Upload folder using huggingface_hub
Browse files- README.md +31 -0
- checkpoint-8991/config.json +33 -0
- checkpoint-8991/model.safetensors +3 -0
- checkpoint-8991/optimizer.pt +3 -0
- checkpoint-8991/rng_state.pth +3 -0
- checkpoint-8991/scheduler.pt +3 -0
- checkpoint-8991/trainer_state.json +2594 -0
- checkpoint-8991/training_args.bin +3 -0
- config.json +33 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- runs/Oct20_19-44-40_Haileys-MacBook-Pro.local/events.out.tfevents.1729478680.Haileys-MacBook-Pro.local.88288.0 +2 -2
- runs/Oct20_19-44-40_Haileys-MacBook-Pro.local/events.out.tfevents.1729483332.Haileys-MacBook-Pro.local.88288.1 +3 -0
- training_args.bin +3 -0
- training_params.json +29 -0
README.md
ADDED
@@ -0,0 +1,31 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
---
|
3 |
+
tags:
|
4 |
+
- autotrain
|
5 |
+
- image-classification
|
6 |
+
base_model: google/vit-base-patch16-384
|
7 |
+
widget:
|
8 |
+
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg
|
9 |
+
example_title: Tiger
|
10 |
+
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg
|
11 |
+
example_title: Teapot
|
12 |
+
- src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg
|
13 |
+
example_title: Palace
|
14 |
+
---
|
15 |
+
|
16 |
+
# Model Trained Using AutoTrain
|
17 |
+
|
18 |
+
- Problem type: Image Classification
|
19 |
+
|
20 |
+
## Validation Metrics
|
21 |
+
loss: 0.003335413057357073
|
22 |
+
|
23 |
+
f1: 0.9523809523809523
|
24 |
+
|
25 |
+
precision: 0.958904109589041
|
26 |
+
|
27 |
+
recall: 0.9459459459459459
|
28 |
+
|
29 |
+
auc: 0.9999543462381301
|
30 |
+
|
31 |
+
accuracy: 0.9988321654988321
|
checkpoint-8991/config.json
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/vit-base-patch16-384",
|
3 |
+
"_num_labels": 2,
|
4 |
+
"architectures": [
|
5 |
+
"ViTForImageClassification"
|
6 |
+
],
|
7 |
+
"attention_probs_dropout_prob": 0.0,
|
8 |
+
"encoder_stride": 16,
|
9 |
+
"hidden_act": "gelu",
|
10 |
+
"hidden_dropout_prob": 0.0,
|
11 |
+
"hidden_size": 768,
|
12 |
+
"id2label": {
|
13 |
+
"0": "not_soyjak",
|
14 |
+
"1": "soyjak"
|
15 |
+
},
|
16 |
+
"image_size": 384,
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 3072,
|
19 |
+
"label2id": {
|
20 |
+
"not_soyjak": 0,
|
21 |
+
"soyjak": 1
|
22 |
+
},
|
23 |
+
"layer_norm_eps": 1e-12,
|
24 |
+
"model_type": "vit",
|
25 |
+
"num_attention_heads": 12,
|
26 |
+
"num_channels": 3,
|
27 |
+
"num_hidden_layers": 12,
|
28 |
+
"patch_size": 16,
|
29 |
+
"problem_type": "single_label_classification",
|
30 |
+
"qkv_bias": true,
|
31 |
+
"torch_dtype": "float32",
|
32 |
+
"transformers_version": "4.42.3"
|
33 |
+
}
|
checkpoint-8991/model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:227d87907aafc9cb1de9990403be5dcfdd6ff412a4009e175a9c7659fa4f6ca3
|
3 |
+
size 344391328
|
checkpoint-8991/optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f613a7efce3d006e4eac04c2f4d3b10343065d4131cd00c8cdcca89f4e659b5b
|
3 |
+
size 688898042
|
checkpoint-8991/rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:06d1386983b4bf5acde4a2c14df332c80abe345a35d2d01f07b2f9767805b49b
|
3 |
+
size 13990
|
checkpoint-8991/scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:947002f3b428f7761d4dfa2a10f3d855152386b1bdf52eb6d1b300e1e22e6572
|
3 |
+
size 1064
|
checkpoint-8991/trainer_state.json
ADDED
@@ -0,0 +1,2594 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.003335413057357073,
|
3 |
+
"best_model_checkpoint": "sj-det-2/checkpoint-8991",
|
4 |
+
"epoch": 3.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 8991,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.008341675008341674,
|
13 |
+
"grad_norm": 16.313989639282227,
|
14 |
+
"learning_rate": 1.388888888888889e-06,
|
15 |
+
"loss": 0.9123,
|
16 |
+
"step": 25
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.01668335001668335,
|
20 |
+
"grad_norm": 8.612642288208008,
|
21 |
+
"learning_rate": 2.777777777777778e-06,
|
22 |
+
"loss": 0.6485,
|
23 |
+
"step": 50
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.025025025025025027,
|
27 |
+
"grad_norm": 4.427529335021973,
|
28 |
+
"learning_rate": 4.166666666666667e-06,
|
29 |
+
"loss": 0.3344,
|
30 |
+
"step": 75
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.0333667000333667,
|
34 |
+
"grad_norm": 1.323181390762329,
|
35 |
+
"learning_rate": 5.555555555555556e-06,
|
36 |
+
"loss": 0.1139,
|
37 |
+
"step": 100
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.04170837504170837,
|
41 |
+
"grad_norm": 1.4804118871688843,
|
42 |
+
"learning_rate": 6.944444444444445e-06,
|
43 |
+
"loss": 0.0661,
|
44 |
+
"step": 125
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.05005005005005005,
|
48 |
+
"grad_norm": 0.09656789153814316,
|
49 |
+
"learning_rate": 8.333333333333334e-06,
|
50 |
+
"loss": 0.0585,
|
51 |
+
"step": 150
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.05839172505839173,
|
55 |
+
"grad_norm": 0.036666255444288254,
|
56 |
+
"learning_rate": 9.722222222222223e-06,
|
57 |
+
"loss": 0.0217,
|
58 |
+
"step": 175
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.0667334000667334,
|
62 |
+
"grad_norm": 0.024522973224520683,
|
63 |
+
"learning_rate": 1.1111111111111112e-05,
|
64 |
+
"loss": 0.0832,
|
65 |
+
"step": 200
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.07507507507507508,
|
69 |
+
"grad_norm": 0.48196515440940857,
|
70 |
+
"learning_rate": 1.25e-05,
|
71 |
+
"loss": 0.027,
|
72 |
+
"step": 225
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.08341675008341674,
|
76 |
+
"grad_norm": 0.08288281410932541,
|
77 |
+
"learning_rate": 1.388888888888889e-05,
|
78 |
+
"loss": 0.0154,
|
79 |
+
"step": 250
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.09175842509175842,
|
83 |
+
"grad_norm": 0.018726350739598274,
|
84 |
+
"learning_rate": 1.527777777777778e-05,
|
85 |
+
"loss": 0.0565,
|
86 |
+
"step": 275
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.1001001001001001,
|
90 |
+
"grad_norm": 2.332498788833618,
|
91 |
+
"learning_rate": 1.6666666666666667e-05,
|
92 |
+
"loss": 0.0314,
|
93 |
+
"step": 300
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.10844177510844177,
|
97 |
+
"grad_norm": 0.6220697164535522,
|
98 |
+
"learning_rate": 1.8055555555555555e-05,
|
99 |
+
"loss": 0.0056,
|
100 |
+
"step": 325
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.11678345011678345,
|
104 |
+
"grad_norm": 0.03775189071893692,
|
105 |
+
"learning_rate": 1.9444444444444445e-05,
|
106 |
+
"loss": 0.0006,
|
107 |
+
"step": 350
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.12512512512512514,
|
111 |
+
"grad_norm": 0.014640103094279766,
|
112 |
+
"learning_rate": 2.0833333333333336e-05,
|
113 |
+
"loss": 0.0559,
|
114 |
+
"step": 375
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.1334668001334668,
|
118 |
+
"grad_norm": 0.00695949187502265,
|
119 |
+
"learning_rate": 2.2222222222222223e-05,
|
120 |
+
"loss": 0.0313,
|
121 |
+
"step": 400
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.14180847514180847,
|
125 |
+
"grad_norm": 0.01726640574634075,
|
126 |
+
"learning_rate": 2.361111111111111e-05,
|
127 |
+
"loss": 0.0142,
|
128 |
+
"step": 425
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.15015015015015015,
|
132 |
+
"grad_norm": 0.0029850194696336985,
|
133 |
+
"learning_rate": 2.5e-05,
|
134 |
+
"loss": 0.0008,
|
135 |
+
"step": 450
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.15849182515849183,
|
139 |
+
"grad_norm": 0.009473559446632862,
|
140 |
+
"learning_rate": 2.6388888888888892e-05,
|
141 |
+
"loss": 0.0627,
|
142 |
+
"step": 475
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.1668335001668335,
|
146 |
+
"grad_norm": 0.01642836630344391,
|
147 |
+
"learning_rate": 2.777777777777778e-05,
|
148 |
+
"loss": 0.0229,
|
149 |
+
"step": 500
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.17517517517517517,
|
153 |
+
"grad_norm": 0.037022557109594345,
|
154 |
+
"learning_rate": 2.916666666666667e-05,
|
155 |
+
"loss": 0.0433,
|
156 |
+
"step": 525
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.18351685018351685,
|
160 |
+
"grad_norm": 0.0169846098870039,
|
161 |
+
"learning_rate": 3.055555555555556e-05,
|
162 |
+
"loss": 0.0387,
|
163 |
+
"step": 550
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.19185852519185853,
|
167 |
+
"grad_norm": 0.0074409060180187225,
|
168 |
+
"learning_rate": 3.194444444444444e-05,
|
169 |
+
"loss": 0.0064,
|
170 |
+
"step": 575
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.2002002002002002,
|
174 |
+
"grad_norm": 0.022537339478731155,
|
175 |
+
"learning_rate": 3.3333333333333335e-05,
|
176 |
+
"loss": 0.0473,
|
177 |
+
"step": 600
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.20854187520854187,
|
181 |
+
"grad_norm": 0.00912486296147108,
|
182 |
+
"learning_rate": 3.472222222222222e-05,
|
183 |
+
"loss": 0.0071,
|
184 |
+
"step": 625
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.21688355021688355,
|
188 |
+
"grad_norm": 0.015622120350599289,
|
189 |
+
"learning_rate": 3.611111111111111e-05,
|
190 |
+
"loss": 0.018,
|
191 |
+
"step": 650
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.22522522522522523,
|
195 |
+
"grad_norm": 0.038481712341308594,
|
196 |
+
"learning_rate": 3.7500000000000003e-05,
|
197 |
+
"loss": 0.0609,
|
198 |
+
"step": 675
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.2335669002335669,
|
202 |
+
"grad_norm": 0.06651923805475235,
|
203 |
+
"learning_rate": 3.888888888888889e-05,
|
204 |
+
"loss": 0.0588,
|
205 |
+
"step": 700
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.24190857524190856,
|
209 |
+
"grad_norm": 0.013173066079616547,
|
210 |
+
"learning_rate": 4.027777777777778e-05,
|
211 |
+
"loss": 0.0006,
|
212 |
+
"step": 725
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.2502502502502503,
|
216 |
+
"grad_norm": 0.002995034446939826,
|
217 |
+
"learning_rate": 4.166666666666667e-05,
|
218 |
+
"loss": 0.0008,
|
219 |
+
"step": 750
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.2585919252585919,
|
223 |
+
"grad_norm": 0.001744964043609798,
|
224 |
+
"learning_rate": 4.305555555555556e-05,
|
225 |
+
"loss": 0.0002,
|
226 |
+
"step": 775
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.2669336002669336,
|
230 |
+
"grad_norm": 0.5510344505310059,
|
231 |
+
"learning_rate": 4.4444444444444447e-05,
|
232 |
+
"loss": 0.0373,
|
233 |
+
"step": 800
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.2752752752752753,
|
237 |
+
"grad_norm": 0.0068357582204043865,
|
238 |
+
"learning_rate": 4.5833333333333334e-05,
|
239 |
+
"loss": 0.0269,
|
240 |
+
"step": 825
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.28361695028361694,
|
244 |
+
"grad_norm": 0.014822622761130333,
|
245 |
+
"learning_rate": 4.722222222222222e-05,
|
246 |
+
"loss": 0.1029,
|
247 |
+
"step": 850
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.29195862529195865,
|
251 |
+
"grad_norm": 0.006352724973112345,
|
252 |
+
"learning_rate": 4.8611111111111115e-05,
|
253 |
+
"loss": 0.0527,
|
254 |
+
"step": 875
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.3003003003003003,
|
258 |
+
"grad_norm": 0.006285039242357016,
|
259 |
+
"learning_rate": 5e-05,
|
260 |
+
"loss": 0.0007,
|
261 |
+
"step": 900
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.30864197530864196,
|
265 |
+
"grad_norm": 0.21172092854976654,
|
266 |
+
"learning_rate": 4.984550735384996e-05,
|
267 |
+
"loss": 0.001,
|
268 |
+
"step": 925
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 0.31698365031698367,
|
272 |
+
"grad_norm": 0.025125497952103615,
|
273 |
+
"learning_rate": 4.969101470769992e-05,
|
274 |
+
"loss": 0.0001,
|
275 |
+
"step": 950
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 0.3253253253253253,
|
279 |
+
"grad_norm": 0.0006277966895140707,
|
280 |
+
"learning_rate": 4.9536522061549875e-05,
|
281 |
+
"loss": 0.0,
|
282 |
+
"step": 975
|
283 |
+
},
|
284 |
+
{
|
285 |
+
"epoch": 0.333667000333667,
|
286 |
+
"grad_norm": 0.0005465027061291039,
|
287 |
+
"learning_rate": 4.9382029415399826e-05,
|
288 |
+
"loss": 0.0,
|
289 |
+
"step": 1000
|
290 |
+
},
|
291 |
+
{
|
292 |
+
"epoch": 0.3420086753420087,
|
293 |
+
"grad_norm": 0.0011532187927514315,
|
294 |
+
"learning_rate": 4.922753676924979e-05,
|
295 |
+
"loss": 0.0303,
|
296 |
+
"step": 1025
|
297 |
+
},
|
298 |
+
{
|
299 |
+
"epoch": 0.35035035035035034,
|
300 |
+
"grad_norm": 0.0010950276628136635,
|
301 |
+
"learning_rate": 4.907304412309974e-05,
|
302 |
+
"loss": 0.0001,
|
303 |
+
"step": 1050
|
304 |
+
},
|
305 |
+
{
|
306 |
+
"epoch": 0.35869202535869205,
|
307 |
+
"grad_norm": 0.0015246294206008315,
|
308 |
+
"learning_rate": 4.89185514769497e-05,
|
309 |
+
"loss": 0.0007,
|
310 |
+
"step": 1075
|
311 |
+
},
|
312 |
+
{
|
313 |
+
"epoch": 0.3670337003670337,
|
314 |
+
"grad_norm": 0.0425708144903183,
|
315 |
+
"learning_rate": 4.876405883079966e-05,
|
316 |
+
"loss": 0.0592,
|
317 |
+
"step": 1100
|
318 |
+
},
|
319 |
+
{
|
320 |
+
"epoch": 0.37537537537537535,
|
321 |
+
"grad_norm": 0.001601433614268899,
|
322 |
+
"learning_rate": 4.8609566184649615e-05,
|
323 |
+
"loss": 0.001,
|
324 |
+
"step": 1125
|
325 |
+
},
|
326 |
+
{
|
327 |
+
"epoch": 0.38371705038371706,
|
328 |
+
"grad_norm": 0.04269692674279213,
|
329 |
+
"learning_rate": 4.8455073538499566e-05,
|
330 |
+
"loss": 0.052,
|
331 |
+
"step": 1150
|
332 |
+
},
|
333 |
+
{
|
334 |
+
"epoch": 0.3920587253920587,
|
335 |
+
"grad_norm": 10.334559440612793,
|
336 |
+
"learning_rate": 4.830058089234953e-05,
|
337 |
+
"loss": 0.0149,
|
338 |
+
"step": 1175
|
339 |
+
},
|
340 |
+
{
|
341 |
+
"epoch": 0.4004004004004004,
|
342 |
+
"grad_norm": 0.0104717668145895,
|
343 |
+
"learning_rate": 4.814608824619948e-05,
|
344 |
+
"loss": 0.0524,
|
345 |
+
"step": 1200
|
346 |
+
},
|
347 |
+
{
|
348 |
+
"epoch": 0.4087420754087421,
|
349 |
+
"grad_norm": 0.004411773756146431,
|
350 |
+
"learning_rate": 4.799159560004944e-05,
|
351 |
+
"loss": 0.0003,
|
352 |
+
"step": 1225
|
353 |
+
},
|
354 |
+
{
|
355 |
+
"epoch": 0.41708375041708373,
|
356 |
+
"grad_norm": 0.0019501554779708385,
|
357 |
+
"learning_rate": 4.7837102953899396e-05,
|
358 |
+
"loss": 0.0001,
|
359 |
+
"step": 1250
|
360 |
+
},
|
361 |
+
{
|
362 |
+
"epoch": 0.42542542542542544,
|
363 |
+
"grad_norm": 0.001354438834823668,
|
364 |
+
"learning_rate": 4.7682610307749354e-05,
|
365 |
+
"loss": 0.0011,
|
366 |
+
"step": 1275
|
367 |
+
},
|
368 |
+
{
|
369 |
+
"epoch": 0.4337671004337671,
|
370 |
+
"grad_norm": 0.0017444408731535077,
|
371 |
+
"learning_rate": 4.752811766159931e-05,
|
372 |
+
"loss": 0.0003,
|
373 |
+
"step": 1300
|
374 |
+
},
|
375 |
+
{
|
376 |
+
"epoch": 0.44210877544210875,
|
377 |
+
"grad_norm": 0.00919169932603836,
|
378 |
+
"learning_rate": 4.737362501544927e-05,
|
379 |
+
"loss": 0.06,
|
380 |
+
"step": 1325
|
381 |
+
},
|
382 |
+
{
|
383 |
+
"epoch": 0.45045045045045046,
|
384 |
+
"grad_norm": 0.001080819871276617,
|
385 |
+
"learning_rate": 4.721913236929923e-05,
|
386 |
+
"loss": 0.0031,
|
387 |
+
"step": 1350
|
388 |
+
},
|
389 |
+
{
|
390 |
+
"epoch": 0.4587921254587921,
|
391 |
+
"grad_norm": 0.03211037814617157,
|
392 |
+
"learning_rate": 4.706463972314918e-05,
|
393 |
+
"loss": 0.0454,
|
394 |
+
"step": 1375
|
395 |
+
},
|
396 |
+
{
|
397 |
+
"epoch": 0.4671338004671338,
|
398 |
+
"grad_norm": 0.025529522448778152,
|
399 |
+
"learning_rate": 4.691014707699914e-05,
|
400 |
+
"loss": 0.0036,
|
401 |
+
"step": 1400
|
402 |
+
},
|
403 |
+
{
|
404 |
+
"epoch": 0.4754754754754755,
|
405 |
+
"grad_norm": 0.0008930154144763947,
|
406 |
+
"learning_rate": 4.675565443084909e-05,
|
407 |
+
"loss": 0.0451,
|
408 |
+
"step": 1425
|
409 |
+
},
|
410 |
+
{
|
411 |
+
"epoch": 0.4838171504838171,
|
412 |
+
"grad_norm": 0.023177633062005043,
|
413 |
+
"learning_rate": 4.660116178469905e-05,
|
414 |
+
"loss": 0.0028,
|
415 |
+
"step": 1450
|
416 |
+
},
|
417 |
+
{
|
418 |
+
"epoch": 0.49215882549215884,
|
419 |
+
"grad_norm": 0.0005259920726530254,
|
420 |
+
"learning_rate": 4.644666913854901e-05,
|
421 |
+
"loss": 0.0461,
|
422 |
+
"step": 1475
|
423 |
+
},
|
424 |
+
{
|
425 |
+
"epoch": 0.5005005005005005,
|
426 |
+
"grad_norm": 0.007525968365371227,
|
427 |
+
"learning_rate": 4.6292176492398966e-05,
|
428 |
+
"loss": 0.0403,
|
429 |
+
"step": 1500
|
430 |
+
},
|
431 |
+
{
|
432 |
+
"epoch": 0.5088421755088421,
|
433 |
+
"grad_norm": 0.005125482566654682,
|
434 |
+
"learning_rate": 4.613768384624892e-05,
|
435 |
+
"loss": 0.0015,
|
436 |
+
"step": 1525
|
437 |
+
},
|
438 |
+
{
|
439 |
+
"epoch": 0.5171838505171839,
|
440 |
+
"grad_norm": 0.00863750372081995,
|
441 |
+
"learning_rate": 4.598319120009888e-05,
|
442 |
+
"loss": 0.0205,
|
443 |
+
"step": 1550
|
444 |
+
},
|
445 |
+
{
|
446 |
+
"epoch": 0.5255255255255256,
|
447 |
+
"grad_norm": 0.0001800910395104438,
|
448 |
+
"learning_rate": 4.582869855394883e-05,
|
449 |
+
"loss": 0.0037,
|
450 |
+
"step": 1575
|
451 |
+
},
|
452 |
+
{
|
453 |
+
"epoch": 0.5338672005338672,
|
454 |
+
"grad_norm": 0.00019970687571913004,
|
455 |
+
"learning_rate": 4.567420590779879e-05,
|
456 |
+
"loss": 0.0152,
|
457 |
+
"step": 1600
|
458 |
+
},
|
459 |
+
{
|
460 |
+
"epoch": 0.5422088755422089,
|
461 |
+
"grad_norm": 0.011676608584821224,
|
462 |
+
"learning_rate": 4.551971326164875e-05,
|
463 |
+
"loss": 0.0635,
|
464 |
+
"step": 1625
|
465 |
+
},
|
466 |
+
{
|
467 |
+
"epoch": 0.5505505505505506,
|
468 |
+
"grad_norm": 0.0008765487000346184,
|
469 |
+
"learning_rate": 4.5365220615498706e-05,
|
470 |
+
"loss": 0.0018,
|
471 |
+
"step": 1650
|
472 |
+
},
|
473 |
+
{
|
474 |
+
"epoch": 0.5588922255588922,
|
475 |
+
"grad_norm": 0.0001437152095604688,
|
476 |
+
"learning_rate": 4.5210727969348656e-05,
|
477 |
+
"loss": 0.0026,
|
478 |
+
"step": 1675
|
479 |
+
},
|
480 |
+
{
|
481 |
+
"epoch": 0.5672339005672339,
|
482 |
+
"grad_norm": 0.00043227593414485455,
|
483 |
+
"learning_rate": 4.505623532319862e-05,
|
484 |
+
"loss": 0.0001,
|
485 |
+
"step": 1700
|
486 |
+
},
|
487 |
+
{
|
488 |
+
"epoch": 0.5755755755755756,
|
489 |
+
"grad_norm": 0.00048412373871542513,
|
490 |
+
"learning_rate": 4.490174267704857e-05,
|
491 |
+
"loss": 0.0164,
|
492 |
+
"step": 1725
|
493 |
+
},
|
494 |
+
{
|
495 |
+
"epoch": 0.5839172505839173,
|
496 |
+
"grad_norm": 7.867912769317627,
|
497 |
+
"learning_rate": 4.474725003089853e-05,
|
498 |
+
"loss": 0.0548,
|
499 |
+
"step": 1750
|
500 |
+
},
|
501 |
+
{
|
502 |
+
"epoch": 0.5922589255922589,
|
503 |
+
"grad_norm": 5.28706169128418,
|
504 |
+
"learning_rate": 4.459275738474849e-05,
|
505 |
+
"loss": 0.0287,
|
506 |
+
"step": 1775
|
507 |
+
},
|
508 |
+
{
|
509 |
+
"epoch": 0.6006006006006006,
|
510 |
+
"grad_norm": 0.005814212840050459,
|
511 |
+
"learning_rate": 4.4438264738598445e-05,
|
512 |
+
"loss": 0.0353,
|
513 |
+
"step": 1800
|
514 |
+
},
|
515 |
+
{
|
516 |
+
"epoch": 0.6089422756089423,
|
517 |
+
"grad_norm": 0.00043659639777615666,
|
518 |
+
"learning_rate": 4.42837720924484e-05,
|
519 |
+
"loss": 0.0005,
|
520 |
+
"step": 1825
|
521 |
+
},
|
522 |
+
{
|
523 |
+
"epoch": 0.6172839506172839,
|
524 |
+
"grad_norm": 0.009556015022099018,
|
525 |
+
"learning_rate": 4.412927944629836e-05,
|
526 |
+
"loss": 0.0004,
|
527 |
+
"step": 1850
|
528 |
+
},
|
529 |
+
{
|
530 |
+
"epoch": 0.6256256256256256,
|
531 |
+
"grad_norm": 0.0006456194678321481,
|
532 |
+
"learning_rate": 4.397478680014832e-05,
|
533 |
+
"loss": 0.0003,
|
534 |
+
"step": 1875
|
535 |
+
},
|
536 |
+
{
|
537 |
+
"epoch": 0.6339673006339673,
|
538 |
+
"grad_norm": 0.0007494451128877699,
|
539 |
+
"learning_rate": 4.382029415399827e-05,
|
540 |
+
"loss": 0.0001,
|
541 |
+
"step": 1900
|
542 |
+
},
|
543 |
+
{
|
544 |
+
"epoch": 0.6423089756423089,
|
545 |
+
"grad_norm": 0.00023574185615871102,
|
546 |
+
"learning_rate": 4.366580150784823e-05,
|
547 |
+
"loss": 0.028,
|
548 |
+
"step": 1925
|
549 |
+
},
|
550 |
+
{
|
551 |
+
"epoch": 0.6506506506506506,
|
552 |
+
"grad_norm": 0.0008282416383735836,
|
553 |
+
"learning_rate": 4.3511308861698184e-05,
|
554 |
+
"loss": 0.0181,
|
555 |
+
"step": 1950
|
556 |
+
},
|
557 |
+
{
|
558 |
+
"epoch": 0.6589923256589924,
|
559 |
+
"grad_norm": 0.0017523517599329352,
|
560 |
+
"learning_rate": 4.335681621554814e-05,
|
561 |
+
"loss": 0.0761,
|
562 |
+
"step": 1975
|
563 |
+
},
|
564 |
+
{
|
565 |
+
"epoch": 0.667334000667334,
|
566 |
+
"grad_norm": 0.0005382935050874949,
|
567 |
+
"learning_rate": 4.32023235693981e-05,
|
568 |
+
"loss": 0.0007,
|
569 |
+
"step": 2000
|
570 |
+
},
|
571 |
+
{
|
572 |
+
"epoch": 0.6756756756756757,
|
573 |
+
"grad_norm": 0.0023508896119892597,
|
574 |
+
"learning_rate": 4.304783092324806e-05,
|
575 |
+
"loss": 0.0001,
|
576 |
+
"step": 2025
|
577 |
+
},
|
578 |
+
{
|
579 |
+
"epoch": 0.6840173506840174,
|
580 |
+
"grad_norm": 0.02000766433775425,
|
581 |
+
"learning_rate": 4.289333827709801e-05,
|
582 |
+
"loss": 0.0461,
|
583 |
+
"step": 2050
|
584 |
+
},
|
585 |
+
{
|
586 |
+
"epoch": 0.6923590256923591,
|
587 |
+
"grad_norm": 0.0016236711526289582,
|
588 |
+
"learning_rate": 4.273884563094797e-05,
|
589 |
+
"loss": 0.0015,
|
590 |
+
"step": 2075
|
591 |
+
},
|
592 |
+
{
|
593 |
+
"epoch": 0.7007007007007007,
|
594 |
+
"grad_norm": 0.01102663204073906,
|
595 |
+
"learning_rate": 4.2584352984797923e-05,
|
596 |
+
"loss": 0.0,
|
597 |
+
"step": 2100
|
598 |
+
},
|
599 |
+
{
|
600 |
+
"epoch": 0.7090423757090424,
|
601 |
+
"grad_norm": 7.1402587890625,
|
602 |
+
"learning_rate": 4.242986033864788e-05,
|
603 |
+
"loss": 0.0458,
|
604 |
+
"step": 2125
|
605 |
+
},
|
606 |
+
{
|
607 |
+
"epoch": 0.7173840507173841,
|
608 |
+
"grad_norm": 0.029623601585626602,
|
609 |
+
"learning_rate": 4.227536769249784e-05,
|
610 |
+
"loss": 0.0265,
|
611 |
+
"step": 2150
|
612 |
+
},
|
613 |
+
{
|
614 |
+
"epoch": 0.7257257257257257,
|
615 |
+
"grad_norm": 0.04872233048081398,
|
616 |
+
"learning_rate": 4.2120875046347796e-05,
|
617 |
+
"loss": 0.0418,
|
618 |
+
"step": 2175
|
619 |
+
},
|
620 |
+
{
|
621 |
+
"epoch": 0.7340674007340674,
|
622 |
+
"grad_norm": 0.01899750903248787,
|
623 |
+
"learning_rate": 4.1966382400197754e-05,
|
624 |
+
"loss": 0.0797,
|
625 |
+
"step": 2200
|
626 |
+
},
|
627 |
+
{
|
628 |
+
"epoch": 0.7424090757424091,
|
629 |
+
"grad_norm": 0.014597664587199688,
|
630 |
+
"learning_rate": 4.181188975404771e-05,
|
631 |
+
"loss": 0.0239,
|
632 |
+
"step": 2225
|
633 |
+
},
|
634 |
+
{
|
635 |
+
"epoch": 0.7507507507507507,
|
636 |
+
"grad_norm": 0.001034777145832777,
|
637 |
+
"learning_rate": 4.165739710789767e-05,
|
638 |
+
"loss": 0.0067,
|
639 |
+
"step": 2250
|
640 |
+
},
|
641 |
+
{
|
642 |
+
"epoch": 0.7590924257590924,
|
643 |
+
"grad_norm": 0.0015550162643194199,
|
644 |
+
"learning_rate": 4.150290446174762e-05,
|
645 |
+
"loss": 0.0482,
|
646 |
+
"step": 2275
|
647 |
+
},
|
648 |
+
{
|
649 |
+
"epoch": 0.7674341007674341,
|
650 |
+
"grad_norm": 0.013912219554185867,
|
651 |
+
"learning_rate": 4.1348411815597585e-05,
|
652 |
+
"loss": 0.0248,
|
653 |
+
"step": 2300
|
654 |
+
},
|
655 |
+
{
|
656 |
+
"epoch": 0.7757757757757757,
|
657 |
+
"grad_norm": 0.0020062343683093786,
|
658 |
+
"learning_rate": 4.1193919169447536e-05,
|
659 |
+
"loss": 0.0156,
|
660 |
+
"step": 2325
|
661 |
+
},
|
662 |
+
{
|
663 |
+
"epoch": 0.7841174507841174,
|
664 |
+
"grad_norm": 6.4558515548706055,
|
665 |
+
"learning_rate": 4.1039426523297493e-05,
|
666 |
+
"loss": 0.0581,
|
667 |
+
"step": 2350
|
668 |
+
},
|
669 |
+
{
|
670 |
+
"epoch": 0.7924591257924591,
|
671 |
+
"grad_norm": 0.020143140107393265,
|
672 |
+
"learning_rate": 4.088493387714745e-05,
|
673 |
+
"loss": 0.0033,
|
674 |
+
"step": 2375
|
675 |
+
},
|
676 |
+
{
|
677 |
+
"epoch": 0.8008008008008008,
|
678 |
+
"grad_norm": 0.00565760675817728,
|
679 |
+
"learning_rate": 4.073044123099741e-05,
|
680 |
+
"loss": 0.0003,
|
681 |
+
"step": 2400
|
682 |
+
},
|
683 |
+
{
|
684 |
+
"epoch": 0.8091424758091424,
|
685 |
+
"grad_norm": 0.0011780583299696445,
|
686 |
+
"learning_rate": 4.057594858484736e-05,
|
687 |
+
"loss": 0.0003,
|
688 |
+
"step": 2425
|
689 |
+
},
|
690 |
+
{
|
691 |
+
"epoch": 0.8174841508174842,
|
692 |
+
"grad_norm": 0.001397665124386549,
|
693 |
+
"learning_rate": 4.0421455938697324e-05,
|
694 |
+
"loss": 0.0033,
|
695 |
+
"step": 2450
|
696 |
+
},
|
697 |
+
{
|
698 |
+
"epoch": 0.8258258258258259,
|
699 |
+
"grad_norm": 0.11108289659023285,
|
700 |
+
"learning_rate": 4.0266963292547275e-05,
|
701 |
+
"loss": 0.0693,
|
702 |
+
"step": 2475
|
703 |
+
},
|
704 |
+
{
|
705 |
+
"epoch": 0.8341675008341675,
|
706 |
+
"grad_norm": 0.0012789678294211626,
|
707 |
+
"learning_rate": 4.011247064639723e-05,
|
708 |
+
"loss": 0.0012,
|
709 |
+
"step": 2500
|
710 |
+
},
|
711 |
+
{
|
712 |
+
"epoch": 0.8425091758425092,
|
713 |
+
"grad_norm": 0.0016298526898026466,
|
714 |
+
"learning_rate": 3.995797800024719e-05,
|
715 |
+
"loss": 0.0001,
|
716 |
+
"step": 2525
|
717 |
+
},
|
718 |
+
{
|
719 |
+
"epoch": 0.8508508508508509,
|
720 |
+
"grad_norm": 0.0012462561717256904,
|
721 |
+
"learning_rate": 3.980348535409715e-05,
|
722 |
+
"loss": 0.0,
|
723 |
+
"step": 2550
|
724 |
+
},
|
725 |
+
{
|
726 |
+
"epoch": 0.8591925258591925,
|
727 |
+
"grad_norm": 0.0011716202134266496,
|
728 |
+
"learning_rate": 3.9648992707947106e-05,
|
729 |
+
"loss": 0.0008,
|
730 |
+
"step": 2575
|
731 |
+
},
|
732 |
+
{
|
733 |
+
"epoch": 0.8675342008675342,
|
734 |
+
"grad_norm": 0.0004461667558643967,
|
735 |
+
"learning_rate": 3.9494500061797063e-05,
|
736 |
+
"loss": 0.0013,
|
737 |
+
"step": 2600
|
738 |
+
},
|
739 |
+
{
|
740 |
+
"epoch": 0.8758758758758759,
|
741 |
+
"grad_norm": 0.004892207216471434,
|
742 |
+
"learning_rate": 3.934000741564702e-05,
|
743 |
+
"loss": 0.0444,
|
744 |
+
"step": 2625
|
745 |
+
},
|
746 |
+
{
|
747 |
+
"epoch": 0.8842175508842175,
|
748 |
+
"grad_norm": 0.0013353752437978983,
|
749 |
+
"learning_rate": 3.918551476949697e-05,
|
750 |
+
"loss": 0.0014,
|
751 |
+
"step": 2650
|
752 |
+
},
|
753 |
+
{
|
754 |
+
"epoch": 0.8925592258925592,
|
755 |
+
"grad_norm": 0.05706316977739334,
|
756 |
+
"learning_rate": 3.9031022123346936e-05,
|
757 |
+
"loss": 0.1006,
|
758 |
+
"step": 2675
|
759 |
+
},
|
760 |
+
{
|
761 |
+
"epoch": 0.9009009009009009,
|
762 |
+
"grad_norm": 0.0032118011731654406,
|
763 |
+
"learning_rate": 3.887652947719689e-05,
|
764 |
+
"loss": 0.0004,
|
765 |
+
"step": 2700
|
766 |
+
},
|
767 |
+
{
|
768 |
+
"epoch": 0.9092425759092426,
|
769 |
+
"grad_norm": 0.002578709740191698,
|
770 |
+
"learning_rate": 3.8722036831046845e-05,
|
771 |
+
"loss": 0.0049,
|
772 |
+
"step": 2725
|
773 |
+
},
|
774 |
+
{
|
775 |
+
"epoch": 0.9175842509175842,
|
776 |
+
"grad_norm": 0.007876419462263584,
|
777 |
+
"learning_rate": 3.85675441848968e-05,
|
778 |
+
"loss": 0.0393,
|
779 |
+
"step": 2750
|
780 |
+
},
|
781 |
+
{
|
782 |
+
"epoch": 0.9259259259259259,
|
783 |
+
"grad_norm": 0.01825695112347603,
|
784 |
+
"learning_rate": 3.841305153874676e-05,
|
785 |
+
"loss": 0.0313,
|
786 |
+
"step": 2775
|
787 |
+
},
|
788 |
+
{
|
789 |
+
"epoch": 0.9342676009342676,
|
790 |
+
"grad_norm": 0.018400903791189194,
|
791 |
+
"learning_rate": 3.825855889259671e-05,
|
792 |
+
"loss": 0.0362,
|
793 |
+
"step": 2800
|
794 |
+
},
|
795 |
+
{
|
796 |
+
"epoch": 0.9426092759426092,
|
797 |
+
"grad_norm": 9.522905349731445,
|
798 |
+
"learning_rate": 3.8104066246446676e-05,
|
799 |
+
"loss": 0.0044,
|
800 |
+
"step": 2825
|
801 |
+
},
|
802 |
+
{
|
803 |
+
"epoch": 0.950950950950951,
|
804 |
+
"grad_norm": 0.055028580129146576,
|
805 |
+
"learning_rate": 3.794957360029663e-05,
|
806 |
+
"loss": 0.0046,
|
807 |
+
"step": 2850
|
808 |
+
},
|
809 |
+
{
|
810 |
+
"epoch": 0.9592926259592927,
|
811 |
+
"grad_norm": 0.0008863616967573762,
|
812 |
+
"learning_rate": 3.7795080954146584e-05,
|
813 |
+
"loss": 0.0007,
|
814 |
+
"step": 2875
|
815 |
+
},
|
816 |
+
{
|
817 |
+
"epoch": 0.9676343009676343,
|
818 |
+
"grad_norm": 0.001261973986402154,
|
819 |
+
"learning_rate": 3.764058830799654e-05,
|
820 |
+
"loss": 0.0001,
|
821 |
+
"step": 2900
|
822 |
+
},
|
823 |
+
{
|
824 |
+
"epoch": 0.975975975975976,
|
825 |
+
"grad_norm": 0.04143986105918884,
|
826 |
+
"learning_rate": 3.74860956618465e-05,
|
827 |
+
"loss": 0.0871,
|
828 |
+
"step": 2925
|
829 |
+
},
|
830 |
+
{
|
831 |
+
"epoch": 0.9843176509843177,
|
832 |
+
"grad_norm": 0.0786428451538086,
|
833 |
+
"learning_rate": 3.733160301569645e-05,
|
834 |
+
"loss": 0.01,
|
835 |
+
"step": 2950
|
836 |
+
},
|
837 |
+
{
|
838 |
+
"epoch": 0.9926593259926594,
|
839 |
+
"grad_norm": 0.0012078924337401986,
|
840 |
+
"learning_rate": 3.7177110369546415e-05,
|
841 |
+
"loss": 0.0392,
|
842 |
+
"step": 2975
|
843 |
+
},
|
844 |
+
{
|
845 |
+
"epoch": 1.0,
|
846 |
+
"eval_accuracy": 0.9978311644978312,
|
847 |
+
"eval_auc": 0.9945649196493791,
|
848 |
+
"eval_f1": 0.9078014184397163,
|
849 |
+
"eval_loss": 0.009599411860108376,
|
850 |
+
"eval_precision": 0.9552238805970149,
|
851 |
+
"eval_recall": 0.8648648648648649,
|
852 |
+
"eval_runtime": 120.7986,
|
853 |
+
"eval_samples_per_second": 49.62,
|
854 |
+
"eval_steps_per_second": 3.104,
|
855 |
+
"step": 2997
|
856 |
+
},
|
857 |
+
{
|
858 |
+
"epoch": 1.001001001001001,
|
859 |
+
"grad_norm": 0.006058184430003166,
|
860 |
+
"learning_rate": 3.7022617723396366e-05,
|
861 |
+
"loss": 0.041,
|
862 |
+
"step": 3000
|
863 |
+
},
|
864 |
+
{
|
865 |
+
"epoch": 1.0093426760093427,
|
866 |
+
"grad_norm": 0.003738154424354434,
|
867 |
+
"learning_rate": 3.6868125077246324e-05,
|
868 |
+
"loss": 0.0005,
|
869 |
+
"step": 3025
|
870 |
+
},
|
871 |
+
{
|
872 |
+
"epoch": 1.0176843510176843,
|
873 |
+
"grad_norm": 0.00978489127010107,
|
874 |
+
"learning_rate": 3.671363243109628e-05,
|
875 |
+
"loss": 0.0452,
|
876 |
+
"step": 3050
|
877 |
+
},
|
878 |
+
{
|
879 |
+
"epoch": 1.026026026026026,
|
880 |
+
"grad_norm": 0.01102358940988779,
|
881 |
+
"learning_rate": 3.655913978494624e-05,
|
882 |
+
"loss": 0.0005,
|
883 |
+
"step": 3075
|
884 |
+
},
|
885 |
+
{
|
886 |
+
"epoch": 1.0343677010343677,
|
887 |
+
"grad_norm": 0.008122111670672894,
|
888 |
+
"learning_rate": 3.64046471387962e-05,
|
889 |
+
"loss": 0.0003,
|
890 |
+
"step": 3100
|
891 |
+
},
|
892 |
+
{
|
893 |
+
"epoch": 1.0427093760427093,
|
894 |
+
"grad_norm": 0.0037865727208554745,
|
895 |
+
"learning_rate": 3.6250154492646154e-05,
|
896 |
+
"loss": 0.0046,
|
897 |
+
"step": 3125
|
898 |
+
},
|
899 |
+
{
|
900 |
+
"epoch": 1.0510510510510511,
|
901 |
+
"grad_norm": 0.0031646236311644316,
|
902 |
+
"learning_rate": 3.609566184649611e-05,
|
903 |
+
"loss": 0.0005,
|
904 |
+
"step": 3150
|
905 |
+
},
|
906 |
+
{
|
907 |
+
"epoch": 1.0593927260593927,
|
908 |
+
"grad_norm": 0.0017720437608659267,
|
909 |
+
"learning_rate": 3.594116920034606e-05,
|
910 |
+
"loss": 0.0001,
|
911 |
+
"step": 3175
|
912 |
+
},
|
913 |
+
{
|
914 |
+
"epoch": 1.0677344010677343,
|
915 |
+
"grad_norm": 0.011880212463438511,
|
916 |
+
"learning_rate": 3.578667655419603e-05,
|
917 |
+
"loss": 0.0286,
|
918 |
+
"step": 3200
|
919 |
+
},
|
920 |
+
{
|
921 |
+
"epoch": 1.0760760760760761,
|
922 |
+
"grad_norm": 0.00772507768124342,
|
923 |
+
"learning_rate": 3.563218390804598e-05,
|
924 |
+
"loss": 0.0001,
|
925 |
+
"step": 3225
|
926 |
+
},
|
927 |
+
{
|
928 |
+
"epoch": 1.0844177510844177,
|
929 |
+
"grad_norm": 0.012478929944336414,
|
930 |
+
"learning_rate": 3.5477691261895936e-05,
|
931 |
+
"loss": 0.0001,
|
932 |
+
"step": 3250
|
933 |
+
},
|
934 |
+
{
|
935 |
+
"epoch": 1.0927594260927593,
|
936 |
+
"grad_norm": 0.004284675233066082,
|
937 |
+
"learning_rate": 3.5323198615745894e-05,
|
938 |
+
"loss": 0.0001,
|
939 |
+
"step": 3275
|
940 |
+
},
|
941 |
+
{
|
942 |
+
"epoch": 1.1011011011011012,
|
943 |
+
"grad_norm": 0.0012670083669945598,
|
944 |
+
"learning_rate": 3.516870596959585e-05,
|
945 |
+
"loss": 0.0001,
|
946 |
+
"step": 3300
|
947 |
+
},
|
948 |
+
{
|
949 |
+
"epoch": 1.1094427761094428,
|
950 |
+
"grad_norm": 0.0011124334996566176,
|
951 |
+
"learning_rate": 3.50142133234458e-05,
|
952 |
+
"loss": 0.0381,
|
953 |
+
"step": 3325
|
954 |
+
},
|
955 |
+
{
|
956 |
+
"epoch": 1.1177844511177844,
|
957 |
+
"grad_norm": 0.35427987575531006,
|
958 |
+
"learning_rate": 3.485972067729576e-05,
|
959 |
+
"loss": 0.0306,
|
960 |
+
"step": 3350
|
961 |
+
},
|
962 |
+
{
|
963 |
+
"epoch": 1.1261261261261262,
|
964 |
+
"grad_norm": 0.0010480162454769015,
|
965 |
+
"learning_rate": 3.470522803114572e-05,
|
966 |
+
"loss": 0.0002,
|
967 |
+
"step": 3375
|
968 |
+
},
|
969 |
+
{
|
970 |
+
"epoch": 1.1344678011344678,
|
971 |
+
"grad_norm": 0.01626741513609886,
|
972 |
+
"learning_rate": 3.4550735384995675e-05,
|
973 |
+
"loss": 0.0049,
|
974 |
+
"step": 3400
|
975 |
+
},
|
976 |
+
{
|
977 |
+
"epoch": 1.1428094761428094,
|
978 |
+
"grad_norm": 0.0009041897719725966,
|
979 |
+
"learning_rate": 3.439624273884563e-05,
|
980 |
+
"loss": 0.0033,
|
981 |
+
"step": 3425
|
982 |
+
},
|
983 |
+
{
|
984 |
+
"epoch": 1.1511511511511512,
|
985 |
+
"grad_norm": 0.005210678093135357,
|
986 |
+
"learning_rate": 3.424175009269559e-05,
|
987 |
+
"loss": 0.0385,
|
988 |
+
"step": 3450
|
989 |
+
},
|
990 |
+
{
|
991 |
+
"epoch": 1.1594928261594928,
|
992 |
+
"grad_norm": 0.00467864703387022,
|
993 |
+
"learning_rate": 3.408725744654555e-05,
|
994 |
+
"loss": 0.0001,
|
995 |
+
"step": 3475
|
996 |
+
},
|
997 |
+
{
|
998 |
+
"epoch": 1.1678345011678344,
|
999 |
+
"grad_norm": 0.019908057525753975,
|
1000 |
+
"learning_rate": 3.39327648003955e-05,
|
1001 |
+
"loss": 0.0001,
|
1002 |
+
"step": 3500
|
1003 |
+
},
|
1004 |
+
{
|
1005 |
+
"epoch": 1.1761761761761762,
|
1006 |
+
"grad_norm": 0.0005183572066016495,
|
1007 |
+
"learning_rate": 3.3778272154245464e-05,
|
1008 |
+
"loss": 0.0004,
|
1009 |
+
"step": 3525
|
1010 |
+
},
|
1011 |
+
{
|
1012 |
+
"epoch": 1.1845178511845178,
|
1013 |
+
"grad_norm": 0.00818272028118372,
|
1014 |
+
"learning_rate": 3.3623779508095414e-05,
|
1015 |
+
"loss": 0.0454,
|
1016 |
+
"step": 3550
|
1017 |
+
},
|
1018 |
+
{
|
1019 |
+
"epoch": 1.1928595261928596,
|
1020 |
+
"grad_norm": 0.0328977033495903,
|
1021 |
+
"learning_rate": 3.346928686194537e-05,
|
1022 |
+
"loss": 0.0003,
|
1023 |
+
"step": 3575
|
1024 |
+
},
|
1025 |
+
{
|
1026 |
+
"epoch": 1.2012012012012012,
|
1027 |
+
"grad_norm": 6.452710151672363,
|
1028 |
+
"learning_rate": 3.331479421579533e-05,
|
1029 |
+
"loss": 0.0432,
|
1030 |
+
"step": 3600
|
1031 |
+
},
|
1032 |
+
{
|
1033 |
+
"epoch": 1.2095428762095428,
|
1034 |
+
"grad_norm": 0.005078097339719534,
|
1035 |
+
"learning_rate": 3.316030156964529e-05,
|
1036 |
+
"loss": 0.0322,
|
1037 |
+
"step": 3625
|
1038 |
+
},
|
1039 |
+
{
|
1040 |
+
"epoch": 1.2178845512178846,
|
1041 |
+
"grad_norm": 0.0003529058594722301,
|
1042 |
+
"learning_rate": 3.300580892349524e-05,
|
1043 |
+
"loss": 0.0003,
|
1044 |
+
"step": 3650
|
1045 |
+
},
|
1046 |
+
{
|
1047 |
+
"epoch": 1.2262262262262262,
|
1048 |
+
"grad_norm": 0.0013959240168333054,
|
1049 |
+
"learning_rate": 3.28513162773452e-05,
|
1050 |
+
"loss": 0.0849,
|
1051 |
+
"step": 3675
|
1052 |
+
},
|
1053 |
+
{
|
1054 |
+
"epoch": 1.2345679012345678,
|
1055 |
+
"grad_norm": 0.016115982085466385,
|
1056 |
+
"learning_rate": 3.2696823631195154e-05,
|
1057 |
+
"loss": 0.056,
|
1058 |
+
"step": 3700
|
1059 |
+
},
|
1060 |
+
{
|
1061 |
+
"epoch": 1.2429095762429097,
|
1062 |
+
"grad_norm": 0.0022523789666593075,
|
1063 |
+
"learning_rate": 3.254233098504511e-05,
|
1064 |
+
"loss": 0.0002,
|
1065 |
+
"step": 3725
|
1066 |
+
},
|
1067 |
+
{
|
1068 |
+
"epoch": 1.2512512512512513,
|
1069 |
+
"grad_norm": 0.016599655151367188,
|
1070 |
+
"learning_rate": 3.238783833889507e-05,
|
1071 |
+
"loss": 0.0319,
|
1072 |
+
"step": 3750
|
1073 |
+
},
|
1074 |
+
{
|
1075 |
+
"epoch": 1.2595929262595928,
|
1076 |
+
"grad_norm": 0.0018683940870687366,
|
1077 |
+
"learning_rate": 3.223334569274503e-05,
|
1078 |
+
"loss": 0.0227,
|
1079 |
+
"step": 3775
|
1080 |
+
},
|
1081 |
+
{
|
1082 |
+
"epoch": 1.2679346012679347,
|
1083 |
+
"grad_norm": 0.002231100806966424,
|
1084 |
+
"learning_rate": 3.207885304659498e-05,
|
1085 |
+
"loss": 0.0001,
|
1086 |
+
"step": 3800
|
1087 |
+
},
|
1088 |
+
{
|
1089 |
+
"epoch": 1.2762762762762763,
|
1090 |
+
"grad_norm": 0.005379370413720608,
|
1091 |
+
"learning_rate": 3.192436040044494e-05,
|
1092 |
+
"loss": 0.0439,
|
1093 |
+
"step": 3825
|
1094 |
+
},
|
1095 |
+
{
|
1096 |
+
"epoch": 1.284617951284618,
|
1097 |
+
"grad_norm": 0.003467306261882186,
|
1098 |
+
"learning_rate": 3.176986775429489e-05,
|
1099 |
+
"loss": 0.0025,
|
1100 |
+
"step": 3850
|
1101 |
+
},
|
1102 |
+
{
|
1103 |
+
"epoch": 1.2929596262929597,
|
1104 |
+
"grad_norm": 0.0015170868718996644,
|
1105 |
+
"learning_rate": 3.161537510814485e-05,
|
1106 |
+
"loss": 0.0037,
|
1107 |
+
"step": 3875
|
1108 |
+
},
|
1109 |
+
{
|
1110 |
+
"epoch": 1.3013013013013013,
|
1111 |
+
"grad_norm": 0.00604633055627346,
|
1112 |
+
"learning_rate": 3.146088246199481e-05,
|
1113 |
+
"loss": 0.0001,
|
1114 |
+
"step": 3900
|
1115 |
+
},
|
1116 |
+
{
|
1117 |
+
"epoch": 1.309642976309643,
|
1118 |
+
"grad_norm": 10.07129955291748,
|
1119 |
+
"learning_rate": 3.1306389815844766e-05,
|
1120 |
+
"loss": 0.0232,
|
1121 |
+
"step": 3925
|
1122 |
+
},
|
1123 |
+
{
|
1124 |
+
"epoch": 1.3179846513179847,
|
1125 |
+
"grad_norm": 0.0015994218410924077,
|
1126 |
+
"learning_rate": 3.1151897169694724e-05,
|
1127 |
+
"loss": 0.0004,
|
1128 |
+
"step": 3950
|
1129 |
+
},
|
1130 |
+
{
|
1131 |
+
"epoch": 1.3263263263263263,
|
1132 |
+
"grad_norm": 0.009616601280868053,
|
1133 |
+
"learning_rate": 3.099740452354468e-05,
|
1134 |
+
"loss": 0.0165,
|
1135 |
+
"step": 3975
|
1136 |
+
},
|
1137 |
+
{
|
1138 |
+
"epoch": 1.3346680013346681,
|
1139 |
+
"grad_norm": 0.009371902793645859,
|
1140 |
+
"learning_rate": 3.084291187739464e-05,
|
1141 |
+
"loss": 0.0008,
|
1142 |
+
"step": 4000
|
1143 |
+
},
|
1144 |
+
{
|
1145 |
+
"epoch": 1.3430096763430097,
|
1146 |
+
"grad_norm": 0.3239502012729645,
|
1147 |
+
"learning_rate": 3.068841923124459e-05,
|
1148 |
+
"loss": 0.0026,
|
1149 |
+
"step": 4025
|
1150 |
+
},
|
1151 |
+
{
|
1152 |
+
"epoch": 1.3513513513513513,
|
1153 |
+
"grad_norm": 0.00045479455729946494,
|
1154 |
+
"learning_rate": 3.0533926585094554e-05,
|
1155 |
+
"loss": 0.0002,
|
1156 |
+
"step": 4050
|
1157 |
+
},
|
1158 |
+
{
|
1159 |
+
"epoch": 1.3596930263596931,
|
1160 |
+
"grad_norm": 6.515894889831543,
|
1161 |
+
"learning_rate": 3.0379433938944505e-05,
|
1162 |
+
"loss": 0.0591,
|
1163 |
+
"step": 4075
|
1164 |
+
},
|
1165 |
+
{
|
1166 |
+
"epoch": 1.3680347013680347,
|
1167 |
+
"grad_norm": 0.0006543731433339417,
|
1168 |
+
"learning_rate": 3.0224941292794463e-05,
|
1169 |
+
"loss": 0.0527,
|
1170 |
+
"step": 4100
|
1171 |
+
},
|
1172 |
+
{
|
1173 |
+
"epoch": 1.3763763763763763,
|
1174 |
+
"grad_norm": 0.0006008436321280897,
|
1175 |
+
"learning_rate": 3.007044864664442e-05,
|
1176 |
+
"loss": 0.0278,
|
1177 |
+
"step": 4125
|
1178 |
+
},
|
1179 |
+
{
|
1180 |
+
"epoch": 1.3847180513847182,
|
1181 |
+
"grad_norm": 0.02133520320057869,
|
1182 |
+
"learning_rate": 2.991595600049438e-05,
|
1183 |
+
"loss": 0.0002,
|
1184 |
+
"step": 4150
|
1185 |
+
},
|
1186 |
+
{
|
1187 |
+
"epoch": 1.3930597263930597,
|
1188 |
+
"grad_norm": 0.0017505575669929385,
|
1189 |
+
"learning_rate": 2.9761463354344333e-05,
|
1190 |
+
"loss": 0.0019,
|
1191 |
+
"step": 4175
|
1192 |
+
},
|
1193 |
+
{
|
1194 |
+
"epoch": 1.4014014014014013,
|
1195 |
+
"grad_norm": 0.0024641742929816246,
|
1196 |
+
"learning_rate": 2.9606970708194294e-05,
|
1197 |
+
"loss": 0.0001,
|
1198 |
+
"step": 4200
|
1199 |
+
},
|
1200 |
+
{
|
1201 |
+
"epoch": 1.4097430764097432,
|
1202 |
+
"grad_norm": 0.0026479989755898714,
|
1203 |
+
"learning_rate": 2.9452478062044248e-05,
|
1204 |
+
"loss": 0.0005,
|
1205 |
+
"step": 4225
|
1206 |
+
},
|
1207 |
+
{
|
1208 |
+
"epoch": 1.4180847514180848,
|
1209 |
+
"grad_norm": 0.0036438798997551203,
|
1210 |
+
"learning_rate": 2.9297985415894202e-05,
|
1211 |
+
"loss": 0.0001,
|
1212 |
+
"step": 4250
|
1213 |
+
},
|
1214 |
+
{
|
1215 |
+
"epoch": 1.4264264264264264,
|
1216 |
+
"grad_norm": 0.005783462896943092,
|
1217 |
+
"learning_rate": 2.9143492769744163e-05,
|
1218 |
+
"loss": 0.0686,
|
1219 |
+
"step": 4275
|
1220 |
+
},
|
1221 |
+
{
|
1222 |
+
"epoch": 1.4347681014347682,
|
1223 |
+
"grad_norm": 0.024165183305740356,
|
1224 |
+
"learning_rate": 2.8989000123594118e-05,
|
1225 |
+
"loss": 0.0198,
|
1226 |
+
"step": 4300
|
1227 |
+
},
|
1228 |
+
{
|
1229 |
+
"epoch": 1.4431097764431098,
|
1230 |
+
"grad_norm": 0.04456571117043495,
|
1231 |
+
"learning_rate": 2.8834507477444072e-05,
|
1232 |
+
"loss": 0.0007,
|
1233 |
+
"step": 4325
|
1234 |
+
},
|
1235 |
+
{
|
1236 |
+
"epoch": 1.4514514514514514,
|
1237 |
+
"grad_norm": 0.0014297548914328218,
|
1238 |
+
"learning_rate": 2.8680014831294033e-05,
|
1239 |
+
"loss": 0.0009,
|
1240 |
+
"step": 4350
|
1241 |
+
},
|
1242 |
+
{
|
1243 |
+
"epoch": 1.4597931264597932,
|
1244 |
+
"grad_norm": 0.002099397825077176,
|
1245 |
+
"learning_rate": 2.8525522185143987e-05,
|
1246 |
+
"loss": 0.0116,
|
1247 |
+
"step": 4375
|
1248 |
+
},
|
1249 |
+
{
|
1250 |
+
"epoch": 1.4681348014681348,
|
1251 |
+
"grad_norm": 0.0008780402131378651,
|
1252 |
+
"learning_rate": 2.837102953899394e-05,
|
1253 |
+
"loss": 0.0001,
|
1254 |
+
"step": 4400
|
1255 |
+
},
|
1256 |
+
{
|
1257 |
+
"epoch": 1.4764764764764764,
|
1258 |
+
"grad_norm": 0.00024115787527989596,
|
1259 |
+
"learning_rate": 2.8216536892843903e-05,
|
1260 |
+
"loss": 0.0216,
|
1261 |
+
"step": 4425
|
1262 |
+
},
|
1263 |
+
{
|
1264 |
+
"epoch": 1.4848181514848182,
|
1265 |
+
"grad_norm": 0.0003840152348857373,
|
1266 |
+
"learning_rate": 2.8062044246693857e-05,
|
1267 |
+
"loss": 0.0361,
|
1268 |
+
"step": 4450
|
1269 |
+
},
|
1270 |
+
{
|
1271 |
+
"epoch": 1.4931598264931598,
|
1272 |
+
"grad_norm": 0.0005036329384893179,
|
1273 |
+
"learning_rate": 2.7907551600543815e-05,
|
1274 |
+
"loss": 0.0001,
|
1275 |
+
"step": 4475
|
1276 |
+
},
|
1277 |
+
{
|
1278 |
+
"epoch": 1.5015015015015014,
|
1279 |
+
"grad_norm": 0.0002450717438478023,
|
1280 |
+
"learning_rate": 2.7753058954393772e-05,
|
1281 |
+
"loss": 0.0001,
|
1282 |
+
"step": 4500
|
1283 |
+
},
|
1284 |
+
{
|
1285 |
+
"epoch": 1.5098431765098432,
|
1286 |
+
"grad_norm": 0.0007805937202647328,
|
1287 |
+
"learning_rate": 2.759856630824373e-05,
|
1288 |
+
"loss": 0.0001,
|
1289 |
+
"step": 4525
|
1290 |
+
},
|
1291 |
+
{
|
1292 |
+
"epoch": 1.5181848515181848,
|
1293 |
+
"grad_norm": 0.0013042781502008438,
|
1294 |
+
"learning_rate": 2.7444073662093684e-05,
|
1295 |
+
"loss": 0.0645,
|
1296 |
+
"step": 4550
|
1297 |
+
},
|
1298 |
+
{
|
1299 |
+
"epoch": 1.5265265265265264,
|
1300 |
+
"grad_norm": 0.0023273963015526533,
|
1301 |
+
"learning_rate": 2.7289581015943645e-05,
|
1302 |
+
"loss": 0.0016,
|
1303 |
+
"step": 4575
|
1304 |
+
},
|
1305 |
+
{
|
1306 |
+
"epoch": 1.5348682015348682,
|
1307 |
+
"grad_norm": 0.0012803663266822696,
|
1308 |
+
"learning_rate": 2.71350883697936e-05,
|
1309 |
+
"loss": 0.0001,
|
1310 |
+
"step": 4600
|
1311 |
+
},
|
1312 |
+
{
|
1313 |
+
"epoch": 1.5432098765432098,
|
1314 |
+
"grad_norm": 0.0039305612444877625,
|
1315 |
+
"learning_rate": 2.6980595723643554e-05,
|
1316 |
+
"loss": 0.0001,
|
1317 |
+
"step": 4625
|
1318 |
+
},
|
1319 |
+
{
|
1320 |
+
"epoch": 1.5515515515515514,
|
1321 |
+
"grad_norm": 0.0023296186700463295,
|
1322 |
+
"learning_rate": 2.6826103077493515e-05,
|
1323 |
+
"loss": 0.0001,
|
1324 |
+
"step": 4650
|
1325 |
+
},
|
1326 |
+
{
|
1327 |
+
"epoch": 1.5598932265598933,
|
1328 |
+
"grad_norm": 0.0008701834594830871,
|
1329 |
+
"learning_rate": 2.667161043134347e-05,
|
1330 |
+
"loss": 0.0306,
|
1331 |
+
"step": 4675
|
1332 |
+
},
|
1333 |
+
{
|
1334 |
+
"epoch": 1.5682349015682349,
|
1335 |
+
"grad_norm": 0.009118028916418552,
|
1336 |
+
"learning_rate": 2.6517117785193424e-05,
|
1337 |
+
"loss": 0.0004,
|
1338 |
+
"step": 4700
|
1339 |
+
},
|
1340 |
+
{
|
1341 |
+
"epoch": 1.5765765765765765,
|
1342 |
+
"grad_norm": 0.004671914968639612,
|
1343 |
+
"learning_rate": 2.6362625139043385e-05,
|
1344 |
+
"loss": 0.0003,
|
1345 |
+
"step": 4725
|
1346 |
+
},
|
1347 |
+
{
|
1348 |
+
"epoch": 1.5849182515849183,
|
1349 |
+
"grad_norm": 0.0014299708418548107,
|
1350 |
+
"learning_rate": 2.620813249289334e-05,
|
1351 |
+
"loss": 0.0001,
|
1352 |
+
"step": 4750
|
1353 |
+
},
|
1354 |
+
{
|
1355 |
+
"epoch": 1.5932599265932599,
|
1356 |
+
"grad_norm": 0.0011580680729821324,
|
1357 |
+
"learning_rate": 2.6053639846743293e-05,
|
1358 |
+
"loss": 0.0001,
|
1359 |
+
"step": 4775
|
1360 |
+
},
|
1361 |
+
{
|
1362 |
+
"epoch": 1.6016016016016015,
|
1363 |
+
"grad_norm": 0.0004150031309109181,
|
1364 |
+
"learning_rate": 2.5899147200593254e-05,
|
1365 |
+
"loss": 0.0201,
|
1366 |
+
"step": 4800
|
1367 |
+
},
|
1368 |
+
{
|
1369 |
+
"epoch": 1.6099432766099433,
|
1370 |
+
"grad_norm": 0.15517137944698334,
|
1371 |
+
"learning_rate": 2.574465455444321e-05,
|
1372 |
+
"loss": 0.0133,
|
1373 |
+
"step": 4825
|
1374 |
+
},
|
1375 |
+
{
|
1376 |
+
"epoch": 1.6182849516182851,
|
1377 |
+
"grad_norm": 0.0004611278127413243,
|
1378 |
+
"learning_rate": 2.5590161908293163e-05,
|
1379 |
+
"loss": 0.0004,
|
1380 |
+
"step": 4850
|
1381 |
+
},
|
1382 |
+
{
|
1383 |
+
"epoch": 1.6266266266266265,
|
1384 |
+
"grad_norm": 7.501170635223389,
|
1385 |
+
"learning_rate": 2.5435669262143124e-05,
|
1386 |
+
"loss": 0.0056,
|
1387 |
+
"step": 4875
|
1388 |
+
},
|
1389 |
+
{
|
1390 |
+
"epoch": 1.6349683016349683,
|
1391 |
+
"grad_norm": 0.0007478469051420689,
|
1392 |
+
"learning_rate": 2.5281176615993078e-05,
|
1393 |
+
"loss": 0.0002,
|
1394 |
+
"step": 4900
|
1395 |
+
},
|
1396 |
+
{
|
1397 |
+
"epoch": 1.6433099766433101,
|
1398 |
+
"grad_norm": 2.6761367321014404,
|
1399 |
+
"learning_rate": 2.5126683969843036e-05,
|
1400 |
+
"loss": 0.0319,
|
1401 |
+
"step": 4925
|
1402 |
+
},
|
1403 |
+
{
|
1404 |
+
"epoch": 1.6516516516516515,
|
1405 |
+
"grad_norm": 0.0006212076405063272,
|
1406 |
+
"learning_rate": 2.4972191323692994e-05,
|
1407 |
+
"loss": 0.0,
|
1408 |
+
"step": 4950
|
1409 |
+
},
|
1410 |
+
{
|
1411 |
+
"epoch": 1.6599933266599933,
|
1412 |
+
"grad_norm": 0.0011125955497846007,
|
1413 |
+
"learning_rate": 2.481769867754295e-05,
|
1414 |
+
"loss": 0.0001,
|
1415 |
+
"step": 4975
|
1416 |
+
},
|
1417 |
+
{
|
1418 |
+
"epoch": 1.6683350016683351,
|
1419 |
+
"grad_norm": 9.793978097150102e-05,
|
1420 |
+
"learning_rate": 2.466320603139291e-05,
|
1421 |
+
"loss": 0.0,
|
1422 |
+
"step": 5000
|
1423 |
+
},
|
1424 |
+
{
|
1425 |
+
"epoch": 1.6766766766766765,
|
1426 |
+
"grad_norm": 0.0029056610073894262,
|
1427 |
+
"learning_rate": 2.4508713385242863e-05,
|
1428 |
+
"loss": 0.0001,
|
1429 |
+
"step": 5025
|
1430 |
+
},
|
1431 |
+
{
|
1432 |
+
"epoch": 1.6850183516850183,
|
1433 |
+
"grad_norm": 0.0007343544857576489,
|
1434 |
+
"learning_rate": 2.435422073909282e-05,
|
1435 |
+
"loss": 0.0521,
|
1436 |
+
"step": 5050
|
1437 |
+
},
|
1438 |
+
{
|
1439 |
+
"epoch": 1.6933600266933602,
|
1440 |
+
"grad_norm": 0.014967703260481358,
|
1441 |
+
"learning_rate": 2.419972809294278e-05,
|
1442 |
+
"loss": 0.0002,
|
1443 |
+
"step": 5075
|
1444 |
+
},
|
1445 |
+
{
|
1446 |
+
"epoch": 1.7017017017017015,
|
1447 |
+
"grad_norm": 0.0071108159609138966,
|
1448 |
+
"learning_rate": 2.4045235446792733e-05,
|
1449 |
+
"loss": 0.0362,
|
1450 |
+
"step": 5100
|
1451 |
+
},
|
1452 |
+
{
|
1453 |
+
"epoch": 1.7100433767100434,
|
1454 |
+
"grad_norm": 0.004571543540805578,
|
1455 |
+
"learning_rate": 2.389074280064269e-05,
|
1456 |
+
"loss": 0.0047,
|
1457 |
+
"step": 5125
|
1458 |
+
},
|
1459 |
+
{
|
1460 |
+
"epoch": 1.7183850517183852,
|
1461 |
+
"grad_norm": 0.007166516967117786,
|
1462 |
+
"learning_rate": 2.3736250154492648e-05,
|
1463 |
+
"loss": 0.0028,
|
1464 |
+
"step": 5150
|
1465 |
+
},
|
1466 |
+
{
|
1467 |
+
"epoch": 1.7267267267267268,
|
1468 |
+
"grad_norm": 0.008930371142923832,
|
1469 |
+
"learning_rate": 2.3581757508342603e-05,
|
1470 |
+
"loss": 0.0396,
|
1471 |
+
"step": 5175
|
1472 |
+
},
|
1473 |
+
{
|
1474 |
+
"epoch": 1.7350684017350684,
|
1475 |
+
"grad_norm": 0.007438777945935726,
|
1476 |
+
"learning_rate": 2.342726486219256e-05,
|
1477 |
+
"loss": 0.0004,
|
1478 |
+
"step": 5200
|
1479 |
+
},
|
1480 |
+
{
|
1481 |
+
"epoch": 1.7434100767434102,
|
1482 |
+
"grad_norm": 0.0038795331493020058,
|
1483 |
+
"learning_rate": 2.3272772216042518e-05,
|
1484 |
+
"loss": 0.0257,
|
1485 |
+
"step": 5225
|
1486 |
+
},
|
1487 |
+
{
|
1488 |
+
"epoch": 1.7517517517517518,
|
1489 |
+
"grad_norm": 0.003188981907442212,
|
1490 |
+
"learning_rate": 2.3118279569892472e-05,
|
1491 |
+
"loss": 0.0003,
|
1492 |
+
"step": 5250
|
1493 |
+
},
|
1494 |
+
{
|
1495 |
+
"epoch": 1.7600934267600934,
|
1496 |
+
"grad_norm": 0.00612386642023921,
|
1497 |
+
"learning_rate": 2.296378692374243e-05,
|
1498 |
+
"loss": 0.0321,
|
1499 |
+
"step": 5275
|
1500 |
+
},
|
1501 |
+
{
|
1502 |
+
"epoch": 1.7684351017684352,
|
1503 |
+
"grad_norm": 0.0007353666005656123,
|
1504 |
+
"learning_rate": 2.2809294277592388e-05,
|
1505 |
+
"loss": 0.0026,
|
1506 |
+
"step": 5300
|
1507 |
+
},
|
1508 |
+
{
|
1509 |
+
"epoch": 1.7767767767767768,
|
1510 |
+
"grad_norm": 0.006964649073779583,
|
1511 |
+
"learning_rate": 2.2654801631442342e-05,
|
1512 |
+
"loss": 0.0009,
|
1513 |
+
"step": 5325
|
1514 |
+
},
|
1515 |
+
{
|
1516 |
+
"epoch": 1.7851184517851184,
|
1517 |
+
"grad_norm": 0.009796655736863613,
|
1518 |
+
"learning_rate": 2.25003089852923e-05,
|
1519 |
+
"loss": 0.0542,
|
1520 |
+
"step": 5350
|
1521 |
+
},
|
1522 |
+
{
|
1523 |
+
"epoch": 1.7934601267934602,
|
1524 |
+
"grad_norm": 0.00797263439744711,
|
1525 |
+
"learning_rate": 2.2345816339142257e-05,
|
1526 |
+
"loss": 0.001,
|
1527 |
+
"step": 5375
|
1528 |
+
},
|
1529 |
+
{
|
1530 |
+
"epoch": 1.8018018018018018,
|
1531 |
+
"grad_norm": 0.0027469745837152004,
|
1532 |
+
"learning_rate": 2.2191323692992215e-05,
|
1533 |
+
"loss": 0.0002,
|
1534 |
+
"step": 5400
|
1535 |
+
},
|
1536 |
+
{
|
1537 |
+
"epoch": 1.8101434768101434,
|
1538 |
+
"grad_norm": 0.0032520880922675133,
|
1539 |
+
"learning_rate": 2.2036831046842172e-05,
|
1540 |
+
"loss": 0.0162,
|
1541 |
+
"step": 5425
|
1542 |
+
},
|
1543 |
+
{
|
1544 |
+
"epoch": 1.8184851518184852,
|
1545 |
+
"grad_norm": 0.0021738584619015455,
|
1546 |
+
"learning_rate": 2.188233840069213e-05,
|
1547 |
+
"loss": 0.0002,
|
1548 |
+
"step": 5450
|
1549 |
+
},
|
1550 |
+
{
|
1551 |
+
"epoch": 1.8268268268268268,
|
1552 |
+
"grad_norm": 0.0015207937685772777,
|
1553 |
+
"learning_rate": 2.1727845754542084e-05,
|
1554 |
+
"loss": 0.0004,
|
1555 |
+
"step": 5475
|
1556 |
+
},
|
1557 |
+
{
|
1558 |
+
"epoch": 1.8351685018351684,
|
1559 |
+
"grad_norm": 0.001575661008246243,
|
1560 |
+
"learning_rate": 2.1573353108392042e-05,
|
1561 |
+
"loss": 0.0314,
|
1562 |
+
"step": 5500
|
1563 |
+
},
|
1564 |
+
{
|
1565 |
+
"epoch": 1.8435101768435103,
|
1566 |
+
"grad_norm": 0.0009630912682041526,
|
1567 |
+
"learning_rate": 2.1418860462242e-05,
|
1568 |
+
"loss": 0.0047,
|
1569 |
+
"step": 5525
|
1570 |
+
},
|
1571 |
+
{
|
1572 |
+
"epoch": 1.8518518518518519,
|
1573 |
+
"grad_norm": 0.001678589964285493,
|
1574 |
+
"learning_rate": 2.1264367816091954e-05,
|
1575 |
+
"loss": 0.0338,
|
1576 |
+
"step": 5550
|
1577 |
+
},
|
1578 |
+
{
|
1579 |
+
"epoch": 1.8601935268601935,
|
1580 |
+
"grad_norm": 0.004351487848907709,
|
1581 |
+
"learning_rate": 2.1109875169941912e-05,
|
1582 |
+
"loss": 0.0063,
|
1583 |
+
"step": 5575
|
1584 |
+
},
|
1585 |
+
{
|
1586 |
+
"epoch": 1.8685352018685353,
|
1587 |
+
"grad_norm": 0.008160965517163277,
|
1588 |
+
"learning_rate": 2.095538252379187e-05,
|
1589 |
+
"loss": 0.0304,
|
1590 |
+
"step": 5600
|
1591 |
+
},
|
1592 |
+
{
|
1593 |
+
"epoch": 1.8768768768768769,
|
1594 |
+
"grad_norm": 0.0037402340676635504,
|
1595 |
+
"learning_rate": 2.0800889877641824e-05,
|
1596 |
+
"loss": 0.0006,
|
1597 |
+
"step": 5625
|
1598 |
+
},
|
1599 |
+
{
|
1600 |
+
"epoch": 1.8852185518852185,
|
1601 |
+
"grad_norm": 0.0016508076805621386,
|
1602 |
+
"learning_rate": 2.064639723149178e-05,
|
1603 |
+
"loss": 0.0001,
|
1604 |
+
"step": 5650
|
1605 |
+
},
|
1606 |
+
{
|
1607 |
+
"epoch": 1.8935602268935603,
|
1608 |
+
"grad_norm": 0.011413435451686382,
|
1609 |
+
"learning_rate": 2.049190458534174e-05,
|
1610 |
+
"loss": 0.0197,
|
1611 |
+
"step": 5675
|
1612 |
+
},
|
1613 |
+
{
|
1614 |
+
"epoch": 1.901901901901902,
|
1615 |
+
"grad_norm": 0.0007151237805373967,
|
1616 |
+
"learning_rate": 2.0337411939191693e-05,
|
1617 |
+
"loss": 0.0002,
|
1618 |
+
"step": 5700
|
1619 |
+
},
|
1620 |
+
{
|
1621 |
+
"epoch": 1.9102435769102435,
|
1622 |
+
"grad_norm": 24.82006072998047,
|
1623 |
+
"learning_rate": 2.018291929304165e-05,
|
1624 |
+
"loss": 0.0506,
|
1625 |
+
"step": 5725
|
1626 |
+
},
|
1627 |
+
{
|
1628 |
+
"epoch": 1.9185852519185853,
|
1629 |
+
"grad_norm": 0.018485259264707565,
|
1630 |
+
"learning_rate": 2.002842664689161e-05,
|
1631 |
+
"loss": 0.0324,
|
1632 |
+
"step": 5750
|
1633 |
+
},
|
1634 |
+
{
|
1635 |
+
"epoch": 1.926926926926927,
|
1636 |
+
"grad_norm": 0.008674918673932552,
|
1637 |
+
"learning_rate": 1.9873934000741566e-05,
|
1638 |
+
"loss": 0.0006,
|
1639 |
+
"step": 5775
|
1640 |
+
},
|
1641 |
+
{
|
1642 |
+
"epoch": 1.9352686019352685,
|
1643 |
+
"grad_norm": 0.0022723872680217028,
|
1644 |
+
"learning_rate": 1.9719441354591524e-05,
|
1645 |
+
"loss": 0.0002,
|
1646 |
+
"step": 5800
|
1647 |
+
},
|
1648 |
+
{
|
1649 |
+
"epoch": 1.9436102769436103,
|
1650 |
+
"grad_norm": 0.0006509014638140798,
|
1651 |
+
"learning_rate": 1.9564948708441482e-05,
|
1652 |
+
"loss": 0.0022,
|
1653 |
+
"step": 5825
|
1654 |
+
},
|
1655 |
+
{
|
1656 |
+
"epoch": 1.951951951951952,
|
1657 |
+
"grad_norm": 0.001537534175440669,
|
1658 |
+
"learning_rate": 1.9410456062291436e-05,
|
1659 |
+
"loss": 0.0002,
|
1660 |
+
"step": 5850
|
1661 |
+
},
|
1662 |
+
{
|
1663 |
+
"epoch": 1.9602936269602935,
|
1664 |
+
"grad_norm": 0.0028848988004028797,
|
1665 |
+
"learning_rate": 1.9255963416141394e-05,
|
1666 |
+
"loss": 0.0312,
|
1667 |
+
"step": 5875
|
1668 |
+
},
|
1669 |
+
{
|
1670 |
+
"epoch": 1.9686353019686353,
|
1671 |
+
"grad_norm": 0.001275117159821093,
|
1672 |
+
"learning_rate": 1.910147076999135e-05,
|
1673 |
+
"loss": 0.0004,
|
1674 |
+
"step": 5900
|
1675 |
+
},
|
1676 |
+
{
|
1677 |
+
"epoch": 1.976976976976977,
|
1678 |
+
"grad_norm": 0.003794446587562561,
|
1679 |
+
"learning_rate": 1.8946978123841306e-05,
|
1680 |
+
"loss": 0.0303,
|
1681 |
+
"step": 5925
|
1682 |
+
},
|
1683 |
+
{
|
1684 |
+
"epoch": 1.9853186519853185,
|
1685 |
+
"grad_norm": 0.0013351526577025652,
|
1686 |
+
"learning_rate": 1.8792485477691263e-05,
|
1687 |
+
"loss": 0.0028,
|
1688 |
+
"step": 5950
|
1689 |
+
},
|
1690 |
+
{
|
1691 |
+
"epoch": 1.9936603269936604,
|
1692 |
+
"grad_norm": 0.0022188324946910143,
|
1693 |
+
"learning_rate": 1.863799283154122e-05,
|
1694 |
+
"loss": 0.0001,
|
1695 |
+
"step": 5975
|
1696 |
+
},
|
1697 |
+
{
|
1698 |
+
"epoch": 2.0,
|
1699 |
+
"eval_accuracy": 0.9983316649983317,
|
1700 |
+
"eval_auc": 0.9996210737764792,
|
1701 |
+
"eval_f1": 0.9285714285714286,
|
1702 |
+
"eval_loss": 0.006147409323602915,
|
1703 |
+
"eval_precision": 0.9848484848484849,
|
1704 |
+
"eval_recall": 0.8783783783783784,
|
1705 |
+
"eval_runtime": 118.7551,
|
1706 |
+
"eval_samples_per_second": 50.474,
|
1707 |
+
"eval_steps_per_second": 3.158,
|
1708 |
+
"step": 5994
|
1709 |
+
},
|
1710 |
+
{
|
1711 |
+
"epoch": 2.002002002002002,
|
1712 |
+
"grad_norm": 0.03426537662744522,
|
1713 |
+
"learning_rate": 1.8483500185391175e-05,
|
1714 |
+
"loss": 0.0001,
|
1715 |
+
"step": 6000
|
1716 |
+
},
|
1717 |
+
{
|
1718 |
+
"epoch": 2.0103436770103436,
|
1719 |
+
"grad_norm": 0.017995424568653107,
|
1720 |
+
"learning_rate": 1.8329007539241133e-05,
|
1721 |
+
"loss": 0.007,
|
1722 |
+
"step": 6025
|
1723 |
+
},
|
1724 |
+
{
|
1725 |
+
"epoch": 2.0186853520186854,
|
1726 |
+
"grad_norm": 0.0008380690705962479,
|
1727 |
+
"learning_rate": 1.817451489309109e-05,
|
1728 |
+
"loss": 0.0163,
|
1729 |
+
"step": 6050
|
1730 |
+
},
|
1731 |
+
{
|
1732 |
+
"epoch": 2.027027027027027,
|
1733 |
+
"grad_norm": 0.0041703651659190655,
|
1734 |
+
"learning_rate": 1.8020022246941045e-05,
|
1735 |
+
"loss": 0.0002,
|
1736 |
+
"step": 6075
|
1737 |
+
},
|
1738 |
+
{
|
1739 |
+
"epoch": 2.0353687020353686,
|
1740 |
+
"grad_norm": 0.0007581566460430622,
|
1741 |
+
"learning_rate": 1.7865529600791003e-05,
|
1742 |
+
"loss": 0.0001,
|
1743 |
+
"step": 6100
|
1744 |
+
},
|
1745 |
+
{
|
1746 |
+
"epoch": 2.0437103770437104,
|
1747 |
+
"grad_norm": 0.000699601077940315,
|
1748 |
+
"learning_rate": 1.771103695464096e-05,
|
1749 |
+
"loss": 0.0003,
|
1750 |
+
"step": 6125
|
1751 |
+
},
|
1752 |
+
{
|
1753 |
+
"epoch": 2.052052052052052,
|
1754 |
+
"grad_norm": 0.021615630015730858,
|
1755 |
+
"learning_rate": 1.7556544308490915e-05,
|
1756 |
+
"loss": 0.0195,
|
1757 |
+
"step": 6150
|
1758 |
+
},
|
1759 |
+
{
|
1760 |
+
"epoch": 2.0603937270603936,
|
1761 |
+
"grad_norm": 0.0024253269657492638,
|
1762 |
+
"learning_rate": 1.7402051662340872e-05,
|
1763 |
+
"loss": 0.0016,
|
1764 |
+
"step": 6175
|
1765 |
+
},
|
1766 |
+
{
|
1767 |
+
"epoch": 2.0687354020687354,
|
1768 |
+
"grad_norm": 8.936987433116883e-05,
|
1769 |
+
"learning_rate": 1.724755901619083e-05,
|
1770 |
+
"loss": 0.0,
|
1771 |
+
"step": 6200
|
1772 |
+
},
|
1773 |
+
{
|
1774 |
+
"epoch": 2.0770770770770772,
|
1775 |
+
"grad_norm": 0.0001456751924706623,
|
1776 |
+
"learning_rate": 1.7093066370040788e-05,
|
1777 |
+
"loss": 0.0,
|
1778 |
+
"step": 6225
|
1779 |
+
},
|
1780 |
+
{
|
1781 |
+
"epoch": 2.0854187520854186,
|
1782 |
+
"grad_norm": 0.0011276202276349068,
|
1783 |
+
"learning_rate": 1.6938573723890745e-05,
|
1784 |
+
"loss": 0.0382,
|
1785 |
+
"step": 6250
|
1786 |
+
},
|
1787 |
+
{
|
1788 |
+
"epoch": 2.0937604270937604,
|
1789 |
+
"grad_norm": 0.00011645181803032756,
|
1790 |
+
"learning_rate": 1.6784081077740703e-05,
|
1791 |
+
"loss": 0.0001,
|
1792 |
+
"step": 6275
|
1793 |
+
},
|
1794 |
+
{
|
1795 |
+
"epoch": 2.1021021021021022,
|
1796 |
+
"grad_norm": 0.00045101341675035655,
|
1797 |
+
"learning_rate": 1.6629588431590657e-05,
|
1798 |
+
"loss": 0.0002,
|
1799 |
+
"step": 6300
|
1800 |
+
},
|
1801 |
+
{
|
1802 |
+
"epoch": 2.1104437771104436,
|
1803 |
+
"grad_norm": 0.0001981251552933827,
|
1804 |
+
"learning_rate": 1.6475095785440615e-05,
|
1805 |
+
"loss": 0.0002,
|
1806 |
+
"step": 6325
|
1807 |
+
},
|
1808 |
+
{
|
1809 |
+
"epoch": 2.1187854521187854,
|
1810 |
+
"grad_norm": 0.0004596684011630714,
|
1811 |
+
"learning_rate": 1.6320603139290573e-05,
|
1812 |
+
"loss": 0.0001,
|
1813 |
+
"step": 6350
|
1814 |
+
},
|
1815 |
+
{
|
1816 |
+
"epoch": 2.1271271271271273,
|
1817 |
+
"grad_norm": 0.0003594656300265342,
|
1818 |
+
"learning_rate": 1.6166110493140527e-05,
|
1819 |
+
"loss": 0.0,
|
1820 |
+
"step": 6375
|
1821 |
+
},
|
1822 |
+
{
|
1823 |
+
"epoch": 2.1354688021354686,
|
1824 |
+
"grad_norm": 0.0015784379793331027,
|
1825 |
+
"learning_rate": 1.6011617846990485e-05,
|
1826 |
+
"loss": 0.0103,
|
1827 |
+
"step": 6400
|
1828 |
+
},
|
1829 |
+
{
|
1830 |
+
"epoch": 2.1438104771438105,
|
1831 |
+
"grad_norm": 0.001940034911967814,
|
1832 |
+
"learning_rate": 1.5857125200840442e-05,
|
1833 |
+
"loss": 0.0019,
|
1834 |
+
"step": 6425
|
1835 |
+
},
|
1836 |
+
{
|
1837 |
+
"epoch": 2.1521521521521523,
|
1838 |
+
"grad_norm": 0.0011929641477763653,
|
1839 |
+
"learning_rate": 1.5702632554690397e-05,
|
1840 |
+
"loss": 0.0,
|
1841 |
+
"step": 6450
|
1842 |
+
},
|
1843 |
+
{
|
1844 |
+
"epoch": 2.1604938271604937,
|
1845 |
+
"grad_norm": 0.0035334331914782524,
|
1846 |
+
"learning_rate": 1.5548139908540354e-05,
|
1847 |
+
"loss": 0.007,
|
1848 |
+
"step": 6475
|
1849 |
+
},
|
1850 |
+
{
|
1851 |
+
"epoch": 2.1688355021688355,
|
1852 |
+
"grad_norm": 0.0003299266390968114,
|
1853 |
+
"learning_rate": 1.5393647262390312e-05,
|
1854 |
+
"loss": 0.0002,
|
1855 |
+
"step": 6500
|
1856 |
+
},
|
1857 |
+
{
|
1858 |
+
"epoch": 2.1771771771771773,
|
1859 |
+
"grad_norm": 0.0005644136690534651,
|
1860 |
+
"learning_rate": 1.5239154616240266e-05,
|
1861 |
+
"loss": 0.0,
|
1862 |
+
"step": 6525
|
1863 |
+
},
|
1864 |
+
{
|
1865 |
+
"epoch": 2.1855188521855187,
|
1866 |
+
"grad_norm": 0.0003227370616514236,
|
1867 |
+
"learning_rate": 1.5084661970090224e-05,
|
1868 |
+
"loss": 0.0,
|
1869 |
+
"step": 6550
|
1870 |
+
},
|
1871 |
+
{
|
1872 |
+
"epoch": 2.1938605271938605,
|
1873 |
+
"grad_norm": 0.0011611609952524304,
|
1874 |
+
"learning_rate": 1.4930169323940182e-05,
|
1875 |
+
"loss": 0.0,
|
1876 |
+
"step": 6575
|
1877 |
+
},
|
1878 |
+
{
|
1879 |
+
"epoch": 2.2022022022022023,
|
1880 |
+
"grad_norm": 0.003065961878746748,
|
1881 |
+
"learning_rate": 1.4775676677790138e-05,
|
1882 |
+
"loss": 0.0291,
|
1883 |
+
"step": 6600
|
1884 |
+
},
|
1885 |
+
{
|
1886 |
+
"epoch": 2.2105438772105437,
|
1887 |
+
"grad_norm": 0.0016901785274967551,
|
1888 |
+
"learning_rate": 1.4621184031640095e-05,
|
1889 |
+
"loss": 0.0,
|
1890 |
+
"step": 6625
|
1891 |
+
},
|
1892 |
+
{
|
1893 |
+
"epoch": 2.2188855522188855,
|
1894 |
+
"grad_norm": 0.003457328537479043,
|
1895 |
+
"learning_rate": 1.4466691385490053e-05,
|
1896 |
+
"loss": 0.0578,
|
1897 |
+
"step": 6650
|
1898 |
+
},
|
1899 |
+
{
|
1900 |
+
"epoch": 2.2272272272272273,
|
1901 |
+
"grad_norm": 0.0015183900250121951,
|
1902 |
+
"learning_rate": 1.4312198739340007e-05,
|
1903 |
+
"loss": 0.0004,
|
1904 |
+
"step": 6675
|
1905 |
+
},
|
1906 |
+
{
|
1907 |
+
"epoch": 2.2355689022355687,
|
1908 |
+
"grad_norm": 0.01456731278449297,
|
1909 |
+
"learning_rate": 1.4157706093189965e-05,
|
1910 |
+
"loss": 0.0001,
|
1911 |
+
"step": 6700
|
1912 |
+
},
|
1913 |
+
{
|
1914 |
+
"epoch": 2.2439105772439105,
|
1915 |
+
"grad_norm": 0.006311188917607069,
|
1916 |
+
"learning_rate": 1.4003213447039923e-05,
|
1917 |
+
"loss": 0.0001,
|
1918 |
+
"step": 6725
|
1919 |
+
},
|
1920 |
+
{
|
1921 |
+
"epoch": 2.2522522522522523,
|
1922 |
+
"grad_norm": 0.00036790382000617683,
|
1923 |
+
"learning_rate": 1.3848720800889879e-05,
|
1924 |
+
"loss": 0.0333,
|
1925 |
+
"step": 6750
|
1926 |
+
},
|
1927 |
+
{
|
1928 |
+
"epoch": 2.2605939272605937,
|
1929 |
+
"grad_norm": 0.002643864369019866,
|
1930 |
+
"learning_rate": 1.3694228154739836e-05,
|
1931 |
+
"loss": 0.0001,
|
1932 |
+
"step": 6775
|
1933 |
+
},
|
1934 |
+
{
|
1935 |
+
"epoch": 2.2689356022689355,
|
1936 |
+
"grad_norm": 0.0004877290630247444,
|
1937 |
+
"learning_rate": 1.3539735508589794e-05,
|
1938 |
+
"loss": 0.0001,
|
1939 |
+
"step": 6800
|
1940 |
+
},
|
1941 |
+
{
|
1942 |
+
"epoch": 2.2772772772772774,
|
1943 |
+
"grad_norm": 0.0006125778309069574,
|
1944 |
+
"learning_rate": 1.3385242862439748e-05,
|
1945 |
+
"loss": 0.0001,
|
1946 |
+
"step": 6825
|
1947 |
+
},
|
1948 |
+
{
|
1949 |
+
"epoch": 2.2856189522856187,
|
1950 |
+
"grad_norm": 0.006749690044671297,
|
1951 |
+
"learning_rate": 1.3230750216289706e-05,
|
1952 |
+
"loss": 0.0295,
|
1953 |
+
"step": 6850
|
1954 |
+
},
|
1955 |
+
{
|
1956 |
+
"epoch": 2.2939606272939606,
|
1957 |
+
"grad_norm": 0.00032511603785678744,
|
1958 |
+
"learning_rate": 1.3076257570139664e-05,
|
1959 |
+
"loss": 0.0003,
|
1960 |
+
"step": 6875
|
1961 |
+
},
|
1962 |
+
{
|
1963 |
+
"epoch": 2.3023023023023024,
|
1964 |
+
"grad_norm": 0.00033115336555056274,
|
1965 |
+
"learning_rate": 1.2921764923989618e-05,
|
1966 |
+
"loss": 0.0001,
|
1967 |
+
"step": 6900
|
1968 |
+
},
|
1969 |
+
{
|
1970 |
+
"epoch": 2.3106439773106437,
|
1971 |
+
"grad_norm": 0.0015968529041856527,
|
1972 |
+
"learning_rate": 1.2767272277839576e-05,
|
1973 |
+
"loss": 0.0001,
|
1974 |
+
"step": 6925
|
1975 |
+
},
|
1976 |
+
{
|
1977 |
+
"epoch": 2.3189856523189856,
|
1978 |
+
"grad_norm": 0.0019475755980238318,
|
1979 |
+
"learning_rate": 1.2612779631689531e-05,
|
1980 |
+
"loss": 0.0001,
|
1981 |
+
"step": 6950
|
1982 |
+
},
|
1983 |
+
{
|
1984 |
+
"epoch": 2.3273273273273274,
|
1985 |
+
"grad_norm": 0.004731306340545416,
|
1986 |
+
"learning_rate": 1.245828698553949e-05,
|
1987 |
+
"loss": 0.0001,
|
1988 |
+
"step": 6975
|
1989 |
+
},
|
1990 |
+
{
|
1991 |
+
"epoch": 2.3356690023356688,
|
1992 |
+
"grad_norm": 0.0013892946299165487,
|
1993 |
+
"learning_rate": 1.2303794339389447e-05,
|
1994 |
+
"loss": 0.0002,
|
1995 |
+
"step": 7000
|
1996 |
+
},
|
1997 |
+
{
|
1998 |
+
"epoch": 2.3440106773440106,
|
1999 |
+
"grad_norm": 0.00041954405605793,
|
2000 |
+
"learning_rate": 1.2149301693239403e-05,
|
2001 |
+
"loss": 0.0215,
|
2002 |
+
"step": 7025
|
2003 |
+
},
|
2004 |
+
{
|
2005 |
+
"epoch": 2.3523523523523524,
|
2006 |
+
"grad_norm": 0.0006229266291484237,
|
2007 |
+
"learning_rate": 1.1994809047089359e-05,
|
2008 |
+
"loss": 0.0001,
|
2009 |
+
"step": 7050
|
2010 |
+
},
|
2011 |
+
{
|
2012 |
+
"epoch": 2.360694027360694,
|
2013 |
+
"grad_norm": 0.0022454725112766027,
|
2014 |
+
"learning_rate": 1.1840316400939316e-05,
|
2015 |
+
"loss": 0.0003,
|
2016 |
+
"step": 7075
|
2017 |
+
},
|
2018 |
+
{
|
2019 |
+
"epoch": 2.3690357023690356,
|
2020 |
+
"grad_norm": 0.0003879007126670331,
|
2021 |
+
"learning_rate": 1.1685823754789272e-05,
|
2022 |
+
"loss": 0.0002,
|
2023 |
+
"step": 7100
|
2024 |
+
},
|
2025 |
+
{
|
2026 |
+
"epoch": 2.3773773773773774,
|
2027 |
+
"grad_norm": 0.0023330808617174625,
|
2028 |
+
"learning_rate": 1.1531331108639228e-05,
|
2029 |
+
"loss": 0.0001,
|
2030 |
+
"step": 7125
|
2031 |
+
},
|
2032 |
+
{
|
2033 |
+
"epoch": 2.3857190523857192,
|
2034 |
+
"grad_norm": 0.009034736081957817,
|
2035 |
+
"learning_rate": 1.1376838462489186e-05,
|
2036 |
+
"loss": 0.0409,
|
2037 |
+
"step": 7150
|
2038 |
+
},
|
2039 |
+
{
|
2040 |
+
"epoch": 2.3940607273940606,
|
2041 |
+
"grad_norm": 0.012751072645187378,
|
2042 |
+
"learning_rate": 1.1222345816339142e-05,
|
2043 |
+
"loss": 0.0001,
|
2044 |
+
"step": 7175
|
2045 |
+
},
|
2046 |
+
{
|
2047 |
+
"epoch": 2.4024024024024024,
|
2048 |
+
"grad_norm": 0.0005218230071477592,
|
2049 |
+
"learning_rate": 1.10678531701891e-05,
|
2050 |
+
"loss": 0.0003,
|
2051 |
+
"step": 7200
|
2052 |
+
},
|
2053 |
+
{
|
2054 |
+
"epoch": 2.4107440774107443,
|
2055 |
+
"grad_norm": 0.005197747144848108,
|
2056 |
+
"learning_rate": 1.0913360524039057e-05,
|
2057 |
+
"loss": 0.0002,
|
2058 |
+
"step": 7225
|
2059 |
+
},
|
2060 |
+
{
|
2061 |
+
"epoch": 2.4190857524190856,
|
2062 |
+
"grad_norm": 0.0004819717141799629,
|
2063 |
+
"learning_rate": 1.0758867877889013e-05,
|
2064 |
+
"loss": 0.0052,
|
2065 |
+
"step": 7250
|
2066 |
+
},
|
2067 |
+
{
|
2068 |
+
"epoch": 2.4274274274274275,
|
2069 |
+
"grad_norm": 0.0005321923526935279,
|
2070 |
+
"learning_rate": 1.060437523173897e-05,
|
2071 |
+
"loss": 0.0,
|
2072 |
+
"step": 7275
|
2073 |
+
},
|
2074 |
+
{
|
2075 |
+
"epoch": 2.4357691024357693,
|
2076 |
+
"grad_norm": 0.0027098068967461586,
|
2077 |
+
"learning_rate": 1.0449882585588927e-05,
|
2078 |
+
"loss": 0.0048,
|
2079 |
+
"step": 7300
|
2080 |
+
},
|
2081 |
+
{
|
2082 |
+
"epoch": 2.4441107774441106,
|
2083 |
+
"grad_norm": 15.976853370666504,
|
2084 |
+
"learning_rate": 1.0295389939438883e-05,
|
2085 |
+
"loss": 0.0131,
|
2086 |
+
"step": 7325
|
2087 |
+
},
|
2088 |
+
{
|
2089 |
+
"epoch": 2.4524524524524525,
|
2090 |
+
"grad_norm": 0.06583759933710098,
|
2091 |
+
"learning_rate": 1.0140897293288839e-05,
|
2092 |
+
"loss": 0.0002,
|
2093 |
+
"step": 7350
|
2094 |
+
},
|
2095 |
+
{
|
2096 |
+
"epoch": 2.4607941274607943,
|
2097 |
+
"grad_norm": 0.0012560704490169883,
|
2098 |
+
"learning_rate": 9.986404647138797e-06,
|
2099 |
+
"loss": 0.0002,
|
2100 |
+
"step": 7375
|
2101 |
+
},
|
2102 |
+
{
|
2103 |
+
"epoch": 2.4691358024691357,
|
2104 |
+
"grad_norm": 0.0004110535664949566,
|
2105 |
+
"learning_rate": 9.831912000988754e-06,
|
2106 |
+
"loss": 0.0001,
|
2107 |
+
"step": 7400
|
2108 |
+
},
|
2109 |
+
{
|
2110 |
+
"epoch": 2.4774774774774775,
|
2111 |
+
"grad_norm": 0.0021631394047290087,
|
2112 |
+
"learning_rate": 9.67741935483871e-06,
|
2113 |
+
"loss": 0.0001,
|
2114 |
+
"step": 7425
|
2115 |
+
},
|
2116 |
+
{
|
2117 |
+
"epoch": 2.4858191524858193,
|
2118 |
+
"grad_norm": 0.002222039271146059,
|
2119 |
+
"learning_rate": 9.522926708688668e-06,
|
2120 |
+
"loss": 0.0169,
|
2121 |
+
"step": 7450
|
2122 |
+
},
|
2123 |
+
{
|
2124 |
+
"epoch": 2.4941608274941607,
|
2125 |
+
"grad_norm": 0.0026871326845139265,
|
2126 |
+
"learning_rate": 9.368434062538624e-06,
|
2127 |
+
"loss": 0.0243,
|
2128 |
+
"step": 7475
|
2129 |
+
},
|
2130 |
+
{
|
2131 |
+
"epoch": 2.5025025025025025,
|
2132 |
+
"grad_norm": 0.00477555999532342,
|
2133 |
+
"learning_rate": 9.21394141638858e-06,
|
2134 |
+
"loss": 0.0005,
|
2135 |
+
"step": 7500
|
2136 |
+
},
|
2137 |
+
{
|
2138 |
+
"epoch": 2.5108441775108443,
|
2139 |
+
"grad_norm": 0.012109145522117615,
|
2140 |
+
"learning_rate": 9.059448770238536e-06,
|
2141 |
+
"loss": 0.0282,
|
2142 |
+
"step": 7525
|
2143 |
+
},
|
2144 |
+
{
|
2145 |
+
"epoch": 2.5191858525191857,
|
2146 |
+
"grad_norm": 0.0005930490442551672,
|
2147 |
+
"learning_rate": 8.904956124088494e-06,
|
2148 |
+
"loss": 0.0002,
|
2149 |
+
"step": 7550
|
2150 |
+
},
|
2151 |
+
{
|
2152 |
+
"epoch": 2.5275275275275275,
|
2153 |
+
"grad_norm": 0.003786754095926881,
|
2154 |
+
"learning_rate": 8.75046347793845e-06,
|
2155 |
+
"loss": 0.0023,
|
2156 |
+
"step": 7575
|
2157 |
+
},
|
2158 |
+
{
|
2159 |
+
"epoch": 2.5358692025358693,
|
2160 |
+
"grad_norm": 0.0012933415127918124,
|
2161 |
+
"learning_rate": 8.595970831788407e-06,
|
2162 |
+
"loss": 0.0119,
|
2163 |
+
"step": 7600
|
2164 |
+
},
|
2165 |
+
{
|
2166 |
+
"epoch": 2.5442108775442107,
|
2167 |
+
"grad_norm": 0.0005488534225150943,
|
2168 |
+
"learning_rate": 8.441478185638365e-06,
|
2169 |
+
"loss": 0.0135,
|
2170 |
+
"step": 7625
|
2171 |
+
},
|
2172 |
+
{
|
2173 |
+
"epoch": 2.5525525525525525,
|
2174 |
+
"grad_norm": 0.0013369097141548991,
|
2175 |
+
"learning_rate": 8.286985539488321e-06,
|
2176 |
+
"loss": 0.0172,
|
2177 |
+
"step": 7650
|
2178 |
+
},
|
2179 |
+
{
|
2180 |
+
"epoch": 2.5608942275608944,
|
2181 |
+
"grad_norm": 0.0037656030617654324,
|
2182 |
+
"learning_rate": 8.132492893338277e-06,
|
2183 |
+
"loss": 0.0163,
|
2184 |
+
"step": 7675
|
2185 |
+
},
|
2186 |
+
{
|
2187 |
+
"epoch": 2.569235902569236,
|
2188 |
+
"grad_norm": 0.0011651143431663513,
|
2189 |
+
"learning_rate": 7.978000247188235e-06,
|
2190 |
+
"loss": 0.0001,
|
2191 |
+
"step": 7700
|
2192 |
+
},
|
2193 |
+
{
|
2194 |
+
"epoch": 2.5775775775775776,
|
2195 |
+
"grad_norm": 0.0026386177632957697,
|
2196 |
+
"learning_rate": 7.82350760103819e-06,
|
2197 |
+
"loss": 0.0002,
|
2198 |
+
"step": 7725
|
2199 |
+
},
|
2200 |
+
{
|
2201 |
+
"epoch": 2.5859192525859194,
|
2202 |
+
"grad_norm": 0.004785032011568546,
|
2203 |
+
"learning_rate": 7.669014954888147e-06,
|
2204 |
+
"loss": 0.0,
|
2205 |
+
"step": 7750
|
2206 |
+
},
|
2207 |
+
{
|
2208 |
+
"epoch": 2.594260927594261,
|
2209 |
+
"grad_norm": 0.00011592906957957894,
|
2210 |
+
"learning_rate": 7.514522308738104e-06,
|
2211 |
+
"loss": 0.0,
|
2212 |
+
"step": 7775
|
2213 |
+
},
|
2214 |
+
{
|
2215 |
+
"epoch": 2.6026026026026026,
|
2216 |
+
"grad_norm": 0.001778353936970234,
|
2217 |
+
"learning_rate": 7.360029662588061e-06,
|
2218 |
+
"loss": 0.0001,
|
2219 |
+
"step": 7800
|
2220 |
+
},
|
2221 |
+
{
|
2222 |
+
"epoch": 2.6109442776109444,
|
2223 |
+
"grad_norm": 0.001059061149135232,
|
2224 |
+
"learning_rate": 7.205537016438017e-06,
|
2225 |
+
"loss": 0.0095,
|
2226 |
+
"step": 7825
|
2227 |
+
},
|
2228 |
+
{
|
2229 |
+
"epoch": 2.619285952619286,
|
2230 |
+
"grad_norm": 0.00041067678830586374,
|
2231 |
+
"learning_rate": 7.051044370287975e-06,
|
2232 |
+
"loss": 0.0001,
|
2233 |
+
"step": 7850
|
2234 |
+
},
|
2235 |
+
{
|
2236 |
+
"epoch": 2.6276276276276276,
|
2237 |
+
"grad_norm": 0.0005333448643796146,
|
2238 |
+
"learning_rate": 6.896551724137932e-06,
|
2239 |
+
"loss": 0.0003,
|
2240 |
+
"step": 7875
|
2241 |
+
},
|
2242 |
+
{
|
2243 |
+
"epoch": 2.6359693026359694,
|
2244 |
+
"grad_norm": 0.03314538672566414,
|
2245 |
+
"learning_rate": 6.742059077987888e-06,
|
2246 |
+
"loss": 0.0019,
|
2247 |
+
"step": 7900
|
2248 |
+
},
|
2249 |
+
{
|
2250 |
+
"epoch": 2.6443109776443112,
|
2251 |
+
"grad_norm": 0.0016822376055642962,
|
2252 |
+
"learning_rate": 6.587566431837845e-06,
|
2253 |
+
"loss": 0.0063,
|
2254 |
+
"step": 7925
|
2255 |
+
},
|
2256 |
+
{
|
2257 |
+
"epoch": 2.6526526526526526,
|
2258 |
+
"grad_norm": 0.0006041646120138466,
|
2259 |
+
"learning_rate": 6.433073785687801e-06,
|
2260 |
+
"loss": 0.0001,
|
2261 |
+
"step": 7950
|
2262 |
+
},
|
2263 |
+
{
|
2264 |
+
"epoch": 2.6609943276609944,
|
2265 |
+
"grad_norm": 0.00021345775166992098,
|
2266 |
+
"learning_rate": 6.278581139537758e-06,
|
2267 |
+
"loss": 0.0001,
|
2268 |
+
"step": 7975
|
2269 |
+
},
|
2270 |
+
{
|
2271 |
+
"epoch": 2.6693360026693362,
|
2272 |
+
"grad_norm": 0.00019755800894927233,
|
2273 |
+
"learning_rate": 6.124088493387715e-06,
|
2274 |
+
"loss": 0.0,
|
2275 |
+
"step": 8000
|
2276 |
+
},
|
2277 |
+
{
|
2278 |
+
"epoch": 2.6776776776776776,
|
2279 |
+
"grad_norm": 0.0004531990271061659,
|
2280 |
+
"learning_rate": 5.969595847237672e-06,
|
2281 |
+
"loss": 0.0,
|
2282 |
+
"step": 8025
|
2283 |
+
},
|
2284 |
+
{
|
2285 |
+
"epoch": 2.6860193526860194,
|
2286 |
+
"grad_norm": 0.002286186208948493,
|
2287 |
+
"learning_rate": 5.815103201087629e-06,
|
2288 |
+
"loss": 0.0,
|
2289 |
+
"step": 8050
|
2290 |
+
},
|
2291 |
+
{
|
2292 |
+
"epoch": 2.6943610276943613,
|
2293 |
+
"grad_norm": 9.491475066170096e-05,
|
2294 |
+
"learning_rate": 5.660610554937585e-06,
|
2295 |
+
"loss": 0.0002,
|
2296 |
+
"step": 8075
|
2297 |
+
},
|
2298 |
+
{
|
2299 |
+
"epoch": 2.7027027027027026,
|
2300 |
+
"grad_norm": 0.00024974957341328263,
|
2301 |
+
"learning_rate": 5.506117908787542e-06,
|
2302 |
+
"loss": 0.0044,
|
2303 |
+
"step": 8100
|
2304 |
+
},
|
2305 |
+
{
|
2306 |
+
"epoch": 2.7110443777110445,
|
2307 |
+
"grad_norm": 0.0010395270073786378,
|
2308 |
+
"learning_rate": 5.351625262637499e-06,
|
2309 |
+
"loss": 0.0001,
|
2310 |
+
"step": 8125
|
2311 |
+
},
|
2312 |
+
{
|
2313 |
+
"epoch": 2.7193860527193863,
|
2314 |
+
"grad_norm": 0.0001526384148746729,
|
2315 |
+
"learning_rate": 5.197132616487455e-06,
|
2316 |
+
"loss": 0.0,
|
2317 |
+
"step": 8150
|
2318 |
+
},
|
2319 |
+
{
|
2320 |
+
"epoch": 2.7277277277277276,
|
2321 |
+
"grad_norm": 0.0006884171743877232,
|
2322 |
+
"learning_rate": 5.042639970337412e-06,
|
2323 |
+
"loss": 0.0001,
|
2324 |
+
"step": 8175
|
2325 |
+
},
|
2326 |
+
{
|
2327 |
+
"epoch": 2.7360694027360695,
|
2328 |
+
"grad_norm": 9.014210809255019e-05,
|
2329 |
+
"learning_rate": 4.88814732418737e-06,
|
2330 |
+
"loss": 0.0001,
|
2331 |
+
"step": 8200
|
2332 |
+
},
|
2333 |
+
{
|
2334 |
+
"epoch": 2.7444110777444113,
|
2335 |
+
"grad_norm": 0.0002122920996043831,
|
2336 |
+
"learning_rate": 4.733654678037326e-06,
|
2337 |
+
"loss": 0.0001,
|
2338 |
+
"step": 8225
|
2339 |
+
},
|
2340 |
+
{
|
2341 |
+
"epoch": 2.7527527527527527,
|
2342 |
+
"grad_norm": 0.0002799911017064005,
|
2343 |
+
"learning_rate": 4.579162031887282e-06,
|
2344 |
+
"loss": 0.0001,
|
2345 |
+
"step": 8250
|
2346 |
+
},
|
2347 |
+
{
|
2348 |
+
"epoch": 2.7610944277610945,
|
2349 |
+
"grad_norm": 0.00325685553252697,
|
2350 |
+
"learning_rate": 4.424669385737239e-06,
|
2351 |
+
"loss": 0.0002,
|
2352 |
+
"step": 8275
|
2353 |
+
},
|
2354 |
+
{
|
2355 |
+
"epoch": 2.7694361027694363,
|
2356 |
+
"grad_norm": 0.00023966682783793658,
|
2357 |
+
"learning_rate": 4.270176739587196e-06,
|
2358 |
+
"loss": 0.0001,
|
2359 |
+
"step": 8300
|
2360 |
+
},
|
2361 |
+
{
|
2362 |
+
"epoch": 2.7777777777777777,
|
2363 |
+
"grad_norm": 0.00024113825929816812,
|
2364 |
+
"learning_rate": 4.115684093437153e-06,
|
2365 |
+
"loss": 0.0,
|
2366 |
+
"step": 8325
|
2367 |
+
},
|
2368 |
+
{
|
2369 |
+
"epoch": 2.7861194527861195,
|
2370 |
+
"grad_norm": 0.00069383904337883,
|
2371 |
+
"learning_rate": 3.96119144728711e-06,
|
2372 |
+
"loss": 0.0282,
|
2373 |
+
"step": 8350
|
2374 |
+
},
|
2375 |
+
{
|
2376 |
+
"epoch": 2.7944611277944613,
|
2377 |
+
"grad_norm": 0.00674516661092639,
|
2378 |
+
"learning_rate": 3.8066988011370657e-06,
|
2379 |
+
"loss": 0.0,
|
2380 |
+
"step": 8375
|
2381 |
+
},
|
2382 |
+
{
|
2383 |
+
"epoch": 2.8028028028028027,
|
2384 |
+
"grad_norm": 0.0006425007013604045,
|
2385 |
+
"learning_rate": 3.652206154987023e-06,
|
2386 |
+
"loss": 0.0001,
|
2387 |
+
"step": 8400
|
2388 |
+
},
|
2389 |
+
{
|
2390 |
+
"epoch": 2.8111444778111445,
|
2391 |
+
"grad_norm": 0.001739656669087708,
|
2392 |
+
"learning_rate": 3.49771350883698e-06,
|
2393 |
+
"loss": 0.0003,
|
2394 |
+
"step": 8425
|
2395 |
+
},
|
2396 |
+
{
|
2397 |
+
"epoch": 2.8194861528194863,
|
2398 |
+
"grad_norm": 0.00028822454623878,
|
2399 |
+
"learning_rate": 3.343220862686936e-06,
|
2400 |
+
"loss": 0.0022,
|
2401 |
+
"step": 8450
|
2402 |
+
},
|
2403 |
+
{
|
2404 |
+
"epoch": 2.8278278278278277,
|
2405 |
+
"grad_norm": 0.0007758860592730343,
|
2406 |
+
"learning_rate": 3.188728216536893e-06,
|
2407 |
+
"loss": 0.0002,
|
2408 |
+
"step": 8475
|
2409 |
+
},
|
2410 |
+
{
|
2411 |
+
"epoch": 2.8361695028361695,
|
2412 |
+
"grad_norm": 6.443789607146755e-05,
|
2413 |
+
"learning_rate": 3.03423557038685e-06,
|
2414 |
+
"loss": 0.0003,
|
2415 |
+
"step": 8500
|
2416 |
+
},
|
2417 |
+
{
|
2418 |
+
"epoch": 2.8445111778445114,
|
2419 |
+
"grad_norm": 0.008539211004972458,
|
2420 |
+
"learning_rate": 2.8797429242368063e-06,
|
2421 |
+
"loss": 0.0009,
|
2422 |
+
"step": 8525
|
2423 |
+
},
|
2424 |
+
{
|
2425 |
+
"epoch": 2.8528528528528527,
|
2426 |
+
"grad_norm": 0.0164475180208683,
|
2427 |
+
"learning_rate": 2.7252502780867635e-06,
|
2428 |
+
"loss": 0.0098,
|
2429 |
+
"step": 8550
|
2430 |
+
},
|
2431 |
+
{
|
2432 |
+
"epoch": 2.8611945278611945,
|
2433 |
+
"grad_norm": 0.0004140229430049658,
|
2434 |
+
"learning_rate": 2.57075763193672e-06,
|
2435 |
+
"loss": 0.0,
|
2436 |
+
"step": 8575
|
2437 |
+
},
|
2438 |
+
{
|
2439 |
+
"epoch": 2.8695362028695364,
|
2440 |
+
"grad_norm": 0.0007194245117716491,
|
2441 |
+
"learning_rate": 2.4162649857866768e-06,
|
2442 |
+
"loss": 0.0002,
|
2443 |
+
"step": 8600
|
2444 |
+
},
|
2445 |
+
{
|
2446 |
+
"epoch": 2.8778778778778777,
|
2447 |
+
"grad_norm": 8.639365114504471e-05,
|
2448 |
+
"learning_rate": 2.261772339636633e-06,
|
2449 |
+
"loss": 0.0,
|
2450 |
+
"step": 8625
|
2451 |
+
},
|
2452 |
+
{
|
2453 |
+
"epoch": 2.8862195528862196,
|
2454 |
+
"grad_norm": 0.00026035771588794887,
|
2455 |
+
"learning_rate": 2.1072796934865904e-06,
|
2456 |
+
"loss": 0.0,
|
2457 |
+
"step": 8650
|
2458 |
+
},
|
2459 |
+
{
|
2460 |
+
"epoch": 2.8945612278945614,
|
2461 |
+
"grad_norm": 0.0003119956236332655,
|
2462 |
+
"learning_rate": 1.952787047336547e-06,
|
2463 |
+
"loss": 0.0164,
|
2464 |
+
"step": 8675
|
2465 |
+
},
|
2466 |
+
{
|
2467 |
+
"epoch": 2.9029029029029028,
|
2468 |
+
"grad_norm": 0.0001721863227430731,
|
2469 |
+
"learning_rate": 1.7982944011865034e-06,
|
2470 |
+
"loss": 0.0,
|
2471 |
+
"step": 8700
|
2472 |
+
},
|
2473 |
+
{
|
2474 |
+
"epoch": 2.9112445779112446,
|
2475 |
+
"grad_norm": 0.0003014954272657633,
|
2476 |
+
"learning_rate": 1.6438017550364605e-06,
|
2477 |
+
"loss": 0.0002,
|
2478 |
+
"step": 8725
|
2479 |
+
},
|
2480 |
+
{
|
2481 |
+
"epoch": 2.9195862529195864,
|
2482 |
+
"grad_norm": 0.002242905320599675,
|
2483 |
+
"learning_rate": 1.489309108886417e-06,
|
2484 |
+
"loss": 0.0001,
|
2485 |
+
"step": 8750
|
2486 |
+
},
|
2487 |
+
{
|
2488 |
+
"epoch": 2.9279279279279278,
|
2489 |
+
"grad_norm": 0.00019342848099768162,
|
2490 |
+
"learning_rate": 1.334816462736374e-06,
|
2491 |
+
"loss": 0.0,
|
2492 |
+
"step": 8775
|
2493 |
+
},
|
2494 |
+
{
|
2495 |
+
"epoch": 2.9362696029362696,
|
2496 |
+
"grad_norm": 0.00041662275907583535,
|
2497 |
+
"learning_rate": 1.1803238165863305e-06,
|
2498 |
+
"loss": 0.0023,
|
2499 |
+
"step": 8800
|
2500 |
+
},
|
2501 |
+
{
|
2502 |
+
"epoch": 2.9446112779446114,
|
2503 |
+
"grad_norm": 8.121671999106184e-05,
|
2504 |
+
"learning_rate": 1.0258311704362874e-06,
|
2505 |
+
"loss": 0.0,
|
2506 |
+
"step": 8825
|
2507 |
+
},
|
2508 |
+
{
|
2509 |
+
"epoch": 2.952952952952953,
|
2510 |
+
"grad_norm": 0.00027466562460176647,
|
2511 |
+
"learning_rate": 8.71338524286244e-07,
|
2512 |
+
"loss": 0.0,
|
2513 |
+
"step": 8850
|
2514 |
+
},
|
2515 |
+
{
|
2516 |
+
"epoch": 2.9612946279612946,
|
2517 |
+
"grad_norm": 0.00015804750728420913,
|
2518 |
+
"learning_rate": 7.168458781362007e-07,
|
2519 |
+
"loss": 0.0,
|
2520 |
+
"step": 8875
|
2521 |
+
},
|
2522 |
+
{
|
2523 |
+
"epoch": 2.9696363029696364,
|
2524 |
+
"grad_norm": 5.4930391343077645e-05,
|
2525 |
+
"learning_rate": 5.623532319861574e-07,
|
2526 |
+
"loss": 0.0011,
|
2527 |
+
"step": 8900
|
2528 |
+
},
|
2529 |
+
{
|
2530 |
+
"epoch": 2.977977977977978,
|
2531 |
+
"grad_norm": 0.0005976063548587263,
|
2532 |
+
"learning_rate": 4.0786058583611417e-07,
|
2533 |
+
"loss": 0.0015,
|
2534 |
+
"step": 8925
|
2535 |
+
},
|
2536 |
+
{
|
2537 |
+
"epoch": 2.9863196529863196,
|
2538 |
+
"grad_norm": 0.0008383295498788357,
|
2539 |
+
"learning_rate": 2.5336793968607095e-07,
|
2540 |
+
"loss": 0.0,
|
2541 |
+
"step": 8950
|
2542 |
+
},
|
2543 |
+
{
|
2544 |
+
"epoch": 2.9946613279946614,
|
2545 |
+
"grad_norm": 0.001575443078763783,
|
2546 |
+
"learning_rate": 9.88752935360277e-08,
|
2547 |
+
"loss": 0.0102,
|
2548 |
+
"step": 8975
|
2549 |
+
},
|
2550 |
+
{
|
2551 |
+
"epoch": 3.0,
|
2552 |
+
"eval_accuracy": 0.9988321654988321,
|
2553 |
+
"eval_auc": 0.9999543462381301,
|
2554 |
+
"eval_f1": 0.9523809523809523,
|
2555 |
+
"eval_loss": 0.003335413057357073,
|
2556 |
+
"eval_precision": 0.958904109589041,
|
2557 |
+
"eval_recall": 0.9459459459459459,
|
2558 |
+
"eval_runtime": 118.1878,
|
2559 |
+
"eval_samples_per_second": 50.716,
|
2560 |
+
"eval_steps_per_second": 3.173,
|
2561 |
+
"step": 8991
|
2562 |
+
}
|
2563 |
+
],
|
2564 |
+
"logging_steps": 25,
|
2565 |
+
"max_steps": 8991,
|
2566 |
+
"num_input_tokens_seen": 0,
|
2567 |
+
"num_train_epochs": 3,
|
2568 |
+
"save_steps": 500,
|
2569 |
+
"stateful_callbacks": {
|
2570 |
+
"EarlyStoppingCallback": {
|
2571 |
+
"args": {
|
2572 |
+
"early_stopping_patience": 5,
|
2573 |
+
"early_stopping_threshold": 0.01
|
2574 |
+
},
|
2575 |
+
"attributes": {
|
2576 |
+
"early_stopping_patience_counter": 0
|
2577 |
+
}
|
2578 |
+
},
|
2579 |
+
"TrainerControl": {
|
2580 |
+
"args": {
|
2581 |
+
"should_epoch_stop": false,
|
2582 |
+
"should_evaluate": false,
|
2583 |
+
"should_log": false,
|
2584 |
+
"should_save": true,
|
2585 |
+
"should_training_stop": true
|
2586 |
+
},
|
2587 |
+
"attributes": {}
|
2588 |
+
}
|
2589 |
+
},
|
2590 |
+
"total_flos": 1.6433249389401342e+19,
|
2591 |
+
"train_batch_size": 8,
|
2592 |
+
"trial_name": null,
|
2593 |
+
"trial_params": null
|
2594 |
+
}
|
checkpoint-8991/training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d638842dd7f1b24e15fbb1462efb8d8f36d068712fe4457a976161e629b3ae1c
|
3 |
+
size 5112
|
config.json
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "google/vit-base-patch16-384",
|
3 |
+
"_num_labels": 2,
|
4 |
+
"architectures": [
|
5 |
+
"ViTForImageClassification"
|
6 |
+
],
|
7 |
+
"attention_probs_dropout_prob": 0.0,
|
8 |
+
"encoder_stride": 16,
|
9 |
+
"hidden_act": "gelu",
|
10 |
+
"hidden_dropout_prob": 0.0,
|
11 |
+
"hidden_size": 768,
|
12 |
+
"id2label": {
|
13 |
+
"0": "not_soyjak",
|
14 |
+
"1": "soyjak"
|
15 |
+
},
|
16 |
+
"image_size": 384,
|
17 |
+
"initializer_range": 0.02,
|
18 |
+
"intermediate_size": 3072,
|
19 |
+
"label2id": {
|
20 |
+
"not_soyjak": 0,
|
21 |
+
"soyjak": 1
|
22 |
+
},
|
23 |
+
"layer_norm_eps": 1e-12,
|
24 |
+
"model_type": "vit",
|
25 |
+
"num_attention_heads": 12,
|
26 |
+
"num_channels": 3,
|
27 |
+
"num_hidden_layers": 12,
|
28 |
+
"patch_size": 16,
|
29 |
+
"problem_type": "single_label_classification",
|
30 |
+
"qkv_bias": true,
|
31 |
+
"torch_dtype": "float32",
|
32 |
+
"transformers_version": "4.42.3"
|
33 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:227d87907aafc9cb1de9990403be5dcfdd6ff412a4009e175a9c7659fa4f6ca3
|
3 |
+
size 344391328
|
preprocessor_config.json
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"do_normalize": true,
|
3 |
+
"do_rescale": true,
|
4 |
+
"do_resize": true,
|
5 |
+
"image_mean": [
|
6 |
+
0.5,
|
7 |
+
0.5,
|
8 |
+
0.5
|
9 |
+
],
|
10 |
+
"image_processor_type": "ViTImageProcessor",
|
11 |
+
"image_std": [
|
12 |
+
0.5,
|
13 |
+
0.5,
|
14 |
+
0.5
|
15 |
+
],
|
16 |
+
"resample": 2,
|
17 |
+
"rescale_factor": 0.00392156862745098,
|
18 |
+
"size": {
|
19 |
+
"height": 384,
|
20 |
+
"width": 384
|
21 |
+
}
|
22 |
+
}
|
runs/Oct20_19-44-40_Haileys-MacBook-Pro.local/events.out.tfevents.1729478680.Haileys-MacBook-Pro.local.88288.0
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:2af8a2c598a7339753e0c2555c3deed6d9f6a567642978d893b72c2ae815f3bc
|
3 |
+
size 82487
|
runs/Oct20_19-44-40_Haileys-MacBook-Pro.local/events.out.tfevents.1729483332.Haileys-MacBook-Pro.local.88288.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e2d9ae7df0da4f8b419d7dd70dd5062128763758e0fd292551299aec2fa2ef60
|
3 |
+
size 607
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d638842dd7f1b24e15fbb1462efb8d8f36d068712fe4457a976161e629b3ae1c
|
3 |
+
size 5112
|
training_params.json
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"data_path": "sj-det-2/autotrain-data",
|
3 |
+
"model": "google/vit-base-patch16-384",
|
4 |
+
"username": "haiefff",
|
5 |
+
"lr": 5e-05,
|
6 |
+
"epochs": 3,
|
7 |
+
"batch_size": 8,
|
8 |
+
"warmup_ratio": 0.1,
|
9 |
+
"gradient_accumulation": 1,
|
10 |
+
"optimizer": "adamw_torch",
|
11 |
+
"scheduler": "linear",
|
12 |
+
"weight_decay": 0.0,
|
13 |
+
"max_grad_norm": 1.0,
|
14 |
+
"seed": 42,
|
15 |
+
"train_split": "train",
|
16 |
+
"valid_split": "validation",
|
17 |
+
"logging_steps": -1,
|
18 |
+
"project_name": "sj-det-2",
|
19 |
+
"auto_find_batch_size": false,
|
20 |
+
"mixed_precision": "none",
|
21 |
+
"save_total_limit": 1,
|
22 |
+
"push_to_hub": true,
|
23 |
+
"eval_strategy": "epoch",
|
24 |
+
"image_column": "autotrain_image",
|
25 |
+
"target_column": "autotrain_label",
|
26 |
+
"log": "tensorboard",
|
27 |
+
"early_stopping_patience": 5,
|
28 |
+
"early_stopping_threshold": 0.01
|
29 |
+
}
|