Upload folder using huggingface_hub
Browse files- README.md +31 -0
- checkpoint-6104/config.json +33 -0
- checkpoint-6104/model.safetensors +3 -0
- checkpoint-6104/optimizer.pt +3 -0
- checkpoint-6104/rng_state.pth +3 -0
- checkpoint-6104/scheduler.pt +3 -0
- checkpoint-6104/trainer_state.json +1776 -0
- checkpoint-6104/training_args.bin +3 -0
- config.json +33 -0
- model.safetensors +3 -0
- preprocessor_config.json +22 -0
- runs/Oct21_16-58-53_r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n/events.out.tfevents.1729529935.r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n.184.0 +2 -2
- runs/Oct21_16-58-53_r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n/events.out.tfevents.1729533203.r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n.184.1 +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: 0.00864658784121275
|
| 22 |
+
|
| 23 |
+
f1: 0.9361702127659575
|
| 24 |
+
|
| 25 |
+
precision: 0.9850746268656716
|
| 26 |
+
|
| 27 |
+
recall: 0.8918918918918919
|
| 28 |
+
|
| 29 |
+
auc: 0.9973051409618573
|
| 30 |
+
|
| 31 |
+
accuracy: 0.9985255570117956
|
checkpoint-6104/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-6104/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:afc98a6e4beae9f2797493e9d5814e9b209923a1ef8abf6b1e47b4a600cab79d
|
| 3 |
+
size 1222485480
|
checkpoint-6104/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c2a7082804ebc6043796b8189d3cf894cfd0a3219e36bdb3c6feb22186d41852
|
| 3 |
+
size 2445205930
|
checkpoint-6104/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c47fd458cc8cdabebbf5eeed7f21b1735e769b3c7972d8387e28384cfb947a4
|
| 3 |
+
size 14244
|
checkpoint-6104/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09aa3c3e969189f099c047afe35c6aa8afd5ac1106b184a854ea466d868930c4
|
| 3 |
+
size 1064
|
checkpoint-6104/trainer_state.json
ADDED
|
@@ -0,0 +1,1776 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.00864658784121275,
|
| 3 |
+
"best_model_checkpoint": "sj-det-5/checkpoint-6104",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 6104,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0081913499344692,
|
| 13 |
+
"grad_norm": 10.024361610412598,
|
| 14 |
+
"learning_rate": 1.3100436681222706e-06,
|
| 15 |
+
"loss": 0.6701,
|
| 16 |
+
"step": 25
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0163826998689384,
|
| 20 |
+
"grad_norm": 5.330548286437988,
|
| 21 |
+
"learning_rate": 2.6746724890829695e-06,
|
| 22 |
+
"loss": 0.4647,
|
| 23 |
+
"step": 50
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0245740498034076,
|
| 27 |
+
"grad_norm": 2.015732765197754,
|
| 28 |
+
"learning_rate": 4.039301310043668e-06,
|
| 29 |
+
"loss": 0.1809,
|
| 30 |
+
"step": 75
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0327653997378768,
|
| 34 |
+
"grad_norm": 6.1893134117126465,
|
| 35 |
+
"learning_rate": 5.4039301310043675e-06,
|
| 36 |
+
"loss": 0.1308,
|
| 37 |
+
"step": 100
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.040956749672346006,
|
| 41 |
+
"grad_norm": 0.21893443167209625,
|
| 42 |
+
"learning_rate": 6.768558951965067e-06,
|
| 43 |
+
"loss": 0.0337,
|
| 44 |
+
"step": 125
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0491480996068152,
|
| 48 |
+
"grad_norm": 5.707635402679443,
|
| 49 |
+
"learning_rate": 8.133187772925763e-06,
|
| 50 |
+
"loss": 0.0952,
|
| 51 |
+
"step": 150
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.05733944954128441,
|
| 55 |
+
"grad_norm": 0.042420562356710434,
|
| 56 |
+
"learning_rate": 9.497816593886463e-06,
|
| 57 |
+
"loss": 0.0225,
|
| 58 |
+
"step": 175
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0655307994757536,
|
| 62 |
+
"grad_norm": 0.07790377736091614,
|
| 63 |
+
"learning_rate": 1.0862445414847162e-05,
|
| 64 |
+
"loss": 0.0961,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0737221494102228,
|
| 69 |
+
"grad_norm": 0.06831495463848114,
|
| 70 |
+
"learning_rate": 1.222707423580786e-05,
|
| 71 |
+
"loss": 0.0029,
|
| 72 |
+
"step": 225
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.08191349934469201,
|
| 76 |
+
"grad_norm": 0.3763171434402466,
|
| 77 |
+
"learning_rate": 1.3591703056768559e-05,
|
| 78 |
+
"loss": 0.0165,
|
| 79 |
+
"step": 250
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.09010484927916121,
|
| 83 |
+
"grad_norm": 0.042719755321741104,
|
| 84 |
+
"learning_rate": 1.4956331877729258e-05,
|
| 85 |
+
"loss": 0.0537,
|
| 86 |
+
"step": 275
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0982961992136304,
|
| 90 |
+
"grad_norm": 1.7841322422027588,
|
| 91 |
+
"learning_rate": 1.632096069868996e-05,
|
| 92 |
+
"loss": 0.0304,
|
| 93 |
+
"step": 300
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1064875491480996,
|
| 97 |
+
"grad_norm": 0.006917485501617193,
|
| 98 |
+
"learning_rate": 1.7685589519650656e-05,
|
| 99 |
+
"loss": 0.0061,
|
| 100 |
+
"step": 325
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.11467889908256881,
|
| 104 |
+
"grad_norm": 0.007079846691340208,
|
| 105 |
+
"learning_rate": 1.9050218340611354e-05,
|
| 106 |
+
"loss": 0.0234,
|
| 107 |
+
"step": 350
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.12287024901703801,
|
| 111 |
+
"grad_norm": 0.013362312689423561,
|
| 112 |
+
"learning_rate": 2.0414847161572055e-05,
|
| 113 |
+
"loss": 0.0214,
|
| 114 |
+
"step": 375
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.1310615989515072,
|
| 118 |
+
"grad_norm": 0.011783940717577934,
|
| 119 |
+
"learning_rate": 2.1779475982532752e-05,
|
| 120 |
+
"loss": 0.0543,
|
| 121 |
+
"step": 400
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.13925294888597642,
|
| 125 |
+
"grad_norm": 0.020665360614657402,
|
| 126 |
+
"learning_rate": 2.3144104803493453e-05,
|
| 127 |
+
"loss": 0.0316,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1474442988204456,
|
| 132 |
+
"grad_norm": 1.4373878240585327,
|
| 133 |
+
"learning_rate": 2.450873362445415e-05,
|
| 134 |
+
"loss": 0.06,
|
| 135 |
+
"step": 450
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1556356487549148,
|
| 139 |
+
"grad_norm": 3.8112599849700928,
|
| 140 |
+
"learning_rate": 2.587336244541485e-05,
|
| 141 |
+
"loss": 0.0182,
|
| 142 |
+
"step": 475
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.16382699868938402,
|
| 146 |
+
"grad_norm": 5.189245223999023,
|
| 147 |
+
"learning_rate": 2.7237991266375546e-05,
|
| 148 |
+
"loss": 0.154,
|
| 149 |
+
"step": 500
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1720183486238532,
|
| 153 |
+
"grad_norm": 0.015202702023088932,
|
| 154 |
+
"learning_rate": 2.8602620087336247e-05,
|
| 155 |
+
"loss": 0.0294,
|
| 156 |
+
"step": 525
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.18020969855832242,
|
| 160 |
+
"grad_norm": 0.015019269660115242,
|
| 161 |
+
"learning_rate": 2.9967248908296948e-05,
|
| 162 |
+
"loss": 0.0591,
|
| 163 |
+
"step": 550
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1884010484927916,
|
| 167 |
+
"grad_norm": 0.0036466640885919333,
|
| 168 |
+
"learning_rate": 3.133187772925764e-05,
|
| 169 |
+
"loss": 0.0034,
|
| 170 |
+
"step": 575
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.1965923984272608,
|
| 174 |
+
"grad_norm": 0.011716458015143871,
|
| 175 |
+
"learning_rate": 3.269650655021834e-05,
|
| 176 |
+
"loss": 0.0015,
|
| 177 |
+
"step": 600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.20478374836173002,
|
| 181 |
+
"grad_norm": 0.0010442484635859728,
|
| 182 |
+
"learning_rate": 3.406113537117904e-05,
|
| 183 |
+
"loss": 0.0002,
|
| 184 |
+
"step": 625
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.2129750982961992,
|
| 188 |
+
"grad_norm": 0.030393408611416817,
|
| 189 |
+
"learning_rate": 3.542576419213974e-05,
|
| 190 |
+
"loss": 0.0555,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.22116644823066842,
|
| 195 |
+
"grad_norm": 0.004823782481253147,
|
| 196 |
+
"learning_rate": 3.6790393013100435e-05,
|
| 197 |
+
"loss": 0.0011,
|
| 198 |
+
"step": 675
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.22935779816513763,
|
| 202 |
+
"grad_norm": 0.02440587244927883,
|
| 203 |
+
"learning_rate": 3.815502183406114e-05,
|
| 204 |
+
"loss": 0.0504,
|
| 205 |
+
"step": 700
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.2375491480996068,
|
| 209 |
+
"grad_norm": 0.05391886085271835,
|
| 210 |
+
"learning_rate": 3.951965065502183e-05,
|
| 211 |
+
"loss": 0.0821,
|
| 212 |
+
"step": 725
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.24574049803407602,
|
| 216 |
+
"grad_norm": 0.006875240709632635,
|
| 217 |
+
"learning_rate": 4.0884279475982535e-05,
|
| 218 |
+
"loss": 0.0008,
|
| 219 |
+
"step": 750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25393184796854523,
|
| 223 |
+
"grad_norm": 0.0016325929900631309,
|
| 224 |
+
"learning_rate": 4.224890829694323e-05,
|
| 225 |
+
"loss": 0.0183,
|
| 226 |
+
"step": 775
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.2621231979030144,
|
| 230 |
+
"grad_norm": 0.029884610325098038,
|
| 231 |
+
"learning_rate": 4.361353711790393e-05,
|
| 232 |
+
"loss": 0.0525,
|
| 233 |
+
"step": 800
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.2703145478374836,
|
| 237 |
+
"grad_norm": 0.05142253637313843,
|
| 238 |
+
"learning_rate": 4.497816593886463e-05,
|
| 239 |
+
"loss": 0.0499,
|
| 240 |
+
"step": 825
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.27850589777195284,
|
| 244 |
+
"grad_norm": 0.0004886944079771638,
|
| 245 |
+
"learning_rate": 4.634279475982533e-05,
|
| 246 |
+
"loss": 0.0401,
|
| 247 |
+
"step": 850
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.286697247706422,
|
| 251 |
+
"grad_norm": 0.003223366802558303,
|
| 252 |
+
"learning_rate": 4.770742358078603e-05,
|
| 253 |
+
"loss": 0.0893,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.2948885976408912,
|
| 258 |
+
"grad_norm": 0.02335241623222828,
|
| 259 |
+
"learning_rate": 4.907205240174673e-05,
|
| 260 |
+
"loss": 0.0003,
|
| 261 |
+
"step": 900
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.30307994757536044,
|
| 265 |
+
"grad_norm": 0.036217641085386276,
|
| 266 |
+
"learning_rate": 4.9951456310679614e-05,
|
| 267 |
+
"loss": 0.0387,
|
| 268 |
+
"step": 925
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.3112712975098296,
|
| 272 |
+
"grad_norm": 0.0005936770467087626,
|
| 273 |
+
"learning_rate": 4.97997572815534e-05,
|
| 274 |
+
"loss": 0.0231,
|
| 275 |
+
"step": 950
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3194626474442988,
|
| 279 |
+
"grad_norm": 0.11047300696372986,
|
| 280 |
+
"learning_rate": 4.964805825242719e-05,
|
| 281 |
+
"loss": 0.0235,
|
| 282 |
+
"step": 975
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.32765399737876805,
|
| 286 |
+
"grad_norm": 0.0038526670541614294,
|
| 287 |
+
"learning_rate": 4.9496359223300976e-05,
|
| 288 |
+
"loss": 0.0804,
|
| 289 |
+
"step": 1000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.33584534731323723,
|
| 293 |
+
"grad_norm": 0.002965217223390937,
|
| 294 |
+
"learning_rate": 4.934466019417476e-05,
|
| 295 |
+
"loss": 0.0019,
|
| 296 |
+
"step": 1025
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.3440366972477064,
|
| 300 |
+
"grad_norm": 1.2460094690322876,
|
| 301 |
+
"learning_rate": 4.9192961165048545e-05,
|
| 302 |
+
"loss": 0.0218,
|
| 303 |
+
"step": 1050
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.3522280471821756,
|
| 307 |
+
"grad_norm": 0.00023055930796545,
|
| 308 |
+
"learning_rate": 4.904126213592233e-05,
|
| 309 |
+
"loss": 0.0824,
|
| 310 |
+
"step": 1075
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.36041939711664484,
|
| 314 |
+
"grad_norm": 0.0008325474918819964,
|
| 315 |
+
"learning_rate": 4.888956310679612e-05,
|
| 316 |
+
"loss": 0.0,
|
| 317 |
+
"step": 1100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.368610747051114,
|
| 321 |
+
"grad_norm": 0.12137041985988617,
|
| 322 |
+
"learning_rate": 4.87378640776699e-05,
|
| 323 |
+
"loss": 0.0429,
|
| 324 |
+
"step": 1125
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.3768020969855832,
|
| 328 |
+
"grad_norm": 0.12224938720464706,
|
| 329 |
+
"learning_rate": 4.858616504854369e-05,
|
| 330 |
+
"loss": 0.0287,
|
| 331 |
+
"step": 1150
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.38499344692005244,
|
| 335 |
+
"grad_norm": 0.00017641298472881317,
|
| 336 |
+
"learning_rate": 4.843446601941748e-05,
|
| 337 |
+
"loss": 0.0937,
|
| 338 |
+
"step": 1175
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.3931847968545216,
|
| 342 |
+
"grad_norm": 0.00043992913560941815,
|
| 343 |
+
"learning_rate": 4.828276699029127e-05,
|
| 344 |
+
"loss": 0.0001,
|
| 345 |
+
"step": 1200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4013761467889908,
|
| 349 |
+
"grad_norm": 0.0067256479524075985,
|
| 350 |
+
"learning_rate": 4.813106796116505e-05,
|
| 351 |
+
"loss": 0.0403,
|
| 352 |
+
"step": 1225
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.40956749672346004,
|
| 356 |
+
"grad_norm": 0.008025743998587132,
|
| 357 |
+
"learning_rate": 4.797936893203884e-05,
|
| 358 |
+
"loss": 0.1236,
|
| 359 |
+
"step": 1250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.41775884665792923,
|
| 363 |
+
"grad_norm": 0.0024881409481167793,
|
| 364 |
+
"learning_rate": 4.7827669902912624e-05,
|
| 365 |
+
"loss": 0.0065,
|
| 366 |
+
"step": 1275
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.4259501965923984,
|
| 370 |
+
"grad_norm": 0.00034138892078772187,
|
| 371 |
+
"learning_rate": 4.767597087378641e-05,
|
| 372 |
+
"loss": 0.0007,
|
| 373 |
+
"step": 1300
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.43414154652686765,
|
| 377 |
+
"grad_norm": 0.09658686816692352,
|
| 378 |
+
"learning_rate": 4.7524271844660193e-05,
|
| 379 |
+
"loss": 0.0557,
|
| 380 |
+
"step": 1325
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.44233289646133683,
|
| 384 |
+
"grad_norm": 0.004737064708024263,
|
| 385 |
+
"learning_rate": 4.7372572815533985e-05,
|
| 386 |
+
"loss": 0.0658,
|
| 387 |
+
"step": 1350
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.450524246395806,
|
| 391 |
+
"grad_norm": 0.014197874814271927,
|
| 392 |
+
"learning_rate": 4.722087378640777e-05,
|
| 393 |
+
"loss": 0.0263,
|
| 394 |
+
"step": 1375
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.45871559633027525,
|
| 398 |
+
"grad_norm": 0.012036195956170559,
|
| 399 |
+
"learning_rate": 4.7069174757281555e-05,
|
| 400 |
+
"loss": 0.0672,
|
| 401 |
+
"step": 1400
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.46690694626474444,
|
| 405 |
+
"grad_norm": 0.006545116659253836,
|
| 406 |
+
"learning_rate": 4.691747572815534e-05,
|
| 407 |
+
"loss": 0.0136,
|
| 408 |
+
"step": 1425
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.4750982961992136,
|
| 412 |
+
"grad_norm": 0.003104938892647624,
|
| 413 |
+
"learning_rate": 4.676577669902913e-05,
|
| 414 |
+
"loss": 0.0412,
|
| 415 |
+
"step": 1450
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.4832896461336828,
|
| 419 |
+
"grad_norm": 6.474489688873291,
|
| 420 |
+
"learning_rate": 4.661407766990292e-05,
|
| 421 |
+
"loss": 0.0564,
|
| 422 |
+
"step": 1475
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.49148099606815204,
|
| 426 |
+
"grad_norm": 0.11798975616693497,
|
| 427 |
+
"learning_rate": 4.64623786407767e-05,
|
| 428 |
+
"loss": 0.0441,
|
| 429 |
+
"step": 1500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.4996723460026212,
|
| 433 |
+
"grad_norm": 0.004167527426034212,
|
| 434 |
+
"learning_rate": 4.631067961165049e-05,
|
| 435 |
+
"loss": 0.0049,
|
| 436 |
+
"step": 1525
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.5078636959370905,
|
| 440 |
+
"grad_norm": 0.023102520033717155,
|
| 441 |
+
"learning_rate": 4.615898058252427e-05,
|
| 442 |
+
"loss": 0.0679,
|
| 443 |
+
"step": 1550
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.5160550458715596,
|
| 447 |
+
"grad_norm": 0.12365124374628067,
|
| 448 |
+
"learning_rate": 4.6007281553398063e-05,
|
| 449 |
+
"loss": 0.0429,
|
| 450 |
+
"step": 1575
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.5242463958060288,
|
| 454 |
+
"grad_norm": 0.02387346886098385,
|
| 455 |
+
"learning_rate": 4.585558252427185e-05,
|
| 456 |
+
"loss": 0.0706,
|
| 457 |
+
"step": 1600
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.5324377457404981,
|
| 461 |
+
"grad_norm": 0.2509637773036957,
|
| 462 |
+
"learning_rate": 4.570388349514563e-05,
|
| 463 |
+
"loss": 0.0108,
|
| 464 |
+
"step": 1625
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.5406290956749672,
|
| 468 |
+
"grad_norm": 0.006082233507186174,
|
| 469 |
+
"learning_rate": 4.555218446601942e-05,
|
| 470 |
+
"loss": 0.1004,
|
| 471 |
+
"step": 1650
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.5488204456094364,
|
| 475 |
+
"grad_norm": 0.002660457044839859,
|
| 476 |
+
"learning_rate": 4.540048543689321e-05,
|
| 477 |
+
"loss": 0.0015,
|
| 478 |
+
"step": 1675
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.5570117955439057,
|
| 482 |
+
"grad_norm": 0.029006479308009148,
|
| 483 |
+
"learning_rate": 4.524878640776699e-05,
|
| 484 |
+
"loss": 0.0698,
|
| 485 |
+
"step": 1700
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.5652031454783748,
|
| 489 |
+
"grad_norm": 2.156792163848877,
|
| 490 |
+
"learning_rate": 4.509708737864078e-05,
|
| 491 |
+
"loss": 0.0054,
|
| 492 |
+
"step": 1725
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.573394495412844,
|
| 496 |
+
"grad_norm": 0.004138014279305935,
|
| 497 |
+
"learning_rate": 4.4945388349514565e-05,
|
| 498 |
+
"loss": 0.0527,
|
| 499 |
+
"step": 1750
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.5815858453473133,
|
| 503 |
+
"grad_norm": 3.908644199371338,
|
| 504 |
+
"learning_rate": 4.479368932038835e-05,
|
| 505 |
+
"loss": 0.0543,
|
| 506 |
+
"step": 1775
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.5897771952817824,
|
| 510 |
+
"grad_norm": 0.4576684236526489,
|
| 511 |
+
"learning_rate": 4.4641990291262135e-05,
|
| 512 |
+
"loss": 0.0386,
|
| 513 |
+
"step": 1800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.5979685452162516,
|
| 517 |
+
"grad_norm": 0.00034501738264225423,
|
| 518 |
+
"learning_rate": 4.4490291262135927e-05,
|
| 519 |
+
"loss": 0.0479,
|
| 520 |
+
"step": 1825
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.6061598951507209,
|
| 524 |
+
"grad_norm": 0.020500104874372482,
|
| 525 |
+
"learning_rate": 4.433859223300971e-05,
|
| 526 |
+
"loss": 0.0522,
|
| 527 |
+
"step": 1850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.61435124508519,
|
| 531 |
+
"grad_norm": 0.07448019832372665,
|
| 532 |
+
"learning_rate": 4.4186893203883496e-05,
|
| 533 |
+
"loss": 0.0044,
|
| 534 |
+
"step": 1875
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.6225425950196593,
|
| 538 |
+
"grad_norm": 0.010664916597306728,
|
| 539 |
+
"learning_rate": 4.403519417475728e-05,
|
| 540 |
+
"loss": 0.0059,
|
| 541 |
+
"step": 1900
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.6307339449541285,
|
| 545 |
+
"grad_norm": 0.0046129715628921986,
|
| 546 |
+
"learning_rate": 4.388349514563107e-05,
|
| 547 |
+
"loss": 0.0575,
|
| 548 |
+
"step": 1925
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.6389252948885976,
|
| 552 |
+
"grad_norm": 0.0065656425431370735,
|
| 553 |
+
"learning_rate": 4.373179611650486e-05,
|
| 554 |
+
"loss": 0.0051,
|
| 555 |
+
"step": 1950
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.6471166448230669,
|
| 559 |
+
"grad_norm": 0.045100320130586624,
|
| 560 |
+
"learning_rate": 4.358009708737864e-05,
|
| 561 |
+
"loss": 0.0765,
|
| 562 |
+
"step": 1975
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.6553079947575361,
|
| 566 |
+
"grad_norm": 0.010744914412498474,
|
| 567 |
+
"learning_rate": 4.342839805825243e-05,
|
| 568 |
+
"loss": 0.0434,
|
| 569 |
+
"step": 2000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.6634993446920052,
|
| 573 |
+
"grad_norm": 0.006751019041985273,
|
| 574 |
+
"learning_rate": 4.327669902912622e-05,
|
| 575 |
+
"loss": 0.0112,
|
| 576 |
+
"step": 2025
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.6716906946264745,
|
| 580 |
+
"grad_norm": 0.0010609191376715899,
|
| 581 |
+
"learning_rate": 4.3125000000000005e-05,
|
| 582 |
+
"loss": 0.0002,
|
| 583 |
+
"step": 2050
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.6798820445609436,
|
| 587 |
+
"grad_norm": 0.0029373185243457556,
|
| 588 |
+
"learning_rate": 4.297330097087379e-05,
|
| 589 |
+
"loss": 0.0457,
|
| 590 |
+
"step": 2075
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.6880733944954128,
|
| 594 |
+
"grad_norm": 0.03929394483566284,
|
| 595 |
+
"learning_rate": 4.2821601941747575e-05,
|
| 596 |
+
"loss": 0.0003,
|
| 597 |
+
"step": 2100
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.6962647444298821,
|
| 601 |
+
"grad_norm": 0.0013411706313490868,
|
| 602 |
+
"learning_rate": 4.266990291262136e-05,
|
| 603 |
+
"loss": 0.0003,
|
| 604 |
+
"step": 2125
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.7044560943643512,
|
| 608 |
+
"grad_norm": 0.15903156995773315,
|
| 609 |
+
"learning_rate": 4.251820388349515e-05,
|
| 610 |
+
"loss": 0.0713,
|
| 611 |
+
"step": 2150
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.7126474442988204,
|
| 615 |
+
"grad_norm": 0.004102173261344433,
|
| 616 |
+
"learning_rate": 4.236650485436893e-05,
|
| 617 |
+
"loss": 0.0071,
|
| 618 |
+
"step": 2175
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.7208387942332897,
|
| 622 |
+
"grad_norm": 0.0019201630493625998,
|
| 623 |
+
"learning_rate": 4.221480582524272e-05,
|
| 624 |
+
"loss": 0.0012,
|
| 625 |
+
"step": 2200
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.7290301441677588,
|
| 629 |
+
"grad_norm": 0.08992388099431992,
|
| 630 |
+
"learning_rate": 4.2063106796116506e-05,
|
| 631 |
+
"loss": 0.0682,
|
| 632 |
+
"step": 2225
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.737221494102228,
|
| 636 |
+
"grad_norm": 0.06924425065517426,
|
| 637 |
+
"learning_rate": 4.19114077669903e-05,
|
| 638 |
+
"loss": 0.0045,
|
| 639 |
+
"step": 2250
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.7454128440366973,
|
| 643 |
+
"grad_norm": 0.006293242797255516,
|
| 644 |
+
"learning_rate": 4.1759708737864076e-05,
|
| 645 |
+
"loss": 0.0328,
|
| 646 |
+
"step": 2275
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.7536041939711664,
|
| 650 |
+
"grad_norm": 0.0008105287561193109,
|
| 651 |
+
"learning_rate": 4.160800970873787e-05,
|
| 652 |
+
"loss": 0.0481,
|
| 653 |
+
"step": 2300
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.7617955439056356,
|
| 657 |
+
"grad_norm": 0.01918267086148262,
|
| 658 |
+
"learning_rate": 4.145631067961165e-05,
|
| 659 |
+
"loss": 0.0529,
|
| 660 |
+
"step": 2325
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.7699868938401049,
|
| 664 |
+
"grad_norm": 0.02567248046398163,
|
| 665 |
+
"learning_rate": 4.130461165048544e-05,
|
| 666 |
+
"loss": 0.0271,
|
| 667 |
+
"step": 2350
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.778178243774574,
|
| 671 |
+
"grad_norm": 0.00797118991613388,
|
| 672 |
+
"learning_rate": 4.115291262135922e-05,
|
| 673 |
+
"loss": 0.0116,
|
| 674 |
+
"step": 2375
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.7863695937090432,
|
| 678 |
+
"grad_norm": 0.034628257155418396,
|
| 679 |
+
"learning_rate": 4.1001213592233014e-05,
|
| 680 |
+
"loss": 0.0389,
|
| 681 |
+
"step": 2400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.7945609436435125,
|
| 685 |
+
"grad_norm": 0.004856760613620281,
|
| 686 |
+
"learning_rate": 4.08495145631068e-05,
|
| 687 |
+
"loss": 0.0076,
|
| 688 |
+
"step": 2425
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.8027522935779816,
|
| 692 |
+
"grad_norm": 0.0018171003321185708,
|
| 693 |
+
"learning_rate": 4.0697815533980584e-05,
|
| 694 |
+
"loss": 0.0001,
|
| 695 |
+
"step": 2450
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.8109436435124509,
|
| 699 |
+
"grad_norm": 0.002607023809105158,
|
| 700 |
+
"learning_rate": 4.054611650485437e-05,
|
| 701 |
+
"loss": 0.0529,
|
| 702 |
+
"step": 2475
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.8191349934469201,
|
| 706 |
+
"grad_norm": 0.00494106812402606,
|
| 707 |
+
"learning_rate": 4.039441747572816e-05,
|
| 708 |
+
"loss": 0.0025,
|
| 709 |
+
"step": 2500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.8273263433813892,
|
| 713 |
+
"grad_norm": 7.6454033851623535,
|
| 714 |
+
"learning_rate": 4.0242718446601946e-05,
|
| 715 |
+
"loss": 0.0506,
|
| 716 |
+
"step": 2525
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.8355176933158585,
|
| 720 |
+
"grad_norm": 0.012740246020257473,
|
| 721 |
+
"learning_rate": 4.009101941747573e-05,
|
| 722 |
+
"loss": 0.0207,
|
| 723 |
+
"step": 2550
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.8437090432503277,
|
| 727 |
+
"grad_norm": 0.19343392550945282,
|
| 728 |
+
"learning_rate": 3.9939320388349516e-05,
|
| 729 |
+
"loss": 0.0018,
|
| 730 |
+
"step": 2575
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.8519003931847968,
|
| 734 |
+
"grad_norm": 0.0015372405759990215,
|
| 735 |
+
"learning_rate": 3.97876213592233e-05,
|
| 736 |
+
"loss": 0.0475,
|
| 737 |
+
"step": 2600
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.8600917431192661,
|
| 741 |
+
"grad_norm": 0.001495638396590948,
|
| 742 |
+
"learning_rate": 3.963592233009709e-05,
|
| 743 |
+
"loss": 0.0204,
|
| 744 |
+
"step": 2625
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.8682830930537353,
|
| 748 |
+
"grad_norm": 0.024810276925563812,
|
| 749 |
+
"learning_rate": 3.948422330097087e-05,
|
| 750 |
+
"loss": 0.0377,
|
| 751 |
+
"step": 2650
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.8764744429882044,
|
| 755 |
+
"grad_norm": 0.001583193545229733,
|
| 756 |
+
"learning_rate": 3.933252427184466e-05,
|
| 757 |
+
"loss": 0.0392,
|
| 758 |
+
"step": 2675
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.8846657929226737,
|
| 762 |
+
"grad_norm": 0.04295048862695694,
|
| 763 |
+
"learning_rate": 3.918082524271845e-05,
|
| 764 |
+
"loss": 0.0349,
|
| 765 |
+
"step": 2700
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.8928571428571429,
|
| 769 |
+
"grad_norm": 0.006643366534262896,
|
| 770 |
+
"learning_rate": 3.902912621359224e-05,
|
| 771 |
+
"loss": 0.0214,
|
| 772 |
+
"step": 2725
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.901048492791612,
|
| 776 |
+
"grad_norm": 0.03927929326891899,
|
| 777 |
+
"learning_rate": 3.887742718446602e-05,
|
| 778 |
+
"loss": 0.0313,
|
| 779 |
+
"step": 2750
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.9092398427260813,
|
| 783 |
+
"grad_norm": 0.0005574201932176948,
|
| 784 |
+
"learning_rate": 3.872572815533981e-05,
|
| 785 |
+
"loss": 0.0167,
|
| 786 |
+
"step": 2775
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.9174311926605505,
|
| 790 |
+
"grad_norm": 0.0002800333604682237,
|
| 791 |
+
"learning_rate": 3.8574029126213594e-05,
|
| 792 |
+
"loss": 0.0038,
|
| 793 |
+
"step": 2800
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.9256225425950196,
|
| 797 |
+
"grad_norm": 0.24904371798038483,
|
| 798 |
+
"learning_rate": 3.842233009708738e-05,
|
| 799 |
+
"loss": 0.0557,
|
| 800 |
+
"step": 2825
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.9338138925294889,
|
| 804 |
+
"grad_norm": 0.009384050033986568,
|
| 805 |
+
"learning_rate": 3.8270631067961164e-05,
|
| 806 |
+
"loss": 0.0349,
|
| 807 |
+
"step": 2850
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.9420052424639581,
|
| 811 |
+
"grad_norm": 0.0032450663857162,
|
| 812 |
+
"learning_rate": 3.8118932038834956e-05,
|
| 813 |
+
"loss": 0.0061,
|
| 814 |
+
"step": 2875
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.9501965923984272,
|
| 818 |
+
"grad_norm": 0.002643973333761096,
|
| 819 |
+
"learning_rate": 3.796723300970874e-05,
|
| 820 |
+
"loss": 0.0011,
|
| 821 |
+
"step": 2900
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.9583879423328965,
|
| 825 |
+
"grad_norm": 0.005450396798551083,
|
| 826 |
+
"learning_rate": 3.7815533980582526e-05,
|
| 827 |
+
"loss": 0.0472,
|
| 828 |
+
"step": 2925
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.9665792922673656,
|
| 832 |
+
"grad_norm": 0.009381350129842758,
|
| 833 |
+
"learning_rate": 3.766383495145631e-05,
|
| 834 |
+
"loss": 0.0221,
|
| 835 |
+
"step": 2950
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.9747706422018348,
|
| 839 |
+
"grad_norm": 0.002026522997766733,
|
| 840 |
+
"learning_rate": 3.75121359223301e-05,
|
| 841 |
+
"loss": 0.0015,
|
| 842 |
+
"step": 2975
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.9829619921363041,
|
| 846 |
+
"grad_norm": 0.0021547062788158655,
|
| 847 |
+
"learning_rate": 3.736043689320389e-05,
|
| 848 |
+
"loss": 0.0346,
|
| 849 |
+
"step": 3000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.9911533420707732,
|
| 853 |
+
"grad_norm": 0.01624121516942978,
|
| 854 |
+
"learning_rate": 3.720873786407767e-05,
|
| 855 |
+
"loss": 0.0446,
|
| 856 |
+
"step": 3025
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.9993446920052425,
|
| 860 |
+
"grad_norm": 0.0049235946498811245,
|
| 861 |
+
"learning_rate": 3.705703883495146e-05,
|
| 862 |
+
"loss": 0.0198,
|
| 863 |
+
"step": 3050
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 1.0,
|
| 867 |
+
"eval_accuracy": 0.9942660550458715,
|
| 868 |
+
"eval_auc": 0.9860887006409395,
|
| 869 |
+
"eval_f1": 0.7058823529411765,
|
| 870 |
+
"eval_loss": 0.02599955163896084,
|
| 871 |
+
"eval_precision": 0.9333333333333333,
|
| 872 |
+
"eval_recall": 0.5675675675675675,
|
| 873 |
+
"eval_runtime": 71.6069,
|
| 874 |
+
"eval_samples_per_second": 85.243,
|
| 875 |
+
"eval_steps_per_second": 5.335,
|
| 876 |
+
"step": 3052
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 1.0075360419397117,
|
| 880 |
+
"grad_norm": 0.008600810542702675,
|
| 881 |
+
"learning_rate": 3.690533980582524e-05,
|
| 882 |
+
"loss": 0.0216,
|
| 883 |
+
"step": 3075
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.015727391874181,
|
| 887 |
+
"grad_norm": 0.730075478553772,
|
| 888 |
+
"learning_rate": 3.6753640776699034e-05,
|
| 889 |
+
"loss": 0.0736,
|
| 890 |
+
"step": 3100
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 1.0239187418086502,
|
| 894 |
+
"grad_norm": 0.0917787104845047,
|
| 895 |
+
"learning_rate": 3.660194174757282e-05,
|
| 896 |
+
"loss": 0.0145,
|
| 897 |
+
"step": 3125
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 1.0321100917431192,
|
| 901 |
+
"grad_norm": 0.012463403865695,
|
| 902 |
+
"learning_rate": 3.6450242718446604e-05,
|
| 903 |
+
"loss": 0.0127,
|
| 904 |
+
"step": 3150
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 1.0403014416775884,
|
| 908 |
+
"grad_norm": 0.0015968201914802194,
|
| 909 |
+
"learning_rate": 3.629854368932039e-05,
|
| 910 |
+
"loss": 0.0014,
|
| 911 |
+
"step": 3175
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 1.0484927916120577,
|
| 915 |
+
"grad_norm": 0.00465328199788928,
|
| 916 |
+
"learning_rate": 3.614684466019418e-05,
|
| 917 |
+
"loss": 0.0266,
|
| 918 |
+
"step": 3200
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 1.056684141546527,
|
| 922 |
+
"grad_norm": 1.6281688213348389,
|
| 923 |
+
"learning_rate": 3.599514563106796e-05,
|
| 924 |
+
"loss": 0.0324,
|
| 925 |
+
"step": 3225
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.0648754914809961,
|
| 929 |
+
"grad_norm": 0.006340661086142063,
|
| 930 |
+
"learning_rate": 3.584344660194175e-05,
|
| 931 |
+
"loss": 0.0197,
|
| 932 |
+
"step": 3250
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 1.0730668414154652,
|
| 936 |
+
"grad_norm": 0.0019152916502207518,
|
| 937 |
+
"learning_rate": 3.5691747572815535e-05,
|
| 938 |
+
"loss": 0.0056,
|
| 939 |
+
"step": 3275
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 1.0812581913499344,
|
| 943 |
+
"grad_norm": 0.0004100031510461122,
|
| 944 |
+
"learning_rate": 3.554004854368932e-05,
|
| 945 |
+
"loss": 0.0003,
|
| 946 |
+
"step": 3300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 1.0894495412844036,
|
| 950 |
+
"grad_norm": 0.0011120839044451714,
|
| 951 |
+
"learning_rate": 3.5388349514563105e-05,
|
| 952 |
+
"loss": 0.0084,
|
| 953 |
+
"step": 3325
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.0976408912188729,
|
| 957 |
+
"grad_norm": 0.0005429446464404464,
|
| 958 |
+
"learning_rate": 3.52366504854369e-05,
|
| 959 |
+
"loss": 0.0048,
|
| 960 |
+
"step": 3350
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 1.105832241153342,
|
| 964 |
+
"grad_norm": 0.0026772853452712297,
|
| 965 |
+
"learning_rate": 3.508495145631068e-05,
|
| 966 |
+
"loss": 0.0449,
|
| 967 |
+
"step": 3375
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.1140235910878113,
|
| 971 |
+
"grad_norm": 0.09353246539831161,
|
| 972 |
+
"learning_rate": 3.493325242718447e-05,
|
| 973 |
+
"loss": 0.0587,
|
| 974 |
+
"step": 3400
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 1.1222149410222806,
|
| 978 |
+
"grad_norm": 0.010271481238305569,
|
| 979 |
+
"learning_rate": 3.478155339805825e-05,
|
| 980 |
+
"loss": 0.0019,
|
| 981 |
+
"step": 3425
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 1.1304062909567496,
|
| 985 |
+
"grad_norm": 0.008390023373067379,
|
| 986 |
+
"learning_rate": 3.4629854368932044e-05,
|
| 987 |
+
"loss": 0.0244,
|
| 988 |
+
"step": 3450
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 1.1385976408912188,
|
| 992 |
+
"grad_norm": 0.0010115652112290263,
|
| 993 |
+
"learning_rate": 3.447815533980583e-05,
|
| 994 |
+
"loss": 0.0416,
|
| 995 |
+
"step": 3475
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.146788990825688,
|
| 999 |
+
"grad_norm": 0.10442027449607849,
|
| 1000 |
+
"learning_rate": 3.4326456310679614e-05,
|
| 1001 |
+
"loss": 0.0777,
|
| 1002 |
+
"step": 3500
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 1.1549803407601573,
|
| 1006 |
+
"grad_norm": 0.005371388513594866,
|
| 1007 |
+
"learning_rate": 3.41747572815534e-05,
|
| 1008 |
+
"loss": 0.0026,
|
| 1009 |
+
"step": 3525
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.1631716906946266,
|
| 1013 |
+
"grad_norm": 0.00910027977079153,
|
| 1014 |
+
"learning_rate": 3.4023058252427183e-05,
|
| 1015 |
+
"loss": 0.0191,
|
| 1016 |
+
"step": 3550
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 1.1713630406290956,
|
| 1020 |
+
"grad_norm": 0.001924333511851728,
|
| 1021 |
+
"learning_rate": 3.3871359223300975e-05,
|
| 1022 |
+
"loss": 0.0037,
|
| 1023 |
+
"step": 3575
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 1.1795543905635648,
|
| 1027 |
+
"grad_norm": 0.006023438181728125,
|
| 1028 |
+
"learning_rate": 3.371966019417476e-05,
|
| 1029 |
+
"loss": 0.061,
|
| 1030 |
+
"step": 3600
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 1.187745740498034,
|
| 1034 |
+
"grad_norm": 0.07021211832761765,
|
| 1035 |
+
"learning_rate": 3.3567961165048545e-05,
|
| 1036 |
+
"loss": 0.0494,
|
| 1037 |
+
"step": 3625
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.1959370904325033,
|
| 1041 |
+
"grad_norm": 0.007707479875534773,
|
| 1042 |
+
"learning_rate": 3.341626213592233e-05,
|
| 1043 |
+
"loss": 0.009,
|
| 1044 |
+
"step": 3650
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 1.2041284403669725,
|
| 1048 |
+
"grad_norm": 0.005671200342476368,
|
| 1049 |
+
"learning_rate": 3.326456310679612e-05,
|
| 1050 |
+
"loss": 0.0006,
|
| 1051 |
+
"step": 3675
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.2123197903014418,
|
| 1055 |
+
"grad_norm": 0.030211837962269783,
|
| 1056 |
+
"learning_rate": 3.31128640776699e-05,
|
| 1057 |
+
"loss": 0.0139,
|
| 1058 |
+
"step": 3700
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 1.2205111402359108,
|
| 1062 |
+
"grad_norm": 0.4061221778392792,
|
| 1063 |
+
"learning_rate": 3.296116504854369e-05,
|
| 1064 |
+
"loss": 0.004,
|
| 1065 |
+
"step": 3725
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 1.22870249017038,
|
| 1069 |
+
"grad_norm": 0.001372786471620202,
|
| 1070 |
+
"learning_rate": 3.280946601941748e-05,
|
| 1071 |
+
"loss": 0.0237,
|
| 1072 |
+
"step": 3750
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 1.2368938401048493,
|
| 1076 |
+
"grad_norm": 0.0021536455024033785,
|
| 1077 |
+
"learning_rate": 3.265776699029127e-05,
|
| 1078 |
+
"loss": 0.0001,
|
| 1079 |
+
"step": 3775
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.2450851900393185,
|
| 1083 |
+
"grad_norm": 0.00104960345197469,
|
| 1084 |
+
"learning_rate": 3.2506067961165047e-05,
|
| 1085 |
+
"loss": 0.0089,
|
| 1086 |
+
"step": 3800
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 1.2532765399737877,
|
| 1090 |
+
"grad_norm": 0.0027104991022497416,
|
| 1091 |
+
"learning_rate": 3.235436893203884e-05,
|
| 1092 |
+
"loss": 0.0007,
|
| 1093 |
+
"step": 3825
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.261467889908257,
|
| 1097 |
+
"grad_norm": 0.012539150193333626,
|
| 1098 |
+
"learning_rate": 3.220266990291262e-05,
|
| 1099 |
+
"loss": 0.0253,
|
| 1100 |
+
"step": 3850
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 1.269659239842726,
|
| 1104 |
+
"grad_norm": 0.00037872151006013155,
|
| 1105 |
+
"learning_rate": 3.205097087378641e-05,
|
| 1106 |
+
"loss": 0.0003,
|
| 1107 |
+
"step": 3875
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 1.2778505897771952,
|
| 1111 |
+
"grad_norm": 0.00394083047285676,
|
| 1112 |
+
"learning_rate": 3.189927184466019e-05,
|
| 1113 |
+
"loss": 0.0005,
|
| 1114 |
+
"step": 3900
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 1.2860419397116645,
|
| 1118 |
+
"grad_norm": 0.00036229859688319266,
|
| 1119 |
+
"learning_rate": 3.1747572815533985e-05,
|
| 1120 |
+
"loss": 0.0001,
|
| 1121 |
+
"step": 3925
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.2942332896461337,
|
| 1125 |
+
"grad_norm": 0.0016652983613312244,
|
| 1126 |
+
"learning_rate": 3.159587378640777e-05,
|
| 1127 |
+
"loss": 0.013,
|
| 1128 |
+
"step": 3950
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 1.302424639580603,
|
| 1132 |
+
"grad_norm": 0.018371159210801125,
|
| 1133 |
+
"learning_rate": 3.1444174757281555e-05,
|
| 1134 |
+
"loss": 0.0301,
|
| 1135 |
+
"step": 3975
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.310615989515072,
|
| 1139 |
+
"grad_norm": 3.56237530708313,
|
| 1140 |
+
"learning_rate": 3.129247572815534e-05,
|
| 1141 |
+
"loss": 0.011,
|
| 1142 |
+
"step": 4000
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 1.3188073394495412,
|
| 1146 |
+
"grad_norm": 0.0032360991463065147,
|
| 1147 |
+
"learning_rate": 3.114077669902913e-05,
|
| 1148 |
+
"loss": 0.0005,
|
| 1149 |
+
"step": 4025
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 1.3269986893840104,
|
| 1153 |
+
"grad_norm": 0.031877271831035614,
|
| 1154 |
+
"learning_rate": 3.099514563106796e-05,
|
| 1155 |
+
"loss": 0.057,
|
| 1156 |
+
"step": 4050
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 1.3351900393184797,
|
| 1160 |
+
"grad_norm": 0.004732233006507158,
|
| 1161 |
+
"learning_rate": 3.084344660194175e-05,
|
| 1162 |
+
"loss": 0.0015,
|
| 1163 |
+
"step": 4075
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 1.343381389252949,
|
| 1167 |
+
"grad_norm": 0.00040538341272622347,
|
| 1168 |
+
"learning_rate": 3.0691747572815536e-05,
|
| 1169 |
+
"loss": 0.0092,
|
| 1170 |
+
"step": 4100
|
| 1171 |
+
},
|
| 1172 |
+
{
|
| 1173 |
+
"epoch": 1.3515727391874182,
|
| 1174 |
+
"grad_norm": 0.0015137548325583339,
|
| 1175 |
+
"learning_rate": 3.054004854368932e-05,
|
| 1176 |
+
"loss": 0.0033,
|
| 1177 |
+
"step": 4125
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.3597640891218874,
|
| 1181 |
+
"grad_norm": 0.00044548933510668576,
|
| 1182 |
+
"learning_rate": 3.038834951456311e-05,
|
| 1183 |
+
"loss": 0.0221,
|
| 1184 |
+
"step": 4150
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 1.3679554390563564,
|
| 1188 |
+
"grad_norm": 0.007979629561305046,
|
| 1189 |
+
"learning_rate": 3.0236650485436897e-05,
|
| 1190 |
+
"loss": 0.0546,
|
| 1191 |
+
"step": 4175
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 1.3761467889908257,
|
| 1195 |
+
"grad_norm": 0.0093601755797863,
|
| 1196 |
+
"learning_rate": 3.0084951456310682e-05,
|
| 1197 |
+
"loss": 0.0057,
|
| 1198 |
+
"step": 4200
|
| 1199 |
+
},
|
| 1200 |
+
{
|
| 1201 |
+
"epoch": 1.384338138925295,
|
| 1202 |
+
"grad_norm": 0.004190749954432249,
|
| 1203 |
+
"learning_rate": 2.9933252427184467e-05,
|
| 1204 |
+
"loss": 0.023,
|
| 1205 |
+
"step": 4225
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 1.3925294888597641,
|
| 1209 |
+
"grad_norm": 0.0023747687228024006,
|
| 1210 |
+
"learning_rate": 2.9781553398058252e-05,
|
| 1211 |
+
"loss": 0.0104,
|
| 1212 |
+
"step": 4250
|
| 1213 |
+
},
|
| 1214 |
+
{
|
| 1215 |
+
"epoch": 1.4007208387942334,
|
| 1216 |
+
"grad_norm": 0.003429379081353545,
|
| 1217 |
+
"learning_rate": 2.962985436893204e-05,
|
| 1218 |
+
"loss": 0.0403,
|
| 1219 |
+
"step": 4275
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.4089121887287024,
|
| 1223 |
+
"grad_norm": 0.0027962957974523306,
|
| 1224 |
+
"learning_rate": 2.9478155339805826e-05,
|
| 1225 |
+
"loss": 0.0002,
|
| 1226 |
+
"step": 4300
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 1.4171035386631716,
|
| 1230 |
+
"grad_norm": 0.0007880463381297886,
|
| 1231 |
+
"learning_rate": 2.932645631067961e-05,
|
| 1232 |
+
"loss": 0.0002,
|
| 1233 |
+
"step": 4325
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 1.4252948885976409,
|
| 1237 |
+
"grad_norm": 0.0016833733534440398,
|
| 1238 |
+
"learning_rate": 2.91747572815534e-05,
|
| 1239 |
+
"loss": 0.0,
|
| 1240 |
+
"step": 4350
|
| 1241 |
+
},
|
| 1242 |
+
{
|
| 1243 |
+
"epoch": 1.43348623853211,
|
| 1244 |
+
"grad_norm": 0.001454943441785872,
|
| 1245 |
+
"learning_rate": 2.9023058252427187e-05,
|
| 1246 |
+
"loss": 0.0001,
|
| 1247 |
+
"step": 4375
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 1.4416775884665793,
|
| 1251 |
+
"grad_norm": 0.0008237131405621767,
|
| 1252 |
+
"learning_rate": 2.887135922330097e-05,
|
| 1253 |
+
"loss": 0.0541,
|
| 1254 |
+
"step": 4400
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 1.4498689384010484,
|
| 1258 |
+
"grad_norm": 0.004223650321364403,
|
| 1259 |
+
"learning_rate": 2.8719660194174757e-05,
|
| 1260 |
+
"loss": 0.0721,
|
| 1261 |
+
"step": 4425
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.4580602883355178,
|
| 1265 |
+
"grad_norm": 0.0012133247219026089,
|
| 1266 |
+
"learning_rate": 2.8567961165048546e-05,
|
| 1267 |
+
"loss": 0.0007,
|
| 1268 |
+
"step": 4450
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.4662516382699868,
|
| 1272 |
+
"grad_norm": 0.004204601515084505,
|
| 1273 |
+
"learning_rate": 2.8416262135922334e-05,
|
| 1274 |
+
"loss": 0.0001,
|
| 1275 |
+
"step": 4475
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 1.474442988204456,
|
| 1279 |
+
"grad_norm": 0.22309060394763947,
|
| 1280 |
+
"learning_rate": 2.8264563106796115e-05,
|
| 1281 |
+
"loss": 0.0446,
|
| 1282 |
+
"step": 4500
|
| 1283 |
+
},
|
| 1284 |
+
{
|
| 1285 |
+
"epoch": 1.4826343381389253,
|
| 1286 |
+
"grad_norm": 0.0030369218438863754,
|
| 1287 |
+
"learning_rate": 2.8112864077669904e-05,
|
| 1288 |
+
"loss": 0.007,
|
| 1289 |
+
"step": 4525
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 1.4908256880733946,
|
| 1293 |
+
"grad_norm": 0.0010259496048092842,
|
| 1294 |
+
"learning_rate": 2.7961165048543692e-05,
|
| 1295 |
+
"loss": 0.0053,
|
| 1296 |
+
"step": 4550
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"epoch": 1.4990170380078638,
|
| 1300 |
+
"grad_norm": 0.00154329021461308,
|
| 1301 |
+
"learning_rate": 2.780946601941748e-05,
|
| 1302 |
+
"loss": 0.0248,
|
| 1303 |
+
"step": 4575
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.5072083879423328,
|
| 1307 |
+
"grad_norm": 0.0024350325111299753,
|
| 1308 |
+
"learning_rate": 2.7657766990291262e-05,
|
| 1309 |
+
"loss": 0.0108,
|
| 1310 |
+
"step": 4600
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.515399737876802,
|
| 1314 |
+
"grad_norm": 0.003887260565534234,
|
| 1315 |
+
"learning_rate": 2.750606796116505e-05,
|
| 1316 |
+
"loss": 0.0017,
|
| 1317 |
+
"step": 4625
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 1.5235910878112713,
|
| 1321 |
+
"grad_norm": 0.030795983970165253,
|
| 1322 |
+
"learning_rate": 2.735436893203884e-05,
|
| 1323 |
+
"loss": 0.038,
|
| 1324 |
+
"step": 4650
|
| 1325 |
+
},
|
| 1326 |
+
{
|
| 1327 |
+
"epoch": 1.5317824377457405,
|
| 1328 |
+
"grad_norm": 0.0022212048061192036,
|
| 1329 |
+
"learning_rate": 2.7202669902912624e-05,
|
| 1330 |
+
"loss": 0.0016,
|
| 1331 |
+
"step": 4675
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 1.5399737876802098,
|
| 1335 |
+
"grad_norm": 0.0006829659687355161,
|
| 1336 |
+
"learning_rate": 2.705097087378641e-05,
|
| 1337 |
+
"loss": 0.0086,
|
| 1338 |
+
"step": 4700
|
| 1339 |
+
},
|
| 1340 |
+
{
|
| 1341 |
+
"epoch": 1.5481651376146788,
|
| 1342 |
+
"grad_norm": 0.0009496193379163742,
|
| 1343 |
+
"learning_rate": 2.6899271844660197e-05,
|
| 1344 |
+
"loss": 0.0035,
|
| 1345 |
+
"step": 4725
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.5563564875491482,
|
| 1349 |
+
"grad_norm": 0.0016711537027731538,
|
| 1350 |
+
"learning_rate": 2.6747572815533982e-05,
|
| 1351 |
+
"loss": 0.004,
|
| 1352 |
+
"step": 4750
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.5645478374836173,
|
| 1356 |
+
"grad_norm": 0.0007192002376541495,
|
| 1357 |
+
"learning_rate": 2.6595873786407767e-05,
|
| 1358 |
+
"loss": 0.0008,
|
| 1359 |
+
"step": 4775
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 1.5727391874180865,
|
| 1363 |
+
"grad_norm": 0.001931645441800356,
|
| 1364 |
+
"learning_rate": 2.6444174757281552e-05,
|
| 1365 |
+
"loss": 0.0402,
|
| 1366 |
+
"step": 4800
|
| 1367 |
+
},
|
| 1368 |
+
{
|
| 1369 |
+
"epoch": 1.5809305373525557,
|
| 1370 |
+
"grad_norm": 0.0018184804357588291,
|
| 1371 |
+
"learning_rate": 2.629247572815534e-05,
|
| 1372 |
+
"loss": 0.0034,
|
| 1373 |
+
"step": 4825
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 1.5891218872870247,
|
| 1377 |
+
"grad_norm": 0.0007973504834808409,
|
| 1378 |
+
"learning_rate": 2.614077669902913e-05,
|
| 1379 |
+
"loss": 0.0007,
|
| 1380 |
+
"step": 4850
|
| 1381 |
+
},
|
| 1382 |
+
{
|
| 1383 |
+
"epoch": 1.5973132372214942,
|
| 1384 |
+
"grad_norm": 0.0021144933998584747,
|
| 1385 |
+
"learning_rate": 2.598907766990291e-05,
|
| 1386 |
+
"loss": 0.0417,
|
| 1387 |
+
"step": 4875
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.6055045871559632,
|
| 1391 |
+
"grad_norm": 0.0050345249474048615,
|
| 1392 |
+
"learning_rate": 2.58373786407767e-05,
|
| 1393 |
+
"loss": 0.0188,
|
| 1394 |
+
"step": 4900
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.6136959370904325,
|
| 1398 |
+
"grad_norm": 0.0019546381663531065,
|
| 1399 |
+
"learning_rate": 2.5685679611650487e-05,
|
| 1400 |
+
"loss": 0.0026,
|
| 1401 |
+
"step": 4925
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 1.6218872870249017,
|
| 1405 |
+
"grad_norm": 0.0003715524508152157,
|
| 1406 |
+
"learning_rate": 2.5533980582524275e-05,
|
| 1407 |
+
"loss": 0.0119,
|
| 1408 |
+
"step": 4950
|
| 1409 |
+
},
|
| 1410 |
+
{
|
| 1411 |
+
"epoch": 1.630078636959371,
|
| 1412 |
+
"grad_norm": 0.0003383459406904876,
|
| 1413 |
+
"learning_rate": 2.5382281553398057e-05,
|
| 1414 |
+
"loss": 0.0,
|
| 1415 |
+
"step": 4975
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 1.6382699868938402,
|
| 1419 |
+
"grad_norm": 0.0004806627694051713,
|
| 1420 |
+
"learning_rate": 2.5230582524271845e-05,
|
| 1421 |
+
"loss": 0.0009,
|
| 1422 |
+
"step": 5000
|
| 1423 |
+
},
|
| 1424 |
+
{
|
| 1425 |
+
"epoch": 1.6464613368283092,
|
| 1426 |
+
"grad_norm": 0.0002839645021595061,
|
| 1427 |
+
"learning_rate": 2.5078883495145633e-05,
|
| 1428 |
+
"loss": 0.0001,
|
| 1429 |
+
"step": 5025
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.6546526867627787,
|
| 1433 |
+
"grad_norm": 0.00042549462523311377,
|
| 1434 |
+
"learning_rate": 2.492718446601942e-05,
|
| 1435 |
+
"loss": 0.0042,
|
| 1436 |
+
"step": 5050
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.6628440366972477,
|
| 1440 |
+
"grad_norm": 0.0005391003796830773,
|
| 1441 |
+
"learning_rate": 2.4775485436893207e-05,
|
| 1442 |
+
"loss": 0.0566,
|
| 1443 |
+
"step": 5075
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 1.671035386631717,
|
| 1447 |
+
"grad_norm": 0.00017977178504224867,
|
| 1448 |
+
"learning_rate": 2.462378640776699e-05,
|
| 1449 |
+
"loss": 0.0591,
|
| 1450 |
+
"step": 5100
|
| 1451 |
+
},
|
| 1452 |
+
{
|
| 1453 |
+
"epoch": 1.6792267365661862,
|
| 1454 |
+
"grad_norm": 0.7689712047576904,
|
| 1455 |
+
"learning_rate": 2.447208737864078e-05,
|
| 1456 |
+
"loss": 0.0574,
|
| 1457 |
+
"step": 5125
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 1.6874180865006552,
|
| 1461 |
+
"grad_norm": 0.5321087837219238,
|
| 1462 |
+
"learning_rate": 2.4320388349514565e-05,
|
| 1463 |
+
"loss": 0.0536,
|
| 1464 |
+
"step": 5150
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"epoch": 1.6956094364351246,
|
| 1468 |
+
"grad_norm": 0.00359991448931396,
|
| 1469 |
+
"learning_rate": 2.4168689320388353e-05,
|
| 1470 |
+
"loss": 0.0181,
|
| 1471 |
+
"step": 5175
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.7038007863695936,
|
| 1475 |
+
"grad_norm": 0.0036572192329913378,
|
| 1476 |
+
"learning_rate": 2.4016990291262138e-05,
|
| 1477 |
+
"loss": 0.0074,
|
| 1478 |
+
"step": 5200
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.7119921363040629,
|
| 1482 |
+
"grad_norm": 0.0010478459298610687,
|
| 1483 |
+
"learning_rate": 2.3865291262135923e-05,
|
| 1484 |
+
"loss": 0.0071,
|
| 1485 |
+
"step": 5225
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 1.7201834862385321,
|
| 1489 |
+
"grad_norm": 0.001847580773755908,
|
| 1490 |
+
"learning_rate": 2.3713592233009708e-05,
|
| 1491 |
+
"loss": 0.0005,
|
| 1492 |
+
"step": 5250
|
| 1493 |
+
},
|
| 1494 |
+
{
|
| 1495 |
+
"epoch": 1.7283748361730014,
|
| 1496 |
+
"grad_norm": 0.00035062164533883333,
|
| 1497 |
+
"learning_rate": 2.3561893203883497e-05,
|
| 1498 |
+
"loss": 0.0005,
|
| 1499 |
+
"step": 5275
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 1.7365661861074706,
|
| 1503 |
+
"grad_norm": 0.006821767892688513,
|
| 1504 |
+
"learning_rate": 2.341019417475728e-05,
|
| 1505 |
+
"loss": 0.0295,
|
| 1506 |
+
"step": 5300
|
| 1507 |
+
},
|
| 1508 |
+
{
|
| 1509 |
+
"epoch": 1.7447575360419396,
|
| 1510 |
+
"grad_norm": 0.02410910837352276,
|
| 1511 |
+
"learning_rate": 2.3258495145631066e-05,
|
| 1512 |
+
"loss": 0.0313,
|
| 1513 |
+
"step": 5325
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.752948885976409,
|
| 1517 |
+
"grad_norm": 0.0028191821184009314,
|
| 1518 |
+
"learning_rate": 2.3106796116504855e-05,
|
| 1519 |
+
"loss": 0.0264,
|
| 1520 |
+
"step": 5350
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.761140235910878,
|
| 1524 |
+
"grad_norm": 0.39254188537597656,
|
| 1525 |
+
"learning_rate": 2.295509708737864e-05,
|
| 1526 |
+
"loss": 0.0009,
|
| 1527 |
+
"step": 5375
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 1.7693315858453473,
|
| 1531 |
+
"grad_norm": 0.001157455495558679,
|
| 1532 |
+
"learning_rate": 2.2803398058252428e-05,
|
| 1533 |
+
"loss": 0.0353,
|
| 1534 |
+
"step": 5400
|
| 1535 |
+
},
|
| 1536 |
+
{
|
| 1537 |
+
"epoch": 1.7775229357798166,
|
| 1538 |
+
"grad_norm": 0.0038004741072654724,
|
| 1539 |
+
"learning_rate": 2.2651699029126213e-05,
|
| 1540 |
+
"loss": 0.0085,
|
| 1541 |
+
"step": 5425
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 1.7857142857142856,
|
| 1545 |
+
"grad_norm": 0.00042372927418909967,
|
| 1546 |
+
"learning_rate": 2.25e-05,
|
| 1547 |
+
"loss": 0.016,
|
| 1548 |
+
"step": 5450
|
| 1549 |
+
},
|
| 1550 |
+
{
|
| 1551 |
+
"epoch": 1.793905635648755,
|
| 1552 |
+
"grad_norm": 0.0004292895901016891,
|
| 1553 |
+
"learning_rate": 2.2348300970873786e-05,
|
| 1554 |
+
"loss": 0.0072,
|
| 1555 |
+
"step": 5475
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.802096985583224,
|
| 1559 |
+
"grad_norm": 0.0011797266779467463,
|
| 1560 |
+
"learning_rate": 2.2196601941747575e-05,
|
| 1561 |
+
"loss": 0.0012,
|
| 1562 |
+
"step": 5500
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 1.8102883355176933,
|
| 1566 |
+
"grad_norm": 0.0008703326457180083,
|
| 1567 |
+
"learning_rate": 2.204490291262136e-05,
|
| 1568 |
+
"loss": 0.0002,
|
| 1569 |
+
"step": 5525
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 1.8184796854521625,
|
| 1573 |
+
"grad_norm": 0.00140041159465909,
|
| 1574 |
+
"learning_rate": 2.1893203883495148e-05,
|
| 1575 |
+
"loss": 0.0433,
|
| 1576 |
+
"step": 5550
|
| 1577 |
+
},
|
| 1578 |
+
{
|
| 1579 |
+
"epoch": 1.8266710353866316,
|
| 1580 |
+
"grad_norm": 3.4882562160491943,
|
| 1581 |
+
"learning_rate": 2.1741504854368933e-05,
|
| 1582 |
+
"loss": 0.0045,
|
| 1583 |
+
"step": 5575
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.834862385321101,
|
| 1587 |
+
"grad_norm": 0.0005097444518469274,
|
| 1588 |
+
"learning_rate": 2.158980582524272e-05,
|
| 1589 |
+
"loss": 0.0001,
|
| 1590 |
+
"step": 5600
|
| 1591 |
+
},
|
| 1592 |
+
{
|
| 1593 |
+
"epoch": 1.84305373525557,
|
| 1594 |
+
"grad_norm": 0.0011758035980165005,
|
| 1595 |
+
"learning_rate": 2.1438106796116506e-05,
|
| 1596 |
+
"loss": 0.0636,
|
| 1597 |
+
"step": 5625
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.8512450851900393,
|
| 1601 |
+
"grad_norm": 0.00332819577306509,
|
| 1602 |
+
"learning_rate": 2.1286407766990295e-05,
|
| 1603 |
+
"loss": 0.0264,
|
| 1604 |
+
"step": 5650
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 1.8594364351245085,
|
| 1608 |
+
"grad_norm": 0.031183086335659027,
|
| 1609 |
+
"learning_rate": 2.113470873786408e-05,
|
| 1610 |
+
"loss": 0.0134,
|
| 1611 |
+
"step": 5675
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 1.8676277850589778,
|
| 1615 |
+
"grad_norm": 0.0163226630538702,
|
| 1616 |
+
"learning_rate": 2.0983009708737868e-05,
|
| 1617 |
+
"loss": 0.0217,
|
| 1618 |
+
"step": 5700
|
| 1619 |
+
},
|
| 1620 |
+
{
|
| 1621 |
+
"epoch": 1.875819134993447,
|
| 1622 |
+
"grad_norm": 0.01119129080325365,
|
| 1623 |
+
"learning_rate": 2.0831310679611653e-05,
|
| 1624 |
+
"loss": 0.0307,
|
| 1625 |
+
"step": 5725
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.884010484927916,
|
| 1629 |
+
"grad_norm": 0.002649197354912758,
|
| 1630 |
+
"learning_rate": 2.0679611650485438e-05,
|
| 1631 |
+
"loss": 0.0043,
|
| 1632 |
+
"step": 5750
|
| 1633 |
+
},
|
| 1634 |
+
{
|
| 1635 |
+
"epoch": 1.8922018348623855,
|
| 1636 |
+
"grad_norm": 0.03346943110227585,
|
| 1637 |
+
"learning_rate": 2.0527912621359223e-05,
|
| 1638 |
+
"loss": 0.0139,
|
| 1639 |
+
"step": 5775
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.9003931847968545,
|
| 1643 |
+
"grad_norm": 0.00223180977627635,
|
| 1644 |
+
"learning_rate": 2.0376213592233008e-05,
|
| 1645 |
+
"loss": 0.0685,
|
| 1646 |
+
"step": 5800
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 1.9085845347313237,
|
| 1650 |
+
"grad_norm": 0.0009488385985605419,
|
| 1651 |
+
"learning_rate": 2.0224514563106796e-05,
|
| 1652 |
+
"loss": 0.0004,
|
| 1653 |
+
"step": 5825
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 1.916775884665793,
|
| 1657 |
+
"grad_norm": 0.0017646062187850475,
|
| 1658 |
+
"learning_rate": 2.007281553398058e-05,
|
| 1659 |
+
"loss": 0.0002,
|
| 1660 |
+
"step": 5850
|
| 1661 |
+
},
|
| 1662 |
+
{
|
| 1663 |
+
"epoch": 1.924967234600262,
|
| 1664 |
+
"grad_norm": 0.0024130092933773994,
|
| 1665 |
+
"learning_rate": 1.992111650485437e-05,
|
| 1666 |
+
"loss": 0.0003,
|
| 1667 |
+
"step": 5875
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.9331585845347314,
|
| 1671 |
+
"grad_norm": 0.0008109256159514189,
|
| 1672 |
+
"learning_rate": 1.9769417475728154e-05,
|
| 1673 |
+
"loss": 0.0223,
|
| 1674 |
+
"step": 5900
|
| 1675 |
+
},
|
| 1676 |
+
{
|
| 1677 |
+
"epoch": 1.9413499344692005,
|
| 1678 |
+
"grad_norm": 0.0029513093177229166,
|
| 1679 |
+
"learning_rate": 1.9617718446601943e-05,
|
| 1680 |
+
"loss": 0.0009,
|
| 1681 |
+
"step": 5925
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.9495412844036697,
|
| 1685 |
+
"grad_norm": 0.006773350294679403,
|
| 1686 |
+
"learning_rate": 1.9466019417475728e-05,
|
| 1687 |
+
"loss": 0.0058,
|
| 1688 |
+
"step": 5950
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 1.957732634338139,
|
| 1692 |
+
"grad_norm": 0.003487163921818137,
|
| 1693 |
+
"learning_rate": 1.9314320388349516e-05,
|
| 1694 |
+
"loss": 0.0003,
|
| 1695 |
+
"step": 5975
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 1.9659239842726082,
|
| 1699 |
+
"grad_norm": 0.002628578804433346,
|
| 1700 |
+
"learning_rate": 1.91626213592233e-05,
|
| 1701 |
+
"loss": 0.0115,
|
| 1702 |
+
"step": 6000
|
| 1703 |
+
},
|
| 1704 |
+
{
|
| 1705 |
+
"epoch": 1.9741153342070774,
|
| 1706 |
+
"grad_norm": 0.001333425403572619,
|
| 1707 |
+
"learning_rate": 1.901092233009709e-05,
|
| 1708 |
+
"loss": 0.0054,
|
| 1709 |
+
"step": 6025
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.9823066841415464,
|
| 1713 |
+
"grad_norm": 0.0018065802287310362,
|
| 1714 |
+
"learning_rate": 1.8859223300970874e-05,
|
| 1715 |
+
"loss": 0.0467,
|
| 1716 |
+
"step": 6050
|
| 1717 |
+
},
|
| 1718 |
+
{
|
| 1719 |
+
"epoch": 1.9904980340760159,
|
| 1720 |
+
"grad_norm": 0.00119776523206383,
|
| 1721 |
+
"learning_rate": 1.8707524271844663e-05,
|
| 1722 |
+
"loss": 0.0194,
|
| 1723 |
+
"step": 6075
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 1.998689384010485,
|
| 1727 |
+
"grad_norm": 0.049537669867277145,
|
| 1728 |
+
"learning_rate": 1.8555825242718448e-05,
|
| 1729 |
+
"loss": 0.0003,
|
| 1730 |
+
"step": 6100
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 2.0,
|
| 1734 |
+
"eval_accuracy": 0.9985255570117956,
|
| 1735 |
+
"eval_auc": 0.9973051409618573,
|
| 1736 |
+
"eval_f1": 0.9361702127659575,
|
| 1737 |
+
"eval_loss": 0.00864658784121275,
|
| 1738 |
+
"eval_precision": 0.9850746268656716,
|
| 1739 |
+
"eval_recall": 0.8918918918918919,
|
| 1740 |
+
"eval_runtime": 71.46,
|
| 1741 |
+
"eval_samples_per_second": 85.418,
|
| 1742 |
+
"eval_steps_per_second": 5.346,
|
| 1743 |
+
"step": 6104
|
| 1744 |
+
}
|
| 1745 |
+
],
|
| 1746 |
+
"logging_steps": 25,
|
| 1747 |
+
"max_steps": 9156,
|
| 1748 |
+
"num_input_tokens_seen": 0,
|
| 1749 |
+
"num_train_epochs": 3,
|
| 1750 |
+
"save_steps": 500,
|
| 1751 |
+
"stateful_callbacks": {
|
| 1752 |
+
"EarlyStoppingCallback": {
|
| 1753 |
+
"args": {
|
| 1754 |
+
"early_stopping_patience": 5,
|
| 1755 |
+
"early_stopping_threshold": 0.01
|
| 1756 |
+
},
|
| 1757 |
+
"attributes": {
|
| 1758 |
+
"early_stopping_patience_counter": 0
|
| 1759 |
+
}
|
| 1760 |
+
},
|
| 1761 |
+
"TrainerControl": {
|
| 1762 |
+
"args": {
|
| 1763 |
+
"should_epoch_stop": false,
|
| 1764 |
+
"should_evaluate": false,
|
| 1765 |
+
"should_log": false,
|
| 1766 |
+
"should_save": true,
|
| 1767 |
+
"should_training_stop": false
|
| 1768 |
+
},
|
| 1769 |
+
"attributes": {}
|
| 1770 |
+
}
|
| 1771 |
+
},
|
| 1772 |
+
"total_flos": 3.9610164221187195e+19,
|
| 1773 |
+
"train_batch_size": 8,
|
| 1774 |
+
"trial_name": null,
|
| 1775 |
+
"trial_params": null
|
| 1776 |
+
}
|
checkpoint-6104/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7303e9041ce44d1d5e283e6daf8dea2823d4fd26bfbd309cbeaa8b62aa54b2f
|
| 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:afc98a6e4beae9f2797493e9d5814e9b209923a1ef8abf6b1e47b4a600cab79d
|
| 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/Oct21_16-58-53_r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n/events.out.tfevents.1729529935.r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n.184.0
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 sha256:2214db88258f9c64ab8a1e9e86cd386560c0e90041ee6222e20c802fb6206fa3
|
| 3 |
+
size 84077
|
runs/Oct21_16-58-53_r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n/events.out.tfevents.1729533203.r-haiefff-sjstuff-2pgy6o3f-f34a7-rgb1n.184.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82f7a12fbc9bcfa0e6fcbba7830d4ad11f5eb00d4af493002044a23fcb0b1c1d
|
| 3 |
+
size 607
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f7303e9041ce44d1d5e283e6daf8dea2823d4fd26bfbd309cbeaa8b62aa54b2f
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "sj-det-5/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-5",
|
| 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 |
+
}
|