End of training
Browse files- README.md +2 -2
- all_results.json +8 -0
- test_results.json +8 -0
- trainer_state.json +1629 -0
README.md
CHANGED
|
@@ -18,8 +18,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [facebook/timesformer-base-finetuned-k400](https://huggingface.co/facebook/timesformer-base-finetuned-k400) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
-
- Loss:
|
| 22 |
-
- Accuracy: 0.
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
|
|
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [facebook/timesformer-base-finetuned-k400](https://huggingface.co/facebook/timesformer-base-finetuned-k400) on an unknown dataset.
|
| 20 |
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 0.9056
|
| 22 |
+
- Accuracy: 0.7750
|
| 23 |
|
| 24 |
## Model description
|
| 25 |
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.1990867579908677,
|
| 3 |
+
"eval_accuracy": 0.7749803304484658,
|
| 4 |
+
"eval_loss": 0.9055941104888916,
|
| 5 |
+
"eval_runtime": 256.8953,
|
| 6 |
+
"eval_samples_per_second": 4.948,
|
| 7 |
+
"eval_steps_per_second": 2.476
|
| 8 |
+
}
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 2.1990867579908677,
|
| 3 |
+
"eval_accuracy": 0.7749803304484658,
|
| 4 |
+
"eval_loss": 0.9055941104888916,
|
| 5 |
+
"eval_runtime": 256.8953,
|
| 6 |
+
"eval_samples_per_second": 4.948,
|
| 7 |
+
"eval_steps_per_second": 2.476
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1629 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6516264428121721,
|
| 3 |
+
"best_model_checkpoint": "/media/cse/HDD/Shawon/shawon/Timesformer_BdSLW60/checkpoint-1754",
|
| 4 |
+
"epoch": 2.1990867579908677,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2190,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0045662100456621,
|
| 13 |
+
"grad_norm": 13.168639183044434,
|
| 14 |
+
"learning_rate": 2.2831050228310503e-06,
|
| 15 |
+
"loss": 4.2832,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0091324200913242,
|
| 20 |
+
"grad_norm": 12.616332054138184,
|
| 21 |
+
"learning_rate": 4.566210045662101e-06,
|
| 22 |
+
"loss": 4.2594,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0136986301369863,
|
| 27 |
+
"grad_norm": 12.632943153381348,
|
| 28 |
+
"learning_rate": 6.849315068493151e-06,
|
| 29 |
+
"loss": 4.1827,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0182648401826484,
|
| 34 |
+
"grad_norm": 12.6991548538208,
|
| 35 |
+
"learning_rate": 9.132420091324201e-06,
|
| 36 |
+
"loss": 4.0751,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.0228310502283105,
|
| 41 |
+
"grad_norm": 13.425009727478027,
|
| 42 |
+
"learning_rate": 1.1415525114155251e-05,
|
| 43 |
+
"loss": 4.0899,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0273972602739726,
|
| 48 |
+
"grad_norm": 13.551825523376465,
|
| 49 |
+
"learning_rate": 1.3698630136986302e-05,
|
| 50 |
+
"loss": 3.9588,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0319634703196347,
|
| 55 |
+
"grad_norm": 12.522140502929688,
|
| 56 |
+
"learning_rate": 1.5981735159817352e-05,
|
| 57 |
+
"loss": 3.9251,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0365296803652968,
|
| 62 |
+
"grad_norm": 12.543455123901367,
|
| 63 |
+
"learning_rate": 1.8264840182648402e-05,
|
| 64 |
+
"loss": 3.8922,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.0410958904109589,
|
| 69 |
+
"grad_norm": 13.488080024719238,
|
| 70 |
+
"learning_rate": 2.0547945205479453e-05,
|
| 71 |
+
"loss": 4.0068,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.045662100456621,
|
| 76 |
+
"grad_norm": 12.935967445373535,
|
| 77 |
+
"learning_rate": 2.2831050228310503e-05,
|
| 78 |
+
"loss": 3.6687,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0502283105022831,
|
| 83 |
+
"grad_norm": 12.15580940246582,
|
| 84 |
+
"learning_rate": 2.5114155251141553e-05,
|
| 85 |
+
"loss": 3.5509,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0547945205479452,
|
| 90 |
+
"grad_norm": 13.309331893920898,
|
| 91 |
+
"learning_rate": 2.7397260273972603e-05,
|
| 92 |
+
"loss": 3.4915,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.0593607305936073,
|
| 97 |
+
"grad_norm": 12.072916030883789,
|
| 98 |
+
"learning_rate": 2.9680365296803654e-05,
|
| 99 |
+
"loss": 3.3421,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.0639269406392694,
|
| 104 |
+
"grad_norm": 13.800947189331055,
|
| 105 |
+
"learning_rate": 3.1963470319634704e-05,
|
| 106 |
+
"loss": 3.3284,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.0684931506849315,
|
| 111 |
+
"grad_norm": 14.214157104492188,
|
| 112 |
+
"learning_rate": 3.424657534246575e-05,
|
| 113 |
+
"loss": 3.1462,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0730593607305936,
|
| 118 |
+
"grad_norm": 12.837231636047363,
|
| 119 |
+
"learning_rate": 3.6529680365296805e-05,
|
| 120 |
+
"loss": 2.9169,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.0776255707762557,
|
| 125 |
+
"grad_norm": 14.62553882598877,
|
| 126 |
+
"learning_rate": 3.881278538812785e-05,
|
| 127 |
+
"loss": 2.9004,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0821917808219178,
|
| 132 |
+
"grad_norm": 14.03178596496582,
|
| 133 |
+
"learning_rate": 4.1095890410958905e-05,
|
| 134 |
+
"loss": 2.9808,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.0867579908675799,
|
| 139 |
+
"grad_norm": 12.56568717956543,
|
| 140 |
+
"learning_rate": 4.337899543378995e-05,
|
| 141 |
+
"loss": 2.6627,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.091324200913242,
|
| 146 |
+
"grad_norm": 14.119081497192383,
|
| 147 |
+
"learning_rate": 4.5662100456621006e-05,
|
| 148 |
+
"loss": 2.4799,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.0958904109589041,
|
| 153 |
+
"grad_norm": 12.76337718963623,
|
| 154 |
+
"learning_rate": 4.794520547945205e-05,
|
| 155 |
+
"loss": 2.318,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1004566210045662,
|
| 160 |
+
"grad_norm": 12.352985382080078,
|
| 161 |
+
"learning_rate": 4.997463216641299e-05,
|
| 162 |
+
"loss": 2.0831,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.1050228310502283,
|
| 167 |
+
"grad_norm": 11.853866577148438,
|
| 168 |
+
"learning_rate": 4.9720953830542875e-05,
|
| 169 |
+
"loss": 2.4107,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.1095890410958904,
|
| 174 |
+
"grad_norm": 13.12740421295166,
|
| 175 |
+
"learning_rate": 4.9467275494672755e-05,
|
| 176 |
+
"loss": 2.0582,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1141552511415525,
|
| 181 |
+
"grad_norm": 11.828140258789062,
|
| 182 |
+
"learning_rate": 4.9213597158802636e-05,
|
| 183 |
+
"loss": 1.8289,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.1187214611872146,
|
| 188 |
+
"grad_norm": 12.555204391479492,
|
| 189 |
+
"learning_rate": 4.895991882293252e-05,
|
| 190 |
+
"loss": 1.5664,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1232876712328767,
|
| 195 |
+
"grad_norm": 10.697933197021484,
|
| 196 |
+
"learning_rate": 4.8706240487062404e-05,
|
| 197 |
+
"loss": 1.5136,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.1278538812785388,
|
| 202 |
+
"grad_norm": 11.606474876403809,
|
| 203 |
+
"learning_rate": 4.8452562151192285e-05,
|
| 204 |
+
"loss": 1.6229,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.1324200913242009,
|
| 209 |
+
"grad_norm": 9.681914329528809,
|
| 210 |
+
"learning_rate": 4.819888381532217e-05,
|
| 211 |
+
"loss": 1.7203,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.136986301369863,
|
| 216 |
+
"grad_norm": 10.477850914001465,
|
| 217 |
+
"learning_rate": 4.794520547945205e-05,
|
| 218 |
+
"loss": 1.5071,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.1415525114155251,
|
| 223 |
+
"grad_norm": 11.697348594665527,
|
| 224 |
+
"learning_rate": 4.769152714358194e-05,
|
| 225 |
+
"loss": 1.6212,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.1461187214611872,
|
| 230 |
+
"grad_norm": 11.230805397033691,
|
| 231 |
+
"learning_rate": 4.743784880771182e-05,
|
| 232 |
+
"loss": 1.3522,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1506849315068493,
|
| 237 |
+
"grad_norm": 8.138943672180176,
|
| 238 |
+
"learning_rate": 4.71841704718417e-05,
|
| 239 |
+
"loss": 1.2338,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.1552511415525114,
|
| 244 |
+
"grad_norm": 10.334748268127441,
|
| 245 |
+
"learning_rate": 4.693049213597159e-05,
|
| 246 |
+
"loss": 1.3393,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.1598173515981735,
|
| 251 |
+
"grad_norm": 9.896265029907227,
|
| 252 |
+
"learning_rate": 4.667681380010147e-05,
|
| 253 |
+
"loss": 1.261,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.1643835616438356,
|
| 258 |
+
"grad_norm": 10.67037582397461,
|
| 259 |
+
"learning_rate": 4.642313546423136e-05,
|
| 260 |
+
"loss": 1.2365,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.1689497716894977,
|
| 265 |
+
"grad_norm": 7.820501327514648,
|
| 266 |
+
"learning_rate": 4.616945712836124e-05,
|
| 267 |
+
"loss": 1.0388,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.1735159817351598,
|
| 272 |
+
"grad_norm": 8.09377670288086,
|
| 273 |
+
"learning_rate": 4.591577879249112e-05,
|
| 274 |
+
"loss": 1.0919,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.1780821917808219,
|
| 279 |
+
"grad_norm": 9.110677719116211,
|
| 280 |
+
"learning_rate": 4.5662100456621006e-05,
|
| 281 |
+
"loss": 1.0683,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.182648401826484,
|
| 286 |
+
"grad_norm": 11.42436695098877,
|
| 287 |
+
"learning_rate": 4.5408422120750886e-05,
|
| 288 |
+
"loss": 1.0649,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1872146118721461,
|
| 293 |
+
"grad_norm": 9.24187183380127,
|
| 294 |
+
"learning_rate": 4.5154743784880774e-05,
|
| 295 |
+
"loss": 0.8407,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.1917808219178082,
|
| 300 |
+
"grad_norm": 8.836689949035645,
|
| 301 |
+
"learning_rate": 4.4901065449010655e-05,
|
| 302 |
+
"loss": 0.8219,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.1963470319634703,
|
| 307 |
+
"grad_norm": 13.152447700500488,
|
| 308 |
+
"learning_rate": 4.4647387113140535e-05,
|
| 309 |
+
"loss": 1.064,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2009132420091324,
|
| 314 |
+
"grad_norm": 10.30240249633789,
|
| 315 |
+
"learning_rate": 4.439370877727042e-05,
|
| 316 |
+
"loss": 0.7285,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.2054794520547945,
|
| 321 |
+
"grad_norm": 6.908230781555176,
|
| 322 |
+
"learning_rate": 4.41400304414003e-05,
|
| 323 |
+
"loss": 0.9826,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2100456621004566,
|
| 328 |
+
"grad_norm": 6.0176920890808105,
|
| 329 |
+
"learning_rate": 4.3886352105530184e-05,
|
| 330 |
+
"loss": 0.7598,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.2146118721461187,
|
| 335 |
+
"grad_norm": 6.647716999053955,
|
| 336 |
+
"learning_rate": 4.363267376966007e-05,
|
| 337 |
+
"loss": 0.8257,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.2191780821917808,
|
| 342 |
+
"grad_norm": 6.749544620513916,
|
| 343 |
+
"learning_rate": 4.337899543378995e-05,
|
| 344 |
+
"loss": 0.8272,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.2237442922374429,
|
| 349 |
+
"grad_norm": 7.512877464294434,
|
| 350 |
+
"learning_rate": 4.312531709791984e-05,
|
| 351 |
+
"loss": 0.8007,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.228310502283105,
|
| 356 |
+
"grad_norm": 6.311679363250732,
|
| 357 |
+
"learning_rate": 4.287163876204972e-05,
|
| 358 |
+
"loss": 0.7219,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.2328767123287671,
|
| 363 |
+
"grad_norm": 6.0948591232299805,
|
| 364 |
+
"learning_rate": 4.26179604261796e-05,
|
| 365 |
+
"loss": 0.5973,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.2374429223744292,
|
| 370 |
+
"grad_norm": 9.038314819335938,
|
| 371 |
+
"learning_rate": 4.236428209030949e-05,
|
| 372 |
+
"loss": 0.784,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.2420091324200913,
|
| 377 |
+
"grad_norm": 5.299352645874023,
|
| 378 |
+
"learning_rate": 4.211060375443937e-05,
|
| 379 |
+
"loss": 0.5438,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.2465753424657534,
|
| 384 |
+
"grad_norm": 7.578807353973389,
|
| 385 |
+
"learning_rate": 4.1856925418569256e-05,
|
| 386 |
+
"loss": 0.7382,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.2511415525114155,
|
| 391 |
+
"grad_norm": 9.358942985534668,
|
| 392 |
+
"learning_rate": 4.160324708269914e-05,
|
| 393 |
+
"loss": 0.7584,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.2557077625570776,
|
| 398 |
+
"grad_norm": 9.18459701538086,
|
| 399 |
+
"learning_rate": 4.134956874682902e-05,
|
| 400 |
+
"loss": 0.8444,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.2602739726027397,
|
| 405 |
+
"grad_norm": 5.982091426849365,
|
| 406 |
+
"learning_rate": 4.1095890410958905e-05,
|
| 407 |
+
"loss": 0.4002,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.2648401826484018,
|
| 412 |
+
"grad_norm": 4.503027439117432,
|
| 413 |
+
"learning_rate": 4.0842212075088786e-05,
|
| 414 |
+
"loss": 0.5723,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.2694063926940639,
|
| 419 |
+
"grad_norm": 7.2720794677734375,
|
| 420 |
+
"learning_rate": 4.058853373921867e-05,
|
| 421 |
+
"loss": 0.5846,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.273972602739726,
|
| 426 |
+
"grad_norm": 6.161769390106201,
|
| 427 |
+
"learning_rate": 4.0334855403348554e-05,
|
| 428 |
+
"loss": 0.5559,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.2785388127853881,
|
| 433 |
+
"grad_norm": 8.040207862854004,
|
| 434 |
+
"learning_rate": 4.0081177067478435e-05,
|
| 435 |
+
"loss": 0.5126,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.2831050228310502,
|
| 440 |
+
"grad_norm": 5.092132568359375,
|
| 441 |
+
"learning_rate": 3.982749873160832e-05,
|
| 442 |
+
"loss": 0.5201,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2876712328767123,
|
| 447 |
+
"grad_norm": 9.829987525939941,
|
| 448 |
+
"learning_rate": 3.95738203957382e-05,
|
| 449 |
+
"loss": 0.4563,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.2922374429223744,
|
| 454 |
+
"grad_norm": 7.451713562011719,
|
| 455 |
+
"learning_rate": 3.932014205986809e-05,
|
| 456 |
+
"loss": 0.7734,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.2968036529680365,
|
| 461 |
+
"grad_norm": 9.06767463684082,
|
| 462 |
+
"learning_rate": 3.906646372399797e-05,
|
| 463 |
+
"loss": 0.4621,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.3013698630136986,
|
| 468 |
+
"grad_norm": 8.293368339538574,
|
| 469 |
+
"learning_rate": 3.881278538812785e-05,
|
| 470 |
+
"loss": 0.5231,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.3059360730593607,
|
| 475 |
+
"grad_norm": 9.901214599609375,
|
| 476 |
+
"learning_rate": 3.855910705225774e-05,
|
| 477 |
+
"loss": 0.6035,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.3105022831050228,
|
| 482 |
+
"grad_norm": 10.380690574645996,
|
| 483 |
+
"learning_rate": 3.830542871638762e-05,
|
| 484 |
+
"loss": 0.3312,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.3150684931506849,
|
| 489 |
+
"grad_norm": 12.21066665649414,
|
| 490 |
+
"learning_rate": 3.80517503805175e-05,
|
| 491 |
+
"loss": 0.5626,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.319634703196347,
|
| 496 |
+
"grad_norm": 8.75660514831543,
|
| 497 |
+
"learning_rate": 3.779807204464739e-05,
|
| 498 |
+
"loss": 0.5877,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.3242009132420091,
|
| 503 |
+
"grad_norm": 5.634203910827637,
|
| 504 |
+
"learning_rate": 3.754439370877727e-05,
|
| 505 |
+
"loss": 0.7606,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.3287671232876712,
|
| 510 |
+
"grad_norm": 4.741566181182861,
|
| 511 |
+
"learning_rate": 3.7290715372907156e-05,
|
| 512 |
+
"loss": 0.5575,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.3333333333333333,
|
| 517 |
+
"grad_norm": 5.109163284301758,
|
| 518 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 519 |
+
"loss": 0.4544,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.3378995433789954,
|
| 524 |
+
"grad_norm": 10.15185832977295,
|
| 525 |
+
"learning_rate": 3.678335870116692e-05,
|
| 526 |
+
"loss": 0.408,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.3424657534246575,
|
| 531 |
+
"grad_norm": 8.459287643432617,
|
| 532 |
+
"learning_rate": 3.6529680365296805e-05,
|
| 533 |
+
"loss": 0.518,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.3470319634703196,
|
| 538 |
+
"grad_norm": 4.05381441116333,
|
| 539 |
+
"learning_rate": 3.6276002029426685e-05,
|
| 540 |
+
"loss": 0.5425,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.3515981735159817,
|
| 545 |
+
"grad_norm": 7.6227335929870605,
|
| 546 |
+
"learning_rate": 3.602232369355657e-05,
|
| 547 |
+
"loss": 0.4797,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.3561643835616438,
|
| 552 |
+
"grad_norm": 4.937866687774658,
|
| 553 |
+
"learning_rate": 3.5768645357686453e-05,
|
| 554 |
+
"loss": 0.447,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3607305936073059,
|
| 559 |
+
"grad_norm": 6.663622856140137,
|
| 560 |
+
"learning_rate": 3.5514967021816334e-05,
|
| 561 |
+
"loss": 0.6484,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.365296803652968,
|
| 566 |
+
"grad_norm": 2.132852792739868,
|
| 567 |
+
"learning_rate": 3.526128868594622e-05,
|
| 568 |
+
"loss": 0.5992,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.3698630136986301,
|
| 573 |
+
"grad_norm": 3.264267921447754,
|
| 574 |
+
"learning_rate": 3.50076103500761e-05,
|
| 575 |
+
"loss": 0.3265,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.3744292237442922,
|
| 580 |
+
"grad_norm": 1.495285153388977,
|
| 581 |
+
"learning_rate": 3.475393201420599e-05,
|
| 582 |
+
"loss": 0.2921,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.3789954337899543,
|
| 587 |
+
"grad_norm": 1.2416194677352905,
|
| 588 |
+
"learning_rate": 3.450025367833587e-05,
|
| 589 |
+
"loss": 0.3851,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.3835616438356164,
|
| 594 |
+
"grad_norm": 6.346595287322998,
|
| 595 |
+
"learning_rate": 3.424657534246575e-05,
|
| 596 |
+
"loss": 0.3337,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.3881278538812785,
|
| 601 |
+
"grad_norm": 3.725011110305786,
|
| 602 |
+
"learning_rate": 3.399289700659564e-05,
|
| 603 |
+
"loss": 0.3301,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.3926940639269406,
|
| 608 |
+
"grad_norm": 5.670325756072998,
|
| 609 |
+
"learning_rate": 3.373921867072552e-05,
|
| 610 |
+
"loss": 0.499,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.3972602739726027,
|
| 615 |
+
"grad_norm": 7.0639424324035645,
|
| 616 |
+
"learning_rate": 3.34855403348554e-05,
|
| 617 |
+
"loss": 0.3419,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.4004566210045662,
|
| 622 |
+
"eval_accuracy": 0.5540398740818469,
|
| 623 |
+
"eval_loss": 1.5632456541061401,
|
| 624 |
+
"eval_runtime": 168.0602,
|
| 625 |
+
"eval_samples_per_second": 5.671,
|
| 626 |
+
"eval_steps_per_second": 2.838,
|
| 627 |
+
"step": 877
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 1.0013698630136987,
|
| 631 |
+
"grad_norm": 2.454108953475952,
|
| 632 |
+
"learning_rate": 3.323186199898529e-05,
|
| 633 |
+
"loss": 0.334,
|
| 634 |
+
"step": 880
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 1.0059360730593607,
|
| 638 |
+
"grad_norm": 12.557369232177734,
|
| 639 |
+
"learning_rate": 3.297818366311517e-05,
|
| 640 |
+
"loss": 0.2568,
|
| 641 |
+
"step": 890
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.0105022831050228,
|
| 645 |
+
"grad_norm": 4.680070400238037,
|
| 646 |
+
"learning_rate": 3.2724505327245055e-05,
|
| 647 |
+
"loss": 0.2696,
|
| 648 |
+
"step": 900
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 1.015068493150685,
|
| 652 |
+
"grad_norm": 2.2027194499969482,
|
| 653 |
+
"learning_rate": 3.2470826991374936e-05,
|
| 654 |
+
"loss": 0.4776,
|
| 655 |
+
"step": 910
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.019634703196347,
|
| 659 |
+
"grad_norm": 8.404707908630371,
|
| 660 |
+
"learning_rate": 3.221714865550482e-05,
|
| 661 |
+
"loss": 0.2588,
|
| 662 |
+
"step": 920
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.0242009132420091,
|
| 666 |
+
"grad_norm": 1.3184863328933716,
|
| 667 |
+
"learning_rate": 3.1963470319634704e-05,
|
| 668 |
+
"loss": 0.2601,
|
| 669 |
+
"step": 930
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 1.0287671232876712,
|
| 673 |
+
"grad_norm": 3.3989500999450684,
|
| 674 |
+
"learning_rate": 3.1709791983764585e-05,
|
| 675 |
+
"loss": 0.3734,
|
| 676 |
+
"step": 940
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 1.0333333333333334,
|
| 680 |
+
"grad_norm": 6.5863142013549805,
|
| 681 |
+
"learning_rate": 3.145611364789447e-05,
|
| 682 |
+
"loss": 0.4799,
|
| 683 |
+
"step": 950
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 1.0378995433789955,
|
| 687 |
+
"grad_norm": 5.605319023132324,
|
| 688 |
+
"learning_rate": 3.120243531202435e-05,
|
| 689 |
+
"loss": 0.2629,
|
| 690 |
+
"step": 960
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 1.0424657534246575,
|
| 694 |
+
"grad_norm": 5.677365779876709,
|
| 695 |
+
"learning_rate": 3.0948756976154234e-05,
|
| 696 |
+
"loss": 0.4172,
|
| 697 |
+
"step": 970
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.0470319634703196,
|
| 701 |
+
"grad_norm": 6.943356037139893,
|
| 702 |
+
"learning_rate": 3.069507864028412e-05,
|
| 703 |
+
"loss": 0.4699,
|
| 704 |
+
"step": 980
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.0515981735159818,
|
| 708 |
+
"grad_norm": 1.7512027025222778,
|
| 709 |
+
"learning_rate": 3.0441400304414e-05,
|
| 710 |
+
"loss": 0.1905,
|
| 711 |
+
"step": 990
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 1.0561643835616439,
|
| 715 |
+
"grad_norm": 3.9878628253936768,
|
| 716 |
+
"learning_rate": 3.0187721968543886e-05,
|
| 717 |
+
"loss": 0.4574,
|
| 718 |
+
"step": 1000
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.060730593607306,
|
| 722 |
+
"grad_norm": 1.668309211730957,
|
| 723 |
+
"learning_rate": 2.993404363267377e-05,
|
| 724 |
+
"loss": 0.3441,
|
| 725 |
+
"step": 1010
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.065296803652968,
|
| 729 |
+
"grad_norm": 14.312126159667969,
|
| 730 |
+
"learning_rate": 2.9680365296803654e-05,
|
| 731 |
+
"loss": 0.4114,
|
| 732 |
+
"step": 1020
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.0698630136986302,
|
| 736 |
+
"grad_norm": 2.793670654296875,
|
| 737 |
+
"learning_rate": 2.9426686960933534e-05,
|
| 738 |
+
"loss": 0.292,
|
| 739 |
+
"step": 1030
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.0744292237442923,
|
| 743 |
+
"grad_norm": 0.9054046273231506,
|
| 744 |
+
"learning_rate": 2.917300862506342e-05,
|
| 745 |
+
"loss": 0.2573,
|
| 746 |
+
"step": 1040
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.0789954337899543,
|
| 750 |
+
"grad_norm": 3.951117515563965,
|
| 751 |
+
"learning_rate": 2.8919330289193303e-05,
|
| 752 |
+
"loss": 0.4457,
|
| 753 |
+
"step": 1050
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 1.0835616438356164,
|
| 757 |
+
"grad_norm": 10.71568775177002,
|
| 758 |
+
"learning_rate": 2.8665651953323187e-05,
|
| 759 |
+
"loss": 0.2332,
|
| 760 |
+
"step": 1060
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.0881278538812786,
|
| 764 |
+
"grad_norm": 5.735054969787598,
|
| 765 |
+
"learning_rate": 2.841197361745307e-05,
|
| 766 |
+
"loss": 0.2876,
|
| 767 |
+
"step": 1070
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 1.0926940639269407,
|
| 771 |
+
"grad_norm": 5.521303176879883,
|
| 772 |
+
"learning_rate": 2.815829528158295e-05,
|
| 773 |
+
"loss": 0.3894,
|
| 774 |
+
"step": 1080
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 1.0972602739726027,
|
| 778 |
+
"grad_norm": 10.683832168579102,
|
| 779 |
+
"learning_rate": 2.7904616945712835e-05,
|
| 780 |
+
"loss": 0.3966,
|
| 781 |
+
"step": 1090
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.1018264840182648,
|
| 785 |
+
"grad_norm": 1.0494991540908813,
|
| 786 |
+
"learning_rate": 2.765093860984272e-05,
|
| 787 |
+
"loss": 0.3233,
|
| 788 |
+
"step": 1100
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.106392694063927,
|
| 792 |
+
"grad_norm": 4.736073017120361,
|
| 793 |
+
"learning_rate": 2.7397260273972603e-05,
|
| 794 |
+
"loss": 0.3742,
|
| 795 |
+
"step": 1110
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 1.110958904109589,
|
| 799 |
+
"grad_norm": 8.027792930603027,
|
| 800 |
+
"learning_rate": 2.7143581938102484e-05,
|
| 801 |
+
"loss": 0.3813,
|
| 802 |
+
"step": 1120
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.1155251141552511,
|
| 806 |
+
"grad_norm": 2.795949935913086,
|
| 807 |
+
"learning_rate": 2.6889903602232368e-05,
|
| 808 |
+
"loss": 0.2194,
|
| 809 |
+
"step": 1130
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 1.1200913242009132,
|
| 813 |
+
"grad_norm": 16.23459815979004,
|
| 814 |
+
"learning_rate": 2.6636225266362252e-05,
|
| 815 |
+
"loss": 0.4115,
|
| 816 |
+
"step": 1140
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 1.1246575342465754,
|
| 820 |
+
"grad_norm": 8.320725440979004,
|
| 821 |
+
"learning_rate": 2.6382546930492136e-05,
|
| 822 |
+
"loss": 0.2755,
|
| 823 |
+
"step": 1150
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.1292237442922375,
|
| 827 |
+
"grad_norm": 2.192859649658203,
|
| 828 |
+
"learning_rate": 2.612886859462202e-05,
|
| 829 |
+
"loss": 0.2373,
|
| 830 |
+
"step": 1160
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.1337899543378995,
|
| 834 |
+
"grad_norm": 6.194462299346924,
|
| 835 |
+
"learning_rate": 2.58751902587519e-05,
|
| 836 |
+
"loss": 0.2048,
|
| 837 |
+
"step": 1170
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 1.1383561643835616,
|
| 841 |
+
"grad_norm": 6.522688388824463,
|
| 842 |
+
"learning_rate": 2.5621511922881785e-05,
|
| 843 |
+
"loss": 0.2106,
|
| 844 |
+
"step": 1180
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.1429223744292236,
|
| 848 |
+
"grad_norm": 7.976909160614014,
|
| 849 |
+
"learning_rate": 2.536783358701167e-05,
|
| 850 |
+
"loss": 0.1977,
|
| 851 |
+
"step": 1190
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 1.1474885844748859,
|
| 855 |
+
"grad_norm": 10.373821258544922,
|
| 856 |
+
"learning_rate": 2.5114155251141553e-05,
|
| 857 |
+
"loss": 0.4609,
|
| 858 |
+
"step": 1200
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.152054794520548,
|
| 862 |
+
"grad_norm": 8.701196670532227,
|
| 863 |
+
"learning_rate": 2.4860476915271437e-05,
|
| 864 |
+
"loss": 0.3093,
|
| 865 |
+
"step": 1210
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.15662100456621,
|
| 869 |
+
"grad_norm": 0.8994976878166199,
|
| 870 |
+
"learning_rate": 2.4606798579401318e-05,
|
| 871 |
+
"loss": 0.1242,
|
| 872 |
+
"step": 1220
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.1611872146118722,
|
| 876 |
+
"grad_norm": 4.23283576965332,
|
| 877 |
+
"learning_rate": 2.4353120243531202e-05,
|
| 878 |
+
"loss": 0.1627,
|
| 879 |
+
"step": 1230
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.1657534246575343,
|
| 883 |
+
"grad_norm": 4.330513000488281,
|
| 884 |
+
"learning_rate": 2.4099441907661086e-05,
|
| 885 |
+
"loss": 0.3118,
|
| 886 |
+
"step": 1240
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.1703196347031963,
|
| 890 |
+
"grad_norm": 2.304389715194702,
|
| 891 |
+
"learning_rate": 2.384576357179097e-05,
|
| 892 |
+
"loss": 0.218,
|
| 893 |
+
"step": 1250
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.1748858447488584,
|
| 897 |
+
"grad_norm": 6.340790271759033,
|
| 898 |
+
"learning_rate": 2.359208523592085e-05,
|
| 899 |
+
"loss": 0.2833,
|
| 900 |
+
"step": 1260
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 1.1794520547945206,
|
| 904 |
+
"grad_norm": 2.6649558544158936,
|
| 905 |
+
"learning_rate": 2.3338406900050735e-05,
|
| 906 |
+
"loss": 0.2335,
|
| 907 |
+
"step": 1270
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.1840182648401827,
|
| 911 |
+
"grad_norm": 0.8412348628044128,
|
| 912 |
+
"learning_rate": 2.308472856418062e-05,
|
| 913 |
+
"loss": 0.1889,
|
| 914 |
+
"step": 1280
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.1885844748858447,
|
| 918 |
+
"grad_norm": 5.932245254516602,
|
| 919 |
+
"learning_rate": 2.2831050228310503e-05,
|
| 920 |
+
"loss": 0.2006,
|
| 921 |
+
"step": 1290
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 1.1931506849315068,
|
| 925 |
+
"grad_norm": 9.092105865478516,
|
| 926 |
+
"learning_rate": 2.2577371892440387e-05,
|
| 927 |
+
"loss": 0.3411,
|
| 928 |
+
"step": 1300
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.197716894977169,
|
| 932 |
+
"grad_norm": 6.397707939147949,
|
| 933 |
+
"learning_rate": 2.2323693556570268e-05,
|
| 934 |
+
"loss": 0.1739,
|
| 935 |
+
"step": 1310
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.202283105022831,
|
| 939 |
+
"grad_norm": 3.637505292892456,
|
| 940 |
+
"learning_rate": 2.207001522070015e-05,
|
| 941 |
+
"loss": 0.2351,
|
| 942 |
+
"step": 1320
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 1.2068493150684931,
|
| 946 |
+
"grad_norm": 3.8445942401885986,
|
| 947 |
+
"learning_rate": 2.1816336884830036e-05,
|
| 948 |
+
"loss": 0.3909,
|
| 949 |
+
"step": 1330
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.2114155251141552,
|
| 953 |
+
"grad_norm": 1.331699252128601,
|
| 954 |
+
"learning_rate": 2.156265854895992e-05,
|
| 955 |
+
"loss": 0.154,
|
| 956 |
+
"step": 1340
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.2159817351598174,
|
| 960 |
+
"grad_norm": 9.704405784606934,
|
| 961 |
+
"learning_rate": 2.13089802130898e-05,
|
| 962 |
+
"loss": 0.1631,
|
| 963 |
+
"step": 1350
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 1.2205479452054795,
|
| 967 |
+
"grad_norm": 2.271298885345459,
|
| 968 |
+
"learning_rate": 2.1055301877219685e-05,
|
| 969 |
+
"loss": 0.1551,
|
| 970 |
+
"step": 1360
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.2251141552511415,
|
| 974 |
+
"grad_norm": 10.894644737243652,
|
| 975 |
+
"learning_rate": 2.080162354134957e-05,
|
| 976 |
+
"loss": 0.3072,
|
| 977 |
+
"step": 1370
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.2296803652968036,
|
| 981 |
+
"grad_norm": 10.65208625793457,
|
| 982 |
+
"learning_rate": 2.0547945205479453e-05,
|
| 983 |
+
"loss": 0.2142,
|
| 984 |
+
"step": 1380
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.2342465753424658,
|
| 988 |
+
"grad_norm": 15.530343055725098,
|
| 989 |
+
"learning_rate": 2.0294266869609337e-05,
|
| 990 |
+
"loss": 0.2107,
|
| 991 |
+
"step": 1390
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.238812785388128,
|
| 995 |
+
"grad_norm": 7.166269779205322,
|
| 996 |
+
"learning_rate": 2.0040588533739217e-05,
|
| 997 |
+
"loss": 0.2599,
|
| 998 |
+
"step": 1400
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.24337899543379,
|
| 1002 |
+
"grad_norm": 1.9615888595581055,
|
| 1003 |
+
"learning_rate": 1.97869101978691e-05,
|
| 1004 |
+
"loss": 0.1511,
|
| 1005 |
+
"step": 1410
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 1.247945205479452,
|
| 1009 |
+
"grad_norm": 6.840981960296631,
|
| 1010 |
+
"learning_rate": 1.9533231861998985e-05,
|
| 1011 |
+
"loss": 0.3887,
|
| 1012 |
+
"step": 1420
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 1.252511415525114,
|
| 1016 |
+
"grad_norm": 5.138172626495361,
|
| 1017 |
+
"learning_rate": 1.927955352612887e-05,
|
| 1018 |
+
"loss": 0.2284,
|
| 1019 |
+
"step": 1430
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.2570776255707763,
|
| 1023 |
+
"grad_norm": 5.770719051361084,
|
| 1024 |
+
"learning_rate": 1.902587519025875e-05,
|
| 1025 |
+
"loss": 0.3119,
|
| 1026 |
+
"step": 1440
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 1.2616438356164383,
|
| 1030 |
+
"grad_norm": 1.5914487838745117,
|
| 1031 |
+
"learning_rate": 1.8772196854388634e-05,
|
| 1032 |
+
"loss": 0.0873,
|
| 1033 |
+
"step": 1450
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.2662100456621004,
|
| 1037 |
+
"grad_norm": 2.5335633754730225,
|
| 1038 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 1039 |
+
"loss": 0.2629,
|
| 1040 |
+
"step": 1460
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 1.2707762557077626,
|
| 1044 |
+
"grad_norm": 10.850088119506836,
|
| 1045 |
+
"learning_rate": 1.8264840182648402e-05,
|
| 1046 |
+
"loss": 0.3098,
|
| 1047 |
+
"step": 1470
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 1.2753424657534247,
|
| 1051 |
+
"grad_norm": 2.9598915576934814,
|
| 1052 |
+
"learning_rate": 1.8011161846778286e-05,
|
| 1053 |
+
"loss": 0.0882,
|
| 1054 |
+
"step": 1480
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 1.2799086757990867,
|
| 1058 |
+
"grad_norm": 2.021665096282959,
|
| 1059 |
+
"learning_rate": 1.7757483510908167e-05,
|
| 1060 |
+
"loss": 0.2079,
|
| 1061 |
+
"step": 1490
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.2844748858447488,
|
| 1065 |
+
"grad_norm": 6.0909833908081055,
|
| 1066 |
+
"learning_rate": 1.750380517503805e-05,
|
| 1067 |
+
"loss": 0.1287,
|
| 1068 |
+
"step": 1500
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 1.2890410958904108,
|
| 1072 |
+
"grad_norm": 7.84592866897583,
|
| 1073 |
+
"learning_rate": 1.7250126839167935e-05,
|
| 1074 |
+
"loss": 0.1406,
|
| 1075 |
+
"step": 1510
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.293607305936073,
|
| 1079 |
+
"grad_norm": 3.269815444946289,
|
| 1080 |
+
"learning_rate": 1.699644850329782e-05,
|
| 1081 |
+
"loss": 0.2997,
|
| 1082 |
+
"step": 1520
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 1.2981735159817351,
|
| 1086 |
+
"grad_norm": 0.7504135966300964,
|
| 1087 |
+
"learning_rate": 1.67427701674277e-05,
|
| 1088 |
+
"loss": 0.1724,
|
| 1089 |
+
"step": 1530
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.3027397260273972,
|
| 1093 |
+
"grad_norm": 2.8693363666534424,
|
| 1094 |
+
"learning_rate": 1.6489091831557584e-05,
|
| 1095 |
+
"loss": 0.1543,
|
| 1096 |
+
"step": 1540
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 1.3073059360730594,
|
| 1100 |
+
"grad_norm": 6.736435890197754,
|
| 1101 |
+
"learning_rate": 1.6235413495687468e-05,
|
| 1102 |
+
"loss": 0.1396,
|
| 1103 |
+
"step": 1550
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 1.3118721461187215,
|
| 1107 |
+
"grad_norm": 0.3472974896430969,
|
| 1108 |
+
"learning_rate": 1.5981735159817352e-05,
|
| 1109 |
+
"loss": 0.2981,
|
| 1110 |
+
"step": 1560
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 1.3164383561643835,
|
| 1114 |
+
"grad_norm": 4.580450057983398,
|
| 1115 |
+
"learning_rate": 1.5728056823947236e-05,
|
| 1116 |
+
"loss": 0.2362,
|
| 1117 |
+
"step": 1570
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.3210045662100456,
|
| 1121 |
+
"grad_norm": 1.7592848539352417,
|
| 1122 |
+
"learning_rate": 1.5474378488077117e-05,
|
| 1123 |
+
"loss": 0.2429,
|
| 1124 |
+
"step": 1580
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.3255707762557076,
|
| 1128 |
+
"grad_norm": 1.0239814519882202,
|
| 1129 |
+
"learning_rate": 1.5220700152207e-05,
|
| 1130 |
+
"loss": 0.1309,
|
| 1131 |
+
"step": 1590
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 1.33013698630137,
|
| 1135 |
+
"grad_norm": 1.68032705783844,
|
| 1136 |
+
"learning_rate": 1.4967021816336885e-05,
|
| 1137 |
+
"loss": 0.2172,
|
| 1138 |
+
"step": 1600
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 1.334703196347032,
|
| 1142 |
+
"grad_norm": 5.960744380950928,
|
| 1143 |
+
"learning_rate": 1.4713343480466767e-05,
|
| 1144 |
+
"loss": 0.2709,
|
| 1145 |
+
"step": 1610
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 1.339269406392694,
|
| 1149 |
+
"grad_norm": 9.478919982910156,
|
| 1150 |
+
"learning_rate": 1.4459665144596651e-05,
|
| 1151 |
+
"loss": 0.2803,
|
| 1152 |
+
"step": 1620
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 1.3438356164383563,
|
| 1156 |
+
"grad_norm": 0.2718620002269745,
|
| 1157 |
+
"learning_rate": 1.4205986808726535e-05,
|
| 1158 |
+
"loss": 0.1481,
|
| 1159 |
+
"step": 1630
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.3484018264840183,
|
| 1163 |
+
"grad_norm": 0.5140727162361145,
|
| 1164 |
+
"learning_rate": 1.3952308472856418e-05,
|
| 1165 |
+
"loss": 0.1621,
|
| 1166 |
+
"step": 1640
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 1.3529680365296803,
|
| 1170 |
+
"grad_norm": 8.044154167175293,
|
| 1171 |
+
"learning_rate": 1.3698630136986302e-05,
|
| 1172 |
+
"loss": 0.2557,
|
| 1173 |
+
"step": 1650
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 1.3575342465753424,
|
| 1177 |
+
"grad_norm": 0.42045390605926514,
|
| 1178 |
+
"learning_rate": 1.3444951801116184e-05,
|
| 1179 |
+
"loss": 0.1946,
|
| 1180 |
+
"step": 1660
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 1.3621004566210044,
|
| 1184 |
+
"grad_norm": 1.3292018175125122,
|
| 1185 |
+
"learning_rate": 1.3191273465246068e-05,
|
| 1186 |
+
"loss": 0.1813,
|
| 1187 |
+
"step": 1670
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 1.3666666666666667,
|
| 1191 |
+
"grad_norm": 0.8583564758300781,
|
| 1192 |
+
"learning_rate": 1.293759512937595e-05,
|
| 1193 |
+
"loss": 0.1753,
|
| 1194 |
+
"step": 1680
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 1.3712328767123287,
|
| 1198 |
+
"grad_norm": 5.7116546630859375,
|
| 1199 |
+
"learning_rate": 1.2683916793505835e-05,
|
| 1200 |
+
"loss": 0.2727,
|
| 1201 |
+
"step": 1690
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.3757990867579908,
|
| 1205 |
+
"grad_norm": 0.348532497882843,
|
| 1206 |
+
"learning_rate": 1.2430238457635719e-05,
|
| 1207 |
+
"loss": 0.1961,
|
| 1208 |
+
"step": 1700
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 1.380365296803653,
|
| 1212 |
+
"grad_norm": 9.204560279846191,
|
| 1213 |
+
"learning_rate": 1.2176560121765601e-05,
|
| 1214 |
+
"loss": 0.257,
|
| 1215 |
+
"step": 1710
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.384931506849315,
|
| 1219 |
+
"grad_norm": 5.398329257965088,
|
| 1220 |
+
"learning_rate": 1.1922881785895485e-05,
|
| 1221 |
+
"loss": 0.1348,
|
| 1222 |
+
"step": 1720
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.3894977168949771,
|
| 1226 |
+
"grad_norm": 0.43983083963394165,
|
| 1227 |
+
"learning_rate": 1.1669203450025367e-05,
|
| 1228 |
+
"loss": 0.2507,
|
| 1229 |
+
"step": 1730
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 1.3940639269406392,
|
| 1233 |
+
"grad_norm": 0.9157253503799438,
|
| 1234 |
+
"learning_rate": 1.1415525114155251e-05,
|
| 1235 |
+
"loss": 0.193,
|
| 1236 |
+
"step": 1740
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 1.3986301369863012,
|
| 1240 |
+
"grad_norm": 0.855165421962738,
|
| 1241 |
+
"learning_rate": 1.1161846778285134e-05,
|
| 1242 |
+
"loss": 0.1291,
|
| 1243 |
+
"step": 1750
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.4004566210045661,
|
| 1247 |
+
"eval_accuracy": 0.6516264428121721,
|
| 1248 |
+
"eval_loss": 1.277839183807373,
|
| 1249 |
+
"eval_runtime": 178.2589,
|
| 1250 |
+
"eval_samples_per_second": 5.346,
|
| 1251 |
+
"eval_steps_per_second": 2.676,
|
| 1252 |
+
"step": 1754
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"epoch": 2.0027397260273974,
|
| 1256 |
+
"grad_norm": 8.00808334350586,
|
| 1257 |
+
"learning_rate": 1.0908168442415018e-05,
|
| 1258 |
+
"loss": 0.1597,
|
| 1259 |
+
"step": 1760
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.007305936073059,
|
| 1263 |
+
"grad_norm": 0.1681767702102661,
|
| 1264 |
+
"learning_rate": 1.06544901065449e-05,
|
| 1265 |
+
"loss": 0.0838,
|
| 1266 |
+
"step": 1770
|
| 1267 |
+
},
|
| 1268 |
+
{
|
| 1269 |
+
"epoch": 2.0118721461187214,
|
| 1270 |
+
"grad_norm": 12.16120719909668,
|
| 1271 |
+
"learning_rate": 1.0400811770674784e-05,
|
| 1272 |
+
"loss": 0.1621,
|
| 1273 |
+
"step": 1780
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 2.0164383561643837,
|
| 1277 |
+
"grad_norm": 1.8593238592147827,
|
| 1278 |
+
"learning_rate": 1.0147133434804668e-05,
|
| 1279 |
+
"loss": 0.0698,
|
| 1280 |
+
"step": 1790
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.0210045662100455,
|
| 1284 |
+
"grad_norm": 1.4579780101776123,
|
| 1285 |
+
"learning_rate": 9.89345509893455e-06,
|
| 1286 |
+
"loss": 0.2562,
|
| 1287 |
+
"step": 1800
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 2.025570776255708,
|
| 1291 |
+
"grad_norm": 2.344836711883545,
|
| 1292 |
+
"learning_rate": 9.639776763064435e-06,
|
| 1293 |
+
"loss": 0.257,
|
| 1294 |
+
"step": 1810
|
| 1295 |
+
},
|
| 1296 |
+
{
|
| 1297 |
+
"epoch": 2.03013698630137,
|
| 1298 |
+
"grad_norm": 8.14854621887207,
|
| 1299 |
+
"learning_rate": 9.386098427194317e-06,
|
| 1300 |
+
"loss": 0.2149,
|
| 1301 |
+
"step": 1820
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.034703196347032,
|
| 1305 |
+
"grad_norm": 11.254741668701172,
|
| 1306 |
+
"learning_rate": 9.132420091324201e-06,
|
| 1307 |
+
"loss": 0.2175,
|
| 1308 |
+
"step": 1830
|
| 1309 |
+
},
|
| 1310 |
+
{
|
| 1311 |
+
"epoch": 2.039269406392694,
|
| 1312 |
+
"grad_norm": 8.984929084777832,
|
| 1313 |
+
"learning_rate": 8.878741755454084e-06,
|
| 1314 |
+
"loss": 0.1927,
|
| 1315 |
+
"step": 1840
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 2.043835616438356,
|
| 1319 |
+
"grad_norm": 13.033331871032715,
|
| 1320 |
+
"learning_rate": 8.625063419583968e-06,
|
| 1321 |
+
"loss": 0.2202,
|
| 1322 |
+
"step": 1850
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.0484018264840183,
|
| 1326 |
+
"grad_norm": 1.0010125637054443,
|
| 1327 |
+
"learning_rate": 8.37138508371385e-06,
|
| 1328 |
+
"loss": 0.1063,
|
| 1329 |
+
"step": 1860
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 2.0529680365296805,
|
| 1333 |
+
"grad_norm": 0.2957708239555359,
|
| 1334 |
+
"learning_rate": 8.117706747843734e-06,
|
| 1335 |
+
"loss": 0.1509,
|
| 1336 |
+
"step": 1870
|
| 1337 |
+
},
|
| 1338 |
+
{
|
| 1339 |
+
"epoch": 2.0575342465753423,
|
| 1340 |
+
"grad_norm": 0.20592088997364044,
|
| 1341 |
+
"learning_rate": 7.864028411973618e-06,
|
| 1342 |
+
"loss": 0.1519,
|
| 1343 |
+
"step": 1880
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.0621004566210046,
|
| 1347 |
+
"grad_norm": 2.5575947761535645,
|
| 1348 |
+
"learning_rate": 7.6103500761035e-06,
|
| 1349 |
+
"loss": 0.1487,
|
| 1350 |
+
"step": 1890
|
| 1351 |
+
},
|
| 1352 |
+
{
|
| 1353 |
+
"epoch": 2.066666666666667,
|
| 1354 |
+
"grad_norm": 3.2097675800323486,
|
| 1355 |
+
"learning_rate": 7.356671740233384e-06,
|
| 1356 |
+
"loss": 0.1506,
|
| 1357 |
+
"step": 1900
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 2.0712328767123287,
|
| 1361 |
+
"grad_norm": 7.255710124969482,
|
| 1362 |
+
"learning_rate": 7.102993404363268e-06,
|
| 1363 |
+
"loss": 0.2091,
|
| 1364 |
+
"step": 1910
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 2.075799086757991,
|
| 1368 |
+
"grad_norm": 0.6733071804046631,
|
| 1369 |
+
"learning_rate": 6.849315068493151e-06,
|
| 1370 |
+
"loss": 0.1179,
|
| 1371 |
+
"step": 1920
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 2.080365296803653,
|
| 1375 |
+
"grad_norm": 5.307835102081299,
|
| 1376 |
+
"learning_rate": 6.595636732623034e-06,
|
| 1377 |
+
"loss": 0.2213,
|
| 1378 |
+
"step": 1930
|
| 1379 |
+
},
|
| 1380 |
+
{
|
| 1381 |
+
"epoch": 2.084931506849315,
|
| 1382 |
+
"grad_norm": 2.307908535003662,
|
| 1383 |
+
"learning_rate": 6.341958396752917e-06,
|
| 1384 |
+
"loss": 0.0695,
|
| 1385 |
+
"step": 1940
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.0894977168949773,
|
| 1389 |
+
"grad_norm": 9.304193496704102,
|
| 1390 |
+
"learning_rate": 6.0882800608828005e-06,
|
| 1391 |
+
"loss": 0.13,
|
| 1392 |
+
"step": 1950
|
| 1393 |
+
},
|
| 1394 |
+
{
|
| 1395 |
+
"epoch": 2.094063926940639,
|
| 1396 |
+
"grad_norm": 0.5198363661766052,
|
| 1397 |
+
"learning_rate": 5.834601725012684e-06,
|
| 1398 |
+
"loss": 0.1376,
|
| 1399 |
+
"step": 1960
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 2.0986301369863014,
|
| 1403 |
+
"grad_norm": 4.95096492767334,
|
| 1404 |
+
"learning_rate": 5.580923389142567e-06,
|
| 1405 |
+
"loss": 0.1322,
|
| 1406 |
+
"step": 1970
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 2.1031963470319637,
|
| 1410 |
+
"grad_norm": 4.969466686248779,
|
| 1411 |
+
"learning_rate": 5.32724505327245e-06,
|
| 1412 |
+
"loss": 0.1839,
|
| 1413 |
+
"step": 1980
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 2.1077625570776255,
|
| 1417 |
+
"grad_norm": 0.14146043360233307,
|
| 1418 |
+
"learning_rate": 5.073566717402334e-06,
|
| 1419 |
+
"loss": 0.2181,
|
| 1420 |
+
"step": 1990
|
| 1421 |
+
},
|
| 1422 |
+
{
|
| 1423 |
+
"epoch": 2.1123287671232878,
|
| 1424 |
+
"grad_norm": 0.31362447142601013,
|
| 1425 |
+
"learning_rate": 4.819888381532217e-06,
|
| 1426 |
+
"loss": 0.2004,
|
| 1427 |
+
"step": 2000
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.1168949771689496,
|
| 1431 |
+
"grad_norm": 9.072457313537598,
|
| 1432 |
+
"learning_rate": 4.566210045662101e-06,
|
| 1433 |
+
"loss": 0.158,
|
| 1434 |
+
"step": 2010
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"epoch": 2.121461187214612,
|
| 1438 |
+
"grad_norm": 6.152866363525391,
|
| 1439 |
+
"learning_rate": 4.312531709791984e-06,
|
| 1440 |
+
"loss": 0.1529,
|
| 1441 |
+
"step": 2020
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 2.126027397260274,
|
| 1445 |
+
"grad_norm": 2.435375452041626,
|
| 1446 |
+
"learning_rate": 4.058853373921867e-06,
|
| 1447 |
+
"loss": 0.172,
|
| 1448 |
+
"step": 2030
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 2.130593607305936,
|
| 1452 |
+
"grad_norm": 7.807186603546143,
|
| 1453 |
+
"learning_rate": 3.80517503805175e-06,
|
| 1454 |
+
"loss": 0.1801,
|
| 1455 |
+
"step": 2040
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 2.135159817351598,
|
| 1459 |
+
"grad_norm": 2.6992197036743164,
|
| 1460 |
+
"learning_rate": 3.551496702181634e-06,
|
| 1461 |
+
"loss": 0.1777,
|
| 1462 |
+
"step": 2050
|
| 1463 |
+
},
|
| 1464 |
+
{
|
| 1465 |
+
"epoch": 2.1397260273972605,
|
| 1466 |
+
"grad_norm": 9.397564888000488,
|
| 1467 |
+
"learning_rate": 3.297818366311517e-06,
|
| 1468 |
+
"loss": 0.0939,
|
| 1469 |
+
"step": 2060
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.1442922374429223,
|
| 1473 |
+
"grad_norm": 3.4112205505371094,
|
| 1474 |
+
"learning_rate": 3.0441400304414002e-06,
|
| 1475 |
+
"loss": 0.1201,
|
| 1476 |
+
"step": 2070
|
| 1477 |
+
},
|
| 1478 |
+
{
|
| 1479 |
+
"epoch": 2.1488584474885846,
|
| 1480 |
+
"grad_norm": 0.36758846044540405,
|
| 1481 |
+
"learning_rate": 2.7904616945712835e-06,
|
| 1482 |
+
"loss": 0.167,
|
| 1483 |
+
"step": 2080
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 2.1534246575342464,
|
| 1487 |
+
"grad_norm": 0.20398499071598053,
|
| 1488 |
+
"learning_rate": 2.536783358701167e-06,
|
| 1489 |
+
"loss": 0.0453,
|
| 1490 |
+
"step": 2090
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 2.1579908675799087,
|
| 1494 |
+
"grad_norm": 0.14668098092079163,
|
| 1495 |
+
"learning_rate": 2.2831050228310503e-06,
|
| 1496 |
+
"loss": 0.0674,
|
| 1497 |
+
"step": 2100
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 2.162557077625571,
|
| 1501 |
+
"grad_norm": 3.2575571537017822,
|
| 1502 |
+
"learning_rate": 2.0294266869609335e-06,
|
| 1503 |
+
"loss": 0.0917,
|
| 1504 |
+
"step": 2110
|
| 1505 |
+
},
|
| 1506 |
+
{
|
| 1507 |
+
"epoch": 2.1671232876712327,
|
| 1508 |
+
"grad_norm": 0.5656366348266602,
|
| 1509 |
+
"learning_rate": 1.775748351090817e-06,
|
| 1510 |
+
"loss": 0.1194,
|
| 1511 |
+
"step": 2120
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.171689497716895,
|
| 1515 |
+
"grad_norm": 7.743866443634033,
|
| 1516 |
+
"learning_rate": 1.5220700152207001e-06,
|
| 1517 |
+
"loss": 0.1199,
|
| 1518 |
+
"step": 2130
|
| 1519 |
+
},
|
| 1520 |
+
{
|
| 1521 |
+
"epoch": 2.1762557077625573,
|
| 1522 |
+
"grad_norm": 1.3446861505508423,
|
| 1523 |
+
"learning_rate": 1.2683916793505835e-06,
|
| 1524 |
+
"loss": 0.0521,
|
| 1525 |
+
"step": 2140
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 2.180821917808219,
|
| 1529 |
+
"grad_norm": 0.33492910861968994,
|
| 1530 |
+
"learning_rate": 1.0147133434804667e-06,
|
| 1531 |
+
"loss": 0.0444,
|
| 1532 |
+
"step": 2150
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 2.1853881278538814,
|
| 1536 |
+
"grad_norm": 0.12891508638858795,
|
| 1537 |
+
"learning_rate": 7.610350076103501e-07,
|
| 1538 |
+
"loss": 0.1172,
|
| 1539 |
+
"step": 2160
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 2.189954337899543,
|
| 1543 |
+
"grad_norm": 5.329843521118164,
|
| 1544 |
+
"learning_rate": 5.073566717402334e-07,
|
| 1545 |
+
"loss": 0.1813,
|
| 1546 |
+
"step": 2170
|
| 1547 |
+
},
|
| 1548 |
+
{
|
| 1549 |
+
"epoch": 2.1945205479452055,
|
| 1550 |
+
"grad_norm": 2.9594337940216064,
|
| 1551 |
+
"learning_rate": 2.536783358701167e-07,
|
| 1552 |
+
"loss": 0.2105,
|
| 1553 |
+
"step": 2180
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.1990867579908677,
|
| 1557 |
+
"grad_norm": 1.7173917293548584,
|
| 1558 |
+
"learning_rate": 0.0,
|
| 1559 |
+
"loss": 0.183,
|
| 1560 |
+
"step": 2190
|
| 1561 |
+
},
|
| 1562 |
+
{
|
| 1563 |
+
"epoch": 2.1990867579908677,
|
| 1564 |
+
"eval_accuracy": 0.6295907660020986,
|
| 1565 |
+
"eval_loss": 1.3431484699249268,
|
| 1566 |
+
"eval_runtime": 185.0677,
|
| 1567 |
+
"eval_samples_per_second": 5.149,
|
| 1568 |
+
"eval_steps_per_second": 2.577,
|
| 1569 |
+
"step": 2190
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 2.1990867579908677,
|
| 1573 |
+
"step": 2190,
|
| 1574 |
+
"total_flos": 1.5353993990024331e+19,
|
| 1575 |
+
"train_loss": 0.7283184543456117,
|
| 1576 |
+
"train_runtime": 5503.9878,
|
| 1577 |
+
"train_samples_per_second": 3.183,
|
| 1578 |
+
"train_steps_per_second": 0.398
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 2.1990867579908677,
|
| 1582 |
+
"eval_accuracy": 0.6516264428121721,
|
| 1583 |
+
"eval_loss": 1.277839183807373,
|
| 1584 |
+
"eval_runtime": 188.0153,
|
| 1585 |
+
"eval_samples_per_second": 5.069,
|
| 1586 |
+
"eval_steps_per_second": 2.537,
|
| 1587 |
+
"step": 2190
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 2.1990867579908677,
|
| 1591 |
+
"eval_accuracy": 0.7749803304484658,
|
| 1592 |
+
"eval_loss": 0.9055941104888916,
|
| 1593 |
+
"eval_runtime": 256.8953,
|
| 1594 |
+
"eval_samples_per_second": 4.948,
|
| 1595 |
+
"eval_steps_per_second": 2.476,
|
| 1596 |
+
"step": 2190
|
| 1597 |
+
}
|
| 1598 |
+
],
|
| 1599 |
+
"logging_steps": 10,
|
| 1600 |
+
"max_steps": 2190,
|
| 1601 |
+
"num_input_tokens_seen": 0,
|
| 1602 |
+
"num_train_epochs": 9223372036854775807,
|
| 1603 |
+
"save_steps": 500,
|
| 1604 |
+
"stateful_callbacks": {
|
| 1605 |
+
"EarlyStoppingCallback": {
|
| 1606 |
+
"args": {
|
| 1607 |
+
"early_stopping_patience": 2,
|
| 1608 |
+
"early_stopping_threshold": 0.0
|
| 1609 |
+
},
|
| 1610 |
+
"attributes": {
|
| 1611 |
+
"early_stopping_patience_counter": 0
|
| 1612 |
+
}
|
| 1613 |
+
},
|
| 1614 |
+
"TrainerControl": {
|
| 1615 |
+
"args": {
|
| 1616 |
+
"should_epoch_stop": false,
|
| 1617 |
+
"should_evaluate": false,
|
| 1618 |
+
"should_log": false,
|
| 1619 |
+
"should_save": true,
|
| 1620 |
+
"should_training_stop": true
|
| 1621 |
+
},
|
| 1622 |
+
"attributes": {}
|
| 1623 |
+
}
|
| 1624 |
+
},
|
| 1625 |
+
"total_flos": 1.5353993990024331e+19,
|
| 1626 |
+
"train_batch_size": 2,
|
| 1627 |
+
"trial_name": null,
|
| 1628 |
+
"trial_params": null
|
| 1629 |
+
}
|