Commit
•
6abcda9
1
Parent(s):
b7c4cf1
End of training
Browse files- README.md +2 -2
- all_results.json +14 -0
- eval_results.json +9 -0
- runs/Feb03_14-18-09_d83aeab94bd9/events.out.tfevents.1706971519.d83aeab94bd9.14705.1 +3 -0
- train_results.json +8 -0
- trainer_state.json +1275 -0
README.md
CHANGED
|
@@ -16,8 +16,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [aubmindlab/bert-base-arabertv2](https://huggingface.co/aubmindlab/bert-base-arabertv2) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss:
|
| 20 |
-
- Accuracy: 0.
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
|
|
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [aubmindlab/bert-base-arabertv2](https://huggingface.co/aubmindlab/bert-base-arabertv2) on an unknown dataset.
|
| 18 |
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.8943
|
| 20 |
+
- Accuracy: 0.65
|
| 21 |
|
| 22 |
## Model description
|
| 23 |
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 22.83,
|
| 3 |
+
"eval_accuracy": 0.65,
|
| 4 |
+
"eval_loss": 0.894333004951477,
|
| 5 |
+
"eval_runtime": 2.7455,
|
| 6 |
+
"eval_samples": 1000,
|
| 7 |
+
"eval_samples_per_second": 364.235,
|
| 8 |
+
"eval_steps_per_second": 45.529,
|
| 9 |
+
"train_loss": 0.2579876671291888,
|
| 10 |
+
"train_runtime": 1594.4621,
|
| 11 |
+
"train_samples": 7000,
|
| 12 |
+
"train_samples_per_second": 100.347,
|
| 13 |
+
"train_steps_per_second": 3.136
|
| 14 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 22.83,
|
| 3 |
+
"eval_accuracy": 0.65,
|
| 4 |
+
"eval_loss": 0.894333004951477,
|
| 5 |
+
"eval_runtime": 2.7455,
|
| 6 |
+
"eval_samples": 1000,
|
| 7 |
+
"eval_samples_per_second": 364.235,
|
| 8 |
+
"eval_steps_per_second": 45.529
|
| 9 |
+
}
|
runs/Feb03_14-18-09_d83aeab94bd9/events.out.tfevents.1706971519.d83aeab94bd9.14705.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:db5a160686c8602803d88e7b0287c00cc92a6e5fe33b4b6053e78c05031f6a38
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 22.83,
|
| 3 |
+
"train_loss": 0.2579876671291888,
|
| 4 |
+
"train_runtime": 1594.4621,
|
| 5 |
+
"train_samples": 7000,
|
| 6 |
+
"train_samples_per_second": 100.347,
|
| 7 |
+
"train_steps_per_second": 3.136
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1275 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.894333004951477,
|
| 3 |
+
"best_model_checkpoint": "bert-base-arabertv2_1/checkpoint-1000",
|
| 4 |
+
"epoch": 22.831050228310502,
|
| 5 |
+
"eval_steps": 1000,
|
| 6 |
+
"global_step": 5000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.11,
|
| 13 |
+
"learning_rate": 9.200000000000001e-07,
|
| 14 |
+
"loss": 1.0908,
|
| 15 |
+
"step": 25
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.23,
|
| 19 |
+
"learning_rate": 1.8400000000000002e-06,
|
| 20 |
+
"loss": 1.0474,
|
| 21 |
+
"step": 50
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.34,
|
| 25 |
+
"learning_rate": 2.7600000000000003e-06,
|
| 26 |
+
"loss": 0.9963,
|
| 27 |
+
"step": 75
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.46,
|
| 31 |
+
"learning_rate": 3.7600000000000004e-06,
|
| 32 |
+
"loss": 1.0041,
|
| 33 |
+
"step": 100
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.57,
|
| 37 |
+
"learning_rate": 4.76e-06,
|
| 38 |
+
"loss": 1.0071,
|
| 39 |
+
"step": 125
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.68,
|
| 43 |
+
"learning_rate": 5.76e-06,
|
| 44 |
+
"loss": 0.9659,
|
| 45 |
+
"step": 150
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8,
|
| 49 |
+
"learning_rate": 6.760000000000001e-06,
|
| 50 |
+
"loss": 0.9522,
|
| 51 |
+
"step": 175
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.91,
|
| 55 |
+
"learning_rate": 7.76e-06,
|
| 56 |
+
"loss": 0.9335,
|
| 57 |
+
"step": 200
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 1.03,
|
| 61 |
+
"learning_rate": 8.76e-06,
|
| 62 |
+
"loss": 0.9048,
|
| 63 |
+
"step": 225
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.14,
|
| 67 |
+
"learning_rate": 9.760000000000001e-06,
|
| 68 |
+
"loss": 0.8749,
|
| 69 |
+
"step": 250
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 1.26,
|
| 73 |
+
"learning_rate": 1.0760000000000002e-05,
|
| 74 |
+
"loss": 0.9108,
|
| 75 |
+
"step": 275
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 1.37,
|
| 79 |
+
"learning_rate": 1.1760000000000001e-05,
|
| 80 |
+
"loss": 0.8703,
|
| 81 |
+
"step": 300
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 1.48,
|
| 85 |
+
"learning_rate": 1.2760000000000001e-05,
|
| 86 |
+
"loss": 0.8957,
|
| 87 |
+
"step": 325
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.6,
|
| 91 |
+
"learning_rate": 1.376e-05,
|
| 92 |
+
"loss": 0.8933,
|
| 93 |
+
"step": 350
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 1.71,
|
| 97 |
+
"learning_rate": 1.4760000000000001e-05,
|
| 98 |
+
"loss": 0.8811,
|
| 99 |
+
"step": 375
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 1.83,
|
| 103 |
+
"learning_rate": 1.576e-05,
|
| 104 |
+
"loss": 0.8982,
|
| 105 |
+
"step": 400
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.94,
|
| 109 |
+
"learning_rate": 1.6760000000000002e-05,
|
| 110 |
+
"loss": 0.857,
|
| 111 |
+
"step": 425
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 2.05,
|
| 115 |
+
"learning_rate": 1.7760000000000003e-05,
|
| 116 |
+
"loss": 0.8538,
|
| 117 |
+
"step": 450
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 2.17,
|
| 121 |
+
"learning_rate": 1.876e-05,
|
| 122 |
+
"loss": 0.8573,
|
| 123 |
+
"step": 475
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.28,
|
| 127 |
+
"learning_rate": 1.976e-05,
|
| 128 |
+
"loss": 0.766,
|
| 129 |
+
"step": 500
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 2.4,
|
| 133 |
+
"learning_rate": 1.9915555555555557e-05,
|
| 134 |
+
"loss": 0.7945,
|
| 135 |
+
"step": 525
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 2.51,
|
| 139 |
+
"learning_rate": 1.9804444444444445e-05,
|
| 140 |
+
"loss": 0.8713,
|
| 141 |
+
"step": 550
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 2.63,
|
| 145 |
+
"learning_rate": 1.9693333333333337e-05,
|
| 146 |
+
"loss": 0.8652,
|
| 147 |
+
"step": 575
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 2.74,
|
| 151 |
+
"learning_rate": 1.9582222222222225e-05,
|
| 152 |
+
"loss": 0.8292,
|
| 153 |
+
"step": 600
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 2.85,
|
| 157 |
+
"learning_rate": 1.9471111111111113e-05,
|
| 158 |
+
"loss": 0.7645,
|
| 159 |
+
"step": 625
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 2.97,
|
| 163 |
+
"learning_rate": 1.936e-05,
|
| 164 |
+
"loss": 0.7594,
|
| 165 |
+
"step": 650
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.08,
|
| 169 |
+
"learning_rate": 1.924888888888889e-05,
|
| 170 |
+
"loss": 0.7144,
|
| 171 |
+
"step": 675
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 3.2,
|
| 175 |
+
"learning_rate": 1.913777777777778e-05,
|
| 176 |
+
"loss": 0.6814,
|
| 177 |
+
"step": 700
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 3.31,
|
| 181 |
+
"learning_rate": 1.902666666666667e-05,
|
| 182 |
+
"loss": 0.7054,
|
| 183 |
+
"step": 725
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.42,
|
| 187 |
+
"learning_rate": 1.8915555555555557e-05,
|
| 188 |
+
"loss": 0.7027,
|
| 189 |
+
"step": 750
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 3.54,
|
| 193 |
+
"learning_rate": 1.8804444444444445e-05,
|
| 194 |
+
"loss": 0.678,
|
| 195 |
+
"step": 775
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 3.65,
|
| 199 |
+
"learning_rate": 1.8693333333333333e-05,
|
| 200 |
+
"loss": 0.7418,
|
| 201 |
+
"step": 800
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 3.77,
|
| 205 |
+
"learning_rate": 1.8582222222222224e-05,
|
| 206 |
+
"loss": 0.7237,
|
| 207 |
+
"step": 825
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 3.88,
|
| 211 |
+
"learning_rate": 1.8471111111111112e-05,
|
| 212 |
+
"loss": 0.688,
|
| 213 |
+
"step": 850
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 4.0,
|
| 217 |
+
"learning_rate": 1.8360000000000004e-05,
|
| 218 |
+
"loss": 0.7115,
|
| 219 |
+
"step": 875
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 4.11,
|
| 223 |
+
"learning_rate": 1.824888888888889e-05,
|
| 224 |
+
"loss": 0.6277,
|
| 225 |
+
"step": 900
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.22,
|
| 229 |
+
"learning_rate": 1.813777777777778e-05,
|
| 230 |
+
"loss": 0.6007,
|
| 231 |
+
"step": 925
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 4.34,
|
| 235 |
+
"learning_rate": 1.8026666666666668e-05,
|
| 236 |
+
"loss": 0.5753,
|
| 237 |
+
"step": 950
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 4.45,
|
| 241 |
+
"learning_rate": 1.7915555555555556e-05,
|
| 242 |
+
"loss": 0.5818,
|
| 243 |
+
"step": 975
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.57,
|
| 247 |
+
"learning_rate": 1.7804444444444448e-05,
|
| 248 |
+
"loss": 0.6125,
|
| 249 |
+
"step": 1000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 4.57,
|
| 253 |
+
"eval_accuracy": 0.65,
|
| 254 |
+
"eval_loss": 0.894333004951477,
|
| 255 |
+
"eval_runtime": 2.6838,
|
| 256 |
+
"eval_samples_per_second": 372.609,
|
| 257 |
+
"eval_steps_per_second": 46.576,
|
| 258 |
+
"step": 1000
|
| 259 |
+
},
|
| 260 |
+
{
|
| 261 |
+
"epoch": 4.68,
|
| 262 |
+
"learning_rate": 1.7693333333333336e-05,
|
| 263 |
+
"loss": 0.5496,
|
| 264 |
+
"step": 1025
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.79,
|
| 268 |
+
"learning_rate": 1.7582222222222224e-05,
|
| 269 |
+
"loss": 0.5603,
|
| 270 |
+
"step": 1050
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 4.91,
|
| 274 |
+
"learning_rate": 1.7471111111111112e-05,
|
| 275 |
+
"loss": 0.5716,
|
| 276 |
+
"step": 1075
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 5.02,
|
| 280 |
+
"learning_rate": 1.736e-05,
|
| 281 |
+
"loss": 0.5522,
|
| 282 |
+
"step": 1100
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 5.14,
|
| 286 |
+
"learning_rate": 1.724888888888889e-05,
|
| 287 |
+
"loss": 0.4755,
|
| 288 |
+
"step": 1125
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 5.25,
|
| 292 |
+
"learning_rate": 1.713777777777778e-05,
|
| 293 |
+
"loss": 0.4464,
|
| 294 |
+
"step": 1150
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 5.37,
|
| 298 |
+
"learning_rate": 1.702666666666667e-05,
|
| 299 |
+
"loss": 0.4405,
|
| 300 |
+
"step": 1175
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 5.48,
|
| 304 |
+
"learning_rate": 1.6915555555555555e-05,
|
| 305 |
+
"loss": 0.4359,
|
| 306 |
+
"step": 1200
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.59,
|
| 310 |
+
"learning_rate": 1.6804444444444447e-05,
|
| 311 |
+
"loss": 0.463,
|
| 312 |
+
"step": 1225
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 5.71,
|
| 316 |
+
"learning_rate": 1.6693333333333335e-05,
|
| 317 |
+
"loss": 0.4788,
|
| 318 |
+
"step": 1250
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 5.82,
|
| 322 |
+
"learning_rate": 1.6582222222222223e-05,
|
| 323 |
+
"loss": 0.4673,
|
| 324 |
+
"step": 1275
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.94,
|
| 328 |
+
"learning_rate": 1.6471111111111115e-05,
|
| 329 |
+
"loss": 0.4315,
|
| 330 |
+
"step": 1300
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 6.05,
|
| 334 |
+
"learning_rate": 1.636e-05,
|
| 335 |
+
"loss": 0.4113,
|
| 336 |
+
"step": 1325
|
| 337 |
+
},
|
| 338 |
+
{
|
| 339 |
+
"epoch": 6.16,
|
| 340 |
+
"learning_rate": 1.624888888888889e-05,
|
| 341 |
+
"loss": 0.3585,
|
| 342 |
+
"step": 1350
|
| 343 |
+
},
|
| 344 |
+
{
|
| 345 |
+
"epoch": 6.28,
|
| 346 |
+
"learning_rate": 1.613777777777778e-05,
|
| 347 |
+
"loss": 0.3595,
|
| 348 |
+
"step": 1375
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 6.39,
|
| 352 |
+
"learning_rate": 1.6026666666666667e-05,
|
| 353 |
+
"loss": 0.3161,
|
| 354 |
+
"step": 1400
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 6.51,
|
| 358 |
+
"learning_rate": 1.5915555555555558e-05,
|
| 359 |
+
"loss": 0.3476,
|
| 360 |
+
"step": 1425
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 6.62,
|
| 364 |
+
"learning_rate": 1.5804444444444446e-05,
|
| 365 |
+
"loss": 0.3784,
|
| 366 |
+
"step": 1450
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.74,
|
| 370 |
+
"learning_rate": 1.5693333333333334e-05,
|
| 371 |
+
"loss": 0.3336,
|
| 372 |
+
"step": 1475
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"epoch": 6.85,
|
| 376 |
+
"learning_rate": 1.5582222222222222e-05,
|
| 377 |
+
"loss": 0.3314,
|
| 378 |
+
"step": 1500
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 6.96,
|
| 382 |
+
"learning_rate": 1.5471111111111114e-05,
|
| 383 |
+
"loss": 0.3501,
|
| 384 |
+
"step": 1525
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 7.08,
|
| 388 |
+
"learning_rate": 1.5360000000000002e-05,
|
| 389 |
+
"loss": 0.274,
|
| 390 |
+
"step": 1550
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 7.19,
|
| 394 |
+
"learning_rate": 1.524888888888889e-05,
|
| 395 |
+
"loss": 0.2473,
|
| 396 |
+
"step": 1575
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 7.31,
|
| 400 |
+
"learning_rate": 1.513777777777778e-05,
|
| 401 |
+
"loss": 0.237,
|
| 402 |
+
"step": 1600
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.42,
|
| 406 |
+
"learning_rate": 1.5026666666666668e-05,
|
| 407 |
+
"loss": 0.2742,
|
| 408 |
+
"step": 1625
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 7.53,
|
| 412 |
+
"learning_rate": 1.4915555555555556e-05,
|
| 413 |
+
"loss": 0.2906,
|
| 414 |
+
"step": 1650
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 7.65,
|
| 418 |
+
"learning_rate": 1.4804444444444446e-05,
|
| 419 |
+
"loss": 0.2499,
|
| 420 |
+
"step": 1675
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 7.76,
|
| 424 |
+
"learning_rate": 1.4693333333333336e-05,
|
| 425 |
+
"loss": 0.2694,
|
| 426 |
+
"step": 1700
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.88,
|
| 430 |
+
"learning_rate": 1.4582222222222224e-05,
|
| 431 |
+
"loss": 0.2465,
|
| 432 |
+
"step": 1725
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 7.99,
|
| 436 |
+
"learning_rate": 1.4471111111111112e-05,
|
| 437 |
+
"loss": 0.2739,
|
| 438 |
+
"step": 1750
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 8.11,
|
| 442 |
+
"learning_rate": 1.4360000000000001e-05,
|
| 443 |
+
"loss": 0.1963,
|
| 444 |
+
"step": 1775
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 8.22,
|
| 448 |
+
"learning_rate": 1.424888888888889e-05,
|
| 449 |
+
"loss": 0.1694,
|
| 450 |
+
"step": 1800
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 8.33,
|
| 454 |
+
"learning_rate": 1.413777777777778e-05,
|
| 455 |
+
"loss": 0.1532,
|
| 456 |
+
"step": 1825
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 8.45,
|
| 460 |
+
"learning_rate": 1.4026666666666669e-05,
|
| 461 |
+
"loss": 0.1692,
|
| 462 |
+
"step": 1850
|
| 463 |
+
},
|
| 464 |
+
{
|
| 465 |
+
"epoch": 8.56,
|
| 466 |
+
"learning_rate": 1.3915555555555557e-05,
|
| 467 |
+
"loss": 0.1687,
|
| 468 |
+
"step": 1875
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 8.68,
|
| 472 |
+
"learning_rate": 1.3804444444444445e-05,
|
| 473 |
+
"loss": 0.2099,
|
| 474 |
+
"step": 1900
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 8.79,
|
| 478 |
+
"learning_rate": 1.3693333333333333e-05,
|
| 479 |
+
"loss": 0.1927,
|
| 480 |
+
"step": 1925
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 8.9,
|
| 484 |
+
"learning_rate": 1.3582222222222223e-05,
|
| 485 |
+
"loss": 0.2065,
|
| 486 |
+
"step": 1950
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 9.02,
|
| 490 |
+
"learning_rate": 1.3471111111111113e-05,
|
| 491 |
+
"loss": 0.2312,
|
| 492 |
+
"step": 1975
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 9.13,
|
| 496 |
+
"learning_rate": 1.3360000000000003e-05,
|
| 497 |
+
"loss": 0.1438,
|
| 498 |
+
"step": 2000
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 9.13,
|
| 502 |
+
"eval_accuracy": 0.622,
|
| 503 |
+
"eval_loss": 1.4417067766189575,
|
| 504 |
+
"eval_runtime": 2.6482,
|
| 505 |
+
"eval_samples_per_second": 377.613,
|
| 506 |
+
"eval_steps_per_second": 47.202,
|
| 507 |
+
"step": 2000
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 9.25,
|
| 511 |
+
"learning_rate": 1.3248888888888889e-05,
|
| 512 |
+
"loss": 0.1875,
|
| 513 |
+
"step": 2025
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 9.36,
|
| 517 |
+
"learning_rate": 1.3137777777777779e-05,
|
| 518 |
+
"loss": 0.1359,
|
| 519 |
+
"step": 2050
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 9.47,
|
| 523 |
+
"learning_rate": 1.3026666666666667e-05,
|
| 524 |
+
"loss": 0.1177,
|
| 525 |
+
"step": 2075
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.59,
|
| 529 |
+
"learning_rate": 1.2915555555555557e-05,
|
| 530 |
+
"loss": 0.1494,
|
| 531 |
+
"step": 2100
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 9.7,
|
| 535 |
+
"learning_rate": 1.2804444444444446e-05,
|
| 536 |
+
"loss": 0.1313,
|
| 537 |
+
"step": 2125
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 9.82,
|
| 541 |
+
"learning_rate": 1.2693333333333336e-05,
|
| 542 |
+
"loss": 0.1571,
|
| 543 |
+
"step": 2150
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 9.93,
|
| 547 |
+
"learning_rate": 1.2582222222222222e-05,
|
| 548 |
+
"loss": 0.1671,
|
| 549 |
+
"step": 2175
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 10.05,
|
| 553 |
+
"learning_rate": 1.2471111111111112e-05,
|
| 554 |
+
"loss": 0.1264,
|
| 555 |
+
"step": 2200
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 10.16,
|
| 559 |
+
"learning_rate": 1.236e-05,
|
| 560 |
+
"loss": 0.0969,
|
| 561 |
+
"step": 2225
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 10.27,
|
| 565 |
+
"learning_rate": 1.224888888888889e-05,
|
| 566 |
+
"loss": 0.0971,
|
| 567 |
+
"step": 2250
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 10.39,
|
| 571 |
+
"learning_rate": 1.213777777777778e-05,
|
| 572 |
+
"loss": 0.1435,
|
| 573 |
+
"step": 2275
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 10.5,
|
| 577 |
+
"learning_rate": 1.202666666666667e-05,
|
| 578 |
+
"loss": 0.0938,
|
| 579 |
+
"step": 2300
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 10.62,
|
| 583 |
+
"learning_rate": 1.1915555555555556e-05,
|
| 584 |
+
"loss": 0.0844,
|
| 585 |
+
"step": 2325
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 10.73,
|
| 589 |
+
"learning_rate": 1.1804444444444446e-05,
|
| 590 |
+
"loss": 0.109,
|
| 591 |
+
"step": 2350
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 10.84,
|
| 595 |
+
"learning_rate": 1.1693333333333334e-05,
|
| 596 |
+
"loss": 0.0935,
|
| 597 |
+
"step": 2375
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 10.96,
|
| 601 |
+
"learning_rate": 1.1582222222222224e-05,
|
| 602 |
+
"loss": 0.1274,
|
| 603 |
+
"step": 2400
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 11.07,
|
| 607 |
+
"learning_rate": 1.1471111111111113e-05,
|
| 608 |
+
"loss": 0.0904,
|
| 609 |
+
"step": 2425
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 11.19,
|
| 613 |
+
"learning_rate": 1.136e-05,
|
| 614 |
+
"loss": 0.0675,
|
| 615 |
+
"step": 2450
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 11.3,
|
| 619 |
+
"learning_rate": 1.124888888888889e-05,
|
| 620 |
+
"loss": 0.0875,
|
| 621 |
+
"step": 2475
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 11.42,
|
| 625 |
+
"learning_rate": 1.113777777777778e-05,
|
| 626 |
+
"loss": 0.0729,
|
| 627 |
+
"step": 2500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 11.53,
|
| 631 |
+
"learning_rate": 1.1026666666666667e-05,
|
| 632 |
+
"loss": 0.0708,
|
| 633 |
+
"step": 2525
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 11.64,
|
| 637 |
+
"learning_rate": 1.0915555555555557e-05,
|
| 638 |
+
"loss": 0.0948,
|
| 639 |
+
"step": 2550
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 11.76,
|
| 643 |
+
"learning_rate": 1.0804444444444447e-05,
|
| 644 |
+
"loss": 0.0795,
|
| 645 |
+
"step": 2575
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 11.87,
|
| 649 |
+
"learning_rate": 1.0693333333333333e-05,
|
| 650 |
+
"loss": 0.0761,
|
| 651 |
+
"step": 2600
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 11.99,
|
| 655 |
+
"learning_rate": 1.0582222222222223e-05,
|
| 656 |
+
"loss": 0.0987,
|
| 657 |
+
"step": 2625
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 12.1,
|
| 661 |
+
"learning_rate": 1.0471111111111113e-05,
|
| 662 |
+
"loss": 0.0723,
|
| 663 |
+
"step": 2650
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 12.21,
|
| 667 |
+
"learning_rate": 1.036e-05,
|
| 668 |
+
"loss": 0.054,
|
| 669 |
+
"step": 2675
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 12.33,
|
| 673 |
+
"learning_rate": 1.024888888888889e-05,
|
| 674 |
+
"loss": 0.0688,
|
| 675 |
+
"step": 2700
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 12.44,
|
| 679 |
+
"learning_rate": 1.013777777777778e-05,
|
| 680 |
+
"loss": 0.079,
|
| 681 |
+
"step": 2725
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 12.56,
|
| 685 |
+
"learning_rate": 1.0026666666666667e-05,
|
| 686 |
+
"loss": 0.0457,
|
| 687 |
+
"step": 2750
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 12.67,
|
| 691 |
+
"learning_rate": 9.915555555555556e-06,
|
| 692 |
+
"loss": 0.0554,
|
| 693 |
+
"step": 2775
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 12.79,
|
| 697 |
+
"learning_rate": 9.804444444444444e-06,
|
| 698 |
+
"loss": 0.066,
|
| 699 |
+
"step": 2800
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 12.9,
|
| 703 |
+
"learning_rate": 9.693333333333334e-06,
|
| 704 |
+
"loss": 0.0432,
|
| 705 |
+
"step": 2825
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 13.01,
|
| 709 |
+
"learning_rate": 9.582222222222222e-06,
|
| 710 |
+
"loss": 0.0783,
|
| 711 |
+
"step": 2850
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 13.13,
|
| 715 |
+
"learning_rate": 9.471111111111112e-06,
|
| 716 |
+
"loss": 0.0645,
|
| 717 |
+
"step": 2875
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 13.24,
|
| 721 |
+
"learning_rate": 9.360000000000002e-06,
|
| 722 |
+
"loss": 0.0483,
|
| 723 |
+
"step": 2900
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 13.36,
|
| 727 |
+
"learning_rate": 9.24888888888889e-06,
|
| 728 |
+
"loss": 0.0524,
|
| 729 |
+
"step": 2925
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 13.47,
|
| 733 |
+
"learning_rate": 9.137777777777778e-06,
|
| 734 |
+
"loss": 0.0497,
|
| 735 |
+
"step": 2950
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 13.58,
|
| 739 |
+
"learning_rate": 9.026666666666666e-06,
|
| 740 |
+
"loss": 0.0492,
|
| 741 |
+
"step": 2975
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 13.7,
|
| 745 |
+
"learning_rate": 8.915555555555556e-06,
|
| 746 |
+
"loss": 0.0537,
|
| 747 |
+
"step": 3000
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 13.7,
|
| 751 |
+
"eval_accuracy": 0.62,
|
| 752 |
+
"eval_loss": 2.1143836975097656,
|
| 753 |
+
"eval_runtime": 2.8406,
|
| 754 |
+
"eval_samples_per_second": 352.038,
|
| 755 |
+
"eval_steps_per_second": 44.005,
|
| 756 |
+
"step": 3000
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 13.81,
|
| 760 |
+
"learning_rate": 8.804444444444446e-06,
|
| 761 |
+
"loss": 0.0493,
|
| 762 |
+
"step": 3025
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 13.93,
|
| 766 |
+
"learning_rate": 8.693333333333334e-06,
|
| 767 |
+
"loss": 0.0601,
|
| 768 |
+
"step": 3050
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 14.04,
|
| 772 |
+
"learning_rate": 8.582222222222223e-06,
|
| 773 |
+
"loss": 0.0522,
|
| 774 |
+
"step": 3075
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 14.16,
|
| 778 |
+
"learning_rate": 8.471111111111112e-06,
|
| 779 |
+
"loss": 0.024,
|
| 780 |
+
"step": 3100
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 14.27,
|
| 784 |
+
"learning_rate": 8.36e-06,
|
| 785 |
+
"loss": 0.0418,
|
| 786 |
+
"step": 3125
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 14.38,
|
| 790 |
+
"learning_rate": 8.24888888888889e-06,
|
| 791 |
+
"loss": 0.0578,
|
| 792 |
+
"step": 3150
|
| 793 |
+
},
|
| 794 |
+
{
|
| 795 |
+
"epoch": 14.5,
|
| 796 |
+
"learning_rate": 8.137777777777779e-06,
|
| 797 |
+
"loss": 0.0423,
|
| 798 |
+
"step": 3175
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 14.61,
|
| 802 |
+
"learning_rate": 8.026666666666667e-06,
|
| 803 |
+
"loss": 0.0478,
|
| 804 |
+
"step": 3200
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 14.73,
|
| 808 |
+
"learning_rate": 7.915555555555557e-06,
|
| 809 |
+
"loss": 0.0447,
|
| 810 |
+
"step": 3225
|
| 811 |
+
},
|
| 812 |
+
{
|
| 813 |
+
"epoch": 14.84,
|
| 814 |
+
"learning_rate": 7.804444444444445e-06,
|
| 815 |
+
"loss": 0.0416,
|
| 816 |
+
"step": 3250
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 14.95,
|
| 820 |
+
"learning_rate": 7.693333333333333e-06,
|
| 821 |
+
"loss": 0.0473,
|
| 822 |
+
"step": 3275
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 15.07,
|
| 826 |
+
"learning_rate": 7.582222222222223e-06,
|
| 827 |
+
"loss": 0.0459,
|
| 828 |
+
"step": 3300
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 15.18,
|
| 832 |
+
"learning_rate": 7.471111111111111e-06,
|
| 833 |
+
"loss": 0.0151,
|
| 834 |
+
"step": 3325
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 15.3,
|
| 838 |
+
"learning_rate": 7.360000000000001e-06,
|
| 839 |
+
"loss": 0.0403,
|
| 840 |
+
"step": 3350
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 15.41,
|
| 844 |
+
"learning_rate": 7.24888888888889e-06,
|
| 845 |
+
"loss": 0.0601,
|
| 846 |
+
"step": 3375
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 15.53,
|
| 850 |
+
"learning_rate": 7.137777777777778e-06,
|
| 851 |
+
"loss": 0.03,
|
| 852 |
+
"step": 3400
|
| 853 |
+
},
|
| 854 |
+
{
|
| 855 |
+
"epoch": 15.64,
|
| 856 |
+
"learning_rate": 7.0266666666666674e-06,
|
| 857 |
+
"loss": 0.0176,
|
| 858 |
+
"step": 3425
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 15.75,
|
| 862 |
+
"learning_rate": 6.915555555555556e-06,
|
| 863 |
+
"loss": 0.0169,
|
| 864 |
+
"step": 3450
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 15.87,
|
| 868 |
+
"learning_rate": 6.8044444444444444e-06,
|
| 869 |
+
"loss": 0.0351,
|
| 870 |
+
"step": 3475
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 15.98,
|
| 874 |
+
"learning_rate": 6.693333333333334e-06,
|
| 875 |
+
"loss": 0.0477,
|
| 876 |
+
"step": 3500
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 16.1,
|
| 880 |
+
"learning_rate": 6.582222222222223e-06,
|
| 881 |
+
"loss": 0.0474,
|
| 882 |
+
"step": 3525
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 16.21,
|
| 886 |
+
"learning_rate": 6.471111111111111e-06,
|
| 887 |
+
"loss": 0.0317,
|
| 888 |
+
"step": 3550
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 16.32,
|
| 892 |
+
"learning_rate": 6.360000000000001e-06,
|
| 893 |
+
"loss": 0.0216,
|
| 894 |
+
"step": 3575
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 16.44,
|
| 898 |
+
"learning_rate": 6.24888888888889e-06,
|
| 899 |
+
"loss": 0.0238,
|
| 900 |
+
"step": 3600
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 16.55,
|
| 904 |
+
"learning_rate": 6.137777777777778e-06,
|
| 905 |
+
"loss": 0.0458,
|
| 906 |
+
"step": 3625
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 16.67,
|
| 910 |
+
"learning_rate": 6.026666666666668e-06,
|
| 911 |
+
"loss": 0.0273,
|
| 912 |
+
"step": 3650
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 16.78,
|
| 916 |
+
"learning_rate": 5.915555555555556e-06,
|
| 917 |
+
"loss": 0.0058,
|
| 918 |
+
"step": 3675
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 16.89,
|
| 922 |
+
"learning_rate": 5.804444444444445e-06,
|
| 923 |
+
"loss": 0.0289,
|
| 924 |
+
"step": 3700
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 17.01,
|
| 928 |
+
"learning_rate": 5.6933333333333344e-06,
|
| 929 |
+
"loss": 0.0227,
|
| 930 |
+
"step": 3725
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 17.12,
|
| 934 |
+
"learning_rate": 5.5822222222222225e-06,
|
| 935 |
+
"loss": 0.0245,
|
| 936 |
+
"step": 3750
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 17.24,
|
| 940 |
+
"learning_rate": 5.4711111111111114e-06,
|
| 941 |
+
"loss": 0.0073,
|
| 942 |
+
"step": 3775
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 17.35,
|
| 946 |
+
"learning_rate": 5.36e-06,
|
| 947 |
+
"loss": 0.024,
|
| 948 |
+
"step": 3800
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 17.47,
|
| 952 |
+
"learning_rate": 5.248888888888889e-06,
|
| 953 |
+
"loss": 0.0094,
|
| 954 |
+
"step": 3825
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 17.58,
|
| 958 |
+
"learning_rate": 5.137777777777778e-06,
|
| 959 |
+
"loss": 0.0309,
|
| 960 |
+
"step": 3850
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 17.69,
|
| 964 |
+
"learning_rate": 5.026666666666667e-06,
|
| 965 |
+
"loss": 0.07,
|
| 966 |
+
"step": 3875
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 17.81,
|
| 970 |
+
"learning_rate": 4.915555555555556e-06,
|
| 971 |
+
"loss": 0.027,
|
| 972 |
+
"step": 3900
|
| 973 |
+
},
|
| 974 |
+
{
|
| 975 |
+
"epoch": 17.92,
|
| 976 |
+
"learning_rate": 4.804444444444445e-06,
|
| 977 |
+
"loss": 0.034,
|
| 978 |
+
"step": 3925
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 18.04,
|
| 982 |
+
"learning_rate": 4.693333333333334e-06,
|
| 983 |
+
"loss": 0.0337,
|
| 984 |
+
"step": 3950
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 18.15,
|
| 988 |
+
"learning_rate": 4.582222222222223e-06,
|
| 989 |
+
"loss": 0.0173,
|
| 990 |
+
"step": 3975
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 18.26,
|
| 994 |
+
"learning_rate": 4.471111111111112e-06,
|
| 995 |
+
"loss": 0.0346,
|
| 996 |
+
"step": 4000
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 18.26,
|
| 1000 |
+
"eval_accuracy": 0.643,
|
| 1001 |
+
"eval_loss": 2.5202674865722656,
|
| 1002 |
+
"eval_runtime": 2.6541,
|
| 1003 |
+
"eval_samples_per_second": 376.769,
|
| 1004 |
+
"eval_steps_per_second": 47.096,
|
| 1005 |
+
"step": 4000
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 18.38,
|
| 1009 |
+
"learning_rate": 4.360000000000001e-06,
|
| 1010 |
+
"loss": 0.0392,
|
| 1011 |
+
"step": 4025
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 18.49,
|
| 1015 |
+
"learning_rate": 4.248888888888889e-06,
|
| 1016 |
+
"loss": 0.0405,
|
| 1017 |
+
"step": 4050
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 18.61,
|
| 1021 |
+
"learning_rate": 4.1377777777777784e-06,
|
| 1022 |
+
"loss": 0.018,
|
| 1023 |
+
"step": 4075
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 18.72,
|
| 1027 |
+
"learning_rate": 4.026666666666667e-06,
|
| 1028 |
+
"loss": 0.0129,
|
| 1029 |
+
"step": 4100
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 18.84,
|
| 1033 |
+
"learning_rate": 3.9155555555555554e-06,
|
| 1034 |
+
"loss": 0.0188,
|
| 1035 |
+
"step": 4125
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 18.95,
|
| 1039 |
+
"learning_rate": 3.8044444444444443e-06,
|
| 1040 |
+
"loss": 0.0141,
|
| 1041 |
+
"step": 4150
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 19.06,
|
| 1045 |
+
"learning_rate": 3.6933333333333337e-06,
|
| 1046 |
+
"loss": 0.0237,
|
| 1047 |
+
"step": 4175
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 19.18,
|
| 1051 |
+
"learning_rate": 3.5822222222222226e-06,
|
| 1052 |
+
"loss": 0.0135,
|
| 1053 |
+
"step": 4200
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 19.29,
|
| 1057 |
+
"learning_rate": 3.471111111111111e-06,
|
| 1058 |
+
"loss": 0.0152,
|
| 1059 |
+
"step": 4225
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 19.41,
|
| 1063 |
+
"learning_rate": 3.3600000000000004e-06,
|
| 1064 |
+
"loss": 0.0286,
|
| 1065 |
+
"step": 4250
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 19.52,
|
| 1069 |
+
"learning_rate": 3.2488888888888894e-06,
|
| 1070 |
+
"loss": 0.0106,
|
| 1071 |
+
"step": 4275
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 19.63,
|
| 1075 |
+
"learning_rate": 3.137777777777778e-06,
|
| 1076 |
+
"loss": 0.0184,
|
| 1077 |
+
"step": 4300
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 19.75,
|
| 1081 |
+
"learning_rate": 3.0266666666666668e-06,
|
| 1082 |
+
"loss": 0.019,
|
| 1083 |
+
"step": 4325
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 19.86,
|
| 1087 |
+
"learning_rate": 2.915555555555556e-06,
|
| 1088 |
+
"loss": 0.0084,
|
| 1089 |
+
"step": 4350
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 19.98,
|
| 1093 |
+
"learning_rate": 2.8044444444444446e-06,
|
| 1094 |
+
"loss": 0.0092,
|
| 1095 |
+
"step": 4375
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 20.09,
|
| 1099 |
+
"learning_rate": 2.6933333333333335e-06,
|
| 1100 |
+
"loss": 0.0138,
|
| 1101 |
+
"step": 4400
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 20.21,
|
| 1105 |
+
"learning_rate": 2.5822222222222224e-06,
|
| 1106 |
+
"loss": 0.0052,
|
| 1107 |
+
"step": 4425
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 20.32,
|
| 1111 |
+
"learning_rate": 2.4711111111111114e-06,
|
| 1112 |
+
"loss": 0.0093,
|
| 1113 |
+
"step": 4450
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 20.43,
|
| 1117 |
+
"learning_rate": 2.3600000000000003e-06,
|
| 1118 |
+
"loss": 0.0114,
|
| 1119 |
+
"step": 4475
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 20.55,
|
| 1123 |
+
"learning_rate": 2.248888888888889e-06,
|
| 1124 |
+
"loss": 0.0047,
|
| 1125 |
+
"step": 4500
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 20.66,
|
| 1129 |
+
"learning_rate": 2.137777777777778e-06,
|
| 1130 |
+
"loss": 0.0304,
|
| 1131 |
+
"step": 4525
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 20.78,
|
| 1135 |
+
"learning_rate": 2.0266666666666666e-06,
|
| 1136 |
+
"loss": 0.0076,
|
| 1137 |
+
"step": 4550
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 20.89,
|
| 1141 |
+
"learning_rate": 1.915555555555556e-06,
|
| 1142 |
+
"loss": 0.0075,
|
| 1143 |
+
"step": 4575
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 21.0,
|
| 1147 |
+
"learning_rate": 1.8044444444444444e-06,
|
| 1148 |
+
"loss": 0.0178,
|
| 1149 |
+
"step": 4600
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 21.12,
|
| 1153 |
+
"learning_rate": 1.6933333333333336e-06,
|
| 1154 |
+
"loss": 0.0104,
|
| 1155 |
+
"step": 4625
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 21.23,
|
| 1159 |
+
"learning_rate": 1.5822222222222223e-06,
|
| 1160 |
+
"loss": 0.0152,
|
| 1161 |
+
"step": 4650
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 21.35,
|
| 1165 |
+
"learning_rate": 1.4711111111111112e-06,
|
| 1166 |
+
"loss": 0.0134,
|
| 1167 |
+
"step": 4675
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 21.46,
|
| 1171 |
+
"learning_rate": 1.3600000000000001e-06,
|
| 1172 |
+
"loss": 0.0252,
|
| 1173 |
+
"step": 4700
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 21.58,
|
| 1177 |
+
"learning_rate": 1.248888888888889e-06,
|
| 1178 |
+
"loss": 0.022,
|
| 1179 |
+
"step": 4725
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 21.69,
|
| 1183 |
+
"learning_rate": 1.137777777777778e-06,
|
| 1184 |
+
"loss": 0.006,
|
| 1185 |
+
"step": 4750
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 21.8,
|
| 1189 |
+
"learning_rate": 1.0266666666666669e-06,
|
| 1190 |
+
"loss": 0.0187,
|
| 1191 |
+
"step": 4775
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 21.92,
|
| 1195 |
+
"learning_rate": 9.155555555555557e-07,
|
| 1196 |
+
"loss": 0.0115,
|
| 1197 |
+
"step": 4800
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 22.03,
|
| 1201 |
+
"learning_rate": 8.044444444444445e-07,
|
| 1202 |
+
"loss": 0.0129,
|
| 1203 |
+
"step": 4825
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 22.15,
|
| 1207 |
+
"learning_rate": 6.933333333333334e-07,
|
| 1208 |
+
"loss": 0.0066,
|
| 1209 |
+
"step": 4850
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 22.26,
|
| 1213 |
+
"learning_rate": 5.822222222222223e-07,
|
| 1214 |
+
"loss": 0.0174,
|
| 1215 |
+
"step": 4875
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 22.37,
|
| 1219 |
+
"learning_rate": 4.7111111111111113e-07,
|
| 1220 |
+
"loss": 0.0181,
|
| 1221 |
+
"step": 4900
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 22.49,
|
| 1225 |
+
"learning_rate": 3.6e-07,
|
| 1226 |
+
"loss": 0.0089,
|
| 1227 |
+
"step": 4925
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 22.6,
|
| 1231 |
+
"learning_rate": 2.488888888888889e-07,
|
| 1232 |
+
"loss": 0.0157,
|
| 1233 |
+
"step": 4950
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 22.72,
|
| 1237 |
+
"learning_rate": 1.3777777777777778e-07,
|
| 1238 |
+
"loss": 0.0219,
|
| 1239 |
+
"step": 4975
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 22.83,
|
| 1243 |
+
"learning_rate": 2.6666666666666667e-08,
|
| 1244 |
+
"loss": 0.0013,
|
| 1245 |
+
"step": 5000
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 22.83,
|
| 1249 |
+
"eval_accuracy": 0.636,
|
| 1250 |
+
"eval_loss": 2.6798980236053467,
|
| 1251 |
+
"eval_runtime": 2.8534,
|
| 1252 |
+
"eval_samples_per_second": 350.458,
|
| 1253 |
+
"eval_steps_per_second": 43.807,
|
| 1254 |
+
"step": 5000
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 22.83,
|
| 1258 |
+
"step": 5000,
|
| 1259 |
+
"total_flos": 1.0512959718739968e+16,
|
| 1260 |
+
"train_loss": 0.2579876671291888,
|
| 1261 |
+
"train_runtime": 1594.4621,
|
| 1262 |
+
"train_samples_per_second": 100.347,
|
| 1263 |
+
"train_steps_per_second": 3.136
|
| 1264 |
+
}
|
| 1265 |
+
],
|
| 1266 |
+
"logging_steps": 25,
|
| 1267 |
+
"max_steps": 5000,
|
| 1268 |
+
"num_input_tokens_seen": 0,
|
| 1269 |
+
"num_train_epochs": 23,
|
| 1270 |
+
"save_steps": 1000,
|
| 1271 |
+
"total_flos": 1.0512959718739968e+16,
|
| 1272 |
+
"train_batch_size": 32,
|
| 1273 |
+
"trial_name": null,
|
| 1274 |
+
"trial_params": null
|
| 1275 |
+
}
|