Upload folder using huggingface_hub
Browse files- README.md +31 -0
- checkpoint-9162/config.json +33 -0
- checkpoint-9162/model.safetensors +3 -0
- checkpoint-9162/optimizer.pt +3 -0
- checkpoint-9162/rng_state.pth +3 -0
- checkpoint-9162/scheduler.pt +3 -0
- checkpoint-9162/trainer_state.json +2643 -0
- checkpoint-9162/training_args.bin +3 -0
- config.json +33 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- runs/Oct22_01-12-49_r-haiefff-sjstuff-186zjrrd-42d15-6st93/events.out.tfevents.1729559571.r-haiefff-sjstuff-186zjrrd-42d15-6st93.222.0 +2 -2
- runs/Oct22_01-12-49_r-haiefff-sjstuff-186zjrrd-42d15-6st93/events.out.tfevents.1729562910.r-haiefff-sjstuff-186zjrrd-42d15-6st93.222.1 +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-large-patch32-384
|
| 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: 0.0027391575276851654
|
| 22 |
+
|
| 23 |
+
f1: 0.9523809523809523
|
| 24 |
+
|
| 25 |
+
precision: 0.958904109589041
|
| 26 |
+
|
| 27 |
+
recall: 0.9459459459459459
|
| 28 |
+
|
| 29 |
+
auc: 0.9999216153508497
|
| 30 |
+
|
| 31 |
+
accuracy: 0.9988539620170268
|
checkpoint-9162/config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-large-patch32-384",
|
| 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": 1024,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "not_soyjak",
|
| 14 |
+
"1": "soyjak"
|
| 15 |
+
},
|
| 16 |
+
"image_size": 384,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 4096,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"not_soyjak": 0,
|
| 21 |
+
"soyjak": 1
|
| 22 |
+
},
|
| 23 |
+
"layer_norm_eps": 1e-12,
|
| 24 |
+
"model_type": "vit",
|
| 25 |
+
"num_attention_heads": 16,
|
| 26 |
+
"num_channels": 3,
|
| 27 |
+
"num_hidden_layers": 24,
|
| 28 |
+
"patch_size": 32,
|
| 29 |
+
"problem_type": "single_label_classification",
|
| 30 |
+
"qkv_bias": true,
|
| 31 |
+
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.45.0"
|
| 33 |
+
}
|
checkpoint-9162/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6370312cf8b3c3c9cdcab23d0c40535752368b0bcb77471852c134abba4add6d
|
| 3 |
+
size 1222485480
|
checkpoint-9162/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ff9d283ec04f6990599d0ae675f95ff9f54feb96dc5674cb5a7a813e121b08d
|
| 3 |
+
size 2445205930
|
checkpoint-9162/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e455470a03ae85cd92795a22919ec1902e11da55fff6d7916647ba425512bc09
|
| 3 |
+
size 14244
|
checkpoint-9162/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9637f4f3050c34b296165f7c171e34f47e27e0faf3a84aaa55c6c80064be806e
|
| 3 |
+
size 1064
|
checkpoint-9162/trainer_state.json
ADDED
|
@@ -0,0 +1,2643 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.0027391575276851654,
|
| 3 |
+
"best_model_checkpoint": "sj-det-6/checkpoint-9162",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 9162,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.008185985592665358,
|
| 13 |
+
"grad_norm": 14.986274719238281,
|
| 14 |
+
"learning_rate": 1.2540894220283533e-06,
|
| 15 |
+
"loss": 1.1682,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.016371971185330715,
|
| 20 |
+
"grad_norm": 8.189007759094238,
|
| 21 |
+
"learning_rate": 2.6172300981461287e-06,
|
| 22 |
+
"loss": 0.7882,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.02455795677799607,
|
| 27 |
+
"grad_norm": 4.545085906982422,
|
| 28 |
+
"learning_rate": 3.980370774263904e-06,
|
| 29 |
+
"loss": 0.3987,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.03274394237066143,
|
| 34 |
+
"grad_norm": 0.8907670974731445,
|
| 35 |
+
"learning_rate": 5.343511450381679e-06,
|
| 36 |
+
"loss": 0.1198,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.040929927963326784,
|
| 41 |
+
"grad_norm": 0.8682488203048706,
|
| 42 |
+
"learning_rate": 6.7066521264994545e-06,
|
| 43 |
+
"loss": 0.1166,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.04911591355599214,
|
| 48 |
+
"grad_norm": 0.1406942456960678,
|
| 49 |
+
"learning_rate": 8.069792802617231e-06,
|
| 50 |
+
"loss": 0.0797,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0573018991486575,
|
| 55 |
+
"grad_norm": 0.030166020616889,
|
| 56 |
+
"learning_rate": 9.432933478735007e-06,
|
| 57 |
+
"loss": 0.0518,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.06548788474132286,
|
| 62 |
+
"grad_norm": 0.0494314506649971,
|
| 63 |
+
"learning_rate": 1.0796074154852782e-05,
|
| 64 |
+
"loss": 0.0505,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.07367387033398821,
|
| 69 |
+
"grad_norm": 0.01392674632370472,
|
| 70 |
+
"learning_rate": 1.2159214830970557e-05,
|
| 71 |
+
"loss": 0.0272,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.08185985592665357,
|
| 76 |
+
"grad_norm": 0.0814874917268753,
|
| 77 |
+
"learning_rate": 1.3522355507088334e-05,
|
| 78 |
+
"loss": 0.0365,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.09004584151931892,
|
| 83 |
+
"grad_norm": 0.04325518757104874,
|
| 84 |
+
"learning_rate": 1.4885496183206107e-05,
|
| 85 |
+
"loss": 0.0576,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.09823182711198428,
|
| 90 |
+
"grad_norm": 0.09735051542520523,
|
| 91 |
+
"learning_rate": 1.624863685932388e-05,
|
| 92 |
+
"loss": 0.0329,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.10641781270464964,
|
| 97 |
+
"grad_norm": 0.64588862657547,
|
| 98 |
+
"learning_rate": 1.761177753544166e-05,
|
| 99 |
+
"loss": 0.0802,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.114603798297315,
|
| 104 |
+
"grad_norm": 0.011760449036955833,
|
| 105 |
+
"learning_rate": 1.8974918211559434e-05,
|
| 106 |
+
"loss": 0.0524,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.12278978388998035,
|
| 111 |
+
"grad_norm": 0.1449550837278366,
|
| 112 |
+
"learning_rate": 2.033805888767721e-05,
|
| 113 |
+
"loss": 0.0347,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.13097576948264572,
|
| 118 |
+
"grad_norm": 0.018402298912405968,
|
| 119 |
+
"learning_rate": 2.1701199563794985e-05,
|
| 120 |
+
"loss": 0.0155,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13916175507531106,
|
| 125 |
+
"grad_norm": 0.016214437782764435,
|
| 126 |
+
"learning_rate": 2.306434023991276e-05,
|
| 127 |
+
"loss": 0.0988,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.14734774066797643,
|
| 132 |
+
"grad_norm": 0.08800540119409561,
|
| 133 |
+
"learning_rate": 2.4427480916030535e-05,
|
| 134 |
+
"loss": 0.0428,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.15553372626064177,
|
| 139 |
+
"grad_norm": 0.005783313885331154,
|
| 140 |
+
"learning_rate": 2.5790621592148313e-05,
|
| 141 |
+
"loss": 0.1005,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.16371971185330714,
|
| 146 |
+
"grad_norm": 0.028863191604614258,
|
| 147 |
+
"learning_rate": 2.7153762268266085e-05,
|
| 148 |
+
"loss": 0.0427,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1719056974459725,
|
| 153 |
+
"grad_norm": 6.242043495178223,
|
| 154 |
+
"learning_rate": 2.851690294438386e-05,
|
| 155 |
+
"loss": 0.0868,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.18009168303863785,
|
| 160 |
+
"grad_norm": 0.013895904645323753,
|
| 161 |
+
"learning_rate": 2.9880043620501635e-05,
|
| 162 |
+
"loss": 0.0898,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1882776686313032,
|
| 167 |
+
"grad_norm": 0.019419686868786812,
|
| 168 |
+
"learning_rate": 3.124318429661941e-05,
|
| 169 |
+
"loss": 0.0351,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.19646365422396855,
|
| 174 |
+
"grad_norm": 0.008130255155265331,
|
| 175 |
+
"learning_rate": 3.260632497273719e-05,
|
| 176 |
+
"loss": 0.0164,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.20464963981663392,
|
| 181 |
+
"grad_norm": 6.5059814453125,
|
| 182 |
+
"learning_rate": 3.3969465648854964e-05,
|
| 183 |
+
"loss": 0.0359,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.2128356254092993,
|
| 188 |
+
"grad_norm": 0.010089697316288948,
|
| 189 |
+
"learning_rate": 3.533260632497274e-05,
|
| 190 |
+
"loss": 0.0305,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.22102161100196463,
|
| 195 |
+
"grad_norm": 0.014347407966852188,
|
| 196 |
+
"learning_rate": 3.6695747001090515e-05,
|
| 197 |
+
"loss": 0.0435,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.22920759659463,
|
| 202 |
+
"grad_norm": 0.017506686970591545,
|
| 203 |
+
"learning_rate": 3.805888767720829e-05,
|
| 204 |
+
"loss": 0.0486,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.23739358218729534,
|
| 209 |
+
"grad_norm": 0.004301242996007204,
|
| 210 |
+
"learning_rate": 3.9422028353326065e-05,
|
| 211 |
+
"loss": 0.0002,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.2455795677799607,
|
| 216 |
+
"grad_norm": 0.00440743425861001,
|
| 217 |
+
"learning_rate": 4.078516902944384e-05,
|
| 218 |
+
"loss": 0.0134,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.2537655533726261,
|
| 223 |
+
"grad_norm": 0.037835538387298584,
|
| 224 |
+
"learning_rate": 4.2148309705561615e-05,
|
| 225 |
+
"loss": 0.0334,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.26195153896529144,
|
| 230 |
+
"grad_norm": 0.11478447914123535,
|
| 231 |
+
"learning_rate": 4.351145038167939e-05,
|
| 232 |
+
"loss": 0.0665,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.27013752455795675,
|
| 237 |
+
"grad_norm": 0.006648090668022633,
|
| 238 |
+
"learning_rate": 4.4874591057797165e-05,
|
| 239 |
+
"loss": 0.0033,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.2783235101506221,
|
| 244 |
+
"grad_norm": 8.948714256286621,
|
| 245 |
+
"learning_rate": 4.623773173391494e-05,
|
| 246 |
+
"loss": 0.0996,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.2865094957432875,
|
| 251 |
+
"grad_norm": 0.012912426143884659,
|
| 252 |
+
"learning_rate": 4.7600872410032716e-05,
|
| 253 |
+
"loss": 0.0102,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.29469548133595286,
|
| 258 |
+
"grad_norm": 0.0026352498680353165,
|
| 259 |
+
"learning_rate": 4.89640130861505e-05,
|
| 260 |
+
"loss": 0.0226,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3028814669286182,
|
| 265 |
+
"grad_norm": 0.15103408694267273,
|
| 266 |
+
"learning_rate": 4.996361431170407e-05,
|
| 267 |
+
"loss": 0.0225,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.31106745252128354,
|
| 272 |
+
"grad_norm": 0.04878251254558563,
|
| 273 |
+
"learning_rate": 4.981200727713766e-05,
|
| 274 |
+
"loss": 0.0005,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3192534381139489,
|
| 279 |
+
"grad_norm": 0.00841938890516758,
|
| 280 |
+
"learning_rate": 4.966040024257126e-05,
|
| 281 |
+
"loss": 0.0513,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.3274394237066143,
|
| 286 |
+
"grad_norm": 0.04339870065450668,
|
| 287 |
+
"learning_rate": 4.9508793208004854e-05,
|
| 288 |
+
"loss": 0.0328,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.33562540929927964,
|
| 293 |
+
"grad_norm": 0.00045626997598446906,
|
| 294 |
+
"learning_rate": 4.935718617343845e-05,
|
| 295 |
+
"loss": 0.0124,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.343811394891945,
|
| 300 |
+
"grad_norm": 0.00013843162741977721,
|
| 301 |
+
"learning_rate": 4.9205579138872046e-05,
|
| 302 |
+
"loss": 0.0,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3519973804846103,
|
| 307 |
+
"grad_norm": 1.9331814050674438,
|
| 308 |
+
"learning_rate": 4.905397210430564e-05,
|
| 309 |
+
"loss": 0.0504,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.3601833660772757,
|
| 314 |
+
"grad_norm": 0.002882858505472541,
|
| 315 |
+
"learning_rate": 4.890236506973924e-05,
|
| 316 |
+
"loss": 0.1157,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.36836935166994106,
|
| 321 |
+
"grad_norm": 0.022938862442970276,
|
| 322 |
+
"learning_rate": 4.875075803517283e-05,
|
| 323 |
+
"loss": 0.0716,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3765553372626064,
|
| 328 |
+
"grad_norm": 0.0010779292788356543,
|
| 329 |
+
"learning_rate": 4.859915100060643e-05,
|
| 330 |
+
"loss": 0.0419,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.3847413228552718,
|
| 335 |
+
"grad_norm": 0.03177574649453163,
|
| 336 |
+
"learning_rate": 4.8447543966040024e-05,
|
| 337 |
+
"loss": 0.0249,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3929273084479371,
|
| 342 |
+
"grad_norm": 0.0006776170921511948,
|
| 343 |
+
"learning_rate": 4.8295936931473626e-05,
|
| 344 |
+
"loss": 0.0415,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4011132940406025,
|
| 349 |
+
"grad_norm": 0.00865075271576643,
|
| 350 |
+
"learning_rate": 4.814432989690722e-05,
|
| 351 |
+
"loss": 0.0304,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.40929927963326784,
|
| 356 |
+
"grad_norm": 0.004142033401876688,
|
| 357 |
+
"learning_rate": 4.799272286234082e-05,
|
| 358 |
+
"loss": 0.066,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4174852652259332,
|
| 363 |
+
"grad_norm": 0.006911768112331629,
|
| 364 |
+
"learning_rate": 4.784111582777441e-05,
|
| 365 |
+
"loss": 0.0345,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.4256712508185986,
|
| 370 |
+
"grad_norm": 0.011829985305666924,
|
| 371 |
+
"learning_rate": 4.768950879320801e-05,
|
| 372 |
+
"loss": 0.0324,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.4338572364112639,
|
| 377 |
+
"grad_norm": 0.024410191923379898,
|
| 378 |
+
"learning_rate": 4.7537901758641604e-05,
|
| 379 |
+
"loss": 0.0412,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.44204322200392926,
|
| 384 |
+
"grad_norm": 0.0020343943033367395,
|
| 385 |
+
"learning_rate": 4.73862947240752e-05,
|
| 386 |
+
"loss": 0.0133,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.4502292075965946,
|
| 391 |
+
"grad_norm": 0.0009943109471350908,
|
| 392 |
+
"learning_rate": 4.7234687689508795e-05,
|
| 393 |
+
"loss": 0.0004,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.45841519318926,
|
| 398 |
+
"grad_norm": 0.00024950483930297196,
|
| 399 |
+
"learning_rate": 4.708308065494239e-05,
|
| 400 |
+
"loss": 0.0,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.46660117878192536,
|
| 405 |
+
"grad_norm": 0.0015270832227542996,
|
| 406 |
+
"learning_rate": 4.6931473620375986e-05,
|
| 407 |
+
"loss": 0.0314,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.4747871643745907,
|
| 412 |
+
"grad_norm": 9.285660743713379,
|
| 413 |
+
"learning_rate": 4.677986658580958e-05,
|
| 414 |
+
"loss": 0.0955,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.48297314996725604,
|
| 419 |
+
"grad_norm": 0.004155229777097702,
|
| 420 |
+
"learning_rate": 4.6628259551243184e-05,
|
| 421 |
+
"loss": 0.003,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.4911591355599214,
|
| 426 |
+
"grad_norm": 0.013534590601921082,
|
| 427 |
+
"learning_rate": 4.647665251667678e-05,
|
| 428 |
+
"loss": 0.0269,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.4993451211525868,
|
| 433 |
+
"grad_norm": 0.004476598929613829,
|
| 434 |
+
"learning_rate": 4.6325045482110375e-05,
|
| 435 |
+
"loss": 0.0011,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5075311067452521,
|
| 440 |
+
"grad_norm": 0.0038523597177118063,
|
| 441 |
+
"learning_rate": 4.617343844754397e-05,
|
| 442 |
+
"loss": 0.0015,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5157170923379175,
|
| 447 |
+
"grad_norm": 0.03839994966983795,
|
| 448 |
+
"learning_rate": 4.602183141297756e-05,
|
| 449 |
+
"loss": 0.1135,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.5239030779305829,
|
| 454 |
+
"grad_norm": 0.000390953995520249,
|
| 455 |
+
"learning_rate": 4.587022437841116e-05,
|
| 456 |
+
"loss": 0.0015,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.5320890635232483,
|
| 461 |
+
"grad_norm": 0.00021860921697225422,
|
| 462 |
+
"learning_rate": 4.571861734384476e-05,
|
| 463 |
+
"loss": 0.0015,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5402750491159135,
|
| 468 |
+
"grad_norm": 0.005596529692411423,
|
| 469 |
+
"learning_rate": 4.556701030927835e-05,
|
| 470 |
+
"loss": 0.1111,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5484610347085789,
|
| 475 |
+
"grad_norm": 5.192307949066162,
|
| 476 |
+
"learning_rate": 4.541540327471195e-05,
|
| 477 |
+
"loss": 0.056,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.5566470203012442,
|
| 482 |
+
"grad_norm": 0.004266222473233938,
|
| 483 |
+
"learning_rate": 4.5263796240145544e-05,
|
| 484 |
+
"loss": 0.0078,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.5648330058939096,
|
| 489 |
+
"grad_norm": 4.940586090087891,
|
| 490 |
+
"learning_rate": 4.511218920557914e-05,
|
| 491 |
+
"loss": 0.0416,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.573018991486575,
|
| 496 |
+
"grad_norm": 0.056938640773296356,
|
| 497 |
+
"learning_rate": 4.4960582171012735e-05,
|
| 498 |
+
"loss": 0.0299,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.5812049770792403,
|
| 503 |
+
"grad_norm": 0.025186289101839066,
|
| 504 |
+
"learning_rate": 4.480897513644634e-05,
|
| 505 |
+
"loss": 0.0046,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5893909626719057,
|
| 510 |
+
"grad_norm": 0.0009007578482851386,
|
| 511 |
+
"learning_rate": 4.465736810187993e-05,
|
| 512 |
+
"loss": 0.0402,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5975769482645711,
|
| 517 |
+
"grad_norm": 0.006263722199946642,
|
| 518 |
+
"learning_rate": 4.450576106731352e-05,
|
| 519 |
+
"loss": 0.0008,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.6057629338572365,
|
| 524 |
+
"grad_norm": 0.002118312055245042,
|
| 525 |
+
"learning_rate": 4.435415403274712e-05,
|
| 526 |
+
"loss": 0.0003,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.6139489194499018,
|
| 531 |
+
"grad_norm": 0.011466097086668015,
|
| 532 |
+
"learning_rate": 4.420254699818071e-05,
|
| 533 |
+
"loss": 0.0025,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.6221349050425671,
|
| 538 |
+
"grad_norm": 0.27655062079429626,
|
| 539 |
+
"learning_rate": 4.4050939963614315e-05,
|
| 540 |
+
"loss": 0.0694,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.6303208906352324,
|
| 545 |
+
"grad_norm": 0.011102208867669106,
|
| 546 |
+
"learning_rate": 4.389933292904791e-05,
|
| 547 |
+
"loss": 0.0771,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.6385068762278978,
|
| 552 |
+
"grad_norm": 0.03391822427511215,
|
| 553 |
+
"learning_rate": 4.3747725894481507e-05,
|
| 554 |
+
"loss": 0.0516,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.6466928618205632,
|
| 559 |
+
"grad_norm": 1.2409979104995728,
|
| 560 |
+
"learning_rate": 4.35961188599151e-05,
|
| 561 |
+
"loss": 0.0052,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.6548788474132285,
|
| 566 |
+
"grad_norm": 0.005223038140684366,
|
| 567 |
+
"learning_rate": 4.34445118253487e-05,
|
| 568 |
+
"loss": 0.0265,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.6630648330058939,
|
| 573 |
+
"grad_norm": 0.015766901895403862,
|
| 574 |
+
"learning_rate": 4.329290479078229e-05,
|
| 575 |
+
"loss": 0.0023,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.6712508185985593,
|
| 580 |
+
"grad_norm": 0.030491165816783905,
|
| 581 |
+
"learning_rate": 4.3141297756215896e-05,
|
| 582 |
+
"loss": 0.0343,
|
| 583 |
+
"step": 2050
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.6794368041912247,
|
| 587 |
+
"grad_norm": 0.00046129990369081497,
|
| 588 |
+
"learning_rate": 4.298969072164949e-05,
|
| 589 |
+
"loss": 0.0121,
|
| 590 |
+
"step": 2075
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.68762278978389,
|
| 594 |
+
"grad_norm": 0.0001839818141888827,
|
| 595 |
+
"learning_rate": 4.283808368708308e-05,
|
| 596 |
+
"loss": 0.0012,
|
| 597 |
+
"step": 2100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.6958087753765554,
|
| 601 |
+
"grad_norm": 0.0003591451095417142,
|
| 602 |
+
"learning_rate": 4.2686476652516676e-05,
|
| 603 |
+
"loss": 0.0374,
|
| 604 |
+
"step": 2125
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.7039947609692206,
|
| 608 |
+
"grad_norm": 0.009124216623604298,
|
| 609 |
+
"learning_rate": 4.253486961795027e-05,
|
| 610 |
+
"loss": 0.0082,
|
| 611 |
+
"step": 2150
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.712180746561886,
|
| 615 |
+
"grad_norm": 0.0014978774124756455,
|
| 616 |
+
"learning_rate": 4.2383262583383874e-05,
|
| 617 |
+
"loss": 0.0306,
|
| 618 |
+
"step": 2175
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.7203667321545514,
|
| 622 |
+
"grad_norm": 1.0534560680389404,
|
| 623 |
+
"learning_rate": 4.223165554881747e-05,
|
| 624 |
+
"loss": 0.0008,
|
| 625 |
+
"step": 2200
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7285527177472167,
|
| 629 |
+
"grad_norm": 0.005840160418301821,
|
| 630 |
+
"learning_rate": 4.2080048514251065e-05,
|
| 631 |
+
"loss": 0.0562,
|
| 632 |
+
"step": 2225
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.7367387033398821,
|
| 636 |
+
"grad_norm": 4.12369966506958,
|
| 637 |
+
"learning_rate": 4.192844147968466e-05,
|
| 638 |
+
"loss": 0.01,
|
| 639 |
+
"step": 2250
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.7449246889325475,
|
| 643 |
+
"grad_norm": 0.0006641410291194916,
|
| 644 |
+
"learning_rate": 4.1776834445118256e-05,
|
| 645 |
+
"loss": 0.0001,
|
| 646 |
+
"step": 2275
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.7531106745252129,
|
| 650 |
+
"grad_norm": 0.0009534868295304477,
|
| 651 |
+
"learning_rate": 4.162522741055185e-05,
|
| 652 |
+
"loss": 0.0854,
|
| 653 |
+
"step": 2300
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.7612966601178782,
|
| 657 |
+
"grad_norm": 0.02312465012073517,
|
| 658 |
+
"learning_rate": 4.147362037598545e-05,
|
| 659 |
+
"loss": 0.0523,
|
| 660 |
+
"step": 2325
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.7694826457105436,
|
| 664 |
+
"grad_norm": 0.009137257933616638,
|
| 665 |
+
"learning_rate": 4.132201334141904e-05,
|
| 666 |
+
"loss": 0.0337,
|
| 667 |
+
"step": 2350
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.777668631303209,
|
| 671 |
+
"grad_norm": 0.01361013948917389,
|
| 672 |
+
"learning_rate": 4.117040630685264e-05,
|
| 673 |
+
"loss": 0.0228,
|
| 674 |
+
"step": 2375
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.7858546168958742,
|
| 678 |
+
"grad_norm": 0.0018830469343811274,
|
| 679 |
+
"learning_rate": 4.1018799272286234e-05,
|
| 680 |
+
"loss": 0.0129,
|
| 681 |
+
"step": 2400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.7940406024885396,
|
| 685 |
+
"grad_norm": 0.6149205565452576,
|
| 686 |
+
"learning_rate": 4.086719223771983e-05,
|
| 687 |
+
"loss": 0.0404,
|
| 688 |
+
"step": 2425
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.802226588081205,
|
| 692 |
+
"grad_norm": 0.005154294427484274,
|
| 693 |
+
"learning_rate": 4.0715585203153425e-05,
|
| 694 |
+
"loss": 0.0011,
|
| 695 |
+
"step": 2450
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.8104125736738703,
|
| 699 |
+
"grad_norm": 0.022492127493023872,
|
| 700 |
+
"learning_rate": 4.056397816858703e-05,
|
| 701 |
+
"loss": 0.042,
|
| 702 |
+
"step": 2475
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.8185985592665357,
|
| 706 |
+
"grad_norm": 0.020115699619054794,
|
| 707 |
+
"learning_rate": 4.041237113402062e-05,
|
| 708 |
+
"loss": 0.0495,
|
| 709 |
+
"step": 2500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.826784544859201,
|
| 713 |
+
"grad_norm": 0.008540102280676365,
|
| 714 |
+
"learning_rate": 4.026076409945422e-05,
|
| 715 |
+
"loss": 0.0074,
|
| 716 |
+
"step": 2525
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.8349705304518664,
|
| 720 |
+
"grad_norm": 0.00022775791876483709,
|
| 721 |
+
"learning_rate": 4.0109157064887814e-05,
|
| 722 |
+
"loss": 0.0001,
|
| 723 |
+
"step": 2550
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.8431565160445318,
|
| 727 |
+
"grad_norm": 0.0014885278651490808,
|
| 728 |
+
"learning_rate": 3.995755003032141e-05,
|
| 729 |
+
"loss": 0.0207,
|
| 730 |
+
"step": 2575
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.8513425016371972,
|
| 734 |
+
"grad_norm": 0.0014814437599852681,
|
| 735 |
+
"learning_rate": 3.9805942995755005e-05,
|
| 736 |
+
"loss": 0.0219,
|
| 737 |
+
"step": 2600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.8595284872298625,
|
| 741 |
+
"grad_norm": 0.0025843679904937744,
|
| 742 |
+
"learning_rate": 3.96543359611886e-05,
|
| 743 |
+
"loss": 0.0137,
|
| 744 |
+
"step": 2625
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.8677144728225278,
|
| 748 |
+
"grad_norm": 0.04107142984867096,
|
| 749 |
+
"learning_rate": 3.9502728926622196e-05,
|
| 750 |
+
"loss": 0.0695,
|
| 751 |
+
"step": 2650
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.8759004584151932,
|
| 755 |
+
"grad_norm": 0.005535195115953684,
|
| 756 |
+
"learning_rate": 3.935112189205579e-05,
|
| 757 |
+
"loss": 0.0157,
|
| 758 |
+
"step": 2675
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.8840864440078585,
|
| 762 |
+
"grad_norm": 5.580929756164551,
|
| 763 |
+
"learning_rate": 3.919951485748939e-05,
|
| 764 |
+
"loss": 0.0806,
|
| 765 |
+
"step": 2700
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.8922724296005239,
|
| 769 |
+
"grad_norm": 0.0086448909714818,
|
| 770 |
+
"learning_rate": 3.904790782292298e-05,
|
| 771 |
+
"loss": 0.046,
|
| 772 |
+
"step": 2725
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.9004584151931893,
|
| 776 |
+
"grad_norm": 0.0019950985442847013,
|
| 777 |
+
"learning_rate": 3.8896300788356585e-05,
|
| 778 |
+
"loss": 0.0004,
|
| 779 |
+
"step": 2750
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.9086444007858546,
|
| 783 |
+
"grad_norm": 0.04631026089191437,
|
| 784 |
+
"learning_rate": 3.874469375379018e-05,
|
| 785 |
+
"loss": 0.1026,
|
| 786 |
+
"step": 2775
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.91683038637852,
|
| 790 |
+
"grad_norm": 0.14316339790821075,
|
| 791 |
+
"learning_rate": 3.8593086719223776e-05,
|
| 792 |
+
"loss": 0.0173,
|
| 793 |
+
"step": 2800
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9250163719711854,
|
| 797 |
+
"grad_norm": 0.002735959365963936,
|
| 798 |
+
"learning_rate": 3.844147968465737e-05,
|
| 799 |
+
"loss": 0.0102,
|
| 800 |
+
"step": 2825
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.9332023575638507,
|
| 804 |
+
"grad_norm": 0.007864673621952534,
|
| 805 |
+
"learning_rate": 3.828987265009097e-05,
|
| 806 |
+
"loss": 0.001,
|
| 807 |
+
"step": 2850
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.9413883431565161,
|
| 811 |
+
"grad_norm": 0.0006219564820639789,
|
| 812 |
+
"learning_rate": 3.813826561552456e-05,
|
| 813 |
+
"loss": 0.0121,
|
| 814 |
+
"step": 2875
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.9495743287491814,
|
| 818 |
+
"grad_norm": 0.20409245789051056,
|
| 819 |
+
"learning_rate": 3.798665858095816e-05,
|
| 820 |
+
"loss": 0.0383,
|
| 821 |
+
"step": 2900
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.9577603143418467,
|
| 825 |
+
"grad_norm": 0.0005641476600430906,
|
| 826 |
+
"learning_rate": 3.7835051546391754e-05,
|
| 827 |
+
"loss": 0.0054,
|
| 828 |
+
"step": 2925
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.9659462999345121,
|
| 832 |
+
"grad_norm": 0.004506452474743128,
|
| 833 |
+
"learning_rate": 3.768344451182535e-05,
|
| 834 |
+
"loss": 0.0167,
|
| 835 |
+
"step": 2950
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.9741322855271775,
|
| 839 |
+
"grad_norm": 0.0007494801538996398,
|
| 840 |
+
"learning_rate": 3.7531837477258945e-05,
|
| 841 |
+
"loss": 0.0016,
|
| 842 |
+
"step": 2975
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.9823182711198428,
|
| 846 |
+
"grad_norm": 0.0016925218515098095,
|
| 847 |
+
"learning_rate": 3.738023044269254e-05,
|
| 848 |
+
"loss": 0.0144,
|
| 849 |
+
"step": 3000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.9905042567125082,
|
| 853 |
+
"grad_norm": 0.007691397797316313,
|
| 854 |
+
"learning_rate": 3.7228623408126137e-05,
|
| 855 |
+
"loss": 0.0264,
|
| 856 |
+
"step": 3025
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.9986902423051736,
|
| 860 |
+
"grad_norm": 0.004836967680603266,
|
| 861 |
+
"learning_rate": 3.707701637355974e-05,
|
| 862 |
+
"loss": 0.0003,
|
| 863 |
+
"step": 3050
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.0,
|
| 867 |
+
"eval_accuracy": 0.9955795677799607,
|
| 868 |
+
"eval_auc": 0.9977481210079818,
|
| 869 |
+
"eval_f1": 0.8187919463087249,
|
| 870 |
+
"eval_loss": 0.014572993852198124,
|
| 871 |
+
"eval_precision": 0.8133333333333334,
|
| 872 |
+
"eval_recall": 0.8243243243243243,
|
| 873 |
+
"eval_runtime": 74.4335,
|
| 874 |
+
"eval_samples_per_second": 82.06,
|
| 875 |
+
"eval_steps_per_second": 5.132,
|
| 876 |
+
"step": 3054
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.006876227897839,
|
| 880 |
+
"grad_norm": 0.0026801538188010454,
|
| 881 |
+
"learning_rate": 3.6925409338993335e-05,
|
| 882 |
+
"loss": 0.0003,
|
| 883 |
+
"step": 3075
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.0150622134905043,
|
| 887 |
+
"grad_norm": 0.0013839630410075188,
|
| 888 |
+
"learning_rate": 3.677380230442693e-05,
|
| 889 |
+
"loss": 0.0415,
|
| 890 |
+
"step": 3100
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.0232481990831697,
|
| 894 |
+
"grad_norm": 0.0052743577398359776,
|
| 895 |
+
"learning_rate": 3.662219526986052e-05,
|
| 896 |
+
"loss": 0.0551,
|
| 897 |
+
"step": 3125
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.031434184675835,
|
| 901 |
+
"grad_norm": 0.001981929875910282,
|
| 902 |
+
"learning_rate": 3.6470588235294114e-05,
|
| 903 |
+
"loss": 0.0011,
|
| 904 |
+
"step": 3150
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.0396201702685004,
|
| 908 |
+
"grad_norm": 0.005122967530041933,
|
| 909 |
+
"learning_rate": 3.631898120072772e-05,
|
| 910 |
+
"loss": 0.0013,
|
| 911 |
+
"step": 3175
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.0478061558611658,
|
| 915 |
+
"grad_norm": 0.0008022825350053608,
|
| 916 |
+
"learning_rate": 3.616737416616131e-05,
|
| 917 |
+
"loss": 0.0001,
|
| 918 |
+
"step": 3200
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.0559921414538311,
|
| 922 |
+
"grad_norm": 0.002690242137759924,
|
| 923 |
+
"learning_rate": 3.601576713159491e-05,
|
| 924 |
+
"loss": 0.0,
|
| 925 |
+
"step": 3225
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.0641781270464965,
|
| 929 |
+
"grad_norm": 0.0003424021415412426,
|
| 930 |
+
"learning_rate": 3.5864160097028504e-05,
|
| 931 |
+
"loss": 0.0066,
|
| 932 |
+
"step": 3250
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.0723641126391619,
|
| 936 |
+
"grad_norm": 0.0005132851074449718,
|
| 937 |
+
"learning_rate": 3.57125530624621e-05,
|
| 938 |
+
"loss": 0.0145,
|
| 939 |
+
"step": 3275
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.080550098231827,
|
| 943 |
+
"grad_norm": 0.0189173873513937,
|
| 944 |
+
"learning_rate": 3.5560946027895695e-05,
|
| 945 |
+
"loss": 0.0465,
|
| 946 |
+
"step": 3300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.0887360838244924,
|
| 950 |
+
"grad_norm": 0.0036357955541461706,
|
| 951 |
+
"learning_rate": 3.54093389933293e-05,
|
| 952 |
+
"loss": 0.0492,
|
| 953 |
+
"step": 3325
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.0969220694171578,
|
| 957 |
+
"grad_norm": 0.003961147274821997,
|
| 958 |
+
"learning_rate": 3.525773195876289e-05,
|
| 959 |
+
"loss": 0.039,
|
| 960 |
+
"step": 3350
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.1051080550098231,
|
| 964 |
+
"grad_norm": 0.003041486954316497,
|
| 965 |
+
"learning_rate": 3.510612492419648e-05,
|
| 966 |
+
"loss": 0.0123,
|
| 967 |
+
"step": 3375
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.1132940406024885,
|
| 971 |
+
"grad_norm": 0.001711357617750764,
|
| 972 |
+
"learning_rate": 3.495451788963008e-05,
|
| 973 |
+
"loss": 0.0003,
|
| 974 |
+
"step": 3400
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.1214800261951539,
|
| 978 |
+
"grad_norm": 0.0013232220662757754,
|
| 979 |
+
"learning_rate": 3.480291085506367e-05,
|
| 980 |
+
"loss": 0.0096,
|
| 981 |
+
"step": 3425
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.1296660117878192,
|
| 985 |
+
"grad_norm": 0.000497565430123359,
|
| 986 |
+
"learning_rate": 3.4651303820497275e-05,
|
| 987 |
+
"loss": 0.0233,
|
| 988 |
+
"step": 3450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 1.1378519973804846,
|
| 992 |
+
"grad_norm": 0.005278216674923897,
|
| 993 |
+
"learning_rate": 3.449969678593087e-05,
|
| 994 |
+
"loss": 0.0399,
|
| 995 |
+
"step": 3475
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.14603798297315,
|
| 999 |
+
"grad_norm": 0.3594399392604828,
|
| 1000 |
+
"learning_rate": 3.4348089751364466e-05,
|
| 1001 |
+
"loss": 0.068,
|
| 1002 |
+
"step": 3500
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.1542239685658153,
|
| 1006 |
+
"grad_norm": 0.013597418554127216,
|
| 1007 |
+
"learning_rate": 3.419648271679806e-05,
|
| 1008 |
+
"loss": 0.0018,
|
| 1009 |
+
"step": 3525
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.1624099541584807,
|
| 1013 |
+
"grad_norm": 0.004102836363017559,
|
| 1014 |
+
"learning_rate": 3.404487568223166e-05,
|
| 1015 |
+
"loss": 0.0292,
|
| 1016 |
+
"step": 3550
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.170595939751146,
|
| 1020 |
+
"grad_norm": 0.0014996347017586231,
|
| 1021 |
+
"learning_rate": 3.389326864766525e-05,
|
| 1022 |
+
"loss": 0.0114,
|
| 1023 |
+
"step": 3575
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.1787819253438114,
|
| 1027 |
+
"grad_norm": 0.018868234008550644,
|
| 1028 |
+
"learning_rate": 3.374166161309885e-05,
|
| 1029 |
+
"loss": 0.0169,
|
| 1030 |
+
"step": 3600
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.1869679109364768,
|
| 1034 |
+
"grad_norm": 0.001259682234376669,
|
| 1035 |
+
"learning_rate": 3.3590054578532444e-05,
|
| 1036 |
+
"loss": 0.0042,
|
| 1037 |
+
"step": 3625
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.1951538965291422,
|
| 1041 |
+
"grad_norm": 0.15819938480854034,
|
| 1042 |
+
"learning_rate": 3.343844754396604e-05,
|
| 1043 |
+
"loss": 0.0007,
|
| 1044 |
+
"step": 3650
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.2033398821218075,
|
| 1048 |
+
"grad_norm": 0.0013165829004719853,
|
| 1049 |
+
"learning_rate": 3.3286840509399635e-05,
|
| 1050 |
+
"loss": 0.0357,
|
| 1051 |
+
"step": 3675
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.211525867714473,
|
| 1055 |
+
"grad_norm": 0.0017084279097616673,
|
| 1056 |
+
"learning_rate": 3.313523347483323e-05,
|
| 1057 |
+
"loss": 0.0034,
|
| 1058 |
+
"step": 3700
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.2197118533071383,
|
| 1062 |
+
"grad_norm": 0.0006795689114369452,
|
| 1063 |
+
"learning_rate": 3.2983626440266826e-05,
|
| 1064 |
+
"loss": 0.0004,
|
| 1065 |
+
"step": 3725
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.2278978388998034,
|
| 1069 |
+
"grad_norm": 0.0009630739223212004,
|
| 1070 |
+
"learning_rate": 3.283201940570043e-05,
|
| 1071 |
+
"loss": 0.0133,
|
| 1072 |
+
"step": 3750
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.2360838244924688,
|
| 1076 |
+
"grad_norm": 0.0106749152764678,
|
| 1077 |
+
"learning_rate": 3.2680412371134024e-05,
|
| 1078 |
+
"loss": 0.0173,
|
| 1079 |
+
"step": 3775
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.2442698100851342,
|
| 1083 |
+
"grad_norm": 0.001351616345345974,
|
| 1084 |
+
"learning_rate": 3.252880533656762e-05,
|
| 1085 |
+
"loss": 0.0023,
|
| 1086 |
+
"step": 3800
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.2524557956777995,
|
| 1090 |
+
"grad_norm": 0.0006208529230207205,
|
| 1091 |
+
"learning_rate": 3.2377198302001215e-05,
|
| 1092 |
+
"loss": 0.0,
|
| 1093 |
+
"step": 3825
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.260641781270465,
|
| 1097 |
+
"grad_norm": 0.0005773603334091604,
|
| 1098 |
+
"learning_rate": 3.222559126743481e-05,
|
| 1099 |
+
"loss": 0.0373,
|
| 1100 |
+
"step": 3850
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.2688277668631303,
|
| 1104 |
+
"grad_norm": 0.0014631077647209167,
|
| 1105 |
+
"learning_rate": 3.2073984232868406e-05,
|
| 1106 |
+
"loss": 0.0025,
|
| 1107 |
+
"step": 3875
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.2770137524557956,
|
| 1111 |
+
"grad_norm": 0.0002054146461887285,
|
| 1112 |
+
"learning_rate": 3.1922377198302e-05,
|
| 1113 |
+
"loss": 0.0013,
|
| 1114 |
+
"step": 3900
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.285199738048461,
|
| 1118 |
+
"grad_norm": 0.0010797285940498114,
|
| 1119 |
+
"learning_rate": 3.17707701637356e-05,
|
| 1120 |
+
"loss": 0.0427,
|
| 1121 |
+
"step": 3925
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.2933857236411264,
|
| 1125 |
+
"grad_norm": 0.022801030427217484,
|
| 1126 |
+
"learning_rate": 3.161916312916919e-05,
|
| 1127 |
+
"loss": 0.0297,
|
| 1128 |
+
"step": 3950
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.3015717092337917,
|
| 1132 |
+
"grad_norm": 0.33411428332328796,
|
| 1133 |
+
"learning_rate": 3.146755609460279e-05,
|
| 1134 |
+
"loss": 0.0412,
|
| 1135 |
+
"step": 3975
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.309757694826457,
|
| 1139 |
+
"grad_norm": 0.014158491976559162,
|
| 1140 |
+
"learning_rate": 3.1315949060036384e-05,
|
| 1141 |
+
"loss": 0.0185,
|
| 1142 |
+
"step": 4000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.3179436804191225,
|
| 1146 |
+
"grad_norm": 0.00048260248149745166,
|
| 1147 |
+
"learning_rate": 3.116434202546999e-05,
|
| 1148 |
+
"loss": 0.0,
|
| 1149 |
+
"step": 4025
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.3261296660117878,
|
| 1153 |
+
"grad_norm": 0.19578149914741516,
|
| 1154 |
+
"learning_rate": 3.101273499090358e-05,
|
| 1155 |
+
"loss": 0.0772,
|
| 1156 |
+
"step": 4050
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.3343156516044532,
|
| 1160 |
+
"grad_norm": 0.0038196328096091747,
|
| 1161 |
+
"learning_rate": 3.086112795633718e-05,
|
| 1162 |
+
"loss": 0.0031,
|
| 1163 |
+
"step": 4075
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 1.3425016371971186,
|
| 1167 |
+
"grad_norm": 0.001698371721431613,
|
| 1168 |
+
"learning_rate": 3.070952092177077e-05,
|
| 1169 |
+
"loss": 0.054,
|
| 1170 |
+
"step": 4100
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.350687622789784,
|
| 1174 |
+
"grad_norm": 0.0078005073592066765,
|
| 1175 |
+
"learning_rate": 3.055791388720437e-05,
|
| 1176 |
+
"loss": 0.0095,
|
| 1177 |
+
"step": 4125
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.3588736083824493,
|
| 1181 |
+
"grad_norm": 0.0008572742226533592,
|
| 1182 |
+
"learning_rate": 3.040630685263796e-05,
|
| 1183 |
+
"loss": 0.0508,
|
| 1184 |
+
"step": 4150
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 1.3670595939751147,
|
| 1188 |
+
"grad_norm": 0.011146394535899162,
|
| 1189 |
+
"learning_rate": 3.0254699818071557e-05,
|
| 1190 |
+
"loss": 0.0004,
|
| 1191 |
+
"step": 4175
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 1.37524557956778,
|
| 1195 |
+
"grad_norm": 0.003408160526305437,
|
| 1196 |
+
"learning_rate": 3.0103092783505156e-05,
|
| 1197 |
+
"loss": 0.0431,
|
| 1198 |
+
"step": 4200
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 1.3834315651604454,
|
| 1202 |
+
"grad_norm": 0.01908206194639206,
|
| 1203 |
+
"learning_rate": 2.995148574893875e-05,
|
| 1204 |
+
"loss": 0.0387,
|
| 1205 |
+
"step": 4225
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 1.3916175507531108,
|
| 1209 |
+
"grad_norm": 0.001946812029927969,
|
| 1210 |
+
"learning_rate": 2.9799878714372347e-05,
|
| 1211 |
+
"loss": 0.0005,
|
| 1212 |
+
"step": 4250
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 1.3998035363457761,
|
| 1216 |
+
"grad_norm": 0.008685130625963211,
|
| 1217 |
+
"learning_rate": 2.9648271679805946e-05,
|
| 1218 |
+
"loss": 0.0786,
|
| 1219 |
+
"step": 4275
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.4079895219384415,
|
| 1223 |
+
"grad_norm": 0.07280147820711136,
|
| 1224 |
+
"learning_rate": 2.949666464523954e-05,
|
| 1225 |
+
"loss": 0.0217,
|
| 1226 |
+
"step": 4300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.4161755075311069,
|
| 1230 |
+
"grad_norm": 0.01758684776723385,
|
| 1231 |
+
"learning_rate": 2.9345057610673137e-05,
|
| 1232 |
+
"loss": 0.0042,
|
| 1233 |
+
"step": 4325
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.424361493123772,
|
| 1237 |
+
"grad_norm": 0.003927039448171854,
|
| 1238 |
+
"learning_rate": 2.9193450576106736e-05,
|
| 1239 |
+
"loss": 0.0033,
|
| 1240 |
+
"step": 4350
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 1.4325474787164374,
|
| 1244 |
+
"grad_norm": 0.003762061009183526,
|
| 1245 |
+
"learning_rate": 2.904184354154033e-05,
|
| 1246 |
+
"loss": 0.0313,
|
| 1247 |
+
"step": 4375
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 1.4407334643091028,
|
| 1251 |
+
"grad_norm": 0.00642773462459445,
|
| 1252 |
+
"learning_rate": 2.8890236506973924e-05,
|
| 1253 |
+
"loss": 0.0003,
|
| 1254 |
+
"step": 4400
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 1.4489194499017681,
|
| 1258 |
+
"grad_norm": 0.005841416772454977,
|
| 1259 |
+
"learning_rate": 2.873862947240752e-05,
|
| 1260 |
+
"loss": 0.0001,
|
| 1261 |
+
"step": 4425
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.4571054354944335,
|
| 1265 |
+
"grad_norm": 0.0015138218877837062,
|
| 1266 |
+
"learning_rate": 2.8587022437841115e-05,
|
| 1267 |
+
"loss": 0.0001,
|
| 1268 |
+
"step": 4450
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.4652914210870989,
|
| 1272 |
+
"grad_norm": 0.005790230352431536,
|
| 1273 |
+
"learning_rate": 2.8435415403274714e-05,
|
| 1274 |
+
"loss": 0.0166,
|
| 1275 |
+
"step": 4475
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.4734774066797642,
|
| 1279 |
+
"grad_norm": 0.002549233613535762,
|
| 1280 |
+
"learning_rate": 2.828380836870831e-05,
|
| 1281 |
+
"loss": 0.0006,
|
| 1282 |
+
"step": 4500
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.4816633922724296,
|
| 1286 |
+
"grad_norm": 0.0019311434589326382,
|
| 1287 |
+
"learning_rate": 2.8132201334141905e-05,
|
| 1288 |
+
"loss": 0.0094,
|
| 1289 |
+
"step": 4525
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.489849377865095,
|
| 1293 |
+
"grad_norm": 0.0018175276927649975,
|
| 1294 |
+
"learning_rate": 2.7980594299575504e-05,
|
| 1295 |
+
"loss": 0.0179,
|
| 1296 |
+
"step": 4550
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 1.4980353634577603,
|
| 1300 |
+
"grad_norm": 0.00037457983125932515,
|
| 1301 |
+
"learning_rate": 2.78289872650091e-05,
|
| 1302 |
+
"loss": 0.0,
|
| 1303 |
+
"step": 4575
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.5062213490504257,
|
| 1307 |
+
"grad_norm": 0.0009425774333067238,
|
| 1308 |
+
"learning_rate": 2.7677380230442695e-05,
|
| 1309 |
+
"loss": 0.0,
|
| 1310 |
+
"step": 4600
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.514407334643091,
|
| 1314 |
+
"grad_norm": 0.00077505485387519,
|
| 1315 |
+
"learning_rate": 2.752577319587629e-05,
|
| 1316 |
+
"loss": 0.0002,
|
| 1317 |
+
"step": 4625
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 1.5225933202357562,
|
| 1321 |
+
"grad_norm": 0.00036527996417135,
|
| 1322 |
+
"learning_rate": 2.737416616130989e-05,
|
| 1323 |
+
"loss": 0.0002,
|
| 1324 |
+
"step": 4650
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 1.5307793058284216,
|
| 1328 |
+
"grad_norm": 0.0006952973199076951,
|
| 1329 |
+
"learning_rate": 2.7222559126743478e-05,
|
| 1330 |
+
"loss": 0.0,
|
| 1331 |
+
"step": 4675
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.538965291421087,
|
| 1335 |
+
"grad_norm": 0.0010176642099395394,
|
| 1336 |
+
"learning_rate": 2.7070952092177077e-05,
|
| 1337 |
+
"loss": 0.0465,
|
| 1338 |
+
"step": 4700
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.5471512770137523,
|
| 1342 |
+
"grad_norm": 0.00043089999235235155,
|
| 1343 |
+
"learning_rate": 2.6919345057610673e-05,
|
| 1344 |
+
"loss": 0.013,
|
| 1345 |
+
"step": 4725
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.5553372626064177,
|
| 1349 |
+
"grad_norm": 0.0003544854116626084,
|
| 1350 |
+
"learning_rate": 2.676773802304427e-05,
|
| 1351 |
+
"loss": 0.0001,
|
| 1352 |
+
"step": 4750
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.563523248199083,
|
| 1356 |
+
"grad_norm": 0.0013717171968892217,
|
| 1357 |
+
"learning_rate": 2.6616130988477867e-05,
|
| 1358 |
+
"loss": 0.0002,
|
| 1359 |
+
"step": 4775
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.5717092337917484,
|
| 1363 |
+
"grad_norm": 0.0011255746940150857,
|
| 1364 |
+
"learning_rate": 2.6464523953911463e-05,
|
| 1365 |
+
"loss": 0.0002,
|
| 1366 |
+
"step": 4800
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 1.5798952193844138,
|
| 1370 |
+
"grad_norm": 0.0009519586456008255,
|
| 1371 |
+
"learning_rate": 2.631291691934506e-05,
|
| 1372 |
+
"loss": 0.0042,
|
| 1373 |
+
"step": 4825
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.5880812049770792,
|
| 1377 |
+
"grad_norm": 0.0012405363377183676,
|
| 1378 |
+
"learning_rate": 2.6161309884778657e-05,
|
| 1379 |
+
"loss": 0.0,
|
| 1380 |
+
"step": 4850
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 1.5962671905697445,
|
| 1384 |
+
"grad_norm": 0.0004118427459616214,
|
| 1385 |
+
"learning_rate": 2.6009702850212253e-05,
|
| 1386 |
+
"loss": 0.0515,
|
| 1387 |
+
"step": 4875
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.60445317616241,
|
| 1391 |
+
"grad_norm": 0.0020054271444678307,
|
| 1392 |
+
"learning_rate": 2.585809581564585e-05,
|
| 1393 |
+
"loss": 0.0162,
|
| 1394 |
+
"step": 4900
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.6126391617550753,
|
| 1398 |
+
"grad_norm": 0.0009532085387036204,
|
| 1399 |
+
"learning_rate": 2.570648878107944e-05,
|
| 1400 |
+
"loss": 0.0003,
|
| 1401 |
+
"step": 4925
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 1.6208251473477406,
|
| 1405 |
+
"grad_norm": 0.009459923952817917,
|
| 1406 |
+
"learning_rate": 2.5554881746513036e-05,
|
| 1407 |
+
"loss": 0.0538,
|
| 1408 |
+
"step": 4950
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.629011132940406,
|
| 1412 |
+
"grad_norm": 0.0009569909307174385,
|
| 1413 |
+
"learning_rate": 2.5403274711946635e-05,
|
| 1414 |
+
"loss": 0.0346,
|
| 1415 |
+
"step": 4975
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 1.6371971185330714,
|
| 1419 |
+
"grad_norm": 0.0007560375961475074,
|
| 1420 |
+
"learning_rate": 2.525166767738023e-05,
|
| 1421 |
+
"loss": 0.0362,
|
| 1422 |
+
"step": 5000
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 1.6453831041257367,
|
| 1426 |
+
"grad_norm": 0.0004632064083125442,
|
| 1427 |
+
"learning_rate": 2.5100060642813826e-05,
|
| 1428 |
+
"loss": 0.0005,
|
| 1429 |
+
"step": 5025
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.653569089718402,
|
| 1433 |
+
"grad_norm": 0.0006956621073186398,
|
| 1434 |
+
"learning_rate": 2.4948453608247425e-05,
|
| 1435 |
+
"loss": 0.0028,
|
| 1436 |
+
"step": 5050
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.6617550753110675,
|
| 1440 |
+
"grad_norm": 0.0004222158167976886,
|
| 1441 |
+
"learning_rate": 2.479684657368102e-05,
|
| 1442 |
+
"loss": 0.0002,
|
| 1443 |
+
"step": 5075
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 1.6699410609037328,
|
| 1447 |
+
"grad_norm": 0.0006862828740850091,
|
| 1448 |
+
"learning_rate": 2.4645239539114617e-05,
|
| 1449 |
+
"loss": 0.0395,
|
| 1450 |
+
"step": 5100
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.6781270464963982,
|
| 1454 |
+
"grad_norm": 0.001836837618611753,
|
| 1455 |
+
"learning_rate": 2.4493632504548212e-05,
|
| 1456 |
+
"loss": 0.0268,
|
| 1457 |
+
"step": 5125
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 1.6863130320890636,
|
| 1461 |
+
"grad_norm": 0.0008472290937788785,
|
| 1462 |
+
"learning_rate": 2.4342025469981808e-05,
|
| 1463 |
+
"loss": 0.0028,
|
| 1464 |
+
"step": 5150
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 1.694499017681729,
|
| 1468 |
+
"grad_norm": 0.0012132123811170459,
|
| 1469 |
+
"learning_rate": 2.4190418435415403e-05,
|
| 1470 |
+
"loss": 0.0001,
|
| 1471 |
+
"step": 5175
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.7026850032743943,
|
| 1475 |
+
"grad_norm": 0.0005179749568924308,
|
| 1476 |
+
"learning_rate": 2.4038811400849002e-05,
|
| 1477 |
+
"loss": 0.0397,
|
| 1478 |
+
"step": 5200
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.7108709888670597,
|
| 1482 |
+
"grad_norm": 0.0008750595734454691,
|
| 1483 |
+
"learning_rate": 2.3887204366282598e-05,
|
| 1484 |
+
"loss": 0.0002,
|
| 1485 |
+
"step": 5225
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 1.719056974459725,
|
| 1489 |
+
"grad_norm": 0.02541407383978367,
|
| 1490 |
+
"learning_rate": 2.373559733171619e-05,
|
| 1491 |
+
"loss": 0.0001,
|
| 1492 |
+
"step": 5250
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.7272429600523904,
|
| 1496 |
+
"grad_norm": 0.02840971015393734,
|
| 1497 |
+
"learning_rate": 2.358399029714979e-05,
|
| 1498 |
+
"loss": 0.0002,
|
| 1499 |
+
"step": 5275
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 1.7354289456450558,
|
| 1503 |
+
"grad_norm": 0.0008489146712236106,
|
| 1504 |
+
"learning_rate": 2.3432383262583385e-05,
|
| 1505 |
+
"loss": 0.0142,
|
| 1506 |
+
"step": 5300
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 1.7436149312377212,
|
| 1510 |
+
"grad_norm": 0.002642701379954815,
|
| 1511 |
+
"learning_rate": 2.328077622801698e-05,
|
| 1512 |
+
"loss": 0.0562,
|
| 1513 |
+
"step": 5325
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.7518009168303865,
|
| 1517 |
+
"grad_norm": 0.004949661903083324,
|
| 1518 |
+
"learning_rate": 2.312916919345058e-05,
|
| 1519 |
+
"loss": 0.0147,
|
| 1520 |
+
"step": 5350
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.759986902423052,
|
| 1524 |
+
"grad_norm": 0.11022938042879105,
|
| 1525 |
+
"learning_rate": 2.297756215888417e-05,
|
| 1526 |
+
"loss": 0.028,
|
| 1527 |
+
"step": 5375
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.768172888015717,
|
| 1531 |
+
"grad_norm": 0.006674617063254118,
|
| 1532 |
+
"learning_rate": 2.282595512431777e-05,
|
| 1533 |
+
"loss": 0.0014,
|
| 1534 |
+
"step": 5400
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.7763588736083824,
|
| 1538 |
+
"grad_norm": 0.0017073052003979683,
|
| 1539 |
+
"learning_rate": 2.2674348089751366e-05,
|
| 1540 |
+
"loss": 0.0028,
|
| 1541 |
+
"step": 5425
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.7845448592010478,
|
| 1545 |
+
"grad_norm": 3.2796061038970947,
|
| 1546 |
+
"learning_rate": 2.252274105518496e-05,
|
| 1547 |
+
"loss": 0.0025,
|
| 1548 |
+
"step": 5450
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 1.7927308447937131,
|
| 1552 |
+
"grad_norm": 0.0006569511606357992,
|
| 1553 |
+
"learning_rate": 2.237113402061856e-05,
|
| 1554 |
+
"loss": 0.0009,
|
| 1555 |
+
"step": 5475
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.8009168303863785,
|
| 1559 |
+
"grad_norm": 0.00020327128004282713,
|
| 1560 |
+
"learning_rate": 2.2219526986052153e-05,
|
| 1561 |
+
"loss": 0.0,
|
| 1562 |
+
"step": 5500
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 1.8091028159790439,
|
| 1566 |
+
"grad_norm": 0.041166484355926514,
|
| 1567 |
+
"learning_rate": 2.2067919951485748e-05,
|
| 1568 |
+
"loss": 0.0937,
|
| 1569 |
+
"step": 5525
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.8172888015717092,
|
| 1573 |
+
"grad_norm": 0.05974828451871872,
|
| 1574 |
+
"learning_rate": 2.1916312916919347e-05,
|
| 1575 |
+
"loss": 0.0147,
|
| 1576 |
+
"step": 5550
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.8254747871643746,
|
| 1580 |
+
"grad_norm": 0.004328828305006027,
|
| 1581 |
+
"learning_rate": 2.1764705882352943e-05,
|
| 1582 |
+
"loss": 0.0041,
|
| 1583 |
+
"step": 5575
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.83366077275704,
|
| 1587 |
+
"grad_norm": 0.004382645711302757,
|
| 1588 |
+
"learning_rate": 2.1613098847786538e-05,
|
| 1589 |
+
"loss": 0.0006,
|
| 1590 |
+
"step": 5600
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 1.8418467583497053,
|
| 1594 |
+
"grad_norm": 0.005558451171964407,
|
| 1595 |
+
"learning_rate": 2.1461491813220134e-05,
|
| 1596 |
+
"loss": 0.0002,
|
| 1597 |
+
"step": 5625
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.8500327439423707,
|
| 1601 |
+
"grad_norm": 0.02043689414858818,
|
| 1602 |
+
"learning_rate": 2.130988477865373e-05,
|
| 1603 |
+
"loss": 0.044,
|
| 1604 |
+
"step": 5650
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 1.8582187295350359,
|
| 1608 |
+
"grad_norm": 0.007403884083032608,
|
| 1609 |
+
"learning_rate": 2.1158277744087325e-05,
|
| 1610 |
+
"loss": 0.0295,
|
| 1611 |
+
"step": 5675
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.8664047151277012,
|
| 1615 |
+
"grad_norm": 0.002108983462676406,
|
| 1616 |
+
"learning_rate": 2.1006670709520924e-05,
|
| 1617 |
+
"loss": 0.0033,
|
| 1618 |
+
"step": 5700
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.8745907007203666,
|
| 1622 |
+
"grad_norm": 0.0540454126894474,
|
| 1623 |
+
"learning_rate": 2.085506367495452e-05,
|
| 1624 |
+
"loss": 0.0247,
|
| 1625 |
+
"step": 5725
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.882776686313032,
|
| 1629 |
+
"grad_norm": 0.02874263934791088,
|
| 1630 |
+
"learning_rate": 2.0703456640388115e-05,
|
| 1631 |
+
"loss": 0.0117,
|
| 1632 |
+
"step": 5750
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 1.8909626719056973,
|
| 1636 |
+
"grad_norm": 0.0010341499000787735,
|
| 1637 |
+
"learning_rate": 2.055184960582171e-05,
|
| 1638 |
+
"loss": 0.0029,
|
| 1639 |
+
"step": 5775
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.8991486574983627,
|
| 1643 |
+
"grad_norm": 0.0022526606917381287,
|
| 1644 |
+
"learning_rate": 2.0400242571255306e-05,
|
| 1645 |
+
"loss": 0.0404,
|
| 1646 |
+
"step": 5800
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 1.907334643091028,
|
| 1650 |
+
"grad_norm": 0.0009332990157417953,
|
| 1651 |
+
"learning_rate": 2.0248635536688905e-05,
|
| 1652 |
+
"loss": 0.0001,
|
| 1653 |
+
"step": 5825
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.9155206286836934,
|
| 1657 |
+
"grad_norm": 0.01605447567999363,
|
| 1658 |
+
"learning_rate": 2.00970285021225e-05,
|
| 1659 |
+
"loss": 0.0744,
|
| 1660 |
+
"step": 5850
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.9237066142763588,
|
| 1664 |
+
"grad_norm": 0.057828161865472794,
|
| 1665 |
+
"learning_rate": 1.9945421467556096e-05,
|
| 1666 |
+
"loss": 0.038,
|
| 1667 |
+
"step": 5875
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.9318925998690242,
|
| 1671 |
+
"grad_norm": 0.01713876612484455,
|
| 1672 |
+
"learning_rate": 1.9793814432989692e-05,
|
| 1673 |
+
"loss": 0.0071,
|
| 1674 |
+
"step": 5900
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 1.9400785854616895,
|
| 1678 |
+
"grad_norm": 0.0016962456284090877,
|
| 1679 |
+
"learning_rate": 1.9642207398423287e-05,
|
| 1680 |
+
"loss": 0.0075,
|
| 1681 |
+
"step": 5925
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.948264571054355,
|
| 1685 |
+
"grad_norm": 0.001768580754287541,
|
| 1686 |
+
"learning_rate": 1.9490600363856883e-05,
|
| 1687 |
+
"loss": 0.0008,
|
| 1688 |
+
"step": 5950
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 1.9564505566470203,
|
| 1692 |
+
"grad_norm": 0.005728750489652157,
|
| 1693 |
+
"learning_rate": 1.9338993329290482e-05,
|
| 1694 |
+
"loss": 0.0627,
|
| 1695 |
+
"step": 5975
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 1.9646365422396856,
|
| 1699 |
+
"grad_norm": 7.606292724609375,
|
| 1700 |
+
"learning_rate": 1.9187386294724078e-05,
|
| 1701 |
+
"loss": 0.0418,
|
| 1702 |
+
"step": 6000
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 1.972822527832351,
|
| 1706 |
+
"grad_norm": 0.0018708339193835855,
|
| 1707 |
+
"learning_rate": 1.903577926015767e-05,
|
| 1708 |
+
"loss": 0.0006,
|
| 1709 |
+
"step": 6025
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.9810085134250164,
|
| 1713 |
+
"grad_norm": 7.210286617279053,
|
| 1714 |
+
"learning_rate": 1.888417222559127e-05,
|
| 1715 |
+
"loss": 0.0324,
|
| 1716 |
+
"step": 6050
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 1.9891944990176817,
|
| 1720 |
+
"grad_norm": 0.021973388269543648,
|
| 1721 |
+
"learning_rate": 1.8732565191024864e-05,
|
| 1722 |
+
"loss": 0.0005,
|
| 1723 |
+
"step": 6075
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 1.9973804846103471,
|
| 1727 |
+
"grad_norm": 0.2803119421005249,
|
| 1728 |
+
"learning_rate": 1.858095815645846e-05,
|
| 1729 |
+
"loss": 0.0454,
|
| 1730 |
+
"step": 6100
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 2.0,
|
| 1734 |
+
"eval_accuracy": 0.9975442043222004,
|
| 1735 |
+
"eval_auc": 0.9995912800437162,
|
| 1736 |
+
"eval_f1": 0.9044585987261147,
|
| 1737 |
+
"eval_loss": 0.008845364674925804,
|
| 1738 |
+
"eval_precision": 0.8554216867469879,
|
| 1739 |
+
"eval_recall": 0.9594594594594594,
|
| 1740 |
+
"eval_runtime": 74.792,
|
| 1741 |
+
"eval_samples_per_second": 81.666,
|
| 1742 |
+
"eval_steps_per_second": 5.107,
|
| 1743 |
+
"step": 6108
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 2.0055664702030125,
|
| 1747 |
+
"grad_norm": 0.022678203880786896,
|
| 1748 |
+
"learning_rate": 1.8435415403274715e-05,
|
| 1749 |
+
"loss": 0.0121,
|
| 1750 |
+
"step": 6125
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 2.013752455795678,
|
| 1754 |
+
"grad_norm": 0.0033217219170182943,
|
| 1755 |
+
"learning_rate": 1.8283808368708307e-05,
|
| 1756 |
+
"loss": 0.0446,
|
| 1757 |
+
"step": 6150
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.021938441388343,
|
| 1761 |
+
"grad_norm": 0.006960027851164341,
|
| 1762 |
+
"learning_rate": 1.8132201334141906e-05,
|
| 1763 |
+
"loss": 0.0067,
|
| 1764 |
+
"step": 6175
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 2.0301244269810086,
|
| 1768 |
+
"grad_norm": 1.543547511100769,
|
| 1769 |
+
"learning_rate": 1.79805942995755e-05,
|
| 1770 |
+
"loss": 0.0201,
|
| 1771 |
+
"step": 6200
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 2.038310412573674,
|
| 1775 |
+
"grad_norm": 0.01694468781352043,
|
| 1776 |
+
"learning_rate": 1.7828987265009097e-05,
|
| 1777 |
+
"loss": 0.0258,
|
| 1778 |
+
"step": 6225
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 2.0464963981663393,
|
| 1782 |
+
"grad_norm": 0.004578434396535158,
|
| 1783 |
+
"learning_rate": 1.7677380230442696e-05,
|
| 1784 |
+
"loss": 0.0006,
|
| 1785 |
+
"step": 6250
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 2.0546823837590047,
|
| 1789 |
+
"grad_norm": 1.5190191268920898,
|
| 1790 |
+
"learning_rate": 1.7525773195876288e-05,
|
| 1791 |
+
"loss": 0.0126,
|
| 1792 |
+
"step": 6275
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 2.06286836935167,
|
| 1796 |
+
"grad_norm": 0.0018938812427222729,
|
| 1797 |
+
"learning_rate": 1.7374166161309884e-05,
|
| 1798 |
+
"loss": 0.0003,
|
| 1799 |
+
"step": 6300
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.0710543549443354,
|
| 1803 |
+
"grad_norm": 0.0011117063695564866,
|
| 1804 |
+
"learning_rate": 1.7222559126743483e-05,
|
| 1805 |
+
"loss": 0.0002,
|
| 1806 |
+
"step": 6325
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 2.079240340537001,
|
| 1810 |
+
"grad_norm": 0.000687290565110743,
|
| 1811 |
+
"learning_rate": 1.7070952092177078e-05,
|
| 1812 |
+
"loss": 0.0154,
|
| 1813 |
+
"step": 6350
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 2.087426326129666,
|
| 1817 |
+
"grad_norm": 0.001672151731327176,
|
| 1818 |
+
"learning_rate": 1.6919345057610674e-05,
|
| 1819 |
+
"loss": 0.0006,
|
| 1820 |
+
"step": 6375
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 2.0956123117223315,
|
| 1824 |
+
"grad_norm": 5.109748363494873,
|
| 1825 |
+
"learning_rate": 1.676773802304427e-05,
|
| 1826 |
+
"loss": 0.0352,
|
| 1827 |
+
"step": 6400
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 2.103798297314997,
|
| 1831 |
+
"grad_norm": 0.020103197544813156,
|
| 1832 |
+
"learning_rate": 1.6616130988477865e-05,
|
| 1833 |
+
"loss": 0.0007,
|
| 1834 |
+
"step": 6425
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 2.1119842829076623,
|
| 1838 |
+
"grad_norm": 0.005654872860759497,
|
| 1839 |
+
"learning_rate": 1.646452395391146e-05,
|
| 1840 |
+
"loss": 0.0004,
|
| 1841 |
+
"step": 6450
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.1201702685003276,
|
| 1845 |
+
"grad_norm": 0.009023274295032024,
|
| 1846 |
+
"learning_rate": 1.631291691934506e-05,
|
| 1847 |
+
"loss": 0.0018,
|
| 1848 |
+
"step": 6475
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 2.128356254092993,
|
| 1852 |
+
"grad_norm": 0.0016646806616336107,
|
| 1853 |
+
"learning_rate": 1.6161309884778655e-05,
|
| 1854 |
+
"loss": 0.0001,
|
| 1855 |
+
"step": 6500
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 2.1365422396856584,
|
| 1859 |
+
"grad_norm": 0.0016474899603053927,
|
| 1860 |
+
"learning_rate": 1.600970285021225e-05,
|
| 1861 |
+
"loss": 0.0005,
|
| 1862 |
+
"step": 6525
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 2.1447282252783237,
|
| 1866 |
+
"grad_norm": 0.014943123795092106,
|
| 1867 |
+
"learning_rate": 1.5858095815645846e-05,
|
| 1868 |
+
"loss": 0.0147,
|
| 1869 |
+
"step": 6550
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 2.1529142108709887,
|
| 1873 |
+
"grad_norm": 0.0021474878303706646,
|
| 1874 |
+
"learning_rate": 1.570648878107944e-05,
|
| 1875 |
+
"loss": 0.0002,
|
| 1876 |
+
"step": 6575
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 2.161100196463654,
|
| 1880 |
+
"grad_norm": 0.0017727293306961656,
|
| 1881 |
+
"learning_rate": 1.555488174651304e-05,
|
| 1882 |
+
"loss": 0.0,
|
| 1883 |
+
"step": 6600
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.1692861820563194,
|
| 1887 |
+
"grad_norm": 0.0010321360314264894,
|
| 1888 |
+
"learning_rate": 1.5403274711946636e-05,
|
| 1889 |
+
"loss": 0.0001,
|
| 1890 |
+
"step": 6625
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 2.1774721676489848,
|
| 1894 |
+
"grad_norm": 0.0004758843861054629,
|
| 1895 |
+
"learning_rate": 1.5251667677380232e-05,
|
| 1896 |
+
"loss": 0.0114,
|
| 1897 |
+
"step": 6650
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 2.18565815324165,
|
| 1901 |
+
"grad_norm": 0.0012892503291368484,
|
| 1902 |
+
"learning_rate": 1.5100060642813826e-05,
|
| 1903 |
+
"loss": 0.0262,
|
| 1904 |
+
"step": 6675
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 2.1938441388343155,
|
| 1908 |
+
"grad_norm": 0.01857004500925541,
|
| 1909 |
+
"learning_rate": 1.4948453608247423e-05,
|
| 1910 |
+
"loss": 0.0003,
|
| 1911 |
+
"step": 6700
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 2.202030124426981,
|
| 1915 |
+
"grad_norm": 0.0008667947258800268,
|
| 1916 |
+
"learning_rate": 1.479684657368102e-05,
|
| 1917 |
+
"loss": 0.0099,
|
| 1918 |
+
"step": 6725
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 2.2102161100196462,
|
| 1922 |
+
"grad_norm": 0.001100408611819148,
|
| 1923 |
+
"learning_rate": 1.4645239539114616e-05,
|
| 1924 |
+
"loss": 0.0341,
|
| 1925 |
+
"step": 6750
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 2.2184020956123116,
|
| 1929 |
+
"grad_norm": 0.001291294815018773,
|
| 1930 |
+
"learning_rate": 1.4493632504548213e-05,
|
| 1931 |
+
"loss": 0.0001,
|
| 1932 |
+
"step": 6775
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 2.226588081204977,
|
| 1936 |
+
"grad_norm": 0.011078906245529652,
|
| 1937 |
+
"learning_rate": 1.4342025469981807e-05,
|
| 1938 |
+
"loss": 0.0479,
|
| 1939 |
+
"step": 6800
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 2.2347740667976423,
|
| 1943 |
+
"grad_norm": 0.0018653717124834657,
|
| 1944 |
+
"learning_rate": 1.4190418435415404e-05,
|
| 1945 |
+
"loss": 0.0026,
|
| 1946 |
+
"step": 6825
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 2.2429600523903077,
|
| 1950 |
+
"grad_norm": 0.0010668339673429728,
|
| 1951 |
+
"learning_rate": 1.4038811400849e-05,
|
| 1952 |
+
"loss": 0.0005,
|
| 1953 |
+
"step": 6850
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 2.251146037982973,
|
| 1957 |
+
"grad_norm": 0.002098933095112443,
|
| 1958 |
+
"learning_rate": 1.3887204366282597e-05,
|
| 1959 |
+
"loss": 0.0005,
|
| 1960 |
+
"step": 6875
|
| 1961 |
+
},
|
| 1962 |
+
{
|
| 1963 |
+
"epoch": 2.2593320235756384,
|
| 1964 |
+
"grad_norm": 0.0008214248227886856,
|
| 1965 |
+
"learning_rate": 1.3735597331716193e-05,
|
| 1966 |
+
"loss": 0.0185,
|
| 1967 |
+
"step": 6900
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 2.267518009168304,
|
| 1971 |
+
"grad_norm": 0.001296183792874217,
|
| 1972 |
+
"learning_rate": 1.3583990297149786e-05,
|
| 1973 |
+
"loss": 0.0354,
|
| 1974 |
+
"step": 6925
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 2.275703994760969,
|
| 1978 |
+
"grad_norm": 0.0009354737703688443,
|
| 1979 |
+
"learning_rate": 1.3432383262583384e-05,
|
| 1980 |
+
"loss": 0.0127,
|
| 1981 |
+
"step": 6950
|
| 1982 |
+
},
|
| 1983 |
+
{
|
| 1984 |
+
"epoch": 2.2838899803536346,
|
| 1985 |
+
"grad_norm": 0.0020546901505440474,
|
| 1986 |
+
"learning_rate": 1.3280776228016981e-05,
|
| 1987 |
+
"loss": 0.0004,
|
| 1988 |
+
"step": 6975
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 2.2920759659463,
|
| 1992 |
+
"grad_norm": 0.02266145683825016,
|
| 1993 |
+
"learning_rate": 1.3129169193450577e-05,
|
| 1994 |
+
"loss": 0.0364,
|
| 1995 |
+
"step": 7000
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 2.3002619515389653,
|
| 1999 |
+
"grad_norm": 0.010976303368806839,
|
| 2000 |
+
"learning_rate": 1.2977562158884174e-05,
|
| 2001 |
+
"loss": 0.0018,
|
| 2002 |
+
"step": 7025
|
| 2003 |
+
},
|
| 2004 |
+
{
|
| 2005 |
+
"epoch": 2.3084479371316307,
|
| 2006 |
+
"grad_norm": 0.006094989832490683,
|
| 2007 |
+
"learning_rate": 1.2825955124317768e-05,
|
| 2008 |
+
"loss": 0.0091,
|
| 2009 |
+
"step": 7050
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 2.316633922724296,
|
| 2013 |
+
"grad_norm": 0.07841998338699341,
|
| 2014 |
+
"learning_rate": 1.2674348089751365e-05,
|
| 2015 |
+
"loss": 0.0094,
|
| 2016 |
+
"step": 7075
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 2.3248199083169614,
|
| 2020 |
+
"grad_norm": 2.630479335784912,
|
| 2021 |
+
"learning_rate": 1.252274105518496e-05,
|
| 2022 |
+
"loss": 0.0029,
|
| 2023 |
+
"step": 7100
|
| 2024 |
+
},
|
| 2025 |
+
{
|
| 2026 |
+
"epoch": 2.3330058939096268,
|
| 2027 |
+
"grad_norm": 0.0039812372997403145,
|
| 2028 |
+
"learning_rate": 1.2371134020618558e-05,
|
| 2029 |
+
"loss": 0.0086,
|
| 2030 |
+
"step": 7125
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 2.341191879502292,
|
| 2034 |
+
"grad_norm": 0.0007643400458618999,
|
| 2035 |
+
"learning_rate": 1.2219526986052153e-05,
|
| 2036 |
+
"loss": 0.0001,
|
| 2037 |
+
"step": 7150
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 2.3493778650949575,
|
| 2041 |
+
"grad_norm": 0.0021422533318400383,
|
| 2042 |
+
"learning_rate": 1.2067919951485749e-05,
|
| 2043 |
+
"loss": 0.0029,
|
| 2044 |
+
"step": 7175
|
| 2045 |
+
},
|
| 2046 |
+
{
|
| 2047 |
+
"epoch": 2.357563850687623,
|
| 2048 |
+
"grad_norm": 0.0017942421836778522,
|
| 2049 |
+
"learning_rate": 1.1916312916919346e-05,
|
| 2050 |
+
"loss": 0.0026,
|
| 2051 |
+
"step": 7200
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 2.3657498362802882,
|
| 2055 |
+
"grad_norm": 0.0007688822224736214,
|
| 2056 |
+
"learning_rate": 1.1764705882352942e-05,
|
| 2057 |
+
"loss": 0.0018,
|
| 2058 |
+
"step": 7225
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 2.3739358218729536,
|
| 2062 |
+
"grad_norm": 0.0004730868386104703,
|
| 2063 |
+
"learning_rate": 1.1613098847786537e-05,
|
| 2064 |
+
"loss": 0.0028,
|
| 2065 |
+
"step": 7250
|
| 2066 |
+
},
|
| 2067 |
+
{
|
| 2068 |
+
"epoch": 2.382121807465619,
|
| 2069 |
+
"grad_norm": 0.0004939533537253737,
|
| 2070 |
+
"learning_rate": 1.1461491813220133e-05,
|
| 2071 |
+
"loss": 0.0003,
|
| 2072 |
+
"step": 7275
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 2.3903077930582843,
|
| 2076 |
+
"grad_norm": 0.0010573838371783495,
|
| 2077 |
+
"learning_rate": 1.130988477865373e-05,
|
| 2078 |
+
"loss": 0.0158,
|
| 2079 |
+
"step": 7300
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 2.3984937786509497,
|
| 2083 |
+
"grad_norm": 0.0003521484904922545,
|
| 2084 |
+
"learning_rate": 1.1158277744087328e-05,
|
| 2085 |
+
"loss": 0.0008,
|
| 2086 |
+
"step": 7325
|
| 2087 |
+
},
|
| 2088 |
+
{
|
| 2089 |
+
"epoch": 2.406679764243615,
|
| 2090 |
+
"grad_norm": 0.00083108467515558,
|
| 2091 |
+
"learning_rate": 1.1006670709520921e-05,
|
| 2092 |
+
"loss": 0.0209,
|
| 2093 |
+
"step": 7350
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 2.4148657498362804,
|
| 2097 |
+
"grad_norm": 0.00046702896361239254,
|
| 2098 |
+
"learning_rate": 1.0855063674954519e-05,
|
| 2099 |
+
"loss": 0.0002,
|
| 2100 |
+
"step": 7375
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 2.423051735428946,
|
| 2104 |
+
"grad_norm": 0.0009466594783589244,
|
| 2105 |
+
"learning_rate": 1.0703456640388114e-05,
|
| 2106 |
+
"loss": 0.0534,
|
| 2107 |
+
"step": 7400
|
| 2108 |
+
},
|
| 2109 |
+
{
|
| 2110 |
+
"epoch": 2.431237721021611,
|
| 2111 |
+
"grad_norm": 0.0005027965526096523,
|
| 2112 |
+
"learning_rate": 1.055184960582171e-05,
|
| 2113 |
+
"loss": 0.0031,
|
| 2114 |
+
"step": 7425
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 2.4394237066142765,
|
| 2118 |
+
"grad_norm": 0.00883590430021286,
|
| 2119 |
+
"learning_rate": 1.0400242571255307e-05,
|
| 2120 |
+
"loss": 0.0078,
|
| 2121 |
+
"step": 7450
|
| 2122 |
+
},
|
| 2123 |
+
{
|
| 2124 |
+
"epoch": 2.4476096922069415,
|
| 2125 |
+
"grad_norm": 0.0009790252661332488,
|
| 2126 |
+
"learning_rate": 1.0248635536688903e-05,
|
| 2127 |
+
"loss": 0.002,
|
| 2128 |
+
"step": 7475
|
| 2129 |
+
},
|
| 2130 |
+
{
|
| 2131 |
+
"epoch": 2.455795677799607,
|
| 2132 |
+
"grad_norm": 0.017697228118777275,
|
| 2133 |
+
"learning_rate": 1.00970285021225e-05,
|
| 2134 |
+
"loss": 0.0049,
|
| 2135 |
+
"step": 7500
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 2.463981663392272,
|
| 2139 |
+
"grad_norm": 0.0005443913978524506,
|
| 2140 |
+
"learning_rate": 9.945421467556094e-06,
|
| 2141 |
+
"loss": 0.0672,
|
| 2142 |
+
"step": 7525
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 2.4721676489849376,
|
| 2146 |
+
"grad_norm": 0.0002733923611231148,
|
| 2147 |
+
"learning_rate": 9.793814432989691e-06,
|
| 2148 |
+
"loss": 0.0089,
|
| 2149 |
+
"step": 7550
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 2.480353634577603,
|
| 2153 |
+
"grad_norm": 0.002381040947511792,
|
| 2154 |
+
"learning_rate": 9.642207398423288e-06,
|
| 2155 |
+
"loss": 0.0001,
|
| 2156 |
+
"step": 7575
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 2.4885396201702683,
|
| 2160 |
+
"grad_norm": 0.0006111008697189391,
|
| 2161 |
+
"learning_rate": 9.490600363856882e-06,
|
| 2162 |
+
"loss": 0.0,
|
| 2163 |
+
"step": 7600
|
| 2164 |
+
},
|
| 2165 |
+
{
|
| 2166 |
+
"epoch": 2.4967256057629337,
|
| 2167 |
+
"grad_norm": 0.0005736036109738052,
|
| 2168 |
+
"learning_rate": 9.33899332929048e-06,
|
| 2169 |
+
"loss": 0.0176,
|
| 2170 |
+
"step": 7625
|
| 2171 |
+
},
|
| 2172 |
+
{
|
| 2173 |
+
"epoch": 2.504911591355599,
|
| 2174 |
+
"grad_norm": 0.00394043792039156,
|
| 2175 |
+
"learning_rate": 9.187386294724077e-06,
|
| 2176 |
+
"loss": 0.0039,
|
| 2177 |
+
"step": 7650
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.5130975769482644,
|
| 2181 |
+
"grad_norm": 0.0005367947742342949,
|
| 2182 |
+
"learning_rate": 9.035779260157672e-06,
|
| 2183 |
+
"loss": 0.0001,
|
| 2184 |
+
"step": 7675
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 2.52128356254093,
|
| 2188 |
+
"grad_norm": 0.004565802868455648,
|
| 2189 |
+
"learning_rate": 8.884172225591268e-06,
|
| 2190 |
+
"loss": 0.0255,
|
| 2191 |
+
"step": 7700
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 2.529469548133595,
|
| 2195 |
+
"grad_norm": 0.0027460469864308834,
|
| 2196 |
+
"learning_rate": 8.732565191024863e-06,
|
| 2197 |
+
"loss": 0.0007,
|
| 2198 |
+
"step": 7725
|
| 2199 |
+
},
|
| 2200 |
+
{
|
| 2201 |
+
"epoch": 2.5376555337262605,
|
| 2202 |
+
"grad_norm": 0.006453353445976973,
|
| 2203 |
+
"learning_rate": 8.58095815645846e-06,
|
| 2204 |
+
"loss": 0.0003,
|
| 2205 |
+
"step": 7750
|
| 2206 |
+
},
|
| 2207 |
+
{
|
| 2208 |
+
"epoch": 2.545841519318926,
|
| 2209 |
+
"grad_norm": 0.05628466606140137,
|
| 2210 |
+
"learning_rate": 8.429351121892056e-06,
|
| 2211 |
+
"loss": 0.0005,
|
| 2212 |
+
"step": 7775
|
| 2213 |
+
},
|
| 2214 |
+
{
|
| 2215 |
+
"epoch": 2.5540275049115913,
|
| 2216 |
+
"grad_norm": 0.0020795781165361404,
|
| 2217 |
+
"learning_rate": 8.277744087325652e-06,
|
| 2218 |
+
"loss": 0.0001,
|
| 2219 |
+
"step": 7800
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.5622134905042566,
|
| 2223 |
+
"grad_norm": 6.260252952575684,
|
| 2224 |
+
"learning_rate": 8.126137052759249e-06,
|
| 2225 |
+
"loss": 0.065,
|
| 2226 |
+
"step": 7825
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 2.570399476096922,
|
| 2230 |
+
"grad_norm": 0.032053545117378235,
|
| 2231 |
+
"learning_rate": 7.974530018192845e-06,
|
| 2232 |
+
"loss": 0.0005,
|
| 2233 |
+
"step": 7850
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 2.5785854616895874,
|
| 2237 |
+
"grad_norm": 0.0007737306877970695,
|
| 2238 |
+
"learning_rate": 7.82292298362644e-06,
|
| 2239 |
+
"loss": 0.0005,
|
| 2240 |
+
"step": 7875
|
| 2241 |
+
},
|
| 2242 |
+
{
|
| 2243 |
+
"epoch": 2.5867714472822527,
|
| 2244 |
+
"grad_norm": 0.0028309288900345564,
|
| 2245 |
+
"learning_rate": 7.671315949060038e-06,
|
| 2246 |
+
"loss": 0.0004,
|
| 2247 |
+
"step": 7900
|
| 2248 |
+
},
|
| 2249 |
+
{
|
| 2250 |
+
"epoch": 2.594957432874918,
|
| 2251 |
+
"grad_norm": 0.002330320654436946,
|
| 2252 |
+
"learning_rate": 7.519708914493632e-06,
|
| 2253 |
+
"loss": 0.0001,
|
| 2254 |
+
"step": 7925
|
| 2255 |
+
},
|
| 2256 |
+
{
|
| 2257 |
+
"epoch": 2.6031434184675835,
|
| 2258 |
+
"grad_norm": 0.002036983147263527,
|
| 2259 |
+
"learning_rate": 7.3681018799272296e-06,
|
| 2260 |
+
"loss": 0.0011,
|
| 2261 |
+
"step": 7950
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.611329404060249,
|
| 2265 |
+
"grad_norm": 0.0005035591893829405,
|
| 2266 |
+
"learning_rate": 7.216494845360824e-06,
|
| 2267 |
+
"loss": 0.0012,
|
| 2268 |
+
"step": 7975
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 2.619515389652914,
|
| 2272 |
+
"grad_norm": 0.011760620400309563,
|
| 2273 |
+
"learning_rate": 7.064887810794421e-06,
|
| 2274 |
+
"loss": 0.0015,
|
| 2275 |
+
"step": 8000
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 2.6277013752455796,
|
| 2279 |
+
"grad_norm": 0.07406982779502869,
|
| 2280 |
+
"learning_rate": 6.913280776228018e-06,
|
| 2281 |
+
"loss": 0.0001,
|
| 2282 |
+
"step": 8025
|
| 2283 |
+
},
|
| 2284 |
+
{
|
| 2285 |
+
"epoch": 2.635887360838245,
|
| 2286 |
+
"grad_norm": 0.002441051648929715,
|
| 2287 |
+
"learning_rate": 6.761673741661613e-06,
|
| 2288 |
+
"loss": 0.0008,
|
| 2289 |
+
"step": 8050
|
| 2290 |
+
},
|
| 2291 |
+
{
|
| 2292 |
+
"epoch": 2.6440733464309103,
|
| 2293 |
+
"grad_norm": 0.001771932584233582,
|
| 2294 |
+
"learning_rate": 6.61006670709521e-06,
|
| 2295 |
+
"loss": 0.0261,
|
| 2296 |
+
"step": 8075
|
| 2297 |
+
},
|
| 2298 |
+
{
|
| 2299 |
+
"epoch": 2.6522593320235757,
|
| 2300 |
+
"grad_norm": 0.003007644321769476,
|
| 2301 |
+
"learning_rate": 6.458459672528806e-06,
|
| 2302 |
+
"loss": 0.0027,
|
| 2303 |
+
"step": 8100
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.660445317616241,
|
| 2307 |
+
"grad_norm": 0.0029620120767503977,
|
| 2308 |
+
"learning_rate": 6.306852637962402e-06,
|
| 2309 |
+
"loss": 0.0001,
|
| 2310 |
+
"step": 8125
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 2.6686313032089064,
|
| 2314 |
+
"grad_norm": 0.002668931847438216,
|
| 2315 |
+
"learning_rate": 6.161309884778654e-06,
|
| 2316 |
+
"loss": 0.0218,
|
| 2317 |
+
"step": 8150
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 2.6768172888015718,
|
| 2321 |
+
"grad_norm": 0.005034355912357569,
|
| 2322 |
+
"learning_rate": 6.0097028502122506e-06,
|
| 2323 |
+
"loss": 0.0002,
|
| 2324 |
+
"step": 8175
|
| 2325 |
+
},
|
| 2326 |
+
{
|
| 2327 |
+
"epoch": 2.685003274394237,
|
| 2328 |
+
"grad_norm": 0.2363610863685608,
|
| 2329 |
+
"learning_rate": 5.858095815645846e-06,
|
| 2330 |
+
"loss": 0.0237,
|
| 2331 |
+
"step": 8200
|
| 2332 |
+
},
|
| 2333 |
+
{
|
| 2334 |
+
"epoch": 2.6931892599869025,
|
| 2335 |
+
"grad_norm": 0.0029050330631434917,
|
| 2336 |
+
"learning_rate": 5.7064887810794426e-06,
|
| 2337 |
+
"loss": 0.0142,
|
| 2338 |
+
"step": 8225
|
| 2339 |
+
},
|
| 2340 |
+
{
|
| 2341 |
+
"epoch": 2.701375245579568,
|
| 2342 |
+
"grad_norm": 0.0002962597936857492,
|
| 2343 |
+
"learning_rate": 5.554881746513038e-06,
|
| 2344 |
+
"loss": 0.0008,
|
| 2345 |
+
"step": 8250
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.7095612311722332,
|
| 2349 |
+
"grad_norm": 0.002445927122607827,
|
| 2350 |
+
"learning_rate": 5.4032747119466346e-06,
|
| 2351 |
+
"loss": 0.0395,
|
| 2352 |
+
"step": 8275
|
| 2353 |
+
},
|
| 2354 |
+
{
|
| 2355 |
+
"epoch": 2.7177472167648986,
|
| 2356 |
+
"grad_norm": 4.508545875549316,
|
| 2357 |
+
"learning_rate": 5.251667677380231e-06,
|
| 2358 |
+
"loss": 0.0135,
|
| 2359 |
+
"step": 8300
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 2.725933202357564,
|
| 2363 |
+
"grad_norm": 0.001652977429330349,
|
| 2364 |
+
"learning_rate": 5.1000606428138266e-06,
|
| 2365 |
+
"loss": 0.0005,
|
| 2366 |
+
"step": 8325
|
| 2367 |
+
},
|
| 2368 |
+
{
|
| 2369 |
+
"epoch": 2.7341191879502293,
|
| 2370 |
+
"grad_norm": 0.007348712533712387,
|
| 2371 |
+
"learning_rate": 4.948453608247423e-06,
|
| 2372 |
+
"loss": 0.0001,
|
| 2373 |
+
"step": 8350
|
| 2374 |
+
},
|
| 2375 |
+
{
|
| 2376 |
+
"epoch": 2.7423051735428947,
|
| 2377 |
+
"grad_norm": 0.0047335317358374596,
|
| 2378 |
+
"learning_rate": 4.796846573681019e-06,
|
| 2379 |
+
"loss": 0.0002,
|
| 2380 |
+
"step": 8375
|
| 2381 |
+
},
|
| 2382 |
+
{
|
| 2383 |
+
"epoch": 2.75049115913556,
|
| 2384 |
+
"grad_norm": 0.0006250338046811521,
|
| 2385 |
+
"learning_rate": 4.645239539114615e-06,
|
| 2386 |
+
"loss": 0.002,
|
| 2387 |
+
"step": 8400
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.7586771447282254,
|
| 2391 |
+
"grad_norm": 0.0020159140694886446,
|
| 2392 |
+
"learning_rate": 4.493632504548211e-06,
|
| 2393 |
+
"loss": 0.0001,
|
| 2394 |
+
"step": 8425
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 2.766863130320891,
|
| 2398 |
+
"grad_norm": 0.005961321294307709,
|
| 2399 |
+
"learning_rate": 4.342025469981807e-06,
|
| 2400 |
+
"loss": 0.0825,
|
| 2401 |
+
"step": 8450
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 2.775049115913556,
|
| 2405 |
+
"grad_norm": 0.004888585302978754,
|
| 2406 |
+
"learning_rate": 4.190418435415403e-06,
|
| 2407 |
+
"loss": 0.0003,
|
| 2408 |
+
"step": 8475
|
| 2409 |
+
},
|
| 2410 |
+
{
|
| 2411 |
+
"epoch": 2.7832351015062216,
|
| 2412 |
+
"grad_norm": 0.0033836057409644127,
|
| 2413 |
+
"learning_rate": 4.038811400849e-06,
|
| 2414 |
+
"loss": 0.0002,
|
| 2415 |
+
"step": 8500
|
| 2416 |
+
},
|
| 2417 |
+
{
|
| 2418 |
+
"epoch": 2.791421087098887,
|
| 2419 |
+
"grad_norm": 0.048891954123973846,
|
| 2420 |
+
"learning_rate": 3.887204366282595e-06,
|
| 2421 |
+
"loss": 0.0002,
|
| 2422 |
+
"step": 8525
|
| 2423 |
+
},
|
| 2424 |
+
{
|
| 2425 |
+
"epoch": 2.7996070726915523,
|
| 2426 |
+
"grad_norm": 0.0019098659977316856,
|
| 2427 |
+
"learning_rate": 3.735597331716192e-06,
|
| 2428 |
+
"loss": 0.0014,
|
| 2429 |
+
"step": 8550
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.8077930582842177,
|
| 2433 |
+
"grad_norm": 0.005250090733170509,
|
| 2434 |
+
"learning_rate": 3.583990297149788e-06,
|
| 2435 |
+
"loss": 0.0019,
|
| 2436 |
+
"step": 8575
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 2.815979043876883,
|
| 2440 |
+
"grad_norm": 0.0005940294940955937,
|
| 2441 |
+
"learning_rate": 3.432383262583384e-06,
|
| 2442 |
+
"loss": 0.0003,
|
| 2443 |
+
"step": 8600
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 2.8241650294695484,
|
| 2447 |
+
"grad_norm": 0.00769090885296464,
|
| 2448 |
+
"learning_rate": 3.28077622801698e-06,
|
| 2449 |
+
"loss": 0.0003,
|
| 2450 |
+
"step": 8625
|
| 2451 |
+
},
|
| 2452 |
+
{
|
| 2453 |
+
"epoch": 2.8323510150622138,
|
| 2454 |
+
"grad_norm": 0.004769071005284786,
|
| 2455 |
+
"learning_rate": 3.129169193450576e-06,
|
| 2456 |
+
"loss": 0.0082,
|
| 2457 |
+
"step": 8650
|
| 2458 |
+
},
|
| 2459 |
+
{
|
| 2460 |
+
"epoch": 2.8405370006548787,
|
| 2461 |
+
"grad_norm": 0.0006792404456064105,
|
| 2462 |
+
"learning_rate": 2.983626440266829e-06,
|
| 2463 |
+
"loss": 0.071,
|
| 2464 |
+
"step": 8675
|
| 2465 |
+
},
|
| 2466 |
+
{
|
| 2467 |
+
"epoch": 2.848722986247544,
|
| 2468 |
+
"grad_norm": 0.0006690117879770696,
|
| 2469 |
+
"learning_rate": 2.832019405700425e-06,
|
| 2470 |
+
"loss": 0.0303,
|
| 2471 |
+
"step": 8700
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.8569089718402094,
|
| 2475 |
+
"grad_norm": 0.003535321680828929,
|
| 2476 |
+
"learning_rate": 2.6804123711340204e-06,
|
| 2477 |
+
"loss": 0.0041,
|
| 2478 |
+
"step": 8725
|
| 2479 |
+
},
|
| 2480 |
+
{
|
| 2481 |
+
"epoch": 2.865094957432875,
|
| 2482 |
+
"grad_norm": 0.007222822401672602,
|
| 2483 |
+
"learning_rate": 2.528805336567617e-06,
|
| 2484 |
+
"loss": 0.0002,
|
| 2485 |
+
"step": 8750
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 2.87328094302554,
|
| 2489 |
+
"grad_norm": 0.001671507372520864,
|
| 2490 |
+
"learning_rate": 2.377198302001213e-06,
|
| 2491 |
+
"loss": 0.0102,
|
| 2492 |
+
"step": 8775
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 2.8814669286182055,
|
| 2496 |
+
"grad_norm": 0.0008472661720588803,
|
| 2497 |
+
"learning_rate": 2.2255912674348092e-06,
|
| 2498 |
+
"loss": 0.0004,
|
| 2499 |
+
"step": 8800
|
| 2500 |
+
},
|
| 2501 |
+
{
|
| 2502 |
+
"epoch": 2.889652914210871,
|
| 2503 |
+
"grad_norm": 0.0037155933678150177,
|
| 2504 |
+
"learning_rate": 2.0739842328684052e-06,
|
| 2505 |
+
"loss": 0.0007,
|
| 2506 |
+
"step": 8825
|
| 2507 |
+
},
|
| 2508 |
+
{
|
| 2509 |
+
"epoch": 2.8978388998035363,
|
| 2510 |
+
"grad_norm": 0.0019001095788553357,
|
| 2511 |
+
"learning_rate": 1.9223771983020012e-06,
|
| 2512 |
+
"loss": 0.0071,
|
| 2513 |
+
"step": 8850
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 2.9060248853962016,
|
| 2517 |
+
"grad_norm": 0.01914367638528347,
|
| 2518 |
+
"learning_rate": 1.7707701637355974e-06,
|
| 2519 |
+
"loss": 0.0003,
|
| 2520 |
+
"step": 8875
|
| 2521 |
+
},
|
| 2522 |
+
{
|
| 2523 |
+
"epoch": 2.914210870988867,
|
| 2524 |
+
"grad_norm": 0.0017183177405968308,
|
| 2525 |
+
"learning_rate": 1.6191631291691936e-06,
|
| 2526 |
+
"loss": 0.0001,
|
| 2527 |
+
"step": 8900
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 2.9223968565815324,
|
| 2531 |
+
"grad_norm": 0.000283712986856699,
|
| 2532 |
+
"learning_rate": 1.4675560946027896e-06,
|
| 2533 |
+
"loss": 0.007,
|
| 2534 |
+
"step": 8925
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 2.9305828421741977,
|
| 2538 |
+
"grad_norm": 0.0013638654490932822,
|
| 2539 |
+
"learning_rate": 1.3159490600363858e-06,
|
| 2540 |
+
"loss": 0.031,
|
| 2541 |
+
"step": 8950
|
| 2542 |
+
},
|
| 2543 |
+
{
|
| 2544 |
+
"epoch": 2.938768827766863,
|
| 2545 |
+
"grad_norm": 0.0024313374888151884,
|
| 2546 |
+
"learning_rate": 1.1643420254699818e-06,
|
| 2547 |
+
"loss": 0.0001,
|
| 2548 |
+
"step": 8975
|
| 2549 |
+
},
|
| 2550 |
+
{
|
| 2551 |
+
"epoch": 2.9469548133595285,
|
| 2552 |
+
"grad_norm": 0.0011194221442565322,
|
| 2553 |
+
"learning_rate": 1.012734990903578e-06,
|
| 2554 |
+
"loss": 0.0001,
|
| 2555 |
+
"step": 9000
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 2.955140798952194,
|
| 2559 |
+
"grad_norm": 0.0018601082265377045,
|
| 2560 |
+
"learning_rate": 8.611279563371741e-07,
|
| 2561 |
+
"loss": 0.0091,
|
| 2562 |
+
"step": 9025
|
| 2563 |
+
},
|
| 2564 |
+
{
|
| 2565 |
+
"epoch": 2.963326784544859,
|
| 2566 |
+
"grad_norm": 0.001517856726422906,
|
| 2567 |
+
"learning_rate": 7.095209217707701e-07,
|
| 2568 |
+
"loss": 0.0003,
|
| 2569 |
+
"step": 9050
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 2.9715127701375246,
|
| 2573 |
+
"grad_norm": 0.00028195424238219857,
|
| 2574 |
+
"learning_rate": 5.579138872043664e-07,
|
| 2575 |
+
"loss": 0.0225,
|
| 2576 |
+
"step": 9075
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 2.97969875573019,
|
| 2580 |
+
"grad_norm": 0.0005502802086994052,
|
| 2581 |
+
"learning_rate": 4.0630685263796246e-07,
|
| 2582 |
+
"loss": 0.0001,
|
| 2583 |
+
"step": 9100
|
| 2584 |
+
},
|
| 2585 |
+
{
|
| 2586 |
+
"epoch": 2.9878847413228553,
|
| 2587 |
+
"grad_norm": 0.0016429908573627472,
|
| 2588 |
+
"learning_rate": 2.5469981807155856e-07,
|
| 2589 |
+
"loss": 0.0003,
|
| 2590 |
+
"step": 9125
|
| 2591 |
+
},
|
| 2592 |
+
{
|
| 2593 |
+
"epoch": 2.9960707269155207,
|
| 2594 |
+
"grad_norm": 0.0061945500783622265,
|
| 2595 |
+
"learning_rate": 1.0309278350515465e-07,
|
| 2596 |
+
"loss": 0.0033,
|
| 2597 |
+
"step": 9150
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 3.0,
|
| 2601 |
+
"eval_accuracy": 0.9988539620170268,
|
| 2602 |
+
"eval_auc": 0.9999216153508497,
|
| 2603 |
+
"eval_f1": 0.9523809523809523,
|
| 2604 |
+
"eval_loss": 0.0027391575276851654,
|
| 2605 |
+
"eval_precision": 0.958904109589041,
|
| 2606 |
+
"eval_recall": 0.9459459459459459,
|
| 2607 |
+
"eval_runtime": 74.4951,
|
| 2608 |
+
"eval_samples_per_second": 81.992,
|
| 2609 |
+
"eval_steps_per_second": 5.128,
|
| 2610 |
+
"step": 9162
|
| 2611 |
+
}
|
| 2612 |
+
],
|
| 2613 |
+
"logging_steps": 25,
|
| 2614 |
+
"max_steps": 9162,
|
| 2615 |
+
"num_input_tokens_seen": 0,
|
| 2616 |
+
"num_train_epochs": 3,
|
| 2617 |
+
"save_steps": 500,
|
| 2618 |
+
"stateful_callbacks": {
|
| 2619 |
+
"EarlyStoppingCallback": {
|
| 2620 |
+
"args": {
|
| 2621 |
+
"early_stopping_patience": 5,
|
| 2622 |
+
"early_stopping_threshold": 0.01
|
| 2623 |
+
},
|
| 2624 |
+
"attributes": {
|
| 2625 |
+
"early_stopping_patience_counter": 2
|
| 2626 |
+
}
|
| 2627 |
+
},
|
| 2628 |
+
"TrainerControl": {
|
| 2629 |
+
"args": {
|
| 2630 |
+
"should_epoch_stop": false,
|
| 2631 |
+
"should_evaluate": false,
|
| 2632 |
+
"should_log": false,
|
| 2633 |
+
"should_save": true,
|
| 2634 |
+
"should_training_stop": true
|
| 2635 |
+
},
|
| 2636 |
+
"attributes": {}
|
| 2637 |
+
}
|
| 2638 |
+
},
|
| 2639 |
+
"total_flos": 5.944931470697103e+19,
|
| 2640 |
+
"train_batch_size": 8,
|
| 2641 |
+
"trial_name": null,
|
| 2642 |
+
"trial_params": null
|
| 2643 |
+
}
|
checkpoint-9162/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad693d0c2f46d11ac2af896306db2390532e5e2824467769cfb05f3952a13840
|
| 3 |
+
size 5240
|
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "google/vit-large-patch32-384",
|
| 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": 1024,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "not_soyjak",
|
| 14 |
+
"1": "soyjak"
|
| 15 |
+
},
|
| 16 |
+
"image_size": 384,
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 4096,
|
| 19 |
+
"label2id": {
|
| 20 |
+
"not_soyjak": 0,
|
| 21 |
+
"soyjak": 1
|
| 22 |
+
},
|
| 23 |
+
"layer_norm_eps": 1e-12,
|
| 24 |
+
"model_type": "vit",
|
| 25 |
+
"num_attention_heads": 16,
|
| 26 |
+
"num_channels": 3,
|
| 27 |
+
"num_hidden_layers": 24,
|
| 28 |
+
"patch_size": 32,
|
| 29 |
+
"problem_type": "single_label_classification",
|
| 30 |
+
"qkv_bias": true,
|
| 31 |
+
"torch_dtype": "float32",
|
| 32 |
+
"transformers_version": "4.45.0"
|
| 33 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6370312cf8b3c3c9cdcab23d0c40535752368b0bcb77471852c134abba4add6d
|
| 3 |
+
size 1222485480
|
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/Oct22_01-12-49_r-haiefff-sjstuff-186zjrrd-42d15-6st93/events.out.tfevents.1729559571.r-haiefff-sjstuff-186zjrrd-42d15-6st93.222.0
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:d6299f3d69112400500f46df5ae6d8fb129e2a700817cd1ae76b8723c6c37b93
|
| 3 |
+
size 84077
|
runs/Oct22_01-12-49_r-haiefff-sjstuff-186zjrrd-42d15-6st93/events.out.tfevents.1729562910.r-haiefff-sjstuff-186zjrrd-42d15-6st93.222.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a63183c2a54fd5c900679bc7ae65350bb811b9bec69b40eef2f0ffe889f4d566
|
| 3 |
+
size 607
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad693d0c2f46d11ac2af896306db2390532e5e2824467769cfb05f3952a13840
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "sj-det-6/autotrain-data",
|
| 3 |
+
"model": "google/vit-large-patch32-384",
|
| 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-6",
|
| 19 |
+
"auto_find_batch_size": false,
|
| 20 |
+
"mixed_precision": "fp16",
|
| 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 |
+
}
|