pushed
Browse files- README.md +5 -2
- all_results.json +13 -0
- eval_results.json +8 -0
- runs/Jun27_20-15-24_7e4d4ebc3019/events.out.tfevents.1719520863.7e4d4ebc3019.579.5 +3 -0
- train_results.json +8 -0
- trainer_state.json +3370 -0
README.md
CHANGED
|
@@ -1,5 +1,8 @@
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
tags:
|
|
|
|
| 3 |
- generated_from_trainer
|
| 4 |
datasets:
|
| 5 |
- imagefolder
|
|
@@ -12,7 +15,7 @@ model-index:
|
|
| 12 |
name: Image Classification
|
| 13 |
type: image-classification
|
| 14 |
dataset:
|
| 15 |
-
name:
|
| 16 |
type: imagefolder
|
| 17 |
config: default
|
| 18 |
split: train
|
|
@@ -28,7 +31,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 28 |
|
| 29 |
# vit_epochs5_batch32_lr5e-05_size224_tiles2_seed3_q3_DA
|
| 30 |
|
| 31 |
-
This model
|
| 32 |
It achieves the following results on the evaluation set:
|
| 33 |
- Loss: 0.0265
|
| 34 |
- Accuracy: 0.9915
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: google/vit-base-patch16-224-in21k
|
| 4 |
tags:
|
| 5 |
+
- image-classification
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- imagefolder
|
|
|
|
| 15 |
name: Image Classification
|
| 16 |
type: image-classification
|
| 17 |
dataset:
|
| 18 |
+
name: Dogs_vs_Cats
|
| 19 |
type: imagefolder
|
| 20 |
config: default
|
| 21 |
split: train
|
|
|
|
| 31 |
|
| 32 |
# vit_epochs5_batch32_lr5e-05_size224_tiles2_seed3_q3_DA
|
| 33 |
|
| 34 |
+
This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the Dogs_vs_Cats dataset.
|
| 35 |
It achieves the following results on the evaluation set:
|
| 36 |
- Loss: 0.0265
|
| 37 |
- Accuracy: 0.9915
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9914666666666667,
|
| 4 |
+
"eval_loss": 0.026531297713518143,
|
| 5 |
+
"eval_runtime": 29.543,
|
| 6 |
+
"eval_samples_per_second": 126.934,
|
| 7 |
+
"eval_steps_per_second": 3.994,
|
| 8 |
+
"total_flos": 5.8118992210944e+18,
|
| 9 |
+
"train_loss": 0.08598557333511585,
|
| 10 |
+
"train_runtime": 1480.9706,
|
| 11 |
+
"train_samples_per_second": 50.642,
|
| 12 |
+
"train_steps_per_second": 1.583
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.9914666666666667,
|
| 4 |
+
"eval_loss": 0.026531297713518143,
|
| 5 |
+
"eval_runtime": 29.543,
|
| 6 |
+
"eval_samples_per_second": 126.934,
|
| 7 |
+
"eval_steps_per_second": 3.994
|
| 8 |
+
}
|
runs/Jun27_20-15-24_7e4d4ebc3019/events.out.tfevents.1719520863.7e4d4ebc3019.579.5
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f0cb503d3d06887b93de6243bd34813af3d6dd3e86c7502c1970a6046c9404ff
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 5.8118992210944e+18,
|
| 4 |
+
"train_loss": 0.08598557333511585,
|
| 5 |
+
"train_runtime": 1480.9706,
|
| 6 |
+
"train_samples_per_second": 50.642,
|
| 7 |
+
"train_steps_per_second": 1.583
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3370 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.026531297713518143,
|
| 3 |
+
"best_model_checkpoint": "vit_epochs5_batch32_lr5e-05_size224_tiles2_seed3_q3_DA/checkpoint-2345",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2345,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.010660980810234541,
|
| 13 |
+
"grad_norm": 2.292423963546753,
|
| 14 |
+
"learning_rate": 4.989339019189766e-05,
|
| 15 |
+
"loss": 0.5943,
|
| 16 |
+
"step": 5
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.021321961620469083,
|
| 20 |
+
"grad_norm": 1.9198814630508423,
|
| 21 |
+
"learning_rate": 4.978678038379531e-05,
|
| 22 |
+
"loss": 0.4038,
|
| 23 |
+
"step": 10
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.031982942430703626,
|
| 27 |
+
"grad_norm": 1.3631443977355957,
|
| 28 |
+
"learning_rate": 4.9680170575692967e-05,
|
| 29 |
+
"loss": 0.3014,
|
| 30 |
+
"step": 15
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.042643923240938165,
|
| 34 |
+
"grad_norm": 2.5426456928253174,
|
| 35 |
+
"learning_rate": 4.957356076759062e-05,
|
| 36 |
+
"loss": 0.285,
|
| 37 |
+
"step": 20
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.053304904051172705,
|
| 41 |
+
"grad_norm": 2.9862465858459473,
|
| 42 |
+
"learning_rate": 4.9466950959488276e-05,
|
| 43 |
+
"loss": 0.2197,
|
| 44 |
+
"step": 25
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.06396588486140725,
|
| 48 |
+
"grad_norm": 0.8719685077667236,
|
| 49 |
+
"learning_rate": 4.936034115138593e-05,
|
| 50 |
+
"loss": 0.185,
|
| 51 |
+
"step": 30
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.07462686567164178,
|
| 55 |
+
"grad_norm": 0.5826507210731506,
|
| 56 |
+
"learning_rate": 4.9253731343283586e-05,
|
| 57 |
+
"loss": 0.1511,
|
| 58 |
+
"step": 35
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.08528784648187633,
|
| 62 |
+
"grad_norm": 3.1477346420288086,
|
| 63 |
+
"learning_rate": 4.914712153518124e-05,
|
| 64 |
+
"loss": 0.1121,
|
| 65 |
+
"step": 40
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.09594882729211088,
|
| 69 |
+
"grad_norm": 4.059586524963379,
|
| 70 |
+
"learning_rate": 4.904051172707889e-05,
|
| 71 |
+
"loss": 0.1852,
|
| 72 |
+
"step": 45
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.10660980810234541,
|
| 76 |
+
"grad_norm": 4.97654390335083,
|
| 77 |
+
"learning_rate": 4.893390191897655e-05,
|
| 78 |
+
"loss": 0.1744,
|
| 79 |
+
"step": 50
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.11727078891257996,
|
| 83 |
+
"grad_norm": 3.6855366230010986,
|
| 84 |
+
"learning_rate": 4.88272921108742e-05,
|
| 85 |
+
"loss": 0.1675,
|
| 86 |
+
"step": 55
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1279317697228145,
|
| 90 |
+
"grad_norm": 0.6016143560409546,
|
| 91 |
+
"learning_rate": 4.872068230277186e-05,
|
| 92 |
+
"loss": 0.1559,
|
| 93 |
+
"step": 60
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.13859275053304904,
|
| 97 |
+
"grad_norm": 1.509653925895691,
|
| 98 |
+
"learning_rate": 4.861407249466951e-05,
|
| 99 |
+
"loss": 0.1744,
|
| 100 |
+
"step": 65
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.14925373134328357,
|
| 104 |
+
"grad_norm": 4.235137939453125,
|
| 105 |
+
"learning_rate": 4.850746268656717e-05,
|
| 106 |
+
"loss": 0.1246,
|
| 107 |
+
"step": 70
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.15991471215351813,
|
| 111 |
+
"grad_norm": 4.131643772125244,
|
| 112 |
+
"learning_rate": 4.840085287846482e-05,
|
| 113 |
+
"loss": 0.1664,
|
| 114 |
+
"step": 75
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.17057569296375266,
|
| 118 |
+
"grad_norm": 1.1380912065505981,
|
| 119 |
+
"learning_rate": 4.829424307036248e-05,
|
| 120 |
+
"loss": 0.0795,
|
| 121 |
+
"step": 80
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.1812366737739872,
|
| 125 |
+
"grad_norm": 1.4260450601577759,
|
| 126 |
+
"learning_rate": 4.8187633262260126e-05,
|
| 127 |
+
"loss": 0.0682,
|
| 128 |
+
"step": 85
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.19189765458422176,
|
| 132 |
+
"grad_norm": 0.28632912039756775,
|
| 133 |
+
"learning_rate": 4.808102345415779e-05,
|
| 134 |
+
"loss": 0.1367,
|
| 135 |
+
"step": 90
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.2025586353944563,
|
| 139 |
+
"grad_norm": 2.2717974185943604,
|
| 140 |
+
"learning_rate": 4.7974413646055436e-05,
|
| 141 |
+
"loss": 0.1406,
|
| 142 |
+
"step": 95
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.21321961620469082,
|
| 146 |
+
"grad_norm": 5.882214069366455,
|
| 147 |
+
"learning_rate": 4.78678038379531e-05,
|
| 148 |
+
"loss": 0.0957,
|
| 149 |
+
"step": 100
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.22388059701492538,
|
| 153 |
+
"grad_norm": 1.9833672046661377,
|
| 154 |
+
"learning_rate": 4.7761194029850745e-05,
|
| 155 |
+
"loss": 0.1171,
|
| 156 |
+
"step": 105
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.2345415778251599,
|
| 160 |
+
"grad_norm": 2.357633352279663,
|
| 161 |
+
"learning_rate": 4.765458422174841e-05,
|
| 162 |
+
"loss": 0.1433,
|
| 163 |
+
"step": 110
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.24520255863539445,
|
| 167 |
+
"grad_norm": 2.758540391921997,
|
| 168 |
+
"learning_rate": 4.7547974413646055e-05,
|
| 169 |
+
"loss": 0.1339,
|
| 170 |
+
"step": 115
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.255863539445629,
|
| 174 |
+
"grad_norm": 3.2502870559692383,
|
| 175 |
+
"learning_rate": 4.7441364605543716e-05,
|
| 176 |
+
"loss": 0.102,
|
| 177 |
+
"step": 120
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.26652452025586354,
|
| 181 |
+
"grad_norm": 3.3213181495666504,
|
| 182 |
+
"learning_rate": 4.7334754797441364e-05,
|
| 183 |
+
"loss": 0.1588,
|
| 184 |
+
"step": 125
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.2771855010660981,
|
| 188 |
+
"grad_norm": 2.2992753982543945,
|
| 189 |
+
"learning_rate": 4.7228144989339026e-05,
|
| 190 |
+
"loss": 0.1041,
|
| 191 |
+
"step": 130
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2878464818763326,
|
| 195 |
+
"grad_norm": 1.3673641681671143,
|
| 196 |
+
"learning_rate": 4.7121535181236674e-05,
|
| 197 |
+
"loss": 0.1831,
|
| 198 |
+
"step": 135
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.29850746268656714,
|
| 202 |
+
"grad_norm": 9.642370223999023,
|
| 203 |
+
"learning_rate": 4.7014925373134335e-05,
|
| 204 |
+
"loss": 0.1501,
|
| 205 |
+
"step": 140
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.3091684434968017,
|
| 209 |
+
"grad_norm": 2.5734798908233643,
|
| 210 |
+
"learning_rate": 4.690831556503198e-05,
|
| 211 |
+
"loss": 0.1258,
|
| 212 |
+
"step": 145
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.31982942430703626,
|
| 216 |
+
"grad_norm": 2.1126625537872314,
|
| 217 |
+
"learning_rate": 4.6801705756929645e-05,
|
| 218 |
+
"loss": 0.1356,
|
| 219 |
+
"step": 150
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.3304904051172708,
|
| 223 |
+
"grad_norm": 2.3090267181396484,
|
| 224 |
+
"learning_rate": 4.669509594882729e-05,
|
| 225 |
+
"loss": 0.1099,
|
| 226 |
+
"step": 155
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.3411513859275053,
|
| 230 |
+
"grad_norm": 3.844304084777832,
|
| 231 |
+
"learning_rate": 4.658848614072495e-05,
|
| 232 |
+
"loss": 0.087,
|
| 233 |
+
"step": 160
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.35181236673773986,
|
| 237 |
+
"grad_norm": 1.5647932291030884,
|
| 238 |
+
"learning_rate": 4.64818763326226e-05,
|
| 239 |
+
"loss": 0.1222,
|
| 240 |
+
"step": 165
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.3624733475479744,
|
| 244 |
+
"grad_norm": 1.1747798919677734,
|
| 245 |
+
"learning_rate": 4.637526652452026e-05,
|
| 246 |
+
"loss": 0.1377,
|
| 247 |
+
"step": 170
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.373134328358209,
|
| 251 |
+
"grad_norm": 3.3800017833709717,
|
| 252 |
+
"learning_rate": 4.626865671641791e-05,
|
| 253 |
+
"loss": 0.0709,
|
| 254 |
+
"step": 175
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.3837953091684435,
|
| 258 |
+
"grad_norm": 0.502141535282135,
|
| 259 |
+
"learning_rate": 4.6162046908315566e-05,
|
| 260 |
+
"loss": 0.1023,
|
| 261 |
+
"step": 180
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.39445628997867804,
|
| 265 |
+
"grad_norm": 4.5354485511779785,
|
| 266 |
+
"learning_rate": 4.605543710021322e-05,
|
| 267 |
+
"loss": 0.1717,
|
| 268 |
+
"step": 185
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.4051172707889126,
|
| 272 |
+
"grad_norm": 0.30768585205078125,
|
| 273 |
+
"learning_rate": 4.5948827292110876e-05,
|
| 274 |
+
"loss": 0.1467,
|
| 275 |
+
"step": 190
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.4157782515991471,
|
| 279 |
+
"grad_norm": 0.34890708327293396,
|
| 280 |
+
"learning_rate": 4.584221748400853e-05,
|
| 281 |
+
"loss": 0.0958,
|
| 282 |
+
"step": 195
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.42643923240938164,
|
| 286 |
+
"grad_norm": 3.7051475048065186,
|
| 287 |
+
"learning_rate": 4.5735607675906185e-05,
|
| 288 |
+
"loss": 0.1543,
|
| 289 |
+
"step": 200
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.43710021321961623,
|
| 293 |
+
"grad_norm": 3.108733892440796,
|
| 294 |
+
"learning_rate": 4.562899786780384e-05,
|
| 295 |
+
"loss": 0.0998,
|
| 296 |
+
"step": 205
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.44776119402985076,
|
| 300 |
+
"grad_norm": 3.377547025680542,
|
| 301 |
+
"learning_rate": 4.5522388059701495e-05,
|
| 302 |
+
"loss": 0.0777,
|
| 303 |
+
"step": 210
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.4584221748400853,
|
| 307 |
+
"grad_norm": 1.252957820892334,
|
| 308 |
+
"learning_rate": 4.541577825159915e-05,
|
| 309 |
+
"loss": 0.104,
|
| 310 |
+
"step": 215
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.4690831556503198,
|
| 314 |
+
"grad_norm": 0.5380769371986389,
|
| 315 |
+
"learning_rate": 4.5309168443496804e-05,
|
| 316 |
+
"loss": 0.134,
|
| 317 |
+
"step": 220
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.47974413646055436,
|
| 321 |
+
"grad_norm": 1.7800302505493164,
|
| 322 |
+
"learning_rate": 4.520255863539446e-05,
|
| 323 |
+
"loss": 0.1136,
|
| 324 |
+
"step": 225
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.4904051172707889,
|
| 328 |
+
"grad_norm": 3.849496364593506,
|
| 329 |
+
"learning_rate": 4.5095948827292114e-05,
|
| 330 |
+
"loss": 0.134,
|
| 331 |
+
"step": 230
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.5010660980810234,
|
| 335 |
+
"grad_norm": 2.1841163635253906,
|
| 336 |
+
"learning_rate": 4.498933901918977e-05,
|
| 337 |
+
"loss": 0.0784,
|
| 338 |
+
"step": 235
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.511727078891258,
|
| 342 |
+
"grad_norm": 1.4081329107284546,
|
| 343 |
+
"learning_rate": 4.488272921108742e-05,
|
| 344 |
+
"loss": 0.1378,
|
| 345 |
+
"step": 240
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.5223880597014925,
|
| 349 |
+
"grad_norm": 0.30957549810409546,
|
| 350 |
+
"learning_rate": 4.477611940298508e-05,
|
| 351 |
+
"loss": 0.1117,
|
| 352 |
+
"step": 245
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.5330490405117271,
|
| 356 |
+
"grad_norm": 3.1934280395507812,
|
| 357 |
+
"learning_rate": 4.466950959488273e-05,
|
| 358 |
+
"loss": 0.0788,
|
| 359 |
+
"step": 250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.5437100213219617,
|
| 363 |
+
"grad_norm": 3.456453800201416,
|
| 364 |
+
"learning_rate": 4.456289978678039e-05,
|
| 365 |
+
"loss": 0.1745,
|
| 366 |
+
"step": 255
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.5543710021321961,
|
| 370 |
+
"grad_norm": 2.897150993347168,
|
| 371 |
+
"learning_rate": 4.445628997867804e-05,
|
| 372 |
+
"loss": 0.1105,
|
| 373 |
+
"step": 260
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.5650319829424307,
|
| 377 |
+
"grad_norm": 2.476592540740967,
|
| 378 |
+
"learning_rate": 4.43496801705757e-05,
|
| 379 |
+
"loss": 0.1417,
|
| 380 |
+
"step": 265
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.5756929637526652,
|
| 384 |
+
"grad_norm": 1.5481300354003906,
|
| 385 |
+
"learning_rate": 4.424307036247335e-05,
|
| 386 |
+
"loss": 0.1231,
|
| 387 |
+
"step": 270
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.5863539445628998,
|
| 391 |
+
"grad_norm": 2.645557403564453,
|
| 392 |
+
"learning_rate": 4.4136460554371006e-05,
|
| 393 |
+
"loss": 0.0667,
|
| 394 |
+
"step": 275
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.5970149253731343,
|
| 398 |
+
"grad_norm": 2.6654458045959473,
|
| 399 |
+
"learning_rate": 4.402985074626866e-05,
|
| 400 |
+
"loss": 0.1449,
|
| 401 |
+
"step": 280
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.6076759061833689,
|
| 405 |
+
"grad_norm": 0.4585008919239044,
|
| 406 |
+
"learning_rate": 4.3923240938166316e-05,
|
| 407 |
+
"loss": 0.0742,
|
| 408 |
+
"step": 285
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.6183368869936035,
|
| 412 |
+
"grad_norm": 1.5186400413513184,
|
| 413 |
+
"learning_rate": 4.381663113006397e-05,
|
| 414 |
+
"loss": 0.0709,
|
| 415 |
+
"step": 290
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.6289978678038379,
|
| 419 |
+
"grad_norm": 3.2309410572052,
|
| 420 |
+
"learning_rate": 4.3710021321961625e-05,
|
| 421 |
+
"loss": 0.1686,
|
| 422 |
+
"step": 295
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.6396588486140725,
|
| 426 |
+
"grad_norm": 0.15200960636138916,
|
| 427 |
+
"learning_rate": 4.360341151385928e-05,
|
| 428 |
+
"loss": 0.1037,
|
| 429 |
+
"step": 300
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.650319829424307,
|
| 433 |
+
"grad_norm": 1.4690189361572266,
|
| 434 |
+
"learning_rate": 4.3496801705756935e-05,
|
| 435 |
+
"loss": 0.0663,
|
| 436 |
+
"step": 305
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.6609808102345416,
|
| 440 |
+
"grad_norm": 2.0081136226654053,
|
| 441 |
+
"learning_rate": 4.339019189765459e-05,
|
| 442 |
+
"loss": 0.1326,
|
| 443 |
+
"step": 310
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6716417910447762,
|
| 447 |
+
"grad_norm": 1.1058429479599,
|
| 448 |
+
"learning_rate": 4.328358208955224e-05,
|
| 449 |
+
"loss": 0.0853,
|
| 450 |
+
"step": 315
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.6823027718550106,
|
| 454 |
+
"grad_norm": 5.975476264953613,
|
| 455 |
+
"learning_rate": 4.31769722814499e-05,
|
| 456 |
+
"loss": 0.1648,
|
| 457 |
+
"step": 320
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.6929637526652452,
|
| 461 |
+
"grad_norm": 0.6358932852745056,
|
| 462 |
+
"learning_rate": 4.307036247334755e-05,
|
| 463 |
+
"loss": 0.0939,
|
| 464 |
+
"step": 325
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.7036247334754797,
|
| 468 |
+
"grad_norm": 1.7175092697143555,
|
| 469 |
+
"learning_rate": 4.29637526652452e-05,
|
| 470 |
+
"loss": 0.1278,
|
| 471 |
+
"step": 330
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.7142857142857143,
|
| 475 |
+
"grad_norm": 0.16089807450771332,
|
| 476 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 477 |
+
"loss": 0.0629,
|
| 478 |
+
"step": 335
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.7249466950959488,
|
| 482 |
+
"grad_norm": 2.416827440261841,
|
| 483 |
+
"learning_rate": 4.275053304904051e-05,
|
| 484 |
+
"loss": 0.1739,
|
| 485 |
+
"step": 340
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.7356076759061834,
|
| 489 |
+
"grad_norm": 3.6283304691314697,
|
| 490 |
+
"learning_rate": 4.2643923240938166e-05,
|
| 491 |
+
"loss": 0.178,
|
| 492 |
+
"step": 345
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.746268656716418,
|
| 496 |
+
"grad_norm": 0.9342382550239563,
|
| 497 |
+
"learning_rate": 4.253731343283582e-05,
|
| 498 |
+
"loss": 0.1262,
|
| 499 |
+
"step": 350
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.7569296375266524,
|
| 503 |
+
"grad_norm": 1.0476703643798828,
|
| 504 |
+
"learning_rate": 4.2430703624733475e-05,
|
| 505 |
+
"loss": 0.1325,
|
| 506 |
+
"step": 355
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.767590618336887,
|
| 510 |
+
"grad_norm": 0.7844331860542297,
|
| 511 |
+
"learning_rate": 4.232409381663113e-05,
|
| 512 |
+
"loss": 0.0939,
|
| 513 |
+
"step": 360
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.7782515991471215,
|
| 517 |
+
"grad_norm": 2.6424405574798584,
|
| 518 |
+
"learning_rate": 4.2217484008528785e-05,
|
| 519 |
+
"loss": 0.0978,
|
| 520 |
+
"step": 365
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.7889125799573561,
|
| 524 |
+
"grad_norm": 0.5661392211914062,
|
| 525 |
+
"learning_rate": 4.211087420042644e-05,
|
| 526 |
+
"loss": 0.0939,
|
| 527 |
+
"step": 370
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.7995735607675906,
|
| 531 |
+
"grad_norm": 4.849286079406738,
|
| 532 |
+
"learning_rate": 4.2004264392324094e-05,
|
| 533 |
+
"loss": 0.1184,
|
| 534 |
+
"step": 375
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.8102345415778252,
|
| 538 |
+
"grad_norm": 2.217475652694702,
|
| 539 |
+
"learning_rate": 4.189765458422175e-05,
|
| 540 |
+
"loss": 0.0323,
|
| 541 |
+
"step": 380
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.8208955223880597,
|
| 545 |
+
"grad_norm": 2.111833095550537,
|
| 546 |
+
"learning_rate": 4.1791044776119404e-05,
|
| 547 |
+
"loss": 0.1067,
|
| 548 |
+
"step": 385
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.8315565031982942,
|
| 552 |
+
"grad_norm": 2.516940116882324,
|
| 553 |
+
"learning_rate": 4.168443496801706e-05,
|
| 554 |
+
"loss": 0.0749,
|
| 555 |
+
"step": 390
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.8422174840085288,
|
| 559 |
+
"grad_norm": 3.889086961746216,
|
| 560 |
+
"learning_rate": 4.157782515991471e-05,
|
| 561 |
+
"loss": 0.1237,
|
| 562 |
+
"step": 395
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.8528784648187633,
|
| 566 |
+
"grad_norm": 2.51766300201416,
|
| 567 |
+
"learning_rate": 4.147121535181237e-05,
|
| 568 |
+
"loss": 0.0893,
|
| 569 |
+
"step": 400
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.8635394456289979,
|
| 573 |
+
"grad_norm": 3.7210638523101807,
|
| 574 |
+
"learning_rate": 4.136460554371002e-05,
|
| 575 |
+
"loss": 0.1224,
|
| 576 |
+
"step": 405
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.8742004264392325,
|
| 580 |
+
"grad_norm": 2.199781656265259,
|
| 581 |
+
"learning_rate": 4.125799573560768e-05,
|
| 582 |
+
"loss": 0.0825,
|
| 583 |
+
"step": 410
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.8848614072494669,
|
| 587 |
+
"grad_norm": 2.6633670330047607,
|
| 588 |
+
"learning_rate": 4.115138592750533e-05,
|
| 589 |
+
"loss": 0.0948,
|
| 590 |
+
"step": 415
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.8955223880597015,
|
| 594 |
+
"grad_norm": 0.8969710469245911,
|
| 595 |
+
"learning_rate": 4.104477611940299e-05,
|
| 596 |
+
"loss": 0.0894,
|
| 597 |
+
"step": 420
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.906183368869936,
|
| 601 |
+
"grad_norm": 0.6319212317466736,
|
| 602 |
+
"learning_rate": 4.093816631130064e-05,
|
| 603 |
+
"loss": 0.0833,
|
| 604 |
+
"step": 425
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.9168443496801706,
|
| 608 |
+
"grad_norm": 5.092466354370117,
|
| 609 |
+
"learning_rate": 4.0831556503198296e-05,
|
| 610 |
+
"loss": 0.1485,
|
| 611 |
+
"step": 430
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.9275053304904051,
|
| 615 |
+
"grad_norm": 7.148108959197998,
|
| 616 |
+
"learning_rate": 4.072494669509595e-05,
|
| 617 |
+
"loss": 0.122,
|
| 618 |
+
"step": 435
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.9381663113006397,
|
| 622 |
+
"grad_norm": 3.512585163116455,
|
| 623 |
+
"learning_rate": 4.0618336886993606e-05,
|
| 624 |
+
"loss": 0.1157,
|
| 625 |
+
"step": 440
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.9488272921108742,
|
| 629 |
+
"grad_norm": 3.404421329498291,
|
| 630 |
+
"learning_rate": 4.051172707889126e-05,
|
| 631 |
+
"loss": 0.0901,
|
| 632 |
+
"step": 445
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.9594882729211087,
|
| 636 |
+
"grad_norm": 4.78226375579834,
|
| 637 |
+
"learning_rate": 4.0405117270788915e-05,
|
| 638 |
+
"loss": 0.0734,
|
| 639 |
+
"step": 450
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.9701492537313433,
|
| 643 |
+
"grad_norm": 0.6065996289253235,
|
| 644 |
+
"learning_rate": 4.029850746268657e-05,
|
| 645 |
+
"loss": 0.052,
|
| 646 |
+
"step": 455
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.9808102345415778,
|
| 650 |
+
"grad_norm": 1.4586412906646729,
|
| 651 |
+
"learning_rate": 4.0191897654584225e-05,
|
| 652 |
+
"loss": 0.183,
|
| 653 |
+
"step": 460
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.9914712153518124,
|
| 657 |
+
"grad_norm": 0.49270099401474,
|
| 658 |
+
"learning_rate": 4.008528784648188e-05,
|
| 659 |
+
"loss": 0.0829,
|
| 660 |
+
"step": 465
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 1.0,
|
| 664 |
+
"eval_accuracy": 0.9866666666666667,
|
| 665 |
+
"eval_loss": 0.03893493488430977,
|
| 666 |
+
"eval_runtime": 30.0139,
|
| 667 |
+
"eval_samples_per_second": 124.942,
|
| 668 |
+
"eval_steps_per_second": 3.932,
|
| 669 |
+
"step": 469
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.0021321961620469,
|
| 673 |
+
"grad_norm": 2.595914840698242,
|
| 674 |
+
"learning_rate": 3.997867803837953e-05,
|
| 675 |
+
"loss": 0.1067,
|
| 676 |
+
"step": 470
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 1.0127931769722816,
|
| 680 |
+
"grad_norm": 1.5301272869110107,
|
| 681 |
+
"learning_rate": 3.987206823027719e-05,
|
| 682 |
+
"loss": 0.1166,
|
| 683 |
+
"step": 475
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 1.023454157782516,
|
| 687 |
+
"grad_norm": 1.9160813093185425,
|
| 688 |
+
"learning_rate": 3.976545842217484e-05,
|
| 689 |
+
"loss": 0.1201,
|
| 690 |
+
"step": 480
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 1.0341151385927505,
|
| 694 |
+
"grad_norm": 2.271329402923584,
|
| 695 |
+
"learning_rate": 3.96588486140725e-05,
|
| 696 |
+
"loss": 0.0435,
|
| 697 |
+
"step": 485
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.044776119402985,
|
| 701 |
+
"grad_norm": 1.339680552482605,
|
| 702 |
+
"learning_rate": 3.9552238805970146e-05,
|
| 703 |
+
"loss": 0.1166,
|
| 704 |
+
"step": 490
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.0554371002132197,
|
| 708 |
+
"grad_norm": 3.727053642272949,
|
| 709 |
+
"learning_rate": 3.944562899786781e-05,
|
| 710 |
+
"loss": 0.0871,
|
| 711 |
+
"step": 495
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 1.0660980810234542,
|
| 715 |
+
"grad_norm": 2.686041831970215,
|
| 716 |
+
"learning_rate": 3.9339019189765456e-05,
|
| 717 |
+
"loss": 0.152,
|
| 718 |
+
"step": 500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.0767590618336886,
|
| 722 |
+
"grad_norm": 3.3375051021575928,
|
| 723 |
+
"learning_rate": 3.923240938166312e-05,
|
| 724 |
+
"loss": 0.1189,
|
| 725 |
+
"step": 505
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.0874200426439233,
|
| 729 |
+
"grad_norm": 6.611885070800781,
|
| 730 |
+
"learning_rate": 3.9125799573560765e-05,
|
| 731 |
+
"loss": 0.1387,
|
| 732 |
+
"step": 510
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.0980810234541578,
|
| 736 |
+
"grad_norm": 3.830481767654419,
|
| 737 |
+
"learning_rate": 3.901918976545843e-05,
|
| 738 |
+
"loss": 0.0859,
|
| 739 |
+
"step": 515
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.1087420042643923,
|
| 743 |
+
"grad_norm": 3.5023386478424072,
|
| 744 |
+
"learning_rate": 3.8912579957356075e-05,
|
| 745 |
+
"loss": 0.1761,
|
| 746 |
+
"step": 520
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.1194029850746268,
|
| 750 |
+
"grad_norm": 2.1248180866241455,
|
| 751 |
+
"learning_rate": 3.8805970149253736e-05,
|
| 752 |
+
"loss": 0.1389,
|
| 753 |
+
"step": 525
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 1.1300639658848615,
|
| 757 |
+
"grad_norm": 2.389033555984497,
|
| 758 |
+
"learning_rate": 3.8699360341151384e-05,
|
| 759 |
+
"loss": 0.0555,
|
| 760 |
+
"step": 530
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.140724946695096,
|
| 764 |
+
"grad_norm": 1.1960824728012085,
|
| 765 |
+
"learning_rate": 3.8592750533049046e-05,
|
| 766 |
+
"loss": 0.0855,
|
| 767 |
+
"step": 535
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 1.1513859275053304,
|
| 771 |
+
"grad_norm": 2.4144887924194336,
|
| 772 |
+
"learning_rate": 3.8486140724946694e-05,
|
| 773 |
+
"loss": 0.0832,
|
| 774 |
+
"step": 540
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 1.1620469083155651,
|
| 778 |
+
"grad_norm": 1.4541116952896118,
|
| 779 |
+
"learning_rate": 3.8379530916844355e-05,
|
| 780 |
+
"loss": 0.0389,
|
| 781 |
+
"step": 545
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.1727078891257996,
|
| 785 |
+
"grad_norm": 1.1857450008392334,
|
| 786 |
+
"learning_rate": 3.8272921108742e-05,
|
| 787 |
+
"loss": 0.0493,
|
| 788 |
+
"step": 550
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.183368869936034,
|
| 792 |
+
"grad_norm": 4.700743198394775,
|
| 793 |
+
"learning_rate": 3.8166311300639665e-05,
|
| 794 |
+
"loss": 0.0582,
|
| 795 |
+
"step": 555
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 1.1940298507462686,
|
| 799 |
+
"grad_norm": 1.876307487487793,
|
| 800 |
+
"learning_rate": 3.805970149253731e-05,
|
| 801 |
+
"loss": 0.0468,
|
| 802 |
+
"step": 560
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.2046908315565032,
|
| 806 |
+
"grad_norm": 2.093970775604248,
|
| 807 |
+
"learning_rate": 3.7953091684434974e-05,
|
| 808 |
+
"loss": 0.11,
|
| 809 |
+
"step": 565
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 1.2153518123667377,
|
| 813 |
+
"grad_norm": 1.2070826292037964,
|
| 814 |
+
"learning_rate": 3.784648187633262e-05,
|
| 815 |
+
"loss": 0.098,
|
| 816 |
+
"step": 570
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 1.2260127931769722,
|
| 820 |
+
"grad_norm": 2.544556140899658,
|
| 821 |
+
"learning_rate": 3.7739872068230284e-05,
|
| 822 |
+
"loss": 0.0926,
|
| 823 |
+
"step": 575
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.236673773987207,
|
| 827 |
+
"grad_norm": 3.9542160034179688,
|
| 828 |
+
"learning_rate": 3.763326226012793e-05,
|
| 829 |
+
"loss": 0.1666,
|
| 830 |
+
"step": 580
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.2473347547974414,
|
| 834 |
+
"grad_norm": 0.10062284022569656,
|
| 835 |
+
"learning_rate": 3.752665245202559e-05,
|
| 836 |
+
"loss": 0.0519,
|
| 837 |
+
"step": 585
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 1.2579957356076759,
|
| 841 |
+
"grad_norm": 3.156245708465576,
|
| 842 |
+
"learning_rate": 3.742004264392324e-05,
|
| 843 |
+
"loss": 0.1144,
|
| 844 |
+
"step": 590
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.2686567164179103,
|
| 848 |
+
"grad_norm": 2.764373779296875,
|
| 849 |
+
"learning_rate": 3.73134328358209e-05,
|
| 850 |
+
"loss": 0.0755,
|
| 851 |
+
"step": 595
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 1.279317697228145,
|
| 855 |
+
"grad_norm": 3.7902517318725586,
|
| 856 |
+
"learning_rate": 3.720682302771855e-05,
|
| 857 |
+
"loss": 0.1456,
|
| 858 |
+
"step": 600
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.2899786780383795,
|
| 862 |
+
"grad_norm": 3.4395623207092285,
|
| 863 |
+
"learning_rate": 3.710021321961621e-05,
|
| 864 |
+
"loss": 0.1562,
|
| 865 |
+
"step": 605
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.3006396588486142,
|
| 869 |
+
"grad_norm": 3.2427480220794678,
|
| 870 |
+
"learning_rate": 3.699360341151386e-05,
|
| 871 |
+
"loss": 0.0809,
|
| 872 |
+
"step": 610
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.3113006396588487,
|
| 876 |
+
"grad_norm": 3.8836746215820312,
|
| 877 |
+
"learning_rate": 3.6886993603411515e-05,
|
| 878 |
+
"loss": 0.1228,
|
| 879 |
+
"step": 615
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.3219616204690832,
|
| 883 |
+
"grad_norm": 0.8409526348114014,
|
| 884 |
+
"learning_rate": 3.678038379530917e-05,
|
| 885 |
+
"loss": 0.1178,
|
| 886 |
+
"step": 620
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.3326226012793176,
|
| 890 |
+
"grad_norm": 2.4948830604553223,
|
| 891 |
+
"learning_rate": 3.6673773987206824e-05,
|
| 892 |
+
"loss": 0.1166,
|
| 893 |
+
"step": 625
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.3432835820895521,
|
| 897 |
+
"grad_norm": 1.1228270530700684,
|
| 898 |
+
"learning_rate": 3.656716417910448e-05,
|
| 899 |
+
"loss": 0.0682,
|
| 900 |
+
"step": 630
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 1.3539445628997868,
|
| 904 |
+
"grad_norm": 0.7787748575210571,
|
| 905 |
+
"learning_rate": 3.6460554371002134e-05,
|
| 906 |
+
"loss": 0.1478,
|
| 907 |
+
"step": 635
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.3646055437100213,
|
| 911 |
+
"grad_norm": 0.7375028729438782,
|
| 912 |
+
"learning_rate": 3.635394456289979e-05,
|
| 913 |
+
"loss": 0.0709,
|
| 914 |
+
"step": 640
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.375266524520256,
|
| 918 |
+
"grad_norm": 1.9496781826019287,
|
| 919 |
+
"learning_rate": 3.624733475479744e-05,
|
| 920 |
+
"loss": 0.1115,
|
| 921 |
+
"step": 645
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 1.3859275053304905,
|
| 925 |
+
"grad_norm": 3.955174446105957,
|
| 926 |
+
"learning_rate": 3.61407249466951e-05,
|
| 927 |
+
"loss": 0.0931,
|
| 928 |
+
"step": 650
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.396588486140725,
|
| 932 |
+
"grad_norm": 1.9258500337600708,
|
| 933 |
+
"learning_rate": 3.603411513859275e-05,
|
| 934 |
+
"loss": 0.0732,
|
| 935 |
+
"step": 655
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.4072494669509594,
|
| 939 |
+
"grad_norm": 0.5763303637504578,
|
| 940 |
+
"learning_rate": 3.592750533049041e-05,
|
| 941 |
+
"loss": 0.0706,
|
| 942 |
+
"step": 660
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 1.417910447761194,
|
| 946 |
+
"grad_norm": 0.6870846152305603,
|
| 947 |
+
"learning_rate": 3.582089552238806e-05,
|
| 948 |
+
"loss": 0.1024,
|
| 949 |
+
"step": 665
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.4285714285714286,
|
| 953 |
+
"grad_norm": 1.7867426872253418,
|
| 954 |
+
"learning_rate": 3.571428571428572e-05,
|
| 955 |
+
"loss": 0.0884,
|
| 956 |
+
"step": 670
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.439232409381663,
|
| 960 |
+
"grad_norm": 4.831230163574219,
|
| 961 |
+
"learning_rate": 3.560767590618337e-05,
|
| 962 |
+
"loss": 0.1242,
|
| 963 |
+
"step": 675
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 1.4498933901918978,
|
| 967 |
+
"grad_norm": 0.47306933999061584,
|
| 968 |
+
"learning_rate": 3.5501066098081026e-05,
|
| 969 |
+
"loss": 0.0791,
|
| 970 |
+
"step": 680
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.4605543710021323,
|
| 974 |
+
"grad_norm": 0.7883843183517456,
|
| 975 |
+
"learning_rate": 3.539445628997868e-05,
|
| 976 |
+
"loss": 0.0745,
|
| 977 |
+
"step": 685
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.4712153518123667,
|
| 981 |
+
"grad_norm": 4.445797920227051,
|
| 982 |
+
"learning_rate": 3.5287846481876336e-05,
|
| 983 |
+
"loss": 0.1074,
|
| 984 |
+
"step": 690
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.4818763326226012,
|
| 988 |
+
"grad_norm": 4.065772533416748,
|
| 989 |
+
"learning_rate": 3.518123667377399e-05,
|
| 990 |
+
"loss": 0.2127,
|
| 991 |
+
"step": 695
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.4925373134328357,
|
| 995 |
+
"grad_norm": 6.337629318237305,
|
| 996 |
+
"learning_rate": 3.5074626865671645e-05,
|
| 997 |
+
"loss": 0.1264,
|
| 998 |
+
"step": 700
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.5031982942430704,
|
| 1002 |
+
"grad_norm": 0.4290447533130646,
|
| 1003 |
+
"learning_rate": 3.496801705756929e-05,
|
| 1004 |
+
"loss": 0.1447,
|
| 1005 |
+
"step": 705
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 1.5138592750533049,
|
| 1009 |
+
"grad_norm": 0.4908345341682434,
|
| 1010 |
+
"learning_rate": 3.4861407249466955e-05,
|
| 1011 |
+
"loss": 0.096,
|
| 1012 |
+
"step": 710
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 1.5245202558635396,
|
| 1016 |
+
"grad_norm": 1.963738203048706,
|
| 1017 |
+
"learning_rate": 3.47547974413646e-05,
|
| 1018 |
+
"loss": 0.0641,
|
| 1019 |
+
"step": 715
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.535181236673774,
|
| 1023 |
+
"grad_norm": 2.048469305038452,
|
| 1024 |
+
"learning_rate": 3.4648187633262264e-05,
|
| 1025 |
+
"loss": 0.0773,
|
| 1026 |
+
"step": 720
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 1.5458422174840085,
|
| 1030 |
+
"grad_norm": 1.6802610158920288,
|
| 1031 |
+
"learning_rate": 3.454157782515991e-05,
|
| 1032 |
+
"loss": 0.0855,
|
| 1033 |
+
"step": 725
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.556503198294243,
|
| 1037 |
+
"grad_norm": 1.0892540216445923,
|
| 1038 |
+
"learning_rate": 3.4434968017057574e-05,
|
| 1039 |
+
"loss": 0.0925,
|
| 1040 |
+
"step": 730
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 1.5671641791044775,
|
| 1044 |
+
"grad_norm": 2.5715396404266357,
|
| 1045 |
+
"learning_rate": 3.432835820895522e-05,
|
| 1046 |
+
"loss": 0.0798,
|
| 1047 |
+
"step": 735
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 1.5778251599147122,
|
| 1051 |
+
"grad_norm": 1.4111549854278564,
|
| 1052 |
+
"learning_rate": 3.422174840085288e-05,
|
| 1053 |
+
"loss": 0.0664,
|
| 1054 |
+
"step": 740
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 1.5884861407249466,
|
| 1058 |
+
"grad_norm": 2.28279709815979,
|
| 1059 |
+
"learning_rate": 3.411513859275053e-05,
|
| 1060 |
+
"loss": 0.0753,
|
| 1061 |
+
"step": 745
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.5991471215351813,
|
| 1065 |
+
"grad_norm": 1.1696134805679321,
|
| 1066 |
+
"learning_rate": 3.400852878464819e-05,
|
| 1067 |
+
"loss": 0.1033,
|
| 1068 |
+
"step": 750
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 1.6098081023454158,
|
| 1072 |
+
"grad_norm": 3.135894298553467,
|
| 1073 |
+
"learning_rate": 3.390191897654584e-05,
|
| 1074 |
+
"loss": 0.1009,
|
| 1075 |
+
"step": 755
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.6204690831556503,
|
| 1079 |
+
"grad_norm": 4.619607925415039,
|
| 1080 |
+
"learning_rate": 3.37953091684435e-05,
|
| 1081 |
+
"loss": 0.0887,
|
| 1082 |
+
"step": 760
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 1.6311300639658848,
|
| 1086 |
+
"grad_norm": 5.1181745529174805,
|
| 1087 |
+
"learning_rate": 3.368869936034115e-05,
|
| 1088 |
+
"loss": 0.1032,
|
| 1089 |
+
"step": 765
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.6417910447761193,
|
| 1093 |
+
"grad_norm": 1.4793082475662231,
|
| 1094 |
+
"learning_rate": 3.358208955223881e-05,
|
| 1095 |
+
"loss": 0.0584,
|
| 1096 |
+
"step": 770
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 1.652452025586354,
|
| 1100 |
+
"grad_norm": 2.8294785022735596,
|
| 1101 |
+
"learning_rate": 3.347547974413646e-05,
|
| 1102 |
+
"loss": 0.069,
|
| 1103 |
+
"step": 775
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 1.6631130063965884,
|
| 1107 |
+
"grad_norm": 0.8293038606643677,
|
| 1108 |
+
"learning_rate": 3.336886993603412e-05,
|
| 1109 |
+
"loss": 0.1027,
|
| 1110 |
+
"step": 780
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 1.6737739872068231,
|
| 1114 |
+
"grad_norm": 2.485779285430908,
|
| 1115 |
+
"learning_rate": 3.326226012793177e-05,
|
| 1116 |
+
"loss": 0.0755,
|
| 1117 |
+
"step": 785
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.6844349680170576,
|
| 1121 |
+
"grad_norm": 1.9261149168014526,
|
| 1122 |
+
"learning_rate": 3.3155650319829424e-05,
|
| 1123 |
+
"loss": 0.0588,
|
| 1124 |
+
"step": 790
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.695095948827292,
|
| 1128 |
+
"grad_norm": 2.853330135345459,
|
| 1129 |
+
"learning_rate": 3.304904051172708e-05,
|
| 1130 |
+
"loss": 0.0611,
|
| 1131 |
+
"step": 795
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 1.7057569296375266,
|
| 1135 |
+
"grad_norm": 4.901320457458496,
|
| 1136 |
+
"learning_rate": 3.294243070362473e-05,
|
| 1137 |
+
"loss": 0.0785,
|
| 1138 |
+
"step": 800
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 1.716417910447761,
|
| 1142 |
+
"grad_norm": 2.9853076934814453,
|
| 1143 |
+
"learning_rate": 3.283582089552239e-05,
|
| 1144 |
+
"loss": 0.0872,
|
| 1145 |
+
"step": 805
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 1.7270788912579957,
|
| 1149 |
+
"grad_norm": 3.004443645477295,
|
| 1150 |
+
"learning_rate": 3.272921108742004e-05,
|
| 1151 |
+
"loss": 0.0767,
|
| 1152 |
+
"step": 810
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 1.7377398720682304,
|
| 1156 |
+
"grad_norm": 3.083705186843872,
|
| 1157 |
+
"learning_rate": 3.26226012793177e-05,
|
| 1158 |
+
"loss": 0.0663,
|
| 1159 |
+
"step": 815
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.748400852878465,
|
| 1163 |
+
"grad_norm": 2.67826771736145,
|
| 1164 |
+
"learning_rate": 3.251599147121535e-05,
|
| 1165 |
+
"loss": 0.1163,
|
| 1166 |
+
"step": 820
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 1.7590618336886994,
|
| 1170 |
+
"grad_norm": 4.304074287414551,
|
| 1171 |
+
"learning_rate": 3.240938166311301e-05,
|
| 1172 |
+
"loss": 0.121,
|
| 1173 |
+
"step": 825
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 1.7697228144989339,
|
| 1177 |
+
"grad_norm": 2.474461078643799,
|
| 1178 |
+
"learning_rate": 3.230277185501066e-05,
|
| 1179 |
+
"loss": 0.1038,
|
| 1180 |
+
"step": 830
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 1.7803837953091683,
|
| 1184 |
+
"grad_norm": 0.7145506143569946,
|
| 1185 |
+
"learning_rate": 3.2196162046908317e-05,
|
| 1186 |
+
"loss": 0.0778,
|
| 1187 |
+
"step": 835
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 1.7910447761194028,
|
| 1191 |
+
"grad_norm": 0.8223341703414917,
|
| 1192 |
+
"learning_rate": 3.208955223880597e-05,
|
| 1193 |
+
"loss": 0.0697,
|
| 1194 |
+
"step": 840
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 1.8017057569296375,
|
| 1198 |
+
"grad_norm": 0.9580242037773132,
|
| 1199 |
+
"learning_rate": 3.1982942430703626e-05,
|
| 1200 |
+
"loss": 0.0935,
|
| 1201 |
+
"step": 845
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.8123667377398722,
|
| 1205 |
+
"grad_norm": 0.1878369152545929,
|
| 1206 |
+
"learning_rate": 3.187633262260128e-05,
|
| 1207 |
+
"loss": 0.0885,
|
| 1208 |
+
"step": 850
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 1.8230277185501067,
|
| 1212 |
+
"grad_norm": 1.7058665752410889,
|
| 1213 |
+
"learning_rate": 3.1769722814498935e-05,
|
| 1214 |
+
"loss": 0.0912,
|
| 1215 |
+
"step": 855
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.8336886993603412,
|
| 1219 |
+
"grad_norm": 2.2890548706054688,
|
| 1220 |
+
"learning_rate": 3.166311300639659e-05,
|
| 1221 |
+
"loss": 0.0609,
|
| 1222 |
+
"step": 860
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.8443496801705757,
|
| 1226 |
+
"grad_norm": 0.17230653762817383,
|
| 1227 |
+
"learning_rate": 3.1556503198294245e-05,
|
| 1228 |
+
"loss": 0.1325,
|
| 1229 |
+
"step": 865
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 1.8550106609808101,
|
| 1233 |
+
"grad_norm": 4.040843963623047,
|
| 1234 |
+
"learning_rate": 3.14498933901919e-05,
|
| 1235 |
+
"loss": 0.0733,
|
| 1236 |
+
"step": 870
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 1.8656716417910446,
|
| 1240 |
+
"grad_norm": 2.0240883827209473,
|
| 1241 |
+
"learning_rate": 3.1343283582089554e-05,
|
| 1242 |
+
"loss": 0.1477,
|
| 1243 |
+
"step": 875
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.8763326226012793,
|
| 1247 |
+
"grad_norm": 1.042112946510315,
|
| 1248 |
+
"learning_rate": 3.123667377398721e-05,
|
| 1249 |
+
"loss": 0.0524,
|
| 1250 |
+
"step": 880
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 1.886993603411514,
|
| 1254 |
+
"grad_norm": 0.9438421130180359,
|
| 1255 |
+
"learning_rate": 3.1130063965884864e-05,
|
| 1256 |
+
"loss": 0.1312,
|
| 1257 |
+
"step": 885
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.8976545842217485,
|
| 1261 |
+
"grad_norm": 2.3604328632354736,
|
| 1262 |
+
"learning_rate": 3.102345415778252e-05,
|
| 1263 |
+
"loss": 0.0628,
|
| 1264 |
+
"step": 890
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 1.908315565031983,
|
| 1268 |
+
"grad_norm": 3.372856616973877,
|
| 1269 |
+
"learning_rate": 3.0916844349680173e-05,
|
| 1270 |
+
"loss": 0.0499,
|
| 1271 |
+
"step": 895
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.9189765458422174,
|
| 1275 |
+
"grad_norm": 2.872025728225708,
|
| 1276 |
+
"learning_rate": 3.081023454157783e-05,
|
| 1277 |
+
"loss": 0.0898,
|
| 1278 |
+
"step": 900
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 1.929637526652452,
|
| 1282 |
+
"grad_norm": 3.2129461765289307,
|
| 1283 |
+
"learning_rate": 3.070362473347548e-05,
|
| 1284 |
+
"loss": 0.1591,
|
| 1285 |
+
"step": 905
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.9402985074626866,
|
| 1289 |
+
"grad_norm": 1.3610501289367676,
|
| 1290 |
+
"learning_rate": 3.059701492537314e-05,
|
| 1291 |
+
"loss": 0.1018,
|
| 1292 |
+
"step": 910
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 1.950959488272921,
|
| 1296 |
+
"grad_norm": 0.7978216409683228,
|
| 1297 |
+
"learning_rate": 3.0490405117270792e-05,
|
| 1298 |
+
"loss": 0.0894,
|
| 1299 |
+
"step": 915
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.9616204690831558,
|
| 1303 |
+
"grad_norm": 4.605905532836914,
|
| 1304 |
+
"learning_rate": 3.0383795309168444e-05,
|
| 1305 |
+
"loss": 0.091,
|
| 1306 |
+
"step": 920
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 1.9722814498933903,
|
| 1310 |
+
"grad_norm": 1.034550666809082,
|
| 1311 |
+
"learning_rate": 3.0277185501066102e-05,
|
| 1312 |
+
"loss": 0.0735,
|
| 1313 |
+
"step": 925
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.9829424307036247,
|
| 1317 |
+
"grad_norm": 2.8517279624938965,
|
| 1318 |
+
"learning_rate": 3.0170575692963753e-05,
|
| 1319 |
+
"loss": 0.0916,
|
| 1320 |
+
"step": 930
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 1.9936034115138592,
|
| 1324 |
+
"grad_norm": 1.5132031440734863,
|
| 1325 |
+
"learning_rate": 3.006396588486141e-05,
|
| 1326 |
+
"loss": 0.1276,
|
| 1327 |
+
"step": 935
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 2.0,
|
| 1331 |
+
"eval_accuracy": 0.9906666666666667,
|
| 1332 |
+
"eval_loss": 0.027700183913111687,
|
| 1333 |
+
"eval_runtime": 30.1194,
|
| 1334 |
+
"eval_samples_per_second": 124.504,
|
| 1335 |
+
"eval_steps_per_second": 3.918,
|
| 1336 |
+
"step": 938
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.0042643923240937,
|
| 1340 |
+
"grad_norm": 2.149052143096924,
|
| 1341 |
+
"learning_rate": 2.9957356076759063e-05,
|
| 1342 |
+
"loss": 0.076,
|
| 1343 |
+
"step": 940
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.014925373134328,
|
| 1347 |
+
"grad_norm": 2.3346307277679443,
|
| 1348 |
+
"learning_rate": 2.9850746268656714e-05,
|
| 1349 |
+
"loss": 0.0897,
|
| 1350 |
+
"step": 945
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.025586353944563,
|
| 1354 |
+
"grad_norm": 2.1446378231048584,
|
| 1355 |
+
"learning_rate": 2.9744136460554372e-05,
|
| 1356 |
+
"loss": 0.0669,
|
| 1357 |
+
"step": 950
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.0362473347547976,
|
| 1361 |
+
"grad_norm": 5.740427494049072,
|
| 1362 |
+
"learning_rate": 2.9637526652452023e-05,
|
| 1363 |
+
"loss": 0.1276,
|
| 1364 |
+
"step": 955
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.046908315565032,
|
| 1368 |
+
"grad_norm": 4.793395519256592,
|
| 1369 |
+
"learning_rate": 2.953091684434968e-05,
|
| 1370 |
+
"loss": 0.0694,
|
| 1371 |
+
"step": 960
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.0575692963752665,
|
| 1375 |
+
"grad_norm": 1.091191291809082,
|
| 1376 |
+
"learning_rate": 2.9424307036247333e-05,
|
| 1377 |
+
"loss": 0.0329,
|
| 1378 |
+
"step": 965
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.068230277185501,
|
| 1382 |
+
"grad_norm": 2.0005042552948,
|
| 1383 |
+
"learning_rate": 2.931769722814499e-05,
|
| 1384 |
+
"loss": 0.0466,
|
| 1385 |
+
"step": 970
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.0788912579957355,
|
| 1389 |
+
"grad_norm": 2.4010941982269287,
|
| 1390 |
+
"learning_rate": 2.9211087420042642e-05,
|
| 1391 |
+
"loss": 0.0721,
|
| 1392 |
+
"step": 975
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.08955223880597,
|
| 1396 |
+
"grad_norm": 0.9128118753433228,
|
| 1397 |
+
"learning_rate": 2.91044776119403e-05,
|
| 1398 |
+
"loss": 0.1143,
|
| 1399 |
+
"step": 980
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 2.100213219616205,
|
| 1403 |
+
"grad_norm": 1.4514065980911255,
|
| 1404 |
+
"learning_rate": 2.8997867803837952e-05,
|
| 1405 |
+
"loss": 0.0317,
|
| 1406 |
+
"step": 985
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.1108742004264394,
|
| 1410 |
+
"grad_norm": 4.103957653045654,
|
| 1411 |
+
"learning_rate": 2.889125799573561e-05,
|
| 1412 |
+
"loss": 0.107,
|
| 1413 |
+
"step": 990
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.121535181236674,
|
| 1417 |
+
"grad_norm": 0.8747802376747131,
|
| 1418 |
+
"learning_rate": 2.878464818763326e-05,
|
| 1419 |
+
"loss": 0.0945,
|
| 1420 |
+
"step": 995
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 2.1321961620469083,
|
| 1424 |
+
"grad_norm": 1.40634024143219,
|
| 1425 |
+
"learning_rate": 2.867803837953092e-05,
|
| 1426 |
+
"loss": 0.0997,
|
| 1427 |
+
"step": 1000
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.142857142857143,
|
| 1431 |
+
"grad_norm": 2.4886908531188965,
|
| 1432 |
+
"learning_rate": 2.857142857142857e-05,
|
| 1433 |
+
"loss": 0.0612,
|
| 1434 |
+
"step": 1005
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 2.1535181236673773,
|
| 1438 |
+
"grad_norm": 2.8721213340759277,
|
| 1439 |
+
"learning_rate": 2.846481876332623e-05,
|
| 1440 |
+
"loss": 0.0494,
|
| 1441 |
+
"step": 1010
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.1641791044776117,
|
| 1445 |
+
"grad_norm": 3.9986650943756104,
|
| 1446 |
+
"learning_rate": 2.835820895522388e-05,
|
| 1447 |
+
"loss": 0.0638,
|
| 1448 |
+
"step": 1015
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.1748400852878467,
|
| 1452 |
+
"grad_norm": 0.8146592378616333,
|
| 1453 |
+
"learning_rate": 2.825159914712154e-05,
|
| 1454 |
+
"loss": 0.0494,
|
| 1455 |
+
"step": 1020
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.185501066098081,
|
| 1459 |
+
"grad_norm": 1.4202768802642822,
|
| 1460 |
+
"learning_rate": 2.814498933901919e-05,
|
| 1461 |
+
"loss": 0.0792,
|
| 1462 |
+
"step": 1025
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 2.1961620469083156,
|
| 1466 |
+
"grad_norm": 0.49137574434280396,
|
| 1467 |
+
"learning_rate": 2.8038379530916848e-05,
|
| 1468 |
+
"loss": 0.0661,
|
| 1469 |
+
"step": 1030
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.20682302771855,
|
| 1473 |
+
"grad_norm": 0.4127224385738373,
|
| 1474 |
+
"learning_rate": 2.79317697228145e-05,
|
| 1475 |
+
"loss": 0.0642,
|
| 1476 |
+
"step": 1035
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 2.2174840085287846,
|
| 1480 |
+
"grad_norm": 5.33475399017334,
|
| 1481 |
+
"learning_rate": 2.7825159914712157e-05,
|
| 1482 |
+
"loss": 0.1819,
|
| 1483 |
+
"step": 1040
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 2.228144989339019,
|
| 1487 |
+
"grad_norm": 0.40939345955848694,
|
| 1488 |
+
"learning_rate": 2.771855010660981e-05,
|
| 1489 |
+
"loss": 0.0254,
|
| 1490 |
+
"step": 1045
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.2388059701492535,
|
| 1494 |
+
"grad_norm": 0.8224121332168579,
|
| 1495 |
+
"learning_rate": 2.7611940298507467e-05,
|
| 1496 |
+
"loss": 0.0959,
|
| 1497 |
+
"step": 1050
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 2.2494669509594885,
|
| 1501 |
+
"grad_norm": 4.530190944671631,
|
| 1502 |
+
"learning_rate": 2.7505330490405118e-05,
|
| 1503 |
+
"loss": 0.1041,
|
| 1504 |
+
"step": 1055
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 2.260127931769723,
|
| 1508 |
+
"grad_norm": 5.169997692108154,
|
| 1509 |
+
"learning_rate": 2.7398720682302776e-05,
|
| 1510 |
+
"loss": 0.0901,
|
| 1511 |
+
"step": 1060
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.2707889125799574,
|
| 1515 |
+
"grad_norm": 4.3644514083862305,
|
| 1516 |
+
"learning_rate": 2.7292110874200428e-05,
|
| 1517 |
+
"loss": 0.0585,
|
| 1518 |
+
"step": 1065
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 2.281449893390192,
|
| 1522 |
+
"grad_norm": 0.8456694483757019,
|
| 1523 |
+
"learning_rate": 2.7185501066098086e-05,
|
| 1524 |
+
"loss": 0.066,
|
| 1525 |
+
"step": 1070
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 2.2921108742004264,
|
| 1529 |
+
"grad_norm": 0.9708316326141357,
|
| 1530 |
+
"learning_rate": 2.7078891257995737e-05,
|
| 1531 |
+
"loss": 0.1004,
|
| 1532 |
+
"step": 1075
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.302771855010661,
|
| 1536 |
+
"grad_norm": 1.521084189414978,
|
| 1537 |
+
"learning_rate": 2.6972281449893395e-05,
|
| 1538 |
+
"loss": 0.0404,
|
| 1539 |
+
"step": 1080
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 2.3134328358208958,
|
| 1543 |
+
"grad_norm": 3.5763978958129883,
|
| 1544 |
+
"learning_rate": 2.6865671641791047e-05,
|
| 1545 |
+
"loss": 0.0779,
|
| 1546 |
+
"step": 1085
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 2.3240938166311302,
|
| 1550 |
+
"grad_norm": 2.3992388248443604,
|
| 1551 |
+
"learning_rate": 2.6759061833688705e-05,
|
| 1552 |
+
"loss": 0.0599,
|
| 1553 |
+
"step": 1090
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.3347547974413647,
|
| 1557 |
+
"grad_norm": 2.1601805686950684,
|
| 1558 |
+
"learning_rate": 2.6652452025586356e-05,
|
| 1559 |
+
"loss": 0.1027,
|
| 1560 |
+
"step": 1095
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 2.345415778251599,
|
| 1564 |
+
"grad_norm": 1.30140221118927,
|
| 1565 |
+
"learning_rate": 2.6545842217484007e-05,
|
| 1566 |
+
"loss": 0.0886,
|
| 1567 |
+
"step": 1100
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 2.3560767590618337,
|
| 1571 |
+
"grad_norm": 0.8566694259643555,
|
| 1572 |
+
"learning_rate": 2.6439232409381666e-05,
|
| 1573 |
+
"loss": 0.0676,
|
| 1574 |
+
"step": 1105
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 2.366737739872068,
|
| 1578 |
+
"grad_norm": 0.5876774191856384,
|
| 1579 |
+
"learning_rate": 2.6332622601279317e-05,
|
| 1580 |
+
"loss": 0.0543,
|
| 1581 |
+
"step": 1110
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 2.3773987206823026,
|
| 1585 |
+
"grad_norm": 2.2665822505950928,
|
| 1586 |
+
"learning_rate": 2.6226012793176975e-05,
|
| 1587 |
+
"loss": 0.0621,
|
| 1588 |
+
"step": 1115
|
| 1589 |
+
},
|
| 1590 |
+
{
|
| 1591 |
+
"epoch": 2.388059701492537,
|
| 1592 |
+
"grad_norm": 2.25077223777771,
|
| 1593 |
+
"learning_rate": 2.6119402985074626e-05,
|
| 1594 |
+
"loss": 0.0813,
|
| 1595 |
+
"step": 1120
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 2.398720682302772,
|
| 1599 |
+
"grad_norm": 0.5967782139778137,
|
| 1600 |
+
"learning_rate": 2.6012793176972285e-05,
|
| 1601 |
+
"loss": 0.0566,
|
| 1602 |
+
"step": 1125
|
| 1603 |
+
},
|
| 1604 |
+
{
|
| 1605 |
+
"epoch": 2.4093816631130065,
|
| 1606 |
+
"grad_norm": 0.27977997064590454,
|
| 1607 |
+
"learning_rate": 2.5906183368869936e-05,
|
| 1608 |
+
"loss": 0.0687,
|
| 1609 |
+
"step": 1130
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 2.420042643923241,
|
| 1613 |
+
"grad_norm": 6.441884517669678,
|
| 1614 |
+
"learning_rate": 2.5799573560767594e-05,
|
| 1615 |
+
"loss": 0.0827,
|
| 1616 |
+
"step": 1135
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 2.4307036247334755,
|
| 1620 |
+
"grad_norm": 2.2369818687438965,
|
| 1621 |
+
"learning_rate": 2.5692963752665245e-05,
|
| 1622 |
+
"loss": 0.1213,
|
| 1623 |
+
"step": 1140
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 2.44136460554371,
|
| 1627 |
+
"grad_norm": 3.0550386905670166,
|
| 1628 |
+
"learning_rate": 2.5586353944562904e-05,
|
| 1629 |
+
"loss": 0.0777,
|
| 1630 |
+
"step": 1145
|
| 1631 |
+
},
|
| 1632 |
+
{
|
| 1633 |
+
"epoch": 2.4520255863539444,
|
| 1634 |
+
"grad_norm": 0.1704282909631729,
|
| 1635 |
+
"learning_rate": 2.5479744136460555e-05,
|
| 1636 |
+
"loss": 0.0429,
|
| 1637 |
+
"step": 1150
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 2.4626865671641793,
|
| 1641 |
+
"grad_norm": 2.5110490322113037,
|
| 1642 |
+
"learning_rate": 2.537313432835821e-05,
|
| 1643 |
+
"loss": 0.0821,
|
| 1644 |
+
"step": 1155
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 2.473347547974414,
|
| 1648 |
+
"grad_norm": 1.6381982564926147,
|
| 1649 |
+
"learning_rate": 2.5266524520255864e-05,
|
| 1650 |
+
"loss": 0.0484,
|
| 1651 |
+
"step": 1160
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 2.4840085287846483,
|
| 1655 |
+
"grad_norm": 0.2897392213344574,
|
| 1656 |
+
"learning_rate": 2.515991471215352e-05,
|
| 1657 |
+
"loss": 0.0545,
|
| 1658 |
+
"step": 1165
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 2.4946695095948828,
|
| 1662 |
+
"grad_norm": 3.0763368606567383,
|
| 1663 |
+
"learning_rate": 2.5053304904051174e-05,
|
| 1664 |
+
"loss": 0.0967,
|
| 1665 |
+
"step": 1170
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 2.5053304904051172,
|
| 1669 |
+
"grad_norm": 1.611001968383789,
|
| 1670 |
+
"learning_rate": 2.494669509594883e-05,
|
| 1671 |
+
"loss": 0.0744,
|
| 1672 |
+
"step": 1175
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 2.5159914712153517,
|
| 1676 |
+
"grad_norm": 0.4237760007381439,
|
| 1677 |
+
"learning_rate": 2.4840085287846483e-05,
|
| 1678 |
+
"loss": 0.0526,
|
| 1679 |
+
"step": 1180
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 2.526652452025586,
|
| 1683 |
+
"grad_norm": 1.4915021657943726,
|
| 1684 |
+
"learning_rate": 2.4733475479744138e-05,
|
| 1685 |
+
"loss": 0.0614,
|
| 1686 |
+
"step": 1185
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 2.5373134328358207,
|
| 1690 |
+
"grad_norm": 3.5276854038238525,
|
| 1691 |
+
"learning_rate": 2.4626865671641793e-05,
|
| 1692 |
+
"loss": 0.0693,
|
| 1693 |
+
"step": 1190
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 2.5479744136460556,
|
| 1697 |
+
"grad_norm": 1.6728593111038208,
|
| 1698 |
+
"learning_rate": 2.4520255863539444e-05,
|
| 1699 |
+
"loss": 0.0711,
|
| 1700 |
+
"step": 1195
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 2.55863539445629,
|
| 1704 |
+
"grad_norm": 0.691937267780304,
|
| 1705 |
+
"learning_rate": 2.44136460554371e-05,
|
| 1706 |
+
"loss": 0.0632,
|
| 1707 |
+
"step": 1200
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 2.5692963752665245,
|
| 1711 |
+
"grad_norm": 1.1585723161697388,
|
| 1712 |
+
"learning_rate": 2.4307036247334754e-05,
|
| 1713 |
+
"loss": 0.1272,
|
| 1714 |
+
"step": 1205
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 2.579957356076759,
|
| 1718 |
+
"grad_norm": 1.4050172567367554,
|
| 1719 |
+
"learning_rate": 2.420042643923241e-05,
|
| 1720 |
+
"loss": 0.0491,
|
| 1721 |
+
"step": 1210
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 2.5906183368869935,
|
| 1725 |
+
"grad_norm": 3.591080904006958,
|
| 1726 |
+
"learning_rate": 2.4093816631130063e-05,
|
| 1727 |
+
"loss": 0.1023,
|
| 1728 |
+
"step": 1215
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 2.6012793176972284,
|
| 1732 |
+
"grad_norm": 0.2823375165462494,
|
| 1733 |
+
"learning_rate": 2.3987206823027718e-05,
|
| 1734 |
+
"loss": 0.054,
|
| 1735 |
+
"step": 1220
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 2.611940298507463,
|
| 1739 |
+
"grad_norm": 2.9172515869140625,
|
| 1740 |
+
"learning_rate": 2.3880597014925373e-05,
|
| 1741 |
+
"loss": 0.1029,
|
| 1742 |
+
"step": 1225
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 2.6226012793176974,
|
| 1746 |
+
"grad_norm": 1.3210735321044922,
|
| 1747 |
+
"learning_rate": 2.3773987206823027e-05,
|
| 1748 |
+
"loss": 0.0425,
|
| 1749 |
+
"step": 1230
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 2.633262260127932,
|
| 1753 |
+
"grad_norm": 1.2610630989074707,
|
| 1754 |
+
"learning_rate": 2.3667377398720682e-05,
|
| 1755 |
+
"loss": 0.0764,
|
| 1756 |
+
"step": 1235
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 2.6439232409381663,
|
| 1760 |
+
"grad_norm": 2.457002878189087,
|
| 1761 |
+
"learning_rate": 2.3560767590618337e-05,
|
| 1762 |
+
"loss": 0.1268,
|
| 1763 |
+
"step": 1240
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 2.654584221748401,
|
| 1767 |
+
"grad_norm": 1.5941702127456665,
|
| 1768 |
+
"learning_rate": 2.345415778251599e-05,
|
| 1769 |
+
"loss": 0.0486,
|
| 1770 |
+
"step": 1245
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 2.6652452025586353,
|
| 1774 |
+
"grad_norm": 1.485596776008606,
|
| 1775 |
+
"learning_rate": 2.3347547974413646e-05,
|
| 1776 |
+
"loss": 0.0504,
|
| 1777 |
+
"step": 1250
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 2.6759061833688698,
|
| 1781 |
+
"grad_norm": 1.2748851776123047,
|
| 1782 |
+
"learning_rate": 2.32409381663113e-05,
|
| 1783 |
+
"loss": 0.0883,
|
| 1784 |
+
"step": 1255
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 2.6865671641791042,
|
| 1788 |
+
"grad_norm": 0.8153276443481445,
|
| 1789 |
+
"learning_rate": 2.3134328358208956e-05,
|
| 1790 |
+
"loss": 0.0964,
|
| 1791 |
+
"step": 1260
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 2.697228144989339,
|
| 1795 |
+
"grad_norm": 0.7963466644287109,
|
| 1796 |
+
"learning_rate": 2.302771855010661e-05,
|
| 1797 |
+
"loss": 0.0454,
|
| 1798 |
+
"step": 1265
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 2.7078891257995736,
|
| 1802 |
+
"grad_norm": 0.27712538838386536,
|
| 1803 |
+
"learning_rate": 2.2921108742004265e-05,
|
| 1804 |
+
"loss": 0.0881,
|
| 1805 |
+
"step": 1270
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.718550106609808,
|
| 1809 |
+
"grad_norm": 2.9359426498413086,
|
| 1810 |
+
"learning_rate": 2.281449893390192e-05,
|
| 1811 |
+
"loss": 0.1146,
|
| 1812 |
+
"step": 1275
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 2.7292110874200426,
|
| 1816 |
+
"grad_norm": 0.8790004849433899,
|
| 1817 |
+
"learning_rate": 2.2707889125799575e-05,
|
| 1818 |
+
"loss": 0.0334,
|
| 1819 |
+
"step": 1280
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 2.739872068230277,
|
| 1823 |
+
"grad_norm": 3.954956293106079,
|
| 1824 |
+
"learning_rate": 2.260127931769723e-05,
|
| 1825 |
+
"loss": 0.1174,
|
| 1826 |
+
"step": 1285
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 2.750533049040512,
|
| 1830 |
+
"grad_norm": 2.426549196243286,
|
| 1831 |
+
"learning_rate": 2.2494669509594884e-05,
|
| 1832 |
+
"loss": 0.1405,
|
| 1833 |
+
"step": 1290
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 2.7611940298507465,
|
| 1837 |
+
"grad_norm": 1.1180130243301392,
|
| 1838 |
+
"learning_rate": 2.238805970149254e-05,
|
| 1839 |
+
"loss": 0.0817,
|
| 1840 |
+
"step": 1295
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 2.771855010660981,
|
| 1844 |
+
"grad_norm": 3.375093460083008,
|
| 1845 |
+
"learning_rate": 2.2281449893390194e-05,
|
| 1846 |
+
"loss": 0.0874,
|
| 1847 |
+
"step": 1300
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.7825159914712154,
|
| 1851 |
+
"grad_norm": 0.8761259913444519,
|
| 1852 |
+
"learning_rate": 2.217484008528785e-05,
|
| 1853 |
+
"loss": 0.0354,
|
| 1854 |
+
"step": 1305
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 2.79317697228145,
|
| 1858 |
+
"grad_norm": 2.2289867401123047,
|
| 1859 |
+
"learning_rate": 2.2068230277185503e-05,
|
| 1860 |
+
"loss": 0.0269,
|
| 1861 |
+
"step": 1310
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 2.8038379530916844,
|
| 1865 |
+
"grad_norm": 5.426002025604248,
|
| 1866 |
+
"learning_rate": 2.1961620469083158e-05,
|
| 1867 |
+
"loss": 0.0408,
|
| 1868 |
+
"step": 1315
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.814498933901919,
|
| 1872 |
+
"grad_norm": 3.4484269618988037,
|
| 1873 |
+
"learning_rate": 2.1855010660980813e-05,
|
| 1874 |
+
"loss": 0.0787,
|
| 1875 |
+
"step": 1320
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 2.8251599147121533,
|
| 1879 |
+
"grad_norm": 1.0502134561538696,
|
| 1880 |
+
"learning_rate": 2.1748400852878467e-05,
|
| 1881 |
+
"loss": 0.0681,
|
| 1882 |
+
"step": 1325
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 2.835820895522388,
|
| 1886 |
+
"grad_norm": 1.3416142463684082,
|
| 1887 |
+
"learning_rate": 2.164179104477612e-05,
|
| 1888 |
+
"loss": 0.0492,
|
| 1889 |
+
"step": 1330
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.8464818763326227,
|
| 1893 |
+
"grad_norm": 3.33859920501709,
|
| 1894 |
+
"learning_rate": 2.1535181236673773e-05,
|
| 1895 |
+
"loss": 0.0827,
|
| 1896 |
+
"step": 1335
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 2.857142857142857,
|
| 1900 |
+
"grad_norm": 0.7152363061904907,
|
| 1901 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 1902 |
+
"loss": 0.0573,
|
| 1903 |
+
"step": 1340
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 2.8678038379530917,
|
| 1907 |
+
"grad_norm": 0.4876638948917389,
|
| 1908 |
+
"learning_rate": 2.1321961620469083e-05,
|
| 1909 |
+
"loss": 0.0989,
|
| 1910 |
+
"step": 1345
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 2.878464818763326,
|
| 1914 |
+
"grad_norm": 4.9128313064575195,
|
| 1915 |
+
"learning_rate": 2.1215351812366738e-05,
|
| 1916 |
+
"loss": 0.0719,
|
| 1917 |
+
"step": 1350
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 2.8891257995735606,
|
| 1921 |
+
"grad_norm": 2.4759974479675293,
|
| 1922 |
+
"learning_rate": 2.1108742004264392e-05,
|
| 1923 |
+
"loss": 0.0811,
|
| 1924 |
+
"step": 1355
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 2.8997867803837956,
|
| 1928 |
+
"grad_norm": 1.174643635749817,
|
| 1929 |
+
"learning_rate": 2.1002132196162047e-05,
|
| 1930 |
+
"loss": 0.0431,
|
| 1931 |
+
"step": 1360
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 2.91044776119403,
|
| 1935 |
+
"grad_norm": 3.599703550338745,
|
| 1936 |
+
"learning_rate": 2.0895522388059702e-05,
|
| 1937 |
+
"loss": 0.1302,
|
| 1938 |
+
"step": 1365
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 2.9211087420042645,
|
| 1942 |
+
"grad_norm": 0.06804340332746506,
|
| 1943 |
+
"learning_rate": 2.0788912579957357e-05,
|
| 1944 |
+
"loss": 0.0664,
|
| 1945 |
+
"step": 1370
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 2.931769722814499,
|
| 1949 |
+
"grad_norm": 5.044620513916016,
|
| 1950 |
+
"learning_rate": 2.068230277185501e-05,
|
| 1951 |
+
"loss": 0.0742,
|
| 1952 |
+
"step": 1375
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 2.9424307036247335,
|
| 1956 |
+
"grad_norm": 0.3658342659473419,
|
| 1957 |
+
"learning_rate": 2.0575692963752666e-05,
|
| 1958 |
+
"loss": 0.0426,
|
| 1959 |
+
"step": 1380
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 2.953091684434968,
|
| 1963 |
+
"grad_norm": 0.06923450529575348,
|
| 1964 |
+
"learning_rate": 2.046908315565032e-05,
|
| 1965 |
+
"loss": 0.0646,
|
| 1966 |
+
"step": 1385
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 2.9637526652452024,
|
| 1970 |
+
"grad_norm": 0.5141579508781433,
|
| 1971 |
+
"learning_rate": 2.0362473347547976e-05,
|
| 1972 |
+
"loss": 0.0479,
|
| 1973 |
+
"step": 1390
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 2.974413646055437,
|
| 1977 |
+
"grad_norm": 2.5570452213287354,
|
| 1978 |
+
"learning_rate": 2.025586353944563e-05,
|
| 1979 |
+
"loss": 0.036,
|
| 1980 |
+
"step": 1395
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 2.9850746268656714,
|
| 1984 |
+
"grad_norm": 1.7159104347229004,
|
| 1985 |
+
"learning_rate": 2.0149253731343285e-05,
|
| 1986 |
+
"loss": 0.055,
|
| 1987 |
+
"step": 1400
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 2.9957356076759063,
|
| 1991 |
+
"grad_norm": 2.963869094848633,
|
| 1992 |
+
"learning_rate": 2.004264392324094e-05,
|
| 1993 |
+
"loss": 0.048,
|
| 1994 |
+
"step": 1405
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 3.0,
|
| 1998 |
+
"eval_accuracy": 0.9906666666666667,
|
| 1999 |
+
"eval_loss": 0.02723684348165989,
|
| 2000 |
+
"eval_runtime": 30.119,
|
| 2001 |
+
"eval_samples_per_second": 124.506,
|
| 2002 |
+
"eval_steps_per_second": 3.918,
|
| 2003 |
+
"step": 1407
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 3.0063965884861408,
|
| 2007 |
+
"grad_norm": 0.34967324137687683,
|
| 2008 |
+
"learning_rate": 1.9936034115138594e-05,
|
| 2009 |
+
"loss": 0.0346,
|
| 2010 |
+
"step": 1410
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 3.0170575692963753,
|
| 2014 |
+
"grad_norm": 0.779079258441925,
|
| 2015 |
+
"learning_rate": 1.982942430703625e-05,
|
| 2016 |
+
"loss": 0.0701,
|
| 2017 |
+
"step": 1415
|
| 2018 |
+
},
|
| 2019 |
+
{
|
| 2020 |
+
"epoch": 3.0277185501066097,
|
| 2021 |
+
"grad_norm": 0.7551254630088806,
|
| 2022 |
+
"learning_rate": 1.9722814498933904e-05,
|
| 2023 |
+
"loss": 0.0542,
|
| 2024 |
+
"step": 1420
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 3.038379530916844,
|
| 2028 |
+
"grad_norm": 1.3363436460494995,
|
| 2029 |
+
"learning_rate": 1.961620469083156e-05,
|
| 2030 |
+
"loss": 0.1327,
|
| 2031 |
+
"step": 1425
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 3.0490405117270787,
|
| 2035 |
+
"grad_norm": 5.568324565887451,
|
| 2036 |
+
"learning_rate": 1.9509594882729213e-05,
|
| 2037 |
+
"loss": 0.1644,
|
| 2038 |
+
"step": 1430
|
| 2039 |
+
},
|
| 2040 |
+
{
|
| 2041 |
+
"epoch": 3.0597014925373136,
|
| 2042 |
+
"grad_norm": 1.4754337072372437,
|
| 2043 |
+
"learning_rate": 1.9402985074626868e-05,
|
| 2044 |
+
"loss": 0.1291,
|
| 2045 |
+
"step": 1435
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.070362473347548,
|
| 2049 |
+
"grad_norm": 1.9287389516830444,
|
| 2050 |
+
"learning_rate": 1.9296375266524523e-05,
|
| 2051 |
+
"loss": 0.0333,
|
| 2052 |
+
"step": 1440
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 3.0810234541577826,
|
| 2056 |
+
"grad_norm": 2.4145143032073975,
|
| 2057 |
+
"learning_rate": 1.9189765458422178e-05,
|
| 2058 |
+
"loss": 0.057,
|
| 2059 |
+
"step": 1445
|
| 2060 |
+
},
|
| 2061 |
+
{
|
| 2062 |
+
"epoch": 3.091684434968017,
|
| 2063 |
+
"grad_norm": 0.05112965777516365,
|
| 2064 |
+
"learning_rate": 1.9083155650319832e-05,
|
| 2065 |
+
"loss": 0.0743,
|
| 2066 |
+
"step": 1450
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 3.1023454157782515,
|
| 2070 |
+
"grad_norm": 4.283263683319092,
|
| 2071 |
+
"learning_rate": 1.8976545842217487e-05,
|
| 2072 |
+
"loss": 0.124,
|
| 2073 |
+
"step": 1455
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 3.113006396588486,
|
| 2077 |
+
"grad_norm": 3.570025682449341,
|
| 2078 |
+
"learning_rate": 1.8869936034115142e-05,
|
| 2079 |
+
"loss": 0.0722,
|
| 2080 |
+
"step": 1460
|
| 2081 |
+
},
|
| 2082 |
+
{
|
| 2083 |
+
"epoch": 3.1236673773987205,
|
| 2084 |
+
"grad_norm": 1.0681952238082886,
|
| 2085 |
+
"learning_rate": 1.8763326226012797e-05,
|
| 2086 |
+
"loss": 0.0437,
|
| 2087 |
+
"step": 1465
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 3.1343283582089554,
|
| 2091 |
+
"grad_norm": 0.901505708694458,
|
| 2092 |
+
"learning_rate": 1.865671641791045e-05,
|
| 2093 |
+
"loss": 0.036,
|
| 2094 |
+
"step": 1470
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 3.14498933901919,
|
| 2098 |
+
"grad_norm": 0.745629072189331,
|
| 2099 |
+
"learning_rate": 1.8550106609808106e-05,
|
| 2100 |
+
"loss": 0.1159,
|
| 2101 |
+
"step": 1475
|
| 2102 |
+
},
|
| 2103 |
+
{
|
| 2104 |
+
"epoch": 3.1556503198294243,
|
| 2105 |
+
"grad_norm": 0.3517145812511444,
|
| 2106 |
+
"learning_rate": 1.8443496801705757e-05,
|
| 2107 |
+
"loss": 0.0786,
|
| 2108 |
+
"step": 1480
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 3.166311300639659,
|
| 2112 |
+
"grad_norm": 2.3695156574249268,
|
| 2113 |
+
"learning_rate": 1.8336886993603412e-05,
|
| 2114 |
+
"loss": 0.0592,
|
| 2115 |
+
"step": 1485
|
| 2116 |
+
},
|
| 2117 |
+
{
|
| 2118 |
+
"epoch": 3.1769722814498933,
|
| 2119 |
+
"grad_norm": 1.377666711807251,
|
| 2120 |
+
"learning_rate": 1.8230277185501067e-05,
|
| 2121 |
+
"loss": 0.0771,
|
| 2122 |
+
"step": 1490
|
| 2123 |
+
},
|
| 2124 |
+
{
|
| 2125 |
+
"epoch": 3.1876332622601278,
|
| 2126 |
+
"grad_norm": 3.0045783519744873,
|
| 2127 |
+
"learning_rate": 1.812366737739872e-05,
|
| 2128 |
+
"loss": 0.0818,
|
| 2129 |
+
"step": 1495
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 3.1982942430703627,
|
| 2133 |
+
"grad_norm": 2.717719078063965,
|
| 2134 |
+
"learning_rate": 1.8017057569296376e-05,
|
| 2135 |
+
"loss": 0.0661,
|
| 2136 |
+
"step": 1500
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 3.208955223880597,
|
| 2140 |
+
"grad_norm": 2.37862491607666,
|
| 2141 |
+
"learning_rate": 1.791044776119403e-05,
|
| 2142 |
+
"loss": 0.108,
|
| 2143 |
+
"step": 1505
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 3.2196162046908317,
|
| 2147 |
+
"grad_norm": 3.22446346282959,
|
| 2148 |
+
"learning_rate": 1.7803837953091686e-05,
|
| 2149 |
+
"loss": 0.0297,
|
| 2150 |
+
"step": 1510
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 3.230277185501066,
|
| 2154 |
+
"grad_norm": 2.4398293495178223,
|
| 2155 |
+
"learning_rate": 1.769722814498934e-05,
|
| 2156 |
+
"loss": 0.0669,
|
| 2157 |
+
"step": 1515
|
| 2158 |
+
},
|
| 2159 |
+
{
|
| 2160 |
+
"epoch": 3.2409381663113006,
|
| 2161 |
+
"grad_norm": 1.0918775796890259,
|
| 2162 |
+
"learning_rate": 1.7590618336886995e-05,
|
| 2163 |
+
"loss": 0.0616,
|
| 2164 |
+
"step": 1520
|
| 2165 |
+
},
|
| 2166 |
+
{
|
| 2167 |
+
"epoch": 3.251599147121535,
|
| 2168 |
+
"grad_norm": 1.3496243953704834,
|
| 2169 |
+
"learning_rate": 1.7484008528784647e-05,
|
| 2170 |
+
"loss": 0.1191,
|
| 2171 |
+
"step": 1525
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 3.2622601279317696,
|
| 2175 |
+
"grad_norm": 5.097357273101807,
|
| 2176 |
+
"learning_rate": 1.73773987206823e-05,
|
| 2177 |
+
"loss": 0.0421,
|
| 2178 |
+
"step": 1530
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 3.272921108742004,
|
| 2182 |
+
"grad_norm": 1.4156861305236816,
|
| 2183 |
+
"learning_rate": 1.7270788912579956e-05,
|
| 2184 |
+
"loss": 0.0376,
|
| 2185 |
+
"step": 1535
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 3.283582089552239,
|
| 2189 |
+
"grad_norm": 4.096904754638672,
|
| 2190 |
+
"learning_rate": 1.716417910447761e-05,
|
| 2191 |
+
"loss": 0.1294,
|
| 2192 |
+
"step": 1540
|
| 2193 |
+
},
|
| 2194 |
+
{
|
| 2195 |
+
"epoch": 3.2942430703624734,
|
| 2196 |
+
"grad_norm": 1.5388942956924438,
|
| 2197 |
+
"learning_rate": 1.7057569296375266e-05,
|
| 2198 |
+
"loss": 0.0653,
|
| 2199 |
+
"step": 1545
|
| 2200 |
+
},
|
| 2201 |
+
{
|
| 2202 |
+
"epoch": 3.304904051172708,
|
| 2203 |
+
"grad_norm": 0.18927036225795746,
|
| 2204 |
+
"learning_rate": 1.695095948827292e-05,
|
| 2205 |
+
"loss": 0.0501,
|
| 2206 |
+
"step": 1550
|
| 2207 |
+
},
|
| 2208 |
+
{
|
| 2209 |
+
"epoch": 3.3155650319829424,
|
| 2210 |
+
"grad_norm": 3.4176881313323975,
|
| 2211 |
+
"learning_rate": 1.6844349680170575e-05,
|
| 2212 |
+
"loss": 0.0828,
|
| 2213 |
+
"step": 1555
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 3.326226012793177,
|
| 2217 |
+
"grad_norm": 4.314210414886475,
|
| 2218 |
+
"learning_rate": 1.673773987206823e-05,
|
| 2219 |
+
"loss": 0.1036,
|
| 2220 |
+
"step": 1560
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 3.3368869936034113,
|
| 2224 |
+
"grad_norm": 0.28427213430404663,
|
| 2225 |
+
"learning_rate": 1.6631130063965885e-05,
|
| 2226 |
+
"loss": 0.0317,
|
| 2227 |
+
"step": 1565
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 3.3475479744136463,
|
| 2231 |
+
"grad_norm": 0.7097741961479187,
|
| 2232 |
+
"learning_rate": 1.652452025586354e-05,
|
| 2233 |
+
"loss": 0.0413,
|
| 2234 |
+
"step": 1570
|
| 2235 |
+
},
|
| 2236 |
+
{
|
| 2237 |
+
"epoch": 3.3582089552238807,
|
| 2238 |
+
"grad_norm": 0.6041834354400635,
|
| 2239 |
+
"learning_rate": 1.6417910447761194e-05,
|
| 2240 |
+
"loss": 0.0539,
|
| 2241 |
+
"step": 1575
|
| 2242 |
+
},
|
| 2243 |
+
{
|
| 2244 |
+
"epoch": 3.368869936034115,
|
| 2245 |
+
"grad_norm": 2.7906687259674072,
|
| 2246 |
+
"learning_rate": 1.631130063965885e-05,
|
| 2247 |
+
"loss": 0.0606,
|
| 2248 |
+
"step": 1580
|
| 2249 |
+
},
|
| 2250 |
+
{
|
| 2251 |
+
"epoch": 3.3795309168443497,
|
| 2252 |
+
"grad_norm": 0.2202073633670807,
|
| 2253 |
+
"learning_rate": 1.6204690831556504e-05,
|
| 2254 |
+
"loss": 0.0274,
|
| 2255 |
+
"step": 1585
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 3.390191897654584,
|
| 2259 |
+
"grad_norm": 0.25416404008865356,
|
| 2260 |
+
"learning_rate": 1.6098081023454158e-05,
|
| 2261 |
+
"loss": 0.0588,
|
| 2262 |
+
"step": 1590
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 3.4008528784648187,
|
| 2266 |
+
"grad_norm": 2.887753486633301,
|
| 2267 |
+
"learning_rate": 1.5991471215351813e-05,
|
| 2268 |
+
"loss": 0.1045,
|
| 2269 |
+
"step": 1595
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 3.411513859275053,
|
| 2273 |
+
"grad_norm": 2.687920331954956,
|
| 2274 |
+
"learning_rate": 1.5884861407249468e-05,
|
| 2275 |
+
"loss": 0.1907,
|
| 2276 |
+
"step": 1600
|
| 2277 |
+
},
|
| 2278 |
+
{
|
| 2279 |
+
"epoch": 3.4221748400852876,
|
| 2280 |
+
"grad_norm": 4.6989665031433105,
|
| 2281 |
+
"learning_rate": 1.5778251599147122e-05,
|
| 2282 |
+
"loss": 0.1268,
|
| 2283 |
+
"step": 1605
|
| 2284 |
+
},
|
| 2285 |
+
{
|
| 2286 |
+
"epoch": 3.4328358208955225,
|
| 2287 |
+
"grad_norm": 1.642794132232666,
|
| 2288 |
+
"learning_rate": 1.5671641791044777e-05,
|
| 2289 |
+
"loss": 0.0512,
|
| 2290 |
+
"step": 1610
|
| 2291 |
+
},
|
| 2292 |
+
{
|
| 2293 |
+
"epoch": 3.443496801705757,
|
| 2294 |
+
"grad_norm": 0.4326121509075165,
|
| 2295 |
+
"learning_rate": 1.5565031982942432e-05,
|
| 2296 |
+
"loss": 0.0603,
|
| 2297 |
+
"step": 1615
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 3.4541577825159915,
|
| 2301 |
+
"grad_norm": 0.6310318112373352,
|
| 2302 |
+
"learning_rate": 1.5458422174840087e-05,
|
| 2303 |
+
"loss": 0.041,
|
| 2304 |
+
"step": 1620
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 3.464818763326226,
|
| 2308 |
+
"grad_norm": 2.798521041870117,
|
| 2309 |
+
"learning_rate": 1.535181236673774e-05,
|
| 2310 |
+
"loss": 0.0559,
|
| 2311 |
+
"step": 1625
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 3.4754797441364604,
|
| 2315 |
+
"grad_norm": 1.9396069049835205,
|
| 2316 |
+
"learning_rate": 1.5245202558635396e-05,
|
| 2317 |
+
"loss": 0.0548,
|
| 2318 |
+
"step": 1630
|
| 2319 |
+
},
|
| 2320 |
+
{
|
| 2321 |
+
"epoch": 3.486140724946695,
|
| 2322 |
+
"grad_norm": 0.3199023902416229,
|
| 2323 |
+
"learning_rate": 1.5138592750533051e-05,
|
| 2324 |
+
"loss": 0.0814,
|
| 2325 |
+
"step": 1635
|
| 2326 |
+
},
|
| 2327 |
+
{
|
| 2328 |
+
"epoch": 3.49680170575693,
|
| 2329 |
+
"grad_norm": 0.27606338262557983,
|
| 2330 |
+
"learning_rate": 1.5031982942430706e-05,
|
| 2331 |
+
"loss": 0.0245,
|
| 2332 |
+
"step": 1640
|
| 2333 |
+
},
|
| 2334 |
+
{
|
| 2335 |
+
"epoch": 3.5074626865671643,
|
| 2336 |
+
"grad_norm": 1.0894615650177002,
|
| 2337 |
+
"learning_rate": 1.4925373134328357e-05,
|
| 2338 |
+
"loss": 0.0908,
|
| 2339 |
+
"step": 1645
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 3.518123667377399,
|
| 2343 |
+
"grad_norm": 0.2977132797241211,
|
| 2344 |
+
"learning_rate": 1.4818763326226012e-05,
|
| 2345 |
+
"loss": 0.0443,
|
| 2346 |
+
"step": 1650
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 3.5287846481876333,
|
| 2350 |
+
"grad_norm": 1.04399836063385,
|
| 2351 |
+
"learning_rate": 1.4712153518123666e-05,
|
| 2352 |
+
"loss": 0.0315,
|
| 2353 |
+
"step": 1655
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 3.5394456289978677,
|
| 2357 |
+
"grad_norm": 1.2887417078018188,
|
| 2358 |
+
"learning_rate": 1.4605543710021321e-05,
|
| 2359 |
+
"loss": 0.0427,
|
| 2360 |
+
"step": 1660
|
| 2361 |
+
},
|
| 2362 |
+
{
|
| 2363 |
+
"epoch": 3.550106609808102,
|
| 2364 |
+
"grad_norm": 4.592863082885742,
|
| 2365 |
+
"learning_rate": 1.4498933901918976e-05,
|
| 2366 |
+
"loss": 0.0692,
|
| 2367 |
+
"step": 1665
|
| 2368 |
+
},
|
| 2369 |
+
{
|
| 2370 |
+
"epoch": 3.5607675906183367,
|
| 2371 |
+
"grad_norm": 2.3803274631500244,
|
| 2372 |
+
"learning_rate": 1.439232409381663e-05,
|
| 2373 |
+
"loss": 0.0531,
|
| 2374 |
+
"step": 1670
|
| 2375 |
+
},
|
| 2376 |
+
{
|
| 2377 |
+
"epoch": 3.571428571428571,
|
| 2378 |
+
"grad_norm": 2.485058069229126,
|
| 2379 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 2380 |
+
"loss": 0.0667,
|
| 2381 |
+
"step": 1675
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 3.582089552238806,
|
| 2385 |
+
"grad_norm": 0.6991864442825317,
|
| 2386 |
+
"learning_rate": 1.417910447761194e-05,
|
| 2387 |
+
"loss": 0.0295,
|
| 2388 |
+
"step": 1680
|
| 2389 |
+
},
|
| 2390 |
+
{
|
| 2391 |
+
"epoch": 3.5927505330490406,
|
| 2392 |
+
"grad_norm": 2.812380313873291,
|
| 2393 |
+
"learning_rate": 1.4072494669509595e-05,
|
| 2394 |
+
"loss": 0.0607,
|
| 2395 |
+
"step": 1685
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 3.603411513859275,
|
| 2399 |
+
"grad_norm": 1.6755821704864502,
|
| 2400 |
+
"learning_rate": 1.396588486140725e-05,
|
| 2401 |
+
"loss": 0.0874,
|
| 2402 |
+
"step": 1690
|
| 2403 |
+
},
|
| 2404 |
+
{
|
| 2405 |
+
"epoch": 3.6140724946695095,
|
| 2406 |
+
"grad_norm": 0.129374697804451,
|
| 2407 |
+
"learning_rate": 1.3859275053304904e-05,
|
| 2408 |
+
"loss": 0.0066,
|
| 2409 |
+
"step": 1695
|
| 2410 |
+
},
|
| 2411 |
+
{
|
| 2412 |
+
"epoch": 3.624733475479744,
|
| 2413 |
+
"grad_norm": 1.7369403839111328,
|
| 2414 |
+
"learning_rate": 1.3752665245202559e-05,
|
| 2415 |
+
"loss": 0.0969,
|
| 2416 |
+
"step": 1700
|
| 2417 |
+
},
|
| 2418 |
+
{
|
| 2419 |
+
"epoch": 3.635394456289979,
|
| 2420 |
+
"grad_norm": 2.260180950164795,
|
| 2421 |
+
"learning_rate": 1.3646055437100214e-05,
|
| 2422 |
+
"loss": 0.0661,
|
| 2423 |
+
"step": 1705
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 3.6460554371002134,
|
| 2427 |
+
"grad_norm": 2.13417911529541,
|
| 2428 |
+
"learning_rate": 1.3539445628997869e-05,
|
| 2429 |
+
"loss": 0.0907,
|
| 2430 |
+
"step": 1710
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 3.656716417910448,
|
| 2434 |
+
"grad_norm": 2.1402523517608643,
|
| 2435 |
+
"learning_rate": 1.3432835820895523e-05,
|
| 2436 |
+
"loss": 0.0539,
|
| 2437 |
+
"step": 1715
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 3.6673773987206824,
|
| 2441 |
+
"grad_norm": 0.05029414966702461,
|
| 2442 |
+
"learning_rate": 1.3326226012793178e-05,
|
| 2443 |
+
"loss": 0.0263,
|
| 2444 |
+
"step": 1720
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 3.678038379530917,
|
| 2448 |
+
"grad_norm": 5.060847759246826,
|
| 2449 |
+
"learning_rate": 1.3219616204690833e-05,
|
| 2450 |
+
"loss": 0.0964,
|
| 2451 |
+
"step": 1725
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 3.6886993603411513,
|
| 2455 |
+
"grad_norm": 0.41038060188293457,
|
| 2456 |
+
"learning_rate": 1.3113006396588488e-05,
|
| 2457 |
+
"loss": 0.0639,
|
| 2458 |
+
"step": 1730
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 3.699360341151386,
|
| 2462 |
+
"grad_norm": 4.709085941314697,
|
| 2463 |
+
"learning_rate": 1.3006396588486142e-05,
|
| 2464 |
+
"loss": 0.1232,
|
| 2465 |
+
"step": 1735
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 3.7100213219616203,
|
| 2469 |
+
"grad_norm": 2.7018556594848633,
|
| 2470 |
+
"learning_rate": 1.2899786780383797e-05,
|
| 2471 |
+
"loss": 0.0961,
|
| 2472 |
+
"step": 1740
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 3.7206823027718547,
|
| 2476 |
+
"grad_norm": 1.618735432624817,
|
| 2477 |
+
"learning_rate": 1.2793176972281452e-05,
|
| 2478 |
+
"loss": 0.0268,
|
| 2479 |
+
"step": 1745
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 3.7313432835820897,
|
| 2483 |
+
"grad_norm": 0.16983424127101898,
|
| 2484 |
+
"learning_rate": 1.2686567164179105e-05,
|
| 2485 |
+
"loss": 0.104,
|
| 2486 |
+
"step": 1750
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 3.742004264392324,
|
| 2490 |
+
"grad_norm": 3.36080002784729,
|
| 2491 |
+
"learning_rate": 1.257995735607676e-05,
|
| 2492 |
+
"loss": 0.0432,
|
| 2493 |
+
"step": 1755
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 3.7526652452025586,
|
| 2497 |
+
"grad_norm": 2.6168394088745117,
|
| 2498 |
+
"learning_rate": 1.2473347547974414e-05,
|
| 2499 |
+
"loss": 0.0363,
|
| 2500 |
+
"step": 1760
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 3.763326226012793,
|
| 2504 |
+
"grad_norm": 4.211267471313477,
|
| 2505 |
+
"learning_rate": 1.2366737739872069e-05,
|
| 2506 |
+
"loss": 0.0544,
|
| 2507 |
+
"step": 1765
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 3.7739872068230276,
|
| 2511 |
+
"grad_norm": 3.4754464626312256,
|
| 2512 |
+
"learning_rate": 1.2260127931769722e-05,
|
| 2513 |
+
"loss": 0.0658,
|
| 2514 |
+
"step": 1770
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 3.7846481876332625,
|
| 2518 |
+
"grad_norm": 1.0480775833129883,
|
| 2519 |
+
"learning_rate": 1.2153518123667377e-05,
|
| 2520 |
+
"loss": 0.0675,
|
| 2521 |
+
"step": 1775
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 3.795309168443497,
|
| 2525 |
+
"grad_norm": 2.6933000087738037,
|
| 2526 |
+
"learning_rate": 1.2046908315565032e-05,
|
| 2527 |
+
"loss": 0.0835,
|
| 2528 |
+
"step": 1780
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 3.8059701492537314,
|
| 2532 |
+
"grad_norm": 1.0857077836990356,
|
| 2533 |
+
"learning_rate": 1.1940298507462686e-05,
|
| 2534 |
+
"loss": 0.0181,
|
| 2535 |
+
"step": 1785
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 3.816631130063966,
|
| 2539 |
+
"grad_norm": 2.9868557453155518,
|
| 2540 |
+
"learning_rate": 1.1833688699360341e-05,
|
| 2541 |
+
"loss": 0.0568,
|
| 2542 |
+
"step": 1790
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 3.8272921108742004,
|
| 2546 |
+
"grad_norm": 1.9914056062698364,
|
| 2547 |
+
"learning_rate": 1.1727078891257996e-05,
|
| 2548 |
+
"loss": 0.0808,
|
| 2549 |
+
"step": 1795
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 3.837953091684435,
|
| 2553 |
+
"grad_norm": 4.660048007965088,
|
| 2554 |
+
"learning_rate": 1.162046908315565e-05,
|
| 2555 |
+
"loss": 0.1384,
|
| 2556 |
+
"step": 1800
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 3.8486140724946694,
|
| 2560 |
+
"grad_norm": 0.5528120398521423,
|
| 2561 |
+
"learning_rate": 1.1513859275053305e-05,
|
| 2562 |
+
"loss": 0.0597,
|
| 2563 |
+
"step": 1805
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 3.859275053304904,
|
| 2567 |
+
"grad_norm": 5.454668045043945,
|
| 2568 |
+
"learning_rate": 1.140724946695096e-05,
|
| 2569 |
+
"loss": 0.1469,
|
| 2570 |
+
"step": 1810
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 3.8699360341151388,
|
| 2574 |
+
"grad_norm": 0.2659088969230652,
|
| 2575 |
+
"learning_rate": 1.1300639658848615e-05,
|
| 2576 |
+
"loss": 0.0326,
|
| 2577 |
+
"step": 1815
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 3.8805970149253732,
|
| 2581 |
+
"grad_norm": 2.5112602710723877,
|
| 2582 |
+
"learning_rate": 1.119402985074627e-05,
|
| 2583 |
+
"loss": 0.1284,
|
| 2584 |
+
"step": 1820
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 3.8912579957356077,
|
| 2588 |
+
"grad_norm": 0.7239062786102295,
|
| 2589 |
+
"learning_rate": 1.1087420042643924e-05,
|
| 2590 |
+
"loss": 0.0575,
|
| 2591 |
+
"step": 1825
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 3.901918976545842,
|
| 2595 |
+
"grad_norm": 2.4359853267669678,
|
| 2596 |
+
"learning_rate": 1.0980810234541579e-05,
|
| 2597 |
+
"loss": 0.0837,
|
| 2598 |
+
"step": 1830
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 3.9125799573560767,
|
| 2602 |
+
"grad_norm": 2.6595561504364014,
|
| 2603 |
+
"learning_rate": 1.0874200426439234e-05,
|
| 2604 |
+
"loss": 0.0385,
|
| 2605 |
+
"step": 1835
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 3.923240938166311,
|
| 2609 |
+
"grad_norm": 3.3885459899902344,
|
| 2610 |
+
"learning_rate": 1.0767590618336887e-05,
|
| 2611 |
+
"loss": 0.076,
|
| 2612 |
+
"step": 1840
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 3.933901918976546,
|
| 2616 |
+
"grad_norm": 1.0346359014511108,
|
| 2617 |
+
"learning_rate": 1.0660980810234541e-05,
|
| 2618 |
+
"loss": 0.1292,
|
| 2619 |
+
"step": 1845
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 3.9445628997867805,
|
| 2623 |
+
"grad_norm": 2.4496281147003174,
|
| 2624 |
+
"learning_rate": 1.0554371002132196e-05,
|
| 2625 |
+
"loss": 0.0643,
|
| 2626 |
+
"step": 1850
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 3.955223880597015,
|
| 2630 |
+
"grad_norm": 3.268887519836426,
|
| 2631 |
+
"learning_rate": 1.0447761194029851e-05,
|
| 2632 |
+
"loss": 0.0506,
|
| 2633 |
+
"step": 1855
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 3.9658848614072495,
|
| 2637 |
+
"grad_norm": 5.954560279846191,
|
| 2638 |
+
"learning_rate": 1.0341151385927506e-05,
|
| 2639 |
+
"loss": 0.0393,
|
| 2640 |
+
"step": 1860
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 3.976545842217484,
|
| 2644 |
+
"grad_norm": 0.27536705136299133,
|
| 2645 |
+
"learning_rate": 1.023454157782516e-05,
|
| 2646 |
+
"loss": 0.0398,
|
| 2647 |
+
"step": 1865
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 3.9872068230277184,
|
| 2651 |
+
"grad_norm": 3.0534212589263916,
|
| 2652 |
+
"learning_rate": 1.0127931769722815e-05,
|
| 2653 |
+
"loss": 0.0612,
|
| 2654 |
+
"step": 1870
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 3.997867803837953,
|
| 2658 |
+
"grad_norm": 0.26579713821411133,
|
| 2659 |
+
"learning_rate": 1.002132196162047e-05,
|
| 2660 |
+
"loss": 0.0332,
|
| 2661 |
+
"step": 1875
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 4.0,
|
| 2665 |
+
"eval_accuracy": 0.9914666666666667,
|
| 2666 |
+
"eval_loss": 0.02808554470539093,
|
| 2667 |
+
"eval_runtime": 29.9462,
|
| 2668 |
+
"eval_samples_per_second": 125.225,
|
| 2669 |
+
"eval_steps_per_second": 3.94,
|
| 2670 |
+
"step": 1876
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 4.008528784648187,
|
| 2674 |
+
"grad_norm": 1.9030885696411133,
|
| 2675 |
+
"learning_rate": 9.914712153518125e-06,
|
| 2676 |
+
"loss": 0.0497,
|
| 2677 |
+
"step": 1880
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 4.019189765458422,
|
| 2681 |
+
"grad_norm": 1.4887663125991821,
|
| 2682 |
+
"learning_rate": 9.80810234541578e-06,
|
| 2683 |
+
"loss": 0.0385,
|
| 2684 |
+
"step": 1885
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 4.029850746268656,
|
| 2688 |
+
"grad_norm": 7.906754493713379,
|
| 2689 |
+
"learning_rate": 9.701492537313434e-06,
|
| 2690 |
+
"loss": 0.0568,
|
| 2691 |
+
"step": 1890
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 4.040511727078891,
|
| 2695 |
+
"grad_norm": 3.2405171394348145,
|
| 2696 |
+
"learning_rate": 9.594882729211089e-06,
|
| 2697 |
+
"loss": 0.0432,
|
| 2698 |
+
"step": 1895
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 4.051172707889126,
|
| 2702 |
+
"grad_norm": 3.8229644298553467,
|
| 2703 |
+
"learning_rate": 9.488272921108744e-06,
|
| 2704 |
+
"loss": 0.1288,
|
| 2705 |
+
"step": 1900
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 4.061833688699361,
|
| 2709 |
+
"grad_norm": 3.769545316696167,
|
| 2710 |
+
"learning_rate": 9.381663113006398e-06,
|
| 2711 |
+
"loss": 0.0716,
|
| 2712 |
+
"step": 1905
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 4.072494669509595,
|
| 2716 |
+
"grad_norm": 1.8549972772598267,
|
| 2717 |
+
"learning_rate": 9.275053304904053e-06,
|
| 2718 |
+
"loss": 0.1221,
|
| 2719 |
+
"step": 1910
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 4.08315565031983,
|
| 2723 |
+
"grad_norm": 1.1318780183792114,
|
| 2724 |
+
"learning_rate": 9.168443496801706e-06,
|
| 2725 |
+
"loss": 0.0842,
|
| 2726 |
+
"step": 1915
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 4.093816631130064,
|
| 2730 |
+
"grad_norm": 2.434079885482788,
|
| 2731 |
+
"learning_rate": 9.06183368869936e-06,
|
| 2732 |
+
"loss": 0.1083,
|
| 2733 |
+
"step": 1920
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 4.104477611940299,
|
| 2737 |
+
"grad_norm": 2.334904193878174,
|
| 2738 |
+
"learning_rate": 8.955223880597016e-06,
|
| 2739 |
+
"loss": 0.0583,
|
| 2740 |
+
"step": 1925
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 4.115138592750533,
|
| 2744 |
+
"grad_norm": 0.5059848427772522,
|
| 2745 |
+
"learning_rate": 8.84861407249467e-06,
|
| 2746 |
+
"loss": 0.0122,
|
| 2747 |
+
"step": 1930
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 4.1257995735607675,
|
| 2751 |
+
"grad_norm": 2.021977663040161,
|
| 2752 |
+
"learning_rate": 8.742004264392323e-06,
|
| 2753 |
+
"loss": 0.0849,
|
| 2754 |
+
"step": 1935
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 4.136460554371002,
|
| 2758 |
+
"grad_norm": 2.2379519939422607,
|
| 2759 |
+
"learning_rate": 8.635394456289978e-06,
|
| 2760 |
+
"loss": 0.0757,
|
| 2761 |
+
"step": 1940
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 4.1471215351812365,
|
| 2765 |
+
"grad_norm": 0.2632559835910797,
|
| 2766 |
+
"learning_rate": 8.528784648187633e-06,
|
| 2767 |
+
"loss": 0.0818,
|
| 2768 |
+
"step": 1945
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 4.157782515991471,
|
| 2772 |
+
"grad_norm": 0.035075847059488297,
|
| 2773 |
+
"learning_rate": 8.422174840085288e-06,
|
| 2774 |
+
"loss": 0.0363,
|
| 2775 |
+
"step": 1950
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 4.1684434968017055,
|
| 2779 |
+
"grad_norm": 0.2989201843738556,
|
| 2780 |
+
"learning_rate": 8.315565031982942e-06,
|
| 2781 |
+
"loss": 0.0487,
|
| 2782 |
+
"step": 1955
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 4.17910447761194,
|
| 2786 |
+
"grad_norm": 2.132059335708618,
|
| 2787 |
+
"learning_rate": 8.208955223880597e-06,
|
| 2788 |
+
"loss": 0.0764,
|
| 2789 |
+
"step": 1960
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 4.189765458422174,
|
| 2793 |
+
"grad_norm": 0.10834915190935135,
|
| 2794 |
+
"learning_rate": 8.102345415778252e-06,
|
| 2795 |
+
"loss": 0.0237,
|
| 2796 |
+
"step": 1965
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 4.20042643923241,
|
| 2800 |
+
"grad_norm": 0.20066118240356445,
|
| 2801 |
+
"learning_rate": 7.995735607675907e-06,
|
| 2802 |
+
"loss": 0.0593,
|
| 2803 |
+
"step": 1970
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 4.211087420042644,
|
| 2807 |
+
"grad_norm": 0.5289183855056763,
|
| 2808 |
+
"learning_rate": 7.889125799573561e-06,
|
| 2809 |
+
"loss": 0.0851,
|
| 2810 |
+
"step": 1975
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 4.221748400852879,
|
| 2814 |
+
"grad_norm": 2.1197450160980225,
|
| 2815 |
+
"learning_rate": 7.782515991471216e-06,
|
| 2816 |
+
"loss": 0.1004,
|
| 2817 |
+
"step": 1980
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 4.232409381663113,
|
| 2821 |
+
"grad_norm": 0.6274572014808655,
|
| 2822 |
+
"learning_rate": 7.67590618336887e-06,
|
| 2823 |
+
"loss": 0.0912,
|
| 2824 |
+
"step": 1985
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 4.243070362473348,
|
| 2828 |
+
"grad_norm": 2.2780940532684326,
|
| 2829 |
+
"learning_rate": 7.5692963752665255e-06,
|
| 2830 |
+
"loss": 0.0551,
|
| 2831 |
+
"step": 1990
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 4.253731343283582,
|
| 2835 |
+
"grad_norm": 2.6508591175079346,
|
| 2836 |
+
"learning_rate": 7.4626865671641785e-06,
|
| 2837 |
+
"loss": 0.1035,
|
| 2838 |
+
"step": 1995
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 4.264392324093817,
|
| 2842 |
+
"grad_norm": 0.21462765336036682,
|
| 2843 |
+
"learning_rate": 7.356076759061833e-06,
|
| 2844 |
+
"loss": 0.1177,
|
| 2845 |
+
"step": 2000
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 4.275053304904051,
|
| 2849 |
+
"grad_norm": 2.3007192611694336,
|
| 2850 |
+
"learning_rate": 7.249466950959488e-06,
|
| 2851 |
+
"loss": 0.0811,
|
| 2852 |
+
"step": 2005
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 4.285714285714286,
|
| 2856 |
+
"grad_norm": 2.547154426574707,
|
| 2857 |
+
"learning_rate": 7.142857142857143e-06,
|
| 2858 |
+
"loss": 0.0588,
|
| 2859 |
+
"step": 2010
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 4.29637526652452,
|
| 2863 |
+
"grad_norm": 2.522618293762207,
|
| 2864 |
+
"learning_rate": 7.0362473347547975e-06,
|
| 2865 |
+
"loss": 0.0466,
|
| 2866 |
+
"step": 2015
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 4.3070362473347545,
|
| 2870 |
+
"grad_norm": 2.3301403522491455,
|
| 2871 |
+
"learning_rate": 6.929637526652452e-06,
|
| 2872 |
+
"loss": 0.0968,
|
| 2873 |
+
"step": 2020
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 4.317697228144989,
|
| 2877 |
+
"grad_norm": 3.820814609527588,
|
| 2878 |
+
"learning_rate": 6.823027718550107e-06,
|
| 2879 |
+
"loss": 0.0687,
|
| 2880 |
+
"step": 2025
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 4.3283582089552235,
|
| 2884 |
+
"grad_norm": 0.2634601593017578,
|
| 2885 |
+
"learning_rate": 6.716417910447762e-06,
|
| 2886 |
+
"loss": 0.0517,
|
| 2887 |
+
"step": 2030
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 4.339019189765459,
|
| 2891 |
+
"grad_norm": 0.469744473695755,
|
| 2892 |
+
"learning_rate": 6.609808102345416e-06,
|
| 2893 |
+
"loss": 0.0623,
|
| 2894 |
+
"step": 2035
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 4.349680170575693,
|
| 2898 |
+
"grad_norm": 0.25715357065200806,
|
| 2899 |
+
"learning_rate": 6.503198294243071e-06,
|
| 2900 |
+
"loss": 0.0548,
|
| 2901 |
+
"step": 2040
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 4.360341151385928,
|
| 2905 |
+
"grad_norm": 1.8255966901779175,
|
| 2906 |
+
"learning_rate": 6.396588486140726e-06,
|
| 2907 |
+
"loss": 0.1007,
|
| 2908 |
+
"step": 2045
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 4.371002132196162,
|
| 2912 |
+
"grad_norm": 1.4979438781738281,
|
| 2913 |
+
"learning_rate": 6.28997867803838e-06,
|
| 2914 |
+
"loss": 0.0855,
|
| 2915 |
+
"step": 2050
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 4.381663113006397,
|
| 2919 |
+
"grad_norm": 0.04768993705511093,
|
| 2920 |
+
"learning_rate": 6.1833688699360345e-06,
|
| 2921 |
+
"loss": 0.0329,
|
| 2922 |
+
"step": 2055
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 4.392324093816631,
|
| 2926 |
+
"grad_norm": 3.1533761024475098,
|
| 2927 |
+
"learning_rate": 6.076759061833688e-06,
|
| 2928 |
+
"loss": 0.0695,
|
| 2929 |
+
"step": 2060
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 4.402985074626866,
|
| 2933 |
+
"grad_norm": 1.5988001823425293,
|
| 2934 |
+
"learning_rate": 5.970149253731343e-06,
|
| 2935 |
+
"loss": 0.0914,
|
| 2936 |
+
"step": 2065
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 4.4136460554371,
|
| 2940 |
+
"grad_norm": 1.4010282754898071,
|
| 2941 |
+
"learning_rate": 5.863539445628998e-06,
|
| 2942 |
+
"loss": 0.0773,
|
| 2943 |
+
"step": 2070
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 4.424307036247335,
|
| 2947 |
+
"grad_norm": 2.2126927375793457,
|
| 2948 |
+
"learning_rate": 5.756929637526653e-06,
|
| 2949 |
+
"loss": 0.0258,
|
| 2950 |
+
"step": 2075
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 4.434968017057569,
|
| 2954 |
+
"grad_norm": 0.33380889892578125,
|
| 2955 |
+
"learning_rate": 5.650319829424307e-06,
|
| 2956 |
+
"loss": 0.057,
|
| 2957 |
+
"step": 2080
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 4.445628997867804,
|
| 2961 |
+
"grad_norm": 1.0894397497177124,
|
| 2962 |
+
"learning_rate": 5.543710021321962e-06,
|
| 2963 |
+
"loss": 0.024,
|
| 2964 |
+
"step": 2085
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 4.456289978678038,
|
| 2968 |
+
"grad_norm": 0.043762437999248505,
|
| 2969 |
+
"learning_rate": 5.437100213219617e-06,
|
| 2970 |
+
"loss": 0.0382,
|
| 2971 |
+
"step": 2090
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 4.466950959488273,
|
| 2975 |
+
"grad_norm": 0.05760975927114487,
|
| 2976 |
+
"learning_rate": 5.330490405117271e-06,
|
| 2977 |
+
"loss": 0.0169,
|
| 2978 |
+
"step": 2095
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 4.477611940298507,
|
| 2982 |
+
"grad_norm": 1.7139006853103638,
|
| 2983 |
+
"learning_rate": 5.2238805970149255e-06,
|
| 2984 |
+
"loss": 0.0297,
|
| 2985 |
+
"step": 2100
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 4.4882729211087415,
|
| 2989 |
+
"grad_norm": 0.7469732761383057,
|
| 2990 |
+
"learning_rate": 5.11727078891258e-06,
|
| 2991 |
+
"loss": 0.0168,
|
| 2992 |
+
"step": 2105
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 4.498933901918977,
|
| 2996 |
+
"grad_norm": 1.375309944152832,
|
| 2997 |
+
"learning_rate": 5.010660980810235e-06,
|
| 2998 |
+
"loss": 0.0491,
|
| 2999 |
+
"step": 2110
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 4.509594882729211,
|
| 3003 |
+
"grad_norm": 1.1908515691757202,
|
| 3004 |
+
"learning_rate": 4.90405117270789e-06,
|
| 3005 |
+
"loss": 0.0135,
|
| 3006 |
+
"step": 2115
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 4.520255863539446,
|
| 3010 |
+
"grad_norm": 2.957414150238037,
|
| 3011 |
+
"learning_rate": 4.797441364605544e-06,
|
| 3012 |
+
"loss": 0.0803,
|
| 3013 |
+
"step": 2120
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 4.53091684434968,
|
| 3017 |
+
"grad_norm": 1.3513374328613281,
|
| 3018 |
+
"learning_rate": 4.690831556503199e-06,
|
| 3019 |
+
"loss": 0.0455,
|
| 3020 |
+
"step": 2125
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 4.541577825159915,
|
| 3024 |
+
"grad_norm": 1.6028034687042236,
|
| 3025 |
+
"learning_rate": 4.584221748400853e-06,
|
| 3026 |
+
"loss": 0.0386,
|
| 3027 |
+
"step": 2130
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 4.552238805970149,
|
| 3031 |
+
"grad_norm": 0.20417140424251556,
|
| 3032 |
+
"learning_rate": 4.477611940298508e-06,
|
| 3033 |
+
"loss": 0.0979,
|
| 3034 |
+
"step": 2135
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 4.562899786780384,
|
| 3038 |
+
"grad_norm": 2.917415142059326,
|
| 3039 |
+
"learning_rate": 4.371002132196162e-06,
|
| 3040 |
+
"loss": 0.0333,
|
| 3041 |
+
"step": 2140
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 4.573560767590618,
|
| 3045 |
+
"grad_norm": 1.522704005241394,
|
| 3046 |
+
"learning_rate": 4.264392324093816e-06,
|
| 3047 |
+
"loss": 0.1079,
|
| 3048 |
+
"step": 2145
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 4.584221748400853,
|
| 3052 |
+
"grad_norm": 1.3484480381011963,
|
| 3053 |
+
"learning_rate": 4.157782515991471e-06,
|
| 3054 |
+
"loss": 0.0272,
|
| 3055 |
+
"step": 2150
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 4.594882729211087,
|
| 3059 |
+
"grad_norm": 0.02694857493042946,
|
| 3060 |
+
"learning_rate": 4.051172707889126e-06,
|
| 3061 |
+
"loss": 0.0474,
|
| 3062 |
+
"step": 2155
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 4.605543710021322,
|
| 3066 |
+
"grad_norm": 0.6975867748260498,
|
| 3067 |
+
"learning_rate": 3.944562899786781e-06,
|
| 3068 |
+
"loss": 0.0367,
|
| 3069 |
+
"step": 2160
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 4.616204690831556,
|
| 3073 |
+
"grad_norm": 1.6975103616714478,
|
| 3074 |
+
"learning_rate": 3.837953091684435e-06,
|
| 3075 |
+
"loss": 0.0585,
|
| 3076 |
+
"step": 2165
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 4.6268656716417915,
|
| 3080 |
+
"grad_norm": 0.9779977798461914,
|
| 3081 |
+
"learning_rate": 3.7313432835820893e-06,
|
| 3082 |
+
"loss": 0.0494,
|
| 3083 |
+
"step": 2170
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 4.637526652452026,
|
| 3087 |
+
"grad_norm": 3.441746473312378,
|
| 3088 |
+
"learning_rate": 3.624733475479744e-06,
|
| 3089 |
+
"loss": 0.0953,
|
| 3090 |
+
"step": 2175
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 4.6481876332622605,
|
| 3094 |
+
"grad_norm": 2.462442398071289,
|
| 3095 |
+
"learning_rate": 3.5181236673773987e-06,
|
| 3096 |
+
"loss": 0.0265,
|
| 3097 |
+
"step": 2180
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 4.658848614072495,
|
| 3101 |
+
"grad_norm": 0.9558789730072021,
|
| 3102 |
+
"learning_rate": 3.4115138592750535e-06,
|
| 3103 |
+
"loss": 0.0609,
|
| 3104 |
+
"step": 2185
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 4.669509594882729,
|
| 3108 |
+
"grad_norm": 1.215696930885315,
|
| 3109 |
+
"learning_rate": 3.304904051172708e-06,
|
| 3110 |
+
"loss": 0.0626,
|
| 3111 |
+
"step": 2190
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 4.680170575692964,
|
| 3115 |
+
"grad_norm": 0.13017113506793976,
|
| 3116 |
+
"learning_rate": 3.198294243070363e-06,
|
| 3117 |
+
"loss": 0.0544,
|
| 3118 |
+
"step": 2195
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 4.690831556503198,
|
| 3122 |
+
"grad_norm": 2.7555811405181885,
|
| 3123 |
+
"learning_rate": 3.0916844349680173e-06,
|
| 3124 |
+
"loss": 0.0313,
|
| 3125 |
+
"step": 2200
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 4.701492537313433,
|
| 3129 |
+
"grad_norm": 2.671470880508423,
|
| 3130 |
+
"learning_rate": 2.9850746268656716e-06,
|
| 3131 |
+
"loss": 0.0436,
|
| 3132 |
+
"step": 2205
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 4.712153518123667,
|
| 3136 |
+
"grad_norm": 0.039603669196367264,
|
| 3137 |
+
"learning_rate": 2.8784648187633263e-06,
|
| 3138 |
+
"loss": 0.0172,
|
| 3139 |
+
"step": 2210
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 4.722814498933902,
|
| 3143 |
+
"grad_norm": 2.098763942718506,
|
| 3144 |
+
"learning_rate": 2.771855010660981e-06,
|
| 3145 |
+
"loss": 0.0724,
|
| 3146 |
+
"step": 2215
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 4.733475479744136,
|
| 3150 |
+
"grad_norm": 0.73785001039505,
|
| 3151 |
+
"learning_rate": 2.6652452025586354e-06,
|
| 3152 |
+
"loss": 0.0503,
|
| 3153 |
+
"step": 2220
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 4.744136460554371,
|
| 3157 |
+
"grad_norm": 1.8888320922851562,
|
| 3158 |
+
"learning_rate": 2.55863539445629e-06,
|
| 3159 |
+
"loss": 0.0893,
|
| 3160 |
+
"step": 2225
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 4.754797441364605,
|
| 3164 |
+
"grad_norm": 1.278704047203064,
|
| 3165 |
+
"learning_rate": 2.452025586353945e-06,
|
| 3166 |
+
"loss": 0.0684,
|
| 3167 |
+
"step": 2230
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 4.76545842217484,
|
| 3171 |
+
"grad_norm": 1.185849905014038,
|
| 3172 |
+
"learning_rate": 2.3454157782515996e-06,
|
| 3173 |
+
"loss": 0.0328,
|
| 3174 |
+
"step": 2235
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 4.776119402985074,
|
| 3178 |
+
"grad_norm": 2.6555964946746826,
|
| 3179 |
+
"learning_rate": 2.238805970149254e-06,
|
| 3180 |
+
"loss": 0.057,
|
| 3181 |
+
"step": 2240
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 4.786780383795309,
|
| 3185 |
+
"grad_norm": 1.6034694910049438,
|
| 3186 |
+
"learning_rate": 2.132196162046908e-06,
|
| 3187 |
+
"loss": 0.0185,
|
| 3188 |
+
"step": 2245
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 4.797441364605544,
|
| 3192 |
+
"grad_norm": 1.5292361974716187,
|
| 3193 |
+
"learning_rate": 2.025586353944563e-06,
|
| 3194 |
+
"loss": 0.1066,
|
| 3195 |
+
"step": 2250
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 4.8081023454157785,
|
| 3199 |
+
"grad_norm": 0.17130614817142487,
|
| 3200 |
+
"learning_rate": 1.9189765458422177e-06,
|
| 3201 |
+
"loss": 0.0402,
|
| 3202 |
+
"step": 2255
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 4.818763326226013,
|
| 3206 |
+
"grad_norm": 3.7359113693237305,
|
| 3207 |
+
"learning_rate": 1.812366737739872e-06,
|
| 3208 |
+
"loss": 0.0682,
|
| 3209 |
+
"step": 2260
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 4.8294243070362475,
|
| 3213 |
+
"grad_norm": 4.377876281738281,
|
| 3214 |
+
"learning_rate": 1.7057569296375267e-06,
|
| 3215 |
+
"loss": 0.1064,
|
| 3216 |
+
"step": 2265
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 4.840085287846482,
|
| 3220 |
+
"grad_norm": 0.30964645743370056,
|
| 3221 |
+
"learning_rate": 1.5991471215351815e-06,
|
| 3222 |
+
"loss": 0.0231,
|
| 3223 |
+
"step": 2270
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 4.850746268656716,
|
| 3227 |
+
"grad_norm": 6.367454528808594,
|
| 3228 |
+
"learning_rate": 1.4925373134328358e-06,
|
| 3229 |
+
"loss": 0.095,
|
| 3230 |
+
"step": 2275
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 4.861407249466951,
|
| 3234 |
+
"grad_norm": 0.25244995951652527,
|
| 3235 |
+
"learning_rate": 1.3859275053304905e-06,
|
| 3236 |
+
"loss": 0.0194,
|
| 3237 |
+
"step": 2280
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 4.872068230277185,
|
| 3241 |
+
"grad_norm": 0.3098326325416565,
|
| 3242 |
+
"learning_rate": 1.279317697228145e-06,
|
| 3243 |
+
"loss": 0.0422,
|
| 3244 |
+
"step": 2285
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 4.88272921108742,
|
| 3248 |
+
"grad_norm": 2.855001449584961,
|
| 3249 |
+
"learning_rate": 1.1727078891257998e-06,
|
| 3250 |
+
"loss": 0.0269,
|
| 3251 |
+
"step": 2290
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 4.893390191897654,
|
| 3255 |
+
"grad_norm": 1.7481770515441895,
|
| 3256 |
+
"learning_rate": 1.066098081023454e-06,
|
| 3257 |
+
"loss": 0.0477,
|
| 3258 |
+
"step": 2295
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 4.904051172707889,
|
| 3262 |
+
"grad_norm": 1.740970253944397,
|
| 3263 |
+
"learning_rate": 9.594882729211088e-07,
|
| 3264 |
+
"loss": 0.0679,
|
| 3265 |
+
"step": 2300
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 4.914712153518123,
|
| 3269 |
+
"grad_norm": 0.6604803204536438,
|
| 3270 |
+
"learning_rate": 8.528784648187634e-07,
|
| 3271 |
+
"loss": 0.0773,
|
| 3272 |
+
"step": 2305
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 4.925373134328359,
|
| 3276 |
+
"grad_norm": 0.6634294390678406,
|
| 3277 |
+
"learning_rate": 7.462686567164179e-07,
|
| 3278 |
+
"loss": 0.0572,
|
| 3279 |
+
"step": 2310
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 4.936034115138593,
|
| 3283 |
+
"grad_norm": 1.056871771812439,
|
| 3284 |
+
"learning_rate": 6.396588486140725e-07,
|
| 3285 |
+
"loss": 0.0167,
|
| 3286 |
+
"step": 2315
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 4.946695095948828,
|
| 3290 |
+
"grad_norm": 0.4825955629348755,
|
| 3291 |
+
"learning_rate": 5.33049040511727e-07,
|
| 3292 |
+
"loss": 0.0391,
|
| 3293 |
+
"step": 2320
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 4.957356076759062,
|
| 3297 |
+
"grad_norm": 5.046590805053711,
|
| 3298 |
+
"learning_rate": 4.264392324093817e-07,
|
| 3299 |
+
"loss": 0.0652,
|
| 3300 |
+
"step": 2325
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 4.968017057569297,
|
| 3304 |
+
"grad_norm": 1.856524109840393,
|
| 3305 |
+
"learning_rate": 3.1982942430703626e-07,
|
| 3306 |
+
"loss": 0.0257,
|
| 3307 |
+
"step": 2330
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 4.978678038379531,
|
| 3311 |
+
"grad_norm": 1.4161804914474487,
|
| 3312 |
+
"learning_rate": 2.1321961620469084e-07,
|
| 3313 |
+
"loss": 0.0662,
|
| 3314 |
+
"step": 2335
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 4.9893390191897655,
|
| 3318 |
+
"grad_norm": 0.24474987387657166,
|
| 3319 |
+
"learning_rate": 1.0660980810234542e-07,
|
| 3320 |
+
"loss": 0.0418,
|
| 3321 |
+
"step": 2340
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 5.0,
|
| 3325 |
+
"grad_norm": 1.1379473209381104,
|
| 3326 |
+
"learning_rate": 0.0,
|
| 3327 |
+
"loss": 0.0733,
|
| 3328 |
+
"step": 2345
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 5.0,
|
| 3332 |
+
"eval_accuracy": 0.9914666666666667,
|
| 3333 |
+
"eval_loss": 0.026531297713518143,
|
| 3334 |
+
"eval_runtime": 29.9872,
|
| 3335 |
+
"eval_samples_per_second": 125.054,
|
| 3336 |
+
"eval_steps_per_second": 3.935,
|
| 3337 |
+
"step": 2345
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 5.0,
|
| 3341 |
+
"step": 2345,
|
| 3342 |
+
"total_flos": 5.8118992210944e+18,
|
| 3343 |
+
"train_loss": 0.08598557333511585,
|
| 3344 |
+
"train_runtime": 1480.9706,
|
| 3345 |
+
"train_samples_per_second": 50.642,
|
| 3346 |
+
"train_steps_per_second": 1.583
|
| 3347 |
+
}
|
| 3348 |
+
],
|
| 3349 |
+
"logging_steps": 5,
|
| 3350 |
+
"max_steps": 2345,
|
| 3351 |
+
"num_input_tokens_seen": 0,
|
| 3352 |
+
"num_train_epochs": 5,
|
| 3353 |
+
"save_steps": 500,
|
| 3354 |
+
"stateful_callbacks": {
|
| 3355 |
+
"TrainerControl": {
|
| 3356 |
+
"args": {
|
| 3357 |
+
"should_epoch_stop": false,
|
| 3358 |
+
"should_evaluate": false,
|
| 3359 |
+
"should_log": false,
|
| 3360 |
+
"should_save": true,
|
| 3361 |
+
"should_training_stop": true
|
| 3362 |
+
},
|
| 3363 |
+
"attributes": {}
|
| 3364 |
+
}
|
| 3365 |
+
},
|
| 3366 |
+
"total_flos": 5.8118992210944e+18,
|
| 3367 |
+
"train_batch_size": 32,
|
| 3368 |
+
"trial_name": null,
|
| 3369 |
+
"trial_params": null
|
| 3370 |
+
}
|