215,
2960
+ "grad_norm": 18.13840675354004,
2961
+ "learning_rate": 1.3343977491323212e-06,
2962
+ "loss": 0.0625,
2963
+ "step": 10550
2964
+ },
2965
+ {
2966
+ "epoch": 0.8851594542562986,
2967
+ "grad_norm": 0.20215554535388947,
2968
+ "learning_rate": 1.2875803349369288e-06,
2969
+ "loss": 0.0481,
2970
+ "step": 10575
2971
+ },
2972
+ {
2973
+ "epoch": 0.8872520297982757,
2974
+ "grad_norm": 16.416807174682617,
2975
+ "learning_rate": 1.2415716262850385e-06,
2976
+ "loss": 0.0559,
2977
+ "step": 10600
2978
+ },
2979
+ {
2980
+ "epoch": 0.8893446053402528,
2981
+ "grad_norm": 0.40492701530456543,
2982
+ "learning_rate": 1.196373611561814e-06,
2983
+ "loss": 0.0478,
2984
+ "step": 10625
2985
+ },
2986
+ {
2987
+ "epoch": 0.8914371808822299,
2988
+ "grad_norm": 0.08658579736948013,
2989
+ "learning_rate": 1.1519882441161933e-06,
2990
+ "loss": 0.0307,
2991
+ "step": 10650
2992
+ },
2993
+ {
2994
+ "epoch": 0.893529756424207,
2995
+ "grad_norm": 6.548172473907471,
2996
+ "learning_rate": 1.108417442176468e-06,
2997
+ "loss": 0.0499,
2998
+ "step": 10675
2999
+ },
3000
+ {
3001
+ "epoch": 0.895622331966184,
3002
+ "grad_norm": 104.8302001953125,
3003
+ "learning_rate": 1.0656630887673635e-06,
3004
+ "loss": 0.0668,
3005
+ "step": 10700
3006
+ },
3007
+ {
3008
+ "epoch": 0.8977149075081611,
3009
+ "grad_norm": 0.05074332281947136,
3010
+ "learning_rate": 1.0237270316286896e-06,
3011
+ "loss": 0.055,
3012
+ "step": 10725
3013
+ },
3014
+ {
3015
+ "epoch": 0.8998074830501381,
3016
+ "grad_norm": 0.9664784669876099,
3017
+ "learning_rate": 9.826110831354562e-07,
3018
+ "loss": 0.0458,
3019
+ "step": 10750
3020
+ },
3021
+ {
3022
+ "epoch": 0.9019000585921152,
3023
+ "grad_norm": 8.746543884277344,
3024
+ "learning_rate": 9.42317020219572e-07,
3025
+ "loss": 0.0525,
3026
+ "step": 10775
3027
+ },
3028
+ {
3029
+ "epoch": 0.9039926341340923,
3030
+ "grad_norm": 8.132631301879883,
3031
+ "learning_rate": 9.028465842930223e-07,
3032
+ "loss": 0.0647,
3033
+ "step": 10800
3034
+ },
3035
+ {
3036
+ "epoch": 0.9060852096760693,
3037
+ "grad_norm": 8.856213569641113,
3038
+ "learning_rate": 8.64201481172644e-07,
3039
+ "loss": 0.0679,
3040
+ "step": 10825
3041
+ },
3042
+ {
3043
+ "epoch": 0.9081777852180464,
3044
+ "grad_norm": 9.962249755859375,
3045
+ "learning_rate": 8.26383381006366e-07,
3046
+ "loss": 0.0541,
3047
+ "step": 10850
3048
+ },
3049
+ {
3050
+ "epoch": 0.9102703607600234,
3051
+ "grad_norm": 0.11603320389986038,
3052
+ "learning_rate": 7.893939182010669e-07,
3053
+ "loss": 0.0484,
3054
+ "step": 10875
3055
+ },
3056
+ {
3057
+ "epoch": 0.9123629363020005,
3058
+ "grad_norm": 0.10191322863101959,
3059
+ "learning_rate": 7.532346913519051e-07,
3060
+ "loss": 0.0217,
3061
+ "step": 10900
3062
+ },
3063
+ {
3064
+ "epoch": 0.9144555118439776,
3065
+ "grad_norm": 0.037384625524282455,
3066
+ "learning_rate": 7.179072631732608e-07,
3067
+ "loss": 0.0461,
3068
+ "step": 10925
3069
+ },
3070
+ {
3071
+ "epoch": 0.9165480873859546,
3072
+ "grad_norm": 4.9774932861328125,
3073
+ "learning_rate": 6.834131604311767e-07,
3074
+ "loss": 0.0635,
3075
+ "step": 10950
3076
+ },
3077
+ {
3078
+ "epoch": 0.9186406629279317,
3079
+ "grad_norm": 0.1667776256799698,
3080
+ "learning_rate": 6.497538738773945e-07,
3081
+ "loss": 0.0632,
3082
+ "step": 10975
3083
+ },
3084
+ {
3085
+ "epoch": 0.9207332384699087,
3086
+ "grad_norm": 0.22654391825199127,
3087
+ "learning_rate": 6.169308581849076e-07,
3088
+ "loss": 0.0358,
3089
+ "step": 11000
3090
+ },
3091
+ {
3092
+ "epoch": 0.9228258140118858,
3093
+ "grad_norm": 0.11612854152917862,
3094
+ "learning_rate": 5.849455318851149e-07,
3095
+ "loss": 0.0542,
3096
+ "step": 11025
3097
+ },
3098
+ {
3099
+ "epoch": 0.9249183895538629,
3100
+ "grad_norm": 35.777793884277344,
3101
+ "learning_rate": 5.537992773064949e-07,
3102
+ "loss": 0.0661,
3103
+ "step": 11050
3104
+ },
3105
+ {
3106
+ "epoch": 0.9270109650958399,
3107
+ "grad_norm": 31.49355125427246,
3108
+ "learning_rate": 5.234934405148818e-07,
3109
+ "loss": 0.0461,
3110
+ "step": 11075
3111
+ },
3112
+ {
3113
+ "epoch": 0.929103540637817,
3114
+ "grad_norm": 0.2999427318572998,
3115
+ "learning_rate": 4.94029331255279e-07,
3116
+ "loss": 0.0336,
3117
+ "step": 11100
3118
+ },
3119
+ {
3120
+ "epoch": 0.9311961161797941,
3121
+ "grad_norm": 1.5943710803985596,
3122
+ "learning_rate": 4.65408222895265e-07,
3123
+ "loss": 0.1199,
3124
+ "step": 11125
3125
+ },
3126
+ {
3127
+ "epoch": 0.9332886917217712,
3128
+ "grad_norm": 0.04734862968325615,
3129
+ "learning_rate": 4.3763135236996045e-07,
3130
+ "loss": 0.0506,
3131
+ "step": 11150
3132
+ },
3133
+ {
3134
+ "epoch": 0.9353812672637483,
3135
+ "grad_norm": 7.017291069030762,
3136
+ "learning_rate": 4.1069992012855664e-07,
3137
+ "loss": 0.0793,
3138
+ "step": 11175
3139
+ },
3140
+ {
3141
+ "epoch": 0.9374738428057253,
3142
+ "grad_norm": 0.1956498920917511,
3143
+ "learning_rate": 3.8461509008245637e-07,
3144
+ "loss": 0.0557,
3145
+ "step": 11200
3146
+ },
3147
+ {
3148
+ "epoch": 0.9395664183477024,
3149
+ "grad_norm": 14.131322860717773,
3150
+ "learning_rate": 3.5937798955495427e-07,
3151
+ "loss": 0.0493,
3152
+ "step": 11225
3153
+ },
3154
+ {
3155
+ "epoch": 0.9416589938896794,
3156
+ "grad_norm": 0.08296482264995575,
3157
+ "learning_rate": 3.349897092325338e-07,
3158
+ "loss": 0.0785,
3159
+ "step": 11250
3160
+ },
3161
+ {
3162
+ "epoch": 0.9437515694316565,
3163
+ "grad_norm": 6.673073768615723,
3164
+ "learning_rate": 3.1145130311771134e-07,
3165
+ "loss": 0.0728,
3166
+ "step": 11275
3167
+ },
3168
+ {
3169
+ "epoch": 0.9458441449736336,
3170
+ "grad_norm": 0.3945884704589844,
3171
+ "learning_rate": 2.887637884834993e-07,
3172
+ "loss": 0.0453,
3173
+ "step": 11300
3174
+ },
3175
+ {
3176
+ "epoch": 0.9479367205156106,
3177
+ "grad_norm": 0.04271361231803894,
3178
+ "learning_rate": 2.6692814582943707e-07,
3179
+ "loss": 0.0378,
3180
+ "step": 11325
3181
+ },
3182
+ {
3183
+ "epoch": 0.9500292960575877,
3184
+ "grad_norm": 6.191662788391113,
3185
+ "learning_rate": 2.45945318839218e-07,
3186
+ "loss": 0.0506,
3187
+ "step": 11350
3188
+ },
3189
+ {
3190
+ "epoch": 0.9521218715995647,
3191
+ "grad_norm": 0.21662096679210663,
3192
+ "learning_rate": 2.2581621433989785e-07,
3193
+ "loss": 0.057,
3194
+ "step": 11375
3195
+ },
3196
+ {
3197
+ "epoch": 0.9542144471415418,
3198
+ "grad_norm": 0.04659692198038101,
3199
+ "learning_rate": 2.0654170226271696e-07,
3200
+ "loss": 0.0632,
3201
+ "step": 11400
3202
+ },
3203
+ {
3204
+ "epoch": 0.9563070226835189,
3205
+ "grad_norm": 7.704138278961182,
3206
+ "learning_rate": 1.8812261560549272e-07,
3207
+ "loss": 0.0862,
3208
+ "step": 11425
3209
+ },
3210
+ {
3211
+ "epoch": 0.9583995982254959,
3212
+ "grad_norm": 0.09132679551839828,
3213
+ "learning_rate": 1.7055975039662387e-07,
3214
+ "loss": 0.0502,
3215
+ "step": 11450
3216
+ },
3217
+ {
3218
+ "epoch": 0.960492173767473,
3219
+ "grad_norm": 12.371918678283691,
3220
+ "learning_rate": 1.5385386566069137e-07,
3221
+ "loss": 0.0534,
3222
+ "step": 11475
3223
+ },
3224
+ {
3225
+ "epoch": 0.96258474930945,
3226
+ "grad_norm": 0.09234387427568436,
3227
+ "learning_rate": 1.3800568338565133e-07,
3228
+ "loss": 0.0764,
3229
+ "step": 11500
3230
+ },
3231
+ {
3232
+ "epoch": 0.9646773248514271,
3233
+ "grad_norm": 0.1503359079360962,
3234
+ "learning_rate": 1.230158884916266e-07,
3235
+ "loss": 0.0589,
3236
+ "step": 11525
3237
+ },
3238
+ {
3239
+ "epoch": 0.9667699003934042,
3240
+ "grad_norm": 7.913686752319336,
3241
+ "learning_rate": 1.088851288013193e-07,
3242
+ "loss": 0.0655,
3243
+ "step": 11550
3244
+ },
3245
+ {
3246
+ "epoch": 0.9688624759353812,
3247
+ "grad_norm": 8.503890991210938,
3248
+ "learning_rate": 9.56140150120044e-08,
3249
+ "loss": 0.0523,
3250
+ "step": 11575
3251
+ },
3252
+ {
3253
+ "epoch": 0.9709550514773583,
3254
+ "grad_norm": 0.07973815500736237,
3255
+ "learning_rate": 8.320312066913972e-08,
3256
+ "loss": 0.0239,
3257
+ "step": 11600
3258
+ },
3259
+ {
3260
+ "epoch": 0.9730476270193354,
3261
+ "grad_norm": 0.23680634796619415,
3262
+ "learning_rate": 7.165298214157457e-08,
3263
+ "loss": 0.0381,
3264
+ "step": 11625
3265
+ },
3266
+ {
3267
+ "epoch": 0.9751402025613125,
3268
+ "grad_norm": 0.10261060297489166,
3269
+ "learning_rate": 6.096409859837727e-08,
3270
+ "loss": 0.1002,
3271
+ "step": 11650
3272
+ },
3273
+ {
3274
+ "epoch": 0.9772327781032896,
3275
+ "grad_norm": 0.22620265185832977,
3276
+ "learning_rate": 5.113693198725456e-08,
3277
+ "loss": 0.0493,
3278
+ "step": 11675
3279
+ },
3280
+ {
3281
+ "epoch": 0.9793253536452666,
3282
+ "grad_norm": 7.392359256744385,
3283
+ "learning_rate": 4.2171907014596504e-08,
3284
+ "loss": 0.0553,
3285
+ "step": 11700
3286
+ },
3287
+ {
3288
+ "epoch": 0.9814179291872437,
3289
+ "grad_norm": 0.09146345406770706,
3290
+ "learning_rate": 3.406941112710449e-08,
3291
+ "loss": 0.0752,
3292
+ "step": 11725
3293
+ },
3294
+ {
3295
+ "epoch": 0.9835105047292207,
3296
+ "grad_norm": 0.03258276730775833,
3297
+ "learning_rate": 2.6829794495066842e-08,
3298
+ "loss": 0.0422,
3299
+ "step": 11750
3300
+ },
3301
+ {
3302
+ "epoch": 0.9856030802711978,
3303
+ "grad_norm": 15.740374565124512,
3304
+ "learning_rate": 2.0453369997215366e-08,
3305
+ "loss": 0.0587,
3306
+ "step": 11775
3307
+ },
3308
+ {
3309
+ "epoch": 0.9876956558131749,
3310
+ "grad_norm": 0.10255517065525055,
3311
+ "learning_rate": 1.4940413207198413e-08,
3312
+ "loss": 0.0378,
3313
+ "step": 11800
3314
+ },
3315
+ {
3316
+ "epoch": 0.9897882313551519,
3317
+ "grad_norm": 28.702392578125,
3318
+ "learning_rate": 1.0291162381685927e-08,
3319
+ "loss": 0.043,
3320
+ "step": 11825
3321
+ },
3322
+ {
3323
+ "epoch": 0.991880806897129,
3324
+ "grad_norm": 0.1069909930229187,
3325
+ "learning_rate": 6.5058184500577104e-09,
3326
+ "loss": 0.0577,
3327
+ "step": 11850
3328
+ },
3329
+ {
3330
+ "epoch": 0.993973382439106,
3331
+ "grad_norm": 62.75130844116211,
3332
+ "learning_rate": 3.5845450057303554e-09,
3333
+ "loss": 0.0672,
3334
+ "step": 11875
3335
+ },
3336
+ {
3337
+ "epoch": 0.9960659579810831,
3338
+ "grad_norm": 1.7898138761520386,
3339
+ "learning_rate": 1.5274682990762402e-09,
3340
+ "loss": 0.0607,
3341
+ "step": 11900
3342
+ },
3343
+ {
3344
+ "epoch": 0.9981585335230602,
3345
+ "grad_norm": 0.3523304760456085,
3346
+ "learning_rate": 3.346772319834379e-10,
3347
+ "loss": 0.0546,
3348
+ "step": 11925
3349
+ },
3350
+ {
3351
+ "epoch": 1.0,
3352
+ "eval_accuracy": 0.9794339738473816,
3353
+ "eval_auc": 0.9957081876919603,
3354
+ "eval_f1": 0.9853826686447335,
3355
+ "eval_loss": 0.09541802108287811,
3356
+ "eval_precision": 0.9808886765408504,
3357
+ "eval_recall": 0.9899180291938807,
3358
+ "eval_runtime": 3282.7457,
3359
+ "eval_samples_per_second": 9.924,
3360
+ "eval_steps_per_second": 0.207,
3361
+ "step": 11947
3362
+ }
3363
+ ],
3364
+ "logging_steps": 25,
3365
+ "max_steps": 11947,
3366
+ "num_input_tokens_seen": 0,
3367
+ "num_train_epochs": 1,
3368
+ "save_steps": 500,
3369
+ "stateful_callbacks": {
3370
+ "EarlyStoppingCallback": {
3371
+ "args": {
3372
+ "early_stopping_patience": 5,
3373
+ "early_stopping_threshold": 0.01
3374
+ },
3375
+ "attributes": {
3376
+ "early_stopping_patience_counter": 0
3377
+ }
3378
+ },
3379
+ "TrainerControl": {
3380
+ "args": {
3381
+ "should_epoch_stop": false,
3382
+ "should_evaluate": false,
3383
+ "should_log": false,
3384
+ "should_save": true,
3385
+ "should_training_stop": true
3386
+ },
3387
+ "attributes": {}
3388
+ }
3389
+ },
3390
+ "total_flos": 6.602150538303662e+19,
3391
+ "train_batch_size": 24,
3392
+ "trial_name": null,
3393
+ "trial_params": null
3394
+ }
checkpoint-11947/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76126427121423c5e1feb03aaa2837a856f56f7281a2de99cc07a881e6c370d9
3
+ size 5368
config.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "haywoodsloan/ai-image-detector-deploy",
3
+ "_num_labels": 2,
4
+ "architectures": [
5
+ "Swinv2ForImageClassification"
6
+ ],
7
+ "attention_probs_dropout_prob": 0.0,
8
+ "depths": [
9
+ 2,
10
+ 2,
11
+ 18,
12
+ 2
13
+ ],
14
+ "drop_path_rate": 0.1,
15
+ "embed_dim": 192,
16
+ "encoder_stride": 32,
17
+ "hidden_act": "gelu",
18
+ "hidden_dropout_prob": 0.0,
19
+ "hidden_size": 1536,
20
+ "id2label": {
21
+ "0": "artificial",
22
+ "1": "real"
23
+ },
24
+ "image_size": 256,
25
+ "initializer_range": 0.02,
26
+ "label2id": {
27
+ "artificial": 0,
28
+ "real": 1
29
+ },
30
+ "layer_norm_eps": 1e-05,
31
+ "mlp_ratio": 4.0,
32
+ "model_type": "swinv2",
33
+ "num_channels": 3,
34
+ "num_heads": [
35
+ 6,
36
+ 12,
37
+ 24,
38
+ 48
39
+ ],
40
+ "num_layers": 4,
41
+ "out_features": [
42
+ "stage4"
43
+ ],
44
+ "out_indices": [
45
+ 4
46
+ ],
47
+ "patch_size": 4,
48
+ "path_norm": true,
49
+ "pretrained_window_sizes": [
50
+ 12,
51
+ 12,
52
+ 12,
53
+ 6
54
+ ],
55
+ "problem_type": "single_label_classification",
56
+ "qkv_bias": true,
57
+ "stage_names": [
58
+ "stem",
59
+ "stage1",
60
+ "stage2",
61
+ "stage3",
62
+ "stage4"
63
+ ],
64
+ "torch_dtype": "float32",
65
+ "transformers_version": "4.48.0",
66
+ "use_absolute_embeddings": false,
67
+ "window_size": 16
68
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:802ba4e65566633ea393c79a8f7a54ab37e312bd6055442274592086c1a8e8f9
3
+ size 780886280
preprocessor_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_rgb": null,
3
+ "do_normalize": true,
4
+ "do_rescale": true,
5
+ "do_resize": true,
6
+ "image_mean": [
7
+ 0.485,
8
+ 0.456,
9
+ 0.406
10
+ ],
11
+ "image_processor_type": "ViTImageProcessor",
12
+ "image_std": [
13
+ 0.229,
14
+ 0.224,
15
+ 0.225
16
+ ],
17
+ "resample": 3,
18
+ "rescale_factor": 0.00392156862745098,
19
+ "size": {
20
+ "height": 256,
21
+ "width": 256
22
+ }
23
+ }
runs/Apr18_01-52-02_2b9b531dbda5/events.out.tfevents.1744941124.2b9b531dbda5.218.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12e84aa57fb41ec88c8f343924f0adf00f1771b855c1a8e04d2df5b0947d8be8
3
+ size 107127
runs/Apr18_01-52-02_2b9b531dbda5/events.out.tfevents.1744985308.2b9b531dbda5.218.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:283cb248196dd5965f70fb189ddd74ad5deded23eb1bfa2de94f0e80400962ee
3
+ size 607
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:76126427121423c5e1feb03aaa2837a856f56f7281a2de99cc07a881e6c370d9
3
+ size 5368
training_params.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "data_path": "haywoodsloan/ai-images",
3
+ "model": "haywoodsloan/ai-image-detector-deploy",
4
+ "username": "haywoodsloan",
5
+ "lr": 4e-05,
6
+ "epochs": 1,
7
+ "batch_size": 24,
8
+ "warmup_ratio": 0.0,
9
+ "gradient_accumulation": 1,
10
+ "optimizer": "adamw_torch",
11
+ "scheduler": "cosine",
12
+ "weight_decay": 0.001,
13
+ "max_grad_norm": 1.0,
14
+ "seed": 77,
15
+ "train_split": "train",
16
+ "valid_split": "test",
17
+ "logging_steps": -1,
18
+ "project_name": "autotrain-ai-image-detect-20250418-0133",
19
+ "auto_find_batch_size": true,
20
+ "mixed_precision": "bf16",
21
+ "save_total_limit": 1,
22
+ "push_to_hub": true,
23
+ "eval_strategy": "epoch",
24
+ "image_column": "image",
25
+ "target_column": "label",
26
+ "log": "tensorboard",
27
+ "early_stopping_patience": 5,
28
+ "early_stopping_threshold": 0.01
29
+ }