End of training
Browse files- all_results.json +8 -0
- test_results.json +8 -0
- trainer_state.json +1865 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.096326530612245,
|
| 3 |
+
"eval_accuracy": 0.8550922761449077,
|
| 4 |
+
"eval_loss": 0.5230759382247925,
|
| 5 |
+
"eval_runtime": 633.2216,
|
| 6 |
+
"eval_samples_per_second": 2.31,
|
| 7 |
+
"eval_steps_per_second": 0.096
|
| 8 |
+
}
|
test_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.096326530612245,
|
| 3 |
+
"eval_accuracy": 0.8550922761449077,
|
| 4 |
+
"eval_loss": 0.5230759382247925,
|
| 5 |
+
"eval_runtime": 633.2216,
|
| 6 |
+
"eval_samples_per_second": 2.31,
|
| 7 |
+
"eval_steps_per_second": 0.096
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1865 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.7807017543859649,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-Badminton_strokes-finetuned-stroke-classification_10\\checkpoint-2214",
|
| 4 |
+
"epoch": 9.096326530612245,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2450,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.004081632653061225,
|
| 13 |
+
"grad_norm": 4.422181129455566,
|
| 14 |
+
"learning_rate": 2.040816326530612e-06,
|
| 15 |
+
"loss": 2.4923,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.00816326530612245,
|
| 20 |
+
"grad_norm": 6.700362682342529,
|
| 21 |
+
"learning_rate": 4.081632653061224e-06,
|
| 22 |
+
"loss": 2.4502,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.012244897959183673,
|
| 27 |
+
"grad_norm": 5.830870628356934,
|
| 28 |
+
"learning_rate": 6.122448979591837e-06,
|
| 29 |
+
"loss": 2.3498,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0163265306122449,
|
| 34 |
+
"grad_norm": 4.021888256072998,
|
| 35 |
+
"learning_rate": 8.163265306122448e-06,
|
| 36 |
+
"loss": 2.2118,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02040816326530612,
|
| 41 |
+
"grad_norm": 3.9514455795288086,
|
| 42 |
+
"learning_rate": 1.0204081632653061e-05,
|
| 43 |
+
"loss": 2.1642,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.024489795918367346,
|
| 48 |
+
"grad_norm": 3.8018665313720703,
|
| 49 |
+
"learning_rate": 1.2244897959183674e-05,
|
| 50 |
+
"loss": 2.0841,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02857142857142857,
|
| 55 |
+
"grad_norm": 6.1774187088012695,
|
| 56 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 57 |
+
"loss": 2.1007,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0326530612244898,
|
| 62 |
+
"grad_norm": 4.433764934539795,
|
| 63 |
+
"learning_rate": 1.6326530612244897e-05,
|
| 64 |
+
"loss": 2.083,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.036734693877551024,
|
| 69 |
+
"grad_norm": 4.373590469360352,
|
| 70 |
+
"learning_rate": 1.836734693877551e-05,
|
| 71 |
+
"loss": 1.9969,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.04081632653061224,
|
| 76 |
+
"grad_norm": 5.8374810218811035,
|
| 77 |
+
"learning_rate": 2.0408163265306123e-05,
|
| 78 |
+
"loss": 2.0792,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.044897959183673466,
|
| 83 |
+
"grad_norm": 4.199350833892822,
|
| 84 |
+
"learning_rate": 2.2448979591836737e-05,
|
| 85 |
+
"loss": 2.0843,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.04897959183673469,
|
| 90 |
+
"grad_norm": 3.5531275272369385,
|
| 91 |
+
"learning_rate": 2.448979591836735e-05,
|
| 92 |
+
"loss": 2.0739,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.053061224489795916,
|
| 97 |
+
"grad_norm": 4.77189826965332,
|
| 98 |
+
"learning_rate": 2.6530612244897963e-05,
|
| 99 |
+
"loss": 1.9767,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.05714285714285714,
|
| 104 |
+
"grad_norm": 6.462856769561768,
|
| 105 |
+
"learning_rate": 2.857142857142857e-05,
|
| 106 |
+
"loss": 1.9281,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.061224489795918366,
|
| 111 |
+
"grad_norm": 5.483211994171143,
|
| 112 |
+
"learning_rate": 3.061224489795919e-05,
|
| 113 |
+
"loss": 1.7622,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0653061224489796,
|
| 118 |
+
"grad_norm": 8.356924057006836,
|
| 119 |
+
"learning_rate": 3.265306122448979e-05,
|
| 120 |
+
"loss": 1.6668,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.06938775510204082,
|
| 125 |
+
"grad_norm": 5.412153720855713,
|
| 126 |
+
"learning_rate": 3.469387755102041e-05,
|
| 127 |
+
"loss": 1.648,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.07346938775510205,
|
| 132 |
+
"grad_norm": 4.69065523147583,
|
| 133 |
+
"learning_rate": 3.673469387755102e-05,
|
| 134 |
+
"loss": 1.5435,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07755102040816327,
|
| 139 |
+
"grad_norm": 5.284121990203857,
|
| 140 |
+
"learning_rate": 3.8775510204081634e-05,
|
| 141 |
+
"loss": 1.5461,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.08163265306122448,
|
| 146 |
+
"grad_norm": 8.057074546813965,
|
| 147 |
+
"learning_rate": 4.0816326530612245e-05,
|
| 148 |
+
"loss": 1.4362,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08571428571428572,
|
| 153 |
+
"grad_norm": 7.888145446777344,
|
| 154 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 155 |
+
"loss": 1.4473,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.08979591836734693,
|
| 160 |
+
"grad_norm": 11.86211109161377,
|
| 161 |
+
"learning_rate": 4.4897959183673474e-05,
|
| 162 |
+
"loss": 1.3792,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.09387755102040816,
|
| 167 |
+
"grad_norm": 6.048102855682373,
|
| 168 |
+
"learning_rate": 4.6938775510204086e-05,
|
| 169 |
+
"loss": 1.4187,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.09795918367346938,
|
| 174 |
+
"grad_norm": 5.002302646636963,
|
| 175 |
+
"learning_rate": 4.89795918367347e-05,
|
| 176 |
+
"loss": 1.3775,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.10040816326530612,
|
| 181 |
+
"eval_accuracy": 0.4783625730994152,
|
| 182 |
+
"eval_loss": 1.5476539134979248,
|
| 183 |
+
"eval_runtime": 781.3422,
|
| 184 |
+
"eval_samples_per_second": 2.189,
|
| 185 |
+
"eval_steps_per_second": 0.092,
|
| 186 |
+
"step": 246
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 1.0016326530612245,
|
| 190 |
+
"grad_norm": 7.02606725692749,
|
| 191 |
+
"learning_rate": 4.9886621315192745e-05,
|
| 192 |
+
"loss": 1.3045,
|
| 193 |
+
"step": 250
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 1.0057142857142858,
|
| 197 |
+
"grad_norm": 6.545066833496094,
|
| 198 |
+
"learning_rate": 4.965986394557823e-05,
|
| 199 |
+
"loss": 1.1286,
|
| 200 |
+
"step": 260
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.0097959183673468,
|
| 204 |
+
"grad_norm": 8.413867950439453,
|
| 205 |
+
"learning_rate": 4.9433106575963725e-05,
|
| 206 |
+
"loss": 1.478,
|
| 207 |
+
"step": 270
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.013877551020408,
|
| 211 |
+
"grad_norm": 5.661056041717529,
|
| 212 |
+
"learning_rate": 4.9206349206349204e-05,
|
| 213 |
+
"loss": 1.1755,
|
| 214 |
+
"step": 280
|
| 215 |
+
},
|
| 216 |
+
{
|
| 217 |
+
"epoch": 1.0179591836734694,
|
| 218 |
+
"grad_norm": 3.3354427814483643,
|
| 219 |
+
"learning_rate": 4.89795918367347e-05,
|
| 220 |
+
"loss": 1.2003,
|
| 221 |
+
"step": 290
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 1.0220408163265307,
|
| 225 |
+
"grad_norm": 8.000494956970215,
|
| 226 |
+
"learning_rate": 4.875283446712018e-05,
|
| 227 |
+
"loss": 1.2312,
|
| 228 |
+
"step": 300
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 1.026122448979592,
|
| 232 |
+
"grad_norm": 7.573913097381592,
|
| 233 |
+
"learning_rate": 4.8526077097505676e-05,
|
| 234 |
+
"loss": 1.2039,
|
| 235 |
+
"step": 310
|
| 236 |
+
},
|
| 237 |
+
{
|
| 238 |
+
"epoch": 1.030204081632653,
|
| 239 |
+
"grad_norm": 11.523215293884277,
|
| 240 |
+
"learning_rate": 4.8299319727891155e-05,
|
| 241 |
+
"loss": 1.1787,
|
| 242 |
+
"step": 320
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.0342857142857143,
|
| 246 |
+
"grad_norm": 5.250390529632568,
|
| 247 |
+
"learning_rate": 4.807256235827665e-05,
|
| 248 |
+
"loss": 1.2585,
|
| 249 |
+
"step": 330
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.0383673469387755,
|
| 253 |
+
"grad_norm": 7.881366729736328,
|
| 254 |
+
"learning_rate": 4.7845804988662134e-05,
|
| 255 |
+
"loss": 1.1318,
|
| 256 |
+
"step": 340
|
| 257 |
+
},
|
| 258 |
+
{
|
| 259 |
+
"epoch": 1.0424489795918368,
|
| 260 |
+
"grad_norm": 9.184869766235352,
|
| 261 |
+
"learning_rate": 4.761904761904762e-05,
|
| 262 |
+
"loss": 1.1961,
|
| 263 |
+
"step": 350
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 1.046530612244898,
|
| 267 |
+
"grad_norm": 5.172229766845703,
|
| 268 |
+
"learning_rate": 4.7392290249433106e-05,
|
| 269 |
+
"loss": 1.107,
|
| 270 |
+
"step": 360
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 1.0506122448979591,
|
| 274 |
+
"grad_norm": 8.629165649414062,
|
| 275 |
+
"learning_rate": 4.71655328798186e-05,
|
| 276 |
+
"loss": 1.1346,
|
| 277 |
+
"step": 370
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 1.0546938775510204,
|
| 281 |
+
"grad_norm": 8.821961402893066,
|
| 282 |
+
"learning_rate": 4.6938775510204086e-05,
|
| 283 |
+
"loss": 1.1584,
|
| 284 |
+
"step": 380
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.0587755102040817,
|
| 288 |
+
"grad_norm": 6.845242023468018,
|
| 289 |
+
"learning_rate": 4.671201814058957e-05,
|
| 290 |
+
"loss": 1.0883,
|
| 291 |
+
"step": 390
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.062857142857143,
|
| 295 |
+
"grad_norm": 6.316799163818359,
|
| 296 |
+
"learning_rate": 4.648526077097506e-05,
|
| 297 |
+
"loss": 1.0708,
|
| 298 |
+
"step": 400
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.066938775510204,
|
| 302 |
+
"grad_norm": 8.14439582824707,
|
| 303 |
+
"learning_rate": 4.625850340136055e-05,
|
| 304 |
+
"loss": 1.0572,
|
| 305 |
+
"step": 410
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.0710204081632653,
|
| 309 |
+
"grad_norm": 6.715014934539795,
|
| 310 |
+
"learning_rate": 4.603174603174603e-05,
|
| 311 |
+
"loss": 0.9144,
|
| 312 |
+
"step": 420
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.0751020408163265,
|
| 316 |
+
"grad_norm": 5.173260688781738,
|
| 317 |
+
"learning_rate": 4.580498866213152e-05,
|
| 318 |
+
"loss": 1.0762,
|
| 319 |
+
"step": 430
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 1.0791836734693878,
|
| 323 |
+
"grad_norm": 9.654570579528809,
|
| 324 |
+
"learning_rate": 4.557823129251701e-05,
|
| 325 |
+
"loss": 1.1108,
|
| 326 |
+
"step": 440
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.083265306122449,
|
| 330 |
+
"grad_norm": 8.643160820007324,
|
| 331 |
+
"learning_rate": 4.53514739229025e-05,
|
| 332 |
+
"loss": 0.9241,
|
| 333 |
+
"step": 450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.0873469387755101,
|
| 337 |
+
"grad_norm": 8.947994232177734,
|
| 338 |
+
"learning_rate": 4.512471655328798e-05,
|
| 339 |
+
"loss": 0.8726,
|
| 340 |
+
"step": 460
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.0914285714285714,
|
| 344 |
+
"grad_norm": 15.66603946685791,
|
| 345 |
+
"learning_rate": 4.4897959183673474e-05,
|
| 346 |
+
"loss": 1.0067,
|
| 347 |
+
"step": 470
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.0955102040816327,
|
| 351 |
+
"grad_norm": 7.348424911499023,
|
| 352 |
+
"learning_rate": 4.467120181405896e-05,
|
| 353 |
+
"loss": 1.0321,
|
| 354 |
+
"step": 480
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.099591836734694,
|
| 358 |
+
"grad_norm": 11.021227836608887,
|
| 359 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 360 |
+
"loss": 0.9098,
|
| 361 |
+
"step": 490
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.1004081632653062,
|
| 365 |
+
"eval_accuracy": 0.6964912280701754,
|
| 366 |
+
"eval_loss": 1.2118452787399292,
|
| 367 |
+
"eval_runtime": 756.1595,
|
| 368 |
+
"eval_samples_per_second": 2.261,
|
| 369 |
+
"eval_steps_per_second": 0.095,
|
| 370 |
+
"step": 492
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 2.003265306122449,
|
| 374 |
+
"grad_norm": 5.652357578277588,
|
| 375 |
+
"learning_rate": 4.421768707482993e-05,
|
| 376 |
+
"loss": 0.9034,
|
| 377 |
+
"step": 500
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 2.0073469387755103,
|
| 381 |
+
"grad_norm": 8.300456047058105,
|
| 382 |
+
"learning_rate": 4.3990929705215426e-05,
|
| 383 |
+
"loss": 0.7905,
|
| 384 |
+
"step": 510
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 2.0114285714285716,
|
| 388 |
+
"grad_norm": 8.094884872436523,
|
| 389 |
+
"learning_rate": 4.376417233560091e-05,
|
| 390 |
+
"loss": 0.6445,
|
| 391 |
+
"step": 520
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 2.015510204081633,
|
| 395 |
+
"grad_norm": 8.103925704956055,
|
| 396 |
+
"learning_rate": 4.35374149659864e-05,
|
| 397 |
+
"loss": 0.8278,
|
| 398 |
+
"step": 530
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.0195918367346937,
|
| 402 |
+
"grad_norm": 6.3901801109313965,
|
| 403 |
+
"learning_rate": 4.3310657596371884e-05,
|
| 404 |
+
"loss": 0.7736,
|
| 405 |
+
"step": 540
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.023673469387755,
|
| 409 |
+
"grad_norm": 11.346807479858398,
|
| 410 |
+
"learning_rate": 4.308390022675737e-05,
|
| 411 |
+
"loss": 1.0031,
|
| 412 |
+
"step": 550
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 2.027755102040816,
|
| 416 |
+
"grad_norm": 10.086901664733887,
|
| 417 |
+
"learning_rate": 4.2857142857142856e-05,
|
| 418 |
+
"loss": 0.9582,
|
| 419 |
+
"step": 560
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 2.0318367346938775,
|
| 423 |
+
"grad_norm": 4.69091796875,
|
| 424 |
+
"learning_rate": 4.263038548752835e-05,
|
| 425 |
+
"loss": 0.8926,
|
| 426 |
+
"step": 570
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 2.0359183673469388,
|
| 430 |
+
"grad_norm": 11.110679626464844,
|
| 431 |
+
"learning_rate": 4.2403628117913835e-05,
|
| 432 |
+
"loss": 0.745,
|
| 433 |
+
"step": 580
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 2.04,
|
| 437 |
+
"grad_norm": 10.121854782104492,
|
| 438 |
+
"learning_rate": 4.217687074829932e-05,
|
| 439 |
+
"loss": 0.794,
|
| 440 |
+
"step": 590
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.0440816326530613,
|
| 444 |
+
"grad_norm": 12.487133026123047,
|
| 445 |
+
"learning_rate": 4.195011337868481e-05,
|
| 446 |
+
"loss": 0.975,
|
| 447 |
+
"step": 600
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 2.0481632653061226,
|
| 451 |
+
"grad_norm": 5.699638843536377,
|
| 452 |
+
"learning_rate": 4.1723356009070294e-05,
|
| 453 |
+
"loss": 0.7269,
|
| 454 |
+
"step": 610
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 2.052244897959184,
|
| 458 |
+
"grad_norm": 6.98757791519165,
|
| 459 |
+
"learning_rate": 4.149659863945579e-05,
|
| 460 |
+
"loss": 0.9338,
|
| 461 |
+
"step": 620
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 2.0563265306122447,
|
| 465 |
+
"grad_norm": 7.258986473083496,
|
| 466 |
+
"learning_rate": 4.126984126984127e-05,
|
| 467 |
+
"loss": 0.7356,
|
| 468 |
+
"step": 630
|
| 469 |
+
},
|
| 470 |
+
{
|
| 471 |
+
"epoch": 2.060408163265306,
|
| 472 |
+
"grad_norm": 8.993372917175293,
|
| 473 |
+
"learning_rate": 4.104308390022676e-05,
|
| 474 |
+
"loss": 0.778,
|
| 475 |
+
"step": 640
|
| 476 |
+
},
|
| 477 |
+
{
|
| 478 |
+
"epoch": 2.0644897959183672,
|
| 479 |
+
"grad_norm": 8.101334571838379,
|
| 480 |
+
"learning_rate": 4.0816326530612245e-05,
|
| 481 |
+
"loss": 0.612,
|
| 482 |
+
"step": 650
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.0685714285714285,
|
| 486 |
+
"grad_norm": 7.610772132873535,
|
| 487 |
+
"learning_rate": 4.058956916099774e-05,
|
| 488 |
+
"loss": 0.7947,
|
| 489 |
+
"step": 660
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 2.07265306122449,
|
| 493 |
+
"grad_norm": 7.818012237548828,
|
| 494 |
+
"learning_rate": 4.036281179138322e-05,
|
| 495 |
+
"loss": 0.7262,
|
| 496 |
+
"step": 670
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 2.076734693877551,
|
| 500 |
+
"grad_norm": 8.464200973510742,
|
| 501 |
+
"learning_rate": 4.013605442176871e-05,
|
| 502 |
+
"loss": 0.8635,
|
| 503 |
+
"step": 680
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 2.0808163265306123,
|
| 507 |
+
"grad_norm": 7.1510233879089355,
|
| 508 |
+
"learning_rate": 3.9909297052154196e-05,
|
| 509 |
+
"loss": 0.7381,
|
| 510 |
+
"step": 690
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 2.0848979591836736,
|
| 514 |
+
"grad_norm": 3.6917271614074707,
|
| 515 |
+
"learning_rate": 3.968253968253968e-05,
|
| 516 |
+
"loss": 0.7398,
|
| 517 |
+
"step": 700
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 2.088979591836735,
|
| 521 |
+
"grad_norm": 10.353333473205566,
|
| 522 |
+
"learning_rate": 3.945578231292517e-05,
|
| 523 |
+
"loss": 0.7111,
|
| 524 |
+
"step": 710
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 2.093061224489796,
|
| 528 |
+
"grad_norm": 5.636819839477539,
|
| 529 |
+
"learning_rate": 3.922902494331066e-05,
|
| 530 |
+
"loss": 0.793,
|
| 531 |
+
"step": 720
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 2.097142857142857,
|
| 535 |
+
"grad_norm": 7.487854957580566,
|
| 536 |
+
"learning_rate": 3.900226757369615e-05,
|
| 537 |
+
"loss": 0.6294,
|
| 538 |
+
"step": 730
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 2.100408163265306,
|
| 542 |
+
"eval_accuracy": 0.716374269005848,
|
| 543 |
+
"eval_loss": 1.1314142942428589,
|
| 544 |
+
"eval_runtime": 759.7521,
|
| 545 |
+
"eval_samples_per_second": 2.251,
|
| 546 |
+
"eval_steps_per_second": 0.095,
|
| 547 |
+
"step": 738
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 3.0008163265306123,
|
| 551 |
+
"grad_norm": 6.67496919631958,
|
| 552 |
+
"learning_rate": 3.8775510204081634e-05,
|
| 553 |
+
"loss": 0.9053,
|
| 554 |
+
"step": 740
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 3.0048979591836735,
|
| 558 |
+
"grad_norm": 10.205514907836914,
|
| 559 |
+
"learning_rate": 3.854875283446712e-05,
|
| 560 |
+
"loss": 0.6987,
|
| 561 |
+
"step": 750
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 3.008979591836735,
|
| 565 |
+
"grad_norm": 5.589556694030762,
|
| 566 |
+
"learning_rate": 3.832199546485261e-05,
|
| 567 |
+
"loss": 0.6399,
|
| 568 |
+
"step": 760
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 3.013061224489796,
|
| 572 |
+
"grad_norm": 13.978158950805664,
|
| 573 |
+
"learning_rate": 3.809523809523809e-05,
|
| 574 |
+
"loss": 0.6441,
|
| 575 |
+
"step": 770
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 3.0171428571428573,
|
| 579 |
+
"grad_norm": 8.94792652130127,
|
| 580 |
+
"learning_rate": 3.7868480725623585e-05,
|
| 581 |
+
"loss": 0.6705,
|
| 582 |
+
"step": 780
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 3.021224489795918,
|
| 586 |
+
"grad_norm": 5.165012359619141,
|
| 587 |
+
"learning_rate": 3.764172335600907e-05,
|
| 588 |
+
"loss": 0.6792,
|
| 589 |
+
"step": 790
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 3.0253061224489795,
|
| 593 |
+
"grad_norm": 4.369375228881836,
|
| 594 |
+
"learning_rate": 3.7414965986394564e-05,
|
| 595 |
+
"loss": 0.599,
|
| 596 |
+
"step": 800
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.0293877551020407,
|
| 600 |
+
"grad_norm": 10.160274505615234,
|
| 601 |
+
"learning_rate": 3.7188208616780044e-05,
|
| 602 |
+
"loss": 0.6651,
|
| 603 |
+
"step": 810
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 3.033469387755102,
|
| 607 |
+
"grad_norm": 11.945816993713379,
|
| 608 |
+
"learning_rate": 3.6961451247165536e-05,
|
| 609 |
+
"loss": 0.6486,
|
| 610 |
+
"step": 820
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 3.0375510204081633,
|
| 614 |
+
"grad_norm": 8.221893310546875,
|
| 615 |
+
"learning_rate": 3.673469387755102e-05,
|
| 616 |
+
"loss": 0.6481,
|
| 617 |
+
"step": 830
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 3.0416326530612245,
|
| 621 |
+
"grad_norm": 8.286681175231934,
|
| 622 |
+
"learning_rate": 3.650793650793651e-05,
|
| 623 |
+
"loss": 0.5535,
|
| 624 |
+
"step": 840
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 3.045714285714286,
|
| 628 |
+
"grad_norm": 5.675285816192627,
|
| 629 |
+
"learning_rate": 3.6281179138321995e-05,
|
| 630 |
+
"loss": 0.6915,
|
| 631 |
+
"step": 850
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 3.049795918367347,
|
| 635 |
+
"grad_norm": 7.121517658233643,
|
| 636 |
+
"learning_rate": 3.605442176870749e-05,
|
| 637 |
+
"loss": 0.5688,
|
| 638 |
+
"step": 860
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.0538775510204084,
|
| 642 |
+
"grad_norm": 7.682671546936035,
|
| 643 |
+
"learning_rate": 3.5827664399092974e-05,
|
| 644 |
+
"loss": 0.6005,
|
| 645 |
+
"step": 870
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 3.057959183673469,
|
| 649 |
+
"grad_norm": 7.741525173187256,
|
| 650 |
+
"learning_rate": 3.560090702947846e-05,
|
| 651 |
+
"loss": 0.5067,
|
| 652 |
+
"step": 880
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 3.0620408163265305,
|
| 656 |
+
"grad_norm": 6.186898231506348,
|
| 657 |
+
"learning_rate": 3.5374149659863946e-05,
|
| 658 |
+
"loss": 0.5619,
|
| 659 |
+
"step": 890
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 3.0661224489795917,
|
| 663 |
+
"grad_norm": 5.949254035949707,
|
| 664 |
+
"learning_rate": 3.514739229024944e-05,
|
| 665 |
+
"loss": 0.5297,
|
| 666 |
+
"step": 900
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 3.070204081632653,
|
| 670 |
+
"grad_norm": 12.493196487426758,
|
| 671 |
+
"learning_rate": 3.492063492063492e-05,
|
| 672 |
+
"loss": 0.6647,
|
| 673 |
+
"step": 910
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 3.0742857142857143,
|
| 677 |
+
"grad_norm": 15.777326583862305,
|
| 678 |
+
"learning_rate": 3.469387755102041e-05,
|
| 679 |
+
"loss": 0.8303,
|
| 680 |
+
"step": 920
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 3.0783673469387756,
|
| 684 |
+
"grad_norm": 5.946107864379883,
|
| 685 |
+
"learning_rate": 3.44671201814059e-05,
|
| 686 |
+
"loss": 0.5871,
|
| 687 |
+
"step": 930
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 3.082448979591837,
|
| 691 |
+
"grad_norm": 6.941315650939941,
|
| 692 |
+
"learning_rate": 3.4240362811791384e-05,
|
| 693 |
+
"loss": 0.5455,
|
| 694 |
+
"step": 940
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 3.086530612244898,
|
| 698 |
+
"grad_norm": 9.79609489440918,
|
| 699 |
+
"learning_rate": 3.401360544217687e-05,
|
| 700 |
+
"loss": 0.5152,
|
| 701 |
+
"step": 950
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 3.0906122448979594,
|
| 705 |
+
"grad_norm": 5.163089275360107,
|
| 706 |
+
"learning_rate": 3.378684807256236e-05,
|
| 707 |
+
"loss": 0.5824,
|
| 708 |
+
"step": 960
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 3.09469387755102,
|
| 712 |
+
"grad_norm": 5.999682426452637,
|
| 713 |
+
"learning_rate": 3.356009070294785e-05,
|
| 714 |
+
"loss": 0.5886,
|
| 715 |
+
"step": 970
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 3.0987755102040815,
|
| 719 |
+
"grad_norm": 6.851431846618652,
|
| 720 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 721 |
+
"loss": 0.6993,
|
| 722 |
+
"step": 980
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 3.100408163265306,
|
| 726 |
+
"eval_accuracy": 0.7333333333333333,
|
| 727 |
+
"eval_loss": 1.1106853485107422,
|
| 728 |
+
"eval_runtime": 755.202,
|
| 729 |
+
"eval_samples_per_second": 2.264,
|
| 730 |
+
"eval_steps_per_second": 0.095,
|
| 731 |
+
"step": 984
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 4.002448979591836,
|
| 735 |
+
"grad_norm": 10.152717590332031,
|
| 736 |
+
"learning_rate": 3.310657596371882e-05,
|
| 737 |
+
"loss": 0.6247,
|
| 738 |
+
"step": 990
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 4.006530612244898,
|
| 742 |
+
"grad_norm": 12.136672973632812,
|
| 743 |
+
"learning_rate": 3.2879818594104314e-05,
|
| 744 |
+
"loss": 0.5202,
|
| 745 |
+
"step": 1000
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 4.010612244897959,
|
| 749 |
+
"grad_norm": 9.652105331420898,
|
| 750 |
+
"learning_rate": 3.265306122448979e-05,
|
| 751 |
+
"loss": 0.4666,
|
| 752 |
+
"step": 1010
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.014693877551021,
|
| 756 |
+
"grad_norm": 16.446027755737305,
|
| 757 |
+
"learning_rate": 3.2426303854875286e-05,
|
| 758 |
+
"loss": 0.4363,
|
| 759 |
+
"step": 1020
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 4.018775510204081,
|
| 763 |
+
"grad_norm": 5.725858688354492,
|
| 764 |
+
"learning_rate": 3.219954648526077e-05,
|
| 765 |
+
"loss": 0.4498,
|
| 766 |
+
"step": 1030
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 4.022857142857143,
|
| 770 |
+
"grad_norm": 9.73865032196045,
|
| 771 |
+
"learning_rate": 3.1972789115646265e-05,
|
| 772 |
+
"loss": 0.4881,
|
| 773 |
+
"step": 1040
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 4.026938775510204,
|
| 777 |
+
"grad_norm": 10.271039962768555,
|
| 778 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 779 |
+
"loss": 0.7649,
|
| 780 |
+
"step": 1050
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 4.031020408163266,
|
| 784 |
+
"grad_norm": 8.36655044555664,
|
| 785 |
+
"learning_rate": 3.151927437641724e-05,
|
| 786 |
+
"loss": 0.4447,
|
| 787 |
+
"step": 1060
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 4.0351020408163265,
|
| 791 |
+
"grad_norm": 9.38914966583252,
|
| 792 |
+
"learning_rate": 3.1292517006802724e-05,
|
| 793 |
+
"loss": 0.4334,
|
| 794 |
+
"step": 1070
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.039183673469387,
|
| 798 |
+
"grad_norm": 6.157915115356445,
|
| 799 |
+
"learning_rate": 3.106575963718821e-05,
|
| 800 |
+
"loss": 0.4684,
|
| 801 |
+
"step": 1080
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 4.043265306122449,
|
| 805 |
+
"grad_norm": 7.4239020347595215,
|
| 806 |
+
"learning_rate": 3.0839002267573696e-05,
|
| 807 |
+
"loss": 0.4357,
|
| 808 |
+
"step": 1090
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 4.04734693877551,
|
| 812 |
+
"grad_norm": 7.730371952056885,
|
| 813 |
+
"learning_rate": 3.061224489795919e-05,
|
| 814 |
+
"loss": 0.5067,
|
| 815 |
+
"step": 1100
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 4.051428571428572,
|
| 819 |
+
"grad_norm": 9.309793472290039,
|
| 820 |
+
"learning_rate": 3.0385487528344675e-05,
|
| 821 |
+
"loss": 0.5358,
|
| 822 |
+
"step": 1110
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 4.055510204081632,
|
| 826 |
+
"grad_norm": 5.453812599182129,
|
| 827 |
+
"learning_rate": 3.0158730158730158e-05,
|
| 828 |
+
"loss": 0.4429,
|
| 829 |
+
"step": 1120
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 4.059591836734694,
|
| 833 |
+
"grad_norm": 6.250613689422607,
|
| 834 |
+
"learning_rate": 2.9931972789115647e-05,
|
| 835 |
+
"loss": 0.5217,
|
| 836 |
+
"step": 1130
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.063673469387755,
|
| 840 |
+
"grad_norm": 12.171501159667969,
|
| 841 |
+
"learning_rate": 2.9705215419501137e-05,
|
| 842 |
+
"loss": 0.5578,
|
| 843 |
+
"step": 1140
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 4.067755102040817,
|
| 847 |
+
"grad_norm": 5.649395942687988,
|
| 848 |
+
"learning_rate": 2.947845804988662e-05,
|
| 849 |
+
"loss": 0.508,
|
| 850 |
+
"step": 1150
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 4.0718367346938775,
|
| 854 |
+
"grad_norm": 7.545301914215088,
|
| 855 |
+
"learning_rate": 2.925170068027211e-05,
|
| 856 |
+
"loss": 0.3983,
|
| 857 |
+
"step": 1160
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 4.075918367346938,
|
| 861 |
+
"grad_norm": 8.015406608581543,
|
| 862 |
+
"learning_rate": 2.90249433106576e-05,
|
| 863 |
+
"loss": 0.4655,
|
| 864 |
+
"step": 1170
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 4.08,
|
| 868 |
+
"grad_norm": 8.508922576904297,
|
| 869 |
+
"learning_rate": 2.8798185941043088e-05,
|
| 870 |
+
"loss": 0.4419,
|
| 871 |
+
"step": 1180
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 4.084081632653061,
|
| 875 |
+
"grad_norm": 7.398991107940674,
|
| 876 |
+
"learning_rate": 2.857142857142857e-05,
|
| 877 |
+
"loss": 0.5536,
|
| 878 |
+
"step": 1190
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 4.088163265306123,
|
| 882 |
+
"grad_norm": 7.41541051864624,
|
| 883 |
+
"learning_rate": 2.834467120181406e-05,
|
| 884 |
+
"loss": 0.5126,
|
| 885 |
+
"step": 1200
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 4.0922448979591834,
|
| 889 |
+
"grad_norm": 9.597835540771484,
|
| 890 |
+
"learning_rate": 2.811791383219955e-05,
|
| 891 |
+
"loss": 0.5008,
|
| 892 |
+
"step": 1210
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 4.096326530612245,
|
| 896 |
+
"grad_norm": 7.991727352142334,
|
| 897 |
+
"learning_rate": 2.7891156462585033e-05,
|
| 898 |
+
"loss": 0.4599,
|
| 899 |
+
"step": 1220
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 4.100408163265306,
|
| 903 |
+
"grad_norm": 11.23975658416748,
|
| 904 |
+
"learning_rate": 2.7664399092970522e-05,
|
| 905 |
+
"loss": 0.4355,
|
| 906 |
+
"step": 1230
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 4.100408163265306,
|
| 910 |
+
"eval_accuracy": 0.7497076023391813,
|
| 911 |
+
"eval_loss": 1.077431559562683,
|
| 912 |
+
"eval_runtime": 742.1867,
|
| 913 |
+
"eval_samples_per_second": 2.304,
|
| 914 |
+
"eval_steps_per_second": 0.097,
|
| 915 |
+
"step": 1230
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 5.004081632653061,
|
| 919 |
+
"grad_norm": 5.663974285125732,
|
| 920 |
+
"learning_rate": 2.7437641723356012e-05,
|
| 921 |
+
"loss": 0.4312,
|
| 922 |
+
"step": 1240
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"epoch": 5.0081632653061225,
|
| 926 |
+
"grad_norm": 11.165666580200195,
|
| 927 |
+
"learning_rate": 2.72108843537415e-05,
|
| 928 |
+
"loss": 0.4112,
|
| 929 |
+
"step": 1250
|
| 930 |
+
},
|
| 931 |
+
{
|
| 932 |
+
"epoch": 5.012244897959183,
|
| 933 |
+
"grad_norm": 3.895768165588379,
|
| 934 |
+
"learning_rate": 2.6984126984126984e-05,
|
| 935 |
+
"loss": 0.4421,
|
| 936 |
+
"step": 1260
|
| 937 |
+
},
|
| 938 |
+
{
|
| 939 |
+
"epoch": 5.016326530612245,
|
| 940 |
+
"grad_norm": 14.455580711364746,
|
| 941 |
+
"learning_rate": 2.6757369614512473e-05,
|
| 942 |
+
"loss": 0.4569,
|
| 943 |
+
"step": 1270
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 5.020408163265306,
|
| 947 |
+
"grad_norm": 7.778404712677002,
|
| 948 |
+
"learning_rate": 2.6530612244897963e-05,
|
| 949 |
+
"loss": 0.4152,
|
| 950 |
+
"step": 1280
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.024489795918368,
|
| 954 |
+
"grad_norm": 20.504398345947266,
|
| 955 |
+
"learning_rate": 2.6303854875283446e-05,
|
| 956 |
+
"loss": 0.5674,
|
| 957 |
+
"step": 1290
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 5.0285714285714285,
|
| 961 |
+
"grad_norm": 10.40569019317627,
|
| 962 |
+
"learning_rate": 2.6077097505668935e-05,
|
| 963 |
+
"loss": 0.421,
|
| 964 |
+
"step": 1300
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 5.03265306122449,
|
| 968 |
+
"grad_norm": 5.579968452453613,
|
| 969 |
+
"learning_rate": 2.5850340136054425e-05,
|
| 970 |
+
"loss": 0.4584,
|
| 971 |
+
"step": 1310
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 5.036734693877551,
|
| 975 |
+
"grad_norm": 8.29798698425293,
|
| 976 |
+
"learning_rate": 2.5623582766439914e-05,
|
| 977 |
+
"loss": 0.4372,
|
| 978 |
+
"step": 1320
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 5.040816326530612,
|
| 982 |
+
"grad_norm": 6.832752227783203,
|
| 983 |
+
"learning_rate": 2.5396825396825397e-05,
|
| 984 |
+
"loss": 0.4213,
|
| 985 |
+
"step": 1330
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 5.044897959183674,
|
| 989 |
+
"grad_norm": 13.543517112731934,
|
| 990 |
+
"learning_rate": 2.5170068027210887e-05,
|
| 991 |
+
"loss": 0.4999,
|
| 992 |
+
"step": 1340
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.048979591836734,
|
| 996 |
+
"grad_norm": 7.032736778259277,
|
| 997 |
+
"learning_rate": 2.4943310657596373e-05,
|
| 998 |
+
"loss": 0.392,
|
| 999 |
+
"step": 1350
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 5.053061224489796,
|
| 1003 |
+
"grad_norm": 7.005382537841797,
|
| 1004 |
+
"learning_rate": 2.4716553287981862e-05,
|
| 1005 |
+
"loss": 0.4503,
|
| 1006 |
+
"step": 1360
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 5.057142857142857,
|
| 1010 |
+
"grad_norm": 7.160999298095703,
|
| 1011 |
+
"learning_rate": 2.448979591836735e-05,
|
| 1012 |
+
"loss": 0.3743,
|
| 1013 |
+
"step": 1370
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 5.061224489795919,
|
| 1017 |
+
"grad_norm": 11.003737449645996,
|
| 1018 |
+
"learning_rate": 2.4263038548752838e-05,
|
| 1019 |
+
"loss": 0.4526,
|
| 1020 |
+
"step": 1380
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 5.0653061224489795,
|
| 1024 |
+
"grad_norm": 7.8346076011657715,
|
| 1025 |
+
"learning_rate": 2.4036281179138324e-05,
|
| 1026 |
+
"loss": 0.4225,
|
| 1027 |
+
"step": 1390
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 5.069387755102041,
|
| 1031 |
+
"grad_norm": 13.947251319885254,
|
| 1032 |
+
"learning_rate": 2.380952380952381e-05,
|
| 1033 |
+
"loss": 0.4708,
|
| 1034 |
+
"step": 1400
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 5.073469387755102,
|
| 1038 |
+
"grad_norm": 18.543668746948242,
|
| 1039 |
+
"learning_rate": 2.35827664399093e-05,
|
| 1040 |
+
"loss": 0.2931,
|
| 1041 |
+
"step": 1410
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 5.077551020408163,
|
| 1045 |
+
"grad_norm": 4.890735149383545,
|
| 1046 |
+
"learning_rate": 2.3356009070294786e-05,
|
| 1047 |
+
"loss": 0.4582,
|
| 1048 |
+
"step": 1420
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 5.081632653061225,
|
| 1052 |
+
"grad_norm": 3.5454611778259277,
|
| 1053 |
+
"learning_rate": 2.3129251700680275e-05,
|
| 1054 |
+
"loss": 0.3117,
|
| 1055 |
+
"step": 1430
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 5.085714285714285,
|
| 1059 |
+
"grad_norm": 19.41689109802246,
|
| 1060 |
+
"learning_rate": 2.290249433106576e-05,
|
| 1061 |
+
"loss": 0.465,
|
| 1062 |
+
"step": 1440
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 5.089795918367347,
|
| 1066 |
+
"grad_norm": 15.91568374633789,
|
| 1067 |
+
"learning_rate": 2.267573696145125e-05,
|
| 1068 |
+
"loss": 0.439,
|
| 1069 |
+
"step": 1450
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 5.093877551020408,
|
| 1073 |
+
"grad_norm": 6.774694442749023,
|
| 1074 |
+
"learning_rate": 2.2448979591836737e-05,
|
| 1075 |
+
"loss": 0.5177,
|
| 1076 |
+
"step": 1460
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 5.09795918367347,
|
| 1080 |
+
"grad_norm": 6.013669490814209,
|
| 1081 |
+
"learning_rate": 2.2222222222222223e-05,
|
| 1082 |
+
"loss": 0.3388,
|
| 1083 |
+
"step": 1470
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 5.100408163265306,
|
| 1087 |
+
"eval_accuracy": 0.7672514619883041,
|
| 1088 |
+
"eval_loss": 1.070756435394287,
|
| 1089 |
+
"eval_runtime": 762.1738,
|
| 1090 |
+
"eval_samples_per_second": 2.244,
|
| 1091 |
+
"eval_steps_per_second": 0.094,
|
| 1092 |
+
"step": 1476
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 6.0016326530612245,
|
| 1096 |
+
"grad_norm": 9.088335990905762,
|
| 1097 |
+
"learning_rate": 2.1995464852607713e-05,
|
| 1098 |
+
"loss": 0.3312,
|
| 1099 |
+
"step": 1480
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 6.005714285714285,
|
| 1103 |
+
"grad_norm": 5.159632205963135,
|
| 1104 |
+
"learning_rate": 2.17687074829932e-05,
|
| 1105 |
+
"loss": 0.2506,
|
| 1106 |
+
"step": 1490
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 6.009795918367347,
|
| 1110 |
+
"grad_norm": 3.322460174560547,
|
| 1111 |
+
"learning_rate": 2.1541950113378685e-05,
|
| 1112 |
+
"loss": 0.2576,
|
| 1113 |
+
"step": 1500
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 6.013877551020408,
|
| 1117 |
+
"grad_norm": 3.4191062450408936,
|
| 1118 |
+
"learning_rate": 2.1315192743764175e-05,
|
| 1119 |
+
"loss": 0.2997,
|
| 1120 |
+
"step": 1510
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 6.01795918367347,
|
| 1124 |
+
"grad_norm": 6.195023059844971,
|
| 1125 |
+
"learning_rate": 2.108843537414966e-05,
|
| 1126 |
+
"loss": 0.3168,
|
| 1127 |
+
"step": 1520
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 6.02204081632653,
|
| 1131 |
+
"grad_norm": 5.718242168426514,
|
| 1132 |
+
"learning_rate": 2.0861678004535147e-05,
|
| 1133 |
+
"loss": 0.3911,
|
| 1134 |
+
"step": 1530
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 6.026122448979592,
|
| 1138 |
+
"grad_norm": 7.038877010345459,
|
| 1139 |
+
"learning_rate": 2.0634920634920636e-05,
|
| 1140 |
+
"loss": 0.3152,
|
| 1141 |
+
"step": 1540
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 6.030204081632653,
|
| 1145 |
+
"grad_norm": 8.966346740722656,
|
| 1146 |
+
"learning_rate": 2.0408163265306123e-05,
|
| 1147 |
+
"loss": 0.3379,
|
| 1148 |
+
"step": 1550
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.034285714285715,
|
| 1152 |
+
"grad_norm": 10.551358222961426,
|
| 1153 |
+
"learning_rate": 2.018140589569161e-05,
|
| 1154 |
+
"loss": 0.3484,
|
| 1155 |
+
"step": 1560
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 6.0383673469387755,
|
| 1159 |
+
"grad_norm": 8.191295623779297,
|
| 1160 |
+
"learning_rate": 1.9954648526077098e-05,
|
| 1161 |
+
"loss": 0.3636,
|
| 1162 |
+
"step": 1570
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 6.042448979591836,
|
| 1166 |
+
"grad_norm": 6.954779148101807,
|
| 1167 |
+
"learning_rate": 1.9727891156462584e-05,
|
| 1168 |
+
"loss": 0.2535,
|
| 1169 |
+
"step": 1580
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 6.046530612244898,
|
| 1173 |
+
"grad_norm": 7.736076354980469,
|
| 1174 |
+
"learning_rate": 1.9501133786848074e-05,
|
| 1175 |
+
"loss": 0.38,
|
| 1176 |
+
"step": 1590
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 6.050612244897959,
|
| 1180 |
+
"grad_norm": 3.5329229831695557,
|
| 1181 |
+
"learning_rate": 1.927437641723356e-05,
|
| 1182 |
+
"loss": 0.2262,
|
| 1183 |
+
"step": 1600
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 6.054693877551021,
|
| 1187 |
+
"grad_norm": 12.468216896057129,
|
| 1188 |
+
"learning_rate": 1.9047619047619046e-05,
|
| 1189 |
+
"loss": 0.3723,
|
| 1190 |
+
"step": 1610
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.0587755102040814,
|
| 1194 |
+
"grad_norm": 6.249111175537109,
|
| 1195 |
+
"learning_rate": 1.8820861678004536e-05,
|
| 1196 |
+
"loss": 0.4139,
|
| 1197 |
+
"step": 1620
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 6.062857142857143,
|
| 1201 |
+
"grad_norm": 6.767730236053467,
|
| 1202 |
+
"learning_rate": 1.8594104308390022e-05,
|
| 1203 |
+
"loss": 0.3185,
|
| 1204 |
+
"step": 1630
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 6.066938775510204,
|
| 1208 |
+
"grad_norm": 6.568230152130127,
|
| 1209 |
+
"learning_rate": 1.836734693877551e-05,
|
| 1210 |
+
"loss": 0.3436,
|
| 1211 |
+
"step": 1640
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 6.071020408163266,
|
| 1215 |
+
"grad_norm": 9.211495399475098,
|
| 1216 |
+
"learning_rate": 1.8140589569160997e-05,
|
| 1217 |
+
"loss": 0.3812,
|
| 1218 |
+
"step": 1650
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 6.0751020408163265,
|
| 1222 |
+
"grad_norm": 2.9120309352874756,
|
| 1223 |
+
"learning_rate": 1.7913832199546487e-05,
|
| 1224 |
+
"loss": 0.3187,
|
| 1225 |
+
"step": 1660
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 6.079183673469387,
|
| 1229 |
+
"grad_norm": 6.394941806793213,
|
| 1230 |
+
"learning_rate": 1.7687074829931973e-05,
|
| 1231 |
+
"loss": 0.3097,
|
| 1232 |
+
"step": 1670
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 6.083265306122449,
|
| 1236 |
+
"grad_norm": 7.2203192710876465,
|
| 1237 |
+
"learning_rate": 1.746031746031746e-05,
|
| 1238 |
+
"loss": 0.3097,
|
| 1239 |
+
"step": 1680
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 6.08734693877551,
|
| 1243 |
+
"grad_norm": 7.860384464263916,
|
| 1244 |
+
"learning_rate": 1.723356009070295e-05,
|
| 1245 |
+
"loss": 0.338,
|
| 1246 |
+
"step": 1690
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 6.091428571428572,
|
| 1250 |
+
"grad_norm": 11.43612289428711,
|
| 1251 |
+
"learning_rate": 1.7006802721088435e-05,
|
| 1252 |
+
"loss": 0.3943,
|
| 1253 |
+
"step": 1700
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 6.0955102040816325,
|
| 1257 |
+
"grad_norm": 3.026212215423584,
|
| 1258 |
+
"learning_rate": 1.6780045351473924e-05,
|
| 1259 |
+
"loss": 0.2806,
|
| 1260 |
+
"step": 1710
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 6.099591836734694,
|
| 1264 |
+
"grad_norm": 6.003236770629883,
|
| 1265 |
+
"learning_rate": 1.655328798185941e-05,
|
| 1266 |
+
"loss": 0.448,
|
| 1267 |
+
"step": 1720
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 6.100408163265306,
|
| 1271 |
+
"eval_accuracy": 0.7678362573099415,
|
| 1272 |
+
"eval_loss": 1.1309159994125366,
|
| 1273 |
+
"eval_runtime": 760.8059,
|
| 1274 |
+
"eval_samples_per_second": 2.248,
|
| 1275 |
+
"eval_steps_per_second": 0.095,
|
| 1276 |
+
"step": 1722
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 7.003265306122449,
|
| 1280 |
+
"grad_norm": 6.932886600494385,
|
| 1281 |
+
"learning_rate": 1.6326530612244897e-05,
|
| 1282 |
+
"loss": 0.2641,
|
| 1283 |
+
"step": 1730
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 7.00734693877551,
|
| 1287 |
+
"grad_norm": 8.136345863342285,
|
| 1288 |
+
"learning_rate": 1.6099773242630386e-05,
|
| 1289 |
+
"loss": 0.3278,
|
| 1290 |
+
"step": 1740
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 7.011428571428572,
|
| 1294 |
+
"grad_norm": 4.856635093688965,
|
| 1295 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 1296 |
+
"loss": 0.2806,
|
| 1297 |
+
"step": 1750
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 7.015510204081632,
|
| 1301 |
+
"grad_norm": 5.214656352996826,
|
| 1302 |
+
"learning_rate": 1.5646258503401362e-05,
|
| 1303 |
+
"loss": 0.3219,
|
| 1304 |
+
"step": 1760
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 7.019591836734694,
|
| 1308 |
+
"grad_norm": 15.459710121154785,
|
| 1309 |
+
"learning_rate": 1.5419501133786848e-05,
|
| 1310 |
+
"loss": 0.3271,
|
| 1311 |
+
"step": 1770
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 7.023673469387755,
|
| 1315 |
+
"grad_norm": 6.767300128936768,
|
| 1316 |
+
"learning_rate": 1.5192743764172338e-05,
|
| 1317 |
+
"loss": 0.2539,
|
| 1318 |
+
"step": 1780
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 7.027755102040817,
|
| 1322 |
+
"grad_norm": 8.534014701843262,
|
| 1323 |
+
"learning_rate": 1.4965986394557824e-05,
|
| 1324 |
+
"loss": 0.3525,
|
| 1325 |
+
"step": 1790
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 7.0318367346938775,
|
| 1329 |
+
"grad_norm": 6.5173187255859375,
|
| 1330 |
+
"learning_rate": 1.473922902494331e-05,
|
| 1331 |
+
"loss": 0.28,
|
| 1332 |
+
"step": 1800
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 7.035918367346939,
|
| 1336 |
+
"grad_norm": 5.321302890777588,
|
| 1337 |
+
"learning_rate": 1.45124716553288e-05,
|
| 1338 |
+
"loss": 0.2639,
|
| 1339 |
+
"step": 1810
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 7.04,
|
| 1343 |
+
"grad_norm": 11.483617782592773,
|
| 1344 |
+
"learning_rate": 1.4285714285714285e-05,
|
| 1345 |
+
"loss": 0.2779,
|
| 1346 |
+
"step": 1820
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.044081632653061,
|
| 1350 |
+
"grad_norm": 11.726727485656738,
|
| 1351 |
+
"learning_rate": 1.4058956916099775e-05,
|
| 1352 |
+
"loss": 0.3351,
|
| 1353 |
+
"step": 1830
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 7.048163265306123,
|
| 1357 |
+
"grad_norm": 8.581147193908691,
|
| 1358 |
+
"learning_rate": 1.3832199546485261e-05,
|
| 1359 |
+
"loss": 0.3345,
|
| 1360 |
+
"step": 1840
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 7.052244897959183,
|
| 1364 |
+
"grad_norm": 10.193452835083008,
|
| 1365 |
+
"learning_rate": 1.360544217687075e-05,
|
| 1366 |
+
"loss": 0.2625,
|
| 1367 |
+
"step": 1850
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 7.056326530612245,
|
| 1371 |
+
"grad_norm": 7.469024658203125,
|
| 1372 |
+
"learning_rate": 1.3378684807256237e-05,
|
| 1373 |
+
"loss": 0.267,
|
| 1374 |
+
"step": 1860
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 7.060408163265306,
|
| 1378 |
+
"grad_norm": 10.515392303466797,
|
| 1379 |
+
"learning_rate": 1.3151927437641723e-05,
|
| 1380 |
+
"loss": 0.2268,
|
| 1381 |
+
"step": 1870
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 7.064489795918368,
|
| 1385 |
+
"grad_norm": 1.660530924797058,
|
| 1386 |
+
"learning_rate": 1.2925170068027212e-05,
|
| 1387 |
+
"loss": 0.2194,
|
| 1388 |
+
"step": 1880
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.0685714285714285,
|
| 1392 |
+
"grad_norm": 14.603185653686523,
|
| 1393 |
+
"learning_rate": 1.2698412698412699e-05,
|
| 1394 |
+
"loss": 0.2231,
|
| 1395 |
+
"step": 1890
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 7.07265306122449,
|
| 1399 |
+
"grad_norm": 6.94572639465332,
|
| 1400 |
+
"learning_rate": 1.2471655328798186e-05,
|
| 1401 |
+
"loss": 0.2184,
|
| 1402 |
+
"step": 1900
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 7.076734693877551,
|
| 1406 |
+
"grad_norm": 9.031844139099121,
|
| 1407 |
+
"learning_rate": 1.2244897959183674e-05,
|
| 1408 |
+
"loss": 0.3445,
|
| 1409 |
+
"step": 1910
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 7.080816326530612,
|
| 1413 |
+
"grad_norm": 11.935467720031738,
|
| 1414 |
+
"learning_rate": 1.2018140589569162e-05,
|
| 1415 |
+
"loss": 0.2813,
|
| 1416 |
+
"step": 1920
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 7.084897959183674,
|
| 1420 |
+
"grad_norm": 16.639739990234375,
|
| 1421 |
+
"learning_rate": 1.179138321995465e-05,
|
| 1422 |
+
"loss": 0.2856,
|
| 1423 |
+
"step": 1930
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 7.088979591836734,
|
| 1427 |
+
"grad_norm": 12.219932556152344,
|
| 1428 |
+
"learning_rate": 1.1564625850340138e-05,
|
| 1429 |
+
"loss": 0.3086,
|
| 1430 |
+
"step": 1940
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.093061224489796,
|
| 1434 |
+
"grad_norm": 6.9921875,
|
| 1435 |
+
"learning_rate": 1.1337868480725626e-05,
|
| 1436 |
+
"loss": 0.3269,
|
| 1437 |
+
"step": 1950
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 7.097142857142857,
|
| 1441 |
+
"grad_norm": 6.7659406661987305,
|
| 1442 |
+
"learning_rate": 1.1111111111111112e-05,
|
| 1443 |
+
"loss": 0.2328,
|
| 1444 |
+
"step": 1960
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 7.100408163265306,
|
| 1448 |
+
"eval_accuracy": 0.7736842105263158,
|
| 1449 |
+
"eval_loss": 1.0998854637145996,
|
| 1450 |
+
"eval_runtime": 755.5903,
|
| 1451 |
+
"eval_samples_per_second": 2.263,
|
| 1452 |
+
"eval_steps_per_second": 0.095,
|
| 1453 |
+
"step": 1968
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 8.000816326530613,
|
| 1457 |
+
"grad_norm": 7.732854843139648,
|
| 1458 |
+
"learning_rate": 1.08843537414966e-05,
|
| 1459 |
+
"loss": 0.4175,
|
| 1460 |
+
"step": 1970
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 8.004897959183673,
|
| 1464 |
+
"grad_norm": 4.466553688049316,
|
| 1465 |
+
"learning_rate": 1.0657596371882087e-05,
|
| 1466 |
+
"loss": 0.2961,
|
| 1467 |
+
"step": 1980
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 8.008979591836734,
|
| 1471 |
+
"grad_norm": 10.974610328674316,
|
| 1472 |
+
"learning_rate": 1.0430839002267573e-05,
|
| 1473 |
+
"loss": 0.257,
|
| 1474 |
+
"step": 1990
|
| 1475 |
+
},
|
| 1476 |
+
{
|
| 1477 |
+
"epoch": 8.013061224489796,
|
| 1478 |
+
"grad_norm": 8.84461784362793,
|
| 1479 |
+
"learning_rate": 1.0204081632653061e-05,
|
| 1480 |
+
"loss": 0.2006,
|
| 1481 |
+
"step": 2000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 8.017142857142858,
|
| 1485 |
+
"grad_norm": 8.123458862304688,
|
| 1486 |
+
"learning_rate": 9.977324263038549e-06,
|
| 1487 |
+
"loss": 0.2158,
|
| 1488 |
+
"step": 2010
|
| 1489 |
+
},
|
| 1490 |
+
{
|
| 1491 |
+
"epoch": 8.021224489795918,
|
| 1492 |
+
"grad_norm": 1.604798436164856,
|
| 1493 |
+
"learning_rate": 9.750566893424037e-06,
|
| 1494 |
+
"loss": 0.1893,
|
| 1495 |
+
"step": 2020
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 8.02530612244898,
|
| 1499 |
+
"grad_norm": 8.343064308166504,
|
| 1500 |
+
"learning_rate": 9.523809523809523e-06,
|
| 1501 |
+
"loss": 0.2289,
|
| 1502 |
+
"step": 2030
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 8.029387755102041,
|
| 1506 |
+
"grad_norm": 1.9065176248550415,
|
| 1507 |
+
"learning_rate": 9.297052154195011e-06,
|
| 1508 |
+
"loss": 0.2887,
|
| 1509 |
+
"step": 2040
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 8.033469387755103,
|
| 1513 |
+
"grad_norm": 10.564911842346191,
|
| 1514 |
+
"learning_rate": 9.070294784580499e-06,
|
| 1515 |
+
"loss": 0.2056,
|
| 1516 |
+
"step": 2050
|
| 1517 |
+
},
|
| 1518 |
+
{
|
| 1519 |
+
"epoch": 8.037551020408163,
|
| 1520 |
+
"grad_norm": 9.196011543273926,
|
| 1521 |
+
"learning_rate": 8.843537414965987e-06,
|
| 1522 |
+
"loss": 0.2425,
|
| 1523 |
+
"step": 2060
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 8.041632653061225,
|
| 1527 |
+
"grad_norm": 7.146062850952148,
|
| 1528 |
+
"learning_rate": 8.616780045351474e-06,
|
| 1529 |
+
"loss": 0.1673,
|
| 1530 |
+
"step": 2070
|
| 1531 |
+
},
|
| 1532 |
+
{
|
| 1533 |
+
"epoch": 8.045714285714286,
|
| 1534 |
+
"grad_norm": 5.323041915893555,
|
| 1535 |
+
"learning_rate": 8.390022675736962e-06,
|
| 1536 |
+
"loss": 0.1949,
|
| 1537 |
+
"step": 2080
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 8.049795918367346,
|
| 1541 |
+
"grad_norm": 10.371600151062012,
|
| 1542 |
+
"learning_rate": 8.163265306122448e-06,
|
| 1543 |
+
"loss": 0.2263,
|
| 1544 |
+
"step": 2090
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 8.053877551020408,
|
| 1548 |
+
"grad_norm": 0.6877457499504089,
|
| 1549 |
+
"learning_rate": 7.936507936507936e-06,
|
| 1550 |
+
"loss": 0.2532,
|
| 1551 |
+
"step": 2100
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 8.05795918367347,
|
| 1555 |
+
"grad_norm": 2.1630640029907227,
|
| 1556 |
+
"learning_rate": 7.709750566893424e-06,
|
| 1557 |
+
"loss": 0.2739,
|
| 1558 |
+
"step": 2110
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 8.062040816326531,
|
| 1562 |
+
"grad_norm": 6.128444194793701,
|
| 1563 |
+
"learning_rate": 7.482993197278912e-06,
|
| 1564 |
+
"loss": 0.2762,
|
| 1565 |
+
"step": 2120
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 8.066122448979591,
|
| 1569 |
+
"grad_norm": 10.738118171691895,
|
| 1570 |
+
"learning_rate": 7.2562358276644e-06,
|
| 1571 |
+
"loss": 0.3025,
|
| 1572 |
+
"step": 2130
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 8.070204081632653,
|
| 1576 |
+
"grad_norm": 9.642598152160645,
|
| 1577 |
+
"learning_rate": 7.0294784580498875e-06,
|
| 1578 |
+
"loss": 0.2756,
|
| 1579 |
+
"step": 2140
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 8.074285714285715,
|
| 1583 |
+
"grad_norm": 6.180073261260986,
|
| 1584 |
+
"learning_rate": 6.802721088435375e-06,
|
| 1585 |
+
"loss": 0.3265,
|
| 1586 |
+
"step": 2150
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 8.078367346938775,
|
| 1590 |
+
"grad_norm": 5.613744735717773,
|
| 1591 |
+
"learning_rate": 6.5759637188208614e-06,
|
| 1592 |
+
"loss": 0.2121,
|
| 1593 |
+
"step": 2160
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 8.082448979591836,
|
| 1597 |
+
"grad_norm": 8.959142684936523,
|
| 1598 |
+
"learning_rate": 6.349206349206349e-06,
|
| 1599 |
+
"loss": 0.1768,
|
| 1600 |
+
"step": 2170
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 8.086530612244898,
|
| 1604 |
+
"grad_norm": 5.29737663269043,
|
| 1605 |
+
"learning_rate": 6.122448979591837e-06,
|
| 1606 |
+
"loss": 0.25,
|
| 1607 |
+
"step": 2180
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 8.09061224489796,
|
| 1611 |
+
"grad_norm": 5.707335948944092,
|
| 1612 |
+
"learning_rate": 5.895691609977325e-06,
|
| 1613 |
+
"loss": 0.2059,
|
| 1614 |
+
"step": 2190
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 8.09469387755102,
|
| 1618 |
+
"grad_norm": 5.561722278594971,
|
| 1619 |
+
"learning_rate": 5.668934240362813e-06,
|
| 1620 |
+
"loss": 0.1966,
|
| 1621 |
+
"step": 2200
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 8.098775510204081,
|
| 1625 |
+
"grad_norm": 5.105309009552002,
|
| 1626 |
+
"learning_rate": 5.4421768707483e-06,
|
| 1627 |
+
"loss": 0.2329,
|
| 1628 |
+
"step": 2210
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 8.100408163265307,
|
| 1632 |
+
"eval_accuracy": 0.7807017543859649,
|
| 1633 |
+
"eval_loss": 1.1141033172607422,
|
| 1634 |
+
"eval_runtime": 761.671,
|
| 1635 |
+
"eval_samples_per_second": 2.245,
|
| 1636 |
+
"eval_steps_per_second": 0.095,
|
| 1637 |
+
"step": 2214
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 9.002448979591836,
|
| 1641 |
+
"grad_norm": 9.717642784118652,
|
| 1642 |
+
"learning_rate": 5.215419501133787e-06,
|
| 1643 |
+
"loss": 0.199,
|
| 1644 |
+
"step": 2220
|
| 1645 |
+
},
|
| 1646 |
+
{
|
| 1647 |
+
"epoch": 9.006530612244898,
|
| 1648 |
+
"grad_norm": 1.3099863529205322,
|
| 1649 |
+
"learning_rate": 4.9886621315192745e-06,
|
| 1650 |
+
"loss": 0.1484,
|
| 1651 |
+
"step": 2230
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 9.01061224489796,
|
| 1655 |
+
"grad_norm": 6.000232696533203,
|
| 1656 |
+
"learning_rate": 4.7619047619047615e-06,
|
| 1657 |
+
"loss": 0.2215,
|
| 1658 |
+
"step": 2240
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 9.01469387755102,
|
| 1662 |
+
"grad_norm": 6.975632667541504,
|
| 1663 |
+
"learning_rate": 4.535147392290249e-06,
|
| 1664 |
+
"loss": 0.2325,
|
| 1665 |
+
"step": 2250
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 9.018775510204081,
|
| 1669 |
+
"grad_norm": 6.907474994659424,
|
| 1670 |
+
"learning_rate": 4.308390022675737e-06,
|
| 1671 |
+
"loss": 0.1989,
|
| 1672 |
+
"step": 2260
|
| 1673 |
+
},
|
| 1674 |
+
{
|
| 1675 |
+
"epoch": 9.022857142857143,
|
| 1676 |
+
"grad_norm": 3.785592794418335,
|
| 1677 |
+
"learning_rate": 4.081632653061224e-06,
|
| 1678 |
+
"loss": 0.2137,
|
| 1679 |
+
"step": 2270
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 9.026938775510205,
|
| 1683 |
+
"grad_norm": 5.131585597991943,
|
| 1684 |
+
"learning_rate": 3.854875283446712e-06,
|
| 1685 |
+
"loss": 0.1331,
|
| 1686 |
+
"step": 2280
|
| 1687 |
+
},
|
| 1688 |
+
{
|
| 1689 |
+
"epoch": 9.031020408163265,
|
| 1690 |
+
"grad_norm": 7.016502380371094,
|
| 1691 |
+
"learning_rate": 3.6281179138322e-06,
|
| 1692 |
+
"loss": 0.2636,
|
| 1693 |
+
"step": 2290
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 9.035102040816327,
|
| 1697 |
+
"grad_norm": 4.56027364730835,
|
| 1698 |
+
"learning_rate": 3.4013605442176877e-06,
|
| 1699 |
+
"loss": 0.1722,
|
| 1700 |
+
"step": 2300
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.039183673469388,
|
| 1704 |
+
"grad_norm": 23.117076873779297,
|
| 1705 |
+
"learning_rate": 3.1746031746031746e-06,
|
| 1706 |
+
"loss": 0.2598,
|
| 1707 |
+
"step": 2310
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 9.043265306122448,
|
| 1711 |
+
"grad_norm": 2.165933132171631,
|
| 1712 |
+
"learning_rate": 2.9478458049886625e-06,
|
| 1713 |
+
"loss": 0.1233,
|
| 1714 |
+
"step": 2320
|
| 1715 |
+
},
|
| 1716 |
+
{
|
| 1717 |
+
"epoch": 9.04734693877551,
|
| 1718 |
+
"grad_norm": 9.032445907592773,
|
| 1719 |
+
"learning_rate": 2.72108843537415e-06,
|
| 1720 |
+
"loss": 0.1919,
|
| 1721 |
+
"step": 2330
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 9.051428571428572,
|
| 1725 |
+
"grad_norm": 3.221755266189575,
|
| 1726 |
+
"learning_rate": 2.4943310657596373e-06,
|
| 1727 |
+
"loss": 0.2282,
|
| 1728 |
+
"step": 2340
|
| 1729 |
+
},
|
| 1730 |
+
{
|
| 1731 |
+
"epoch": 9.055510204081633,
|
| 1732 |
+
"grad_norm": 2.723907709121704,
|
| 1733 |
+
"learning_rate": 2.2675736961451247e-06,
|
| 1734 |
+
"loss": 0.2176,
|
| 1735 |
+
"step": 2350
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 9.059591836734693,
|
| 1739 |
+
"grad_norm": 7.58160924911499,
|
| 1740 |
+
"learning_rate": 2.040816326530612e-06,
|
| 1741 |
+
"loss": 0.1815,
|
| 1742 |
+
"step": 2360
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 9.063673469387755,
|
| 1746 |
+
"grad_norm": 11.342548370361328,
|
| 1747 |
+
"learning_rate": 1.8140589569161e-06,
|
| 1748 |
+
"loss": 0.2318,
|
| 1749 |
+
"step": 2370
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 9.067755102040817,
|
| 1753 |
+
"grad_norm": 6.335953235626221,
|
| 1754 |
+
"learning_rate": 1.5873015873015873e-06,
|
| 1755 |
+
"loss": 0.1899,
|
| 1756 |
+
"step": 2380
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 9.071836734693878,
|
| 1760 |
+
"grad_norm": 12.549899101257324,
|
| 1761 |
+
"learning_rate": 1.360544217687075e-06,
|
| 1762 |
+
"loss": 0.2156,
|
| 1763 |
+
"step": 2390
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 9.075918367346938,
|
| 1767 |
+
"grad_norm": 13.524937629699707,
|
| 1768 |
+
"learning_rate": 1.1337868480725623e-06,
|
| 1769 |
+
"loss": 0.1842,
|
| 1770 |
+
"step": 2400
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 9.08,
|
| 1774 |
+
"grad_norm": 3.897021532058716,
|
| 1775 |
+
"learning_rate": 9.0702947845805e-07,
|
| 1776 |
+
"loss": 0.2026,
|
| 1777 |
+
"step": 2410
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 9.084081632653062,
|
| 1781 |
+
"grad_norm": 16.518413543701172,
|
| 1782 |
+
"learning_rate": 6.802721088435375e-07,
|
| 1783 |
+
"loss": 0.2534,
|
| 1784 |
+
"step": 2420
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 9.088163265306122,
|
| 1788 |
+
"grad_norm": 4.3819580078125,
|
| 1789 |
+
"learning_rate": 4.53514739229025e-07,
|
| 1790 |
+
"loss": 0.2089,
|
| 1791 |
+
"step": 2430
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 9.092244897959183,
|
| 1795 |
+
"grad_norm": 9.91723918914795,
|
| 1796 |
+
"learning_rate": 2.267573696145125e-07,
|
| 1797 |
+
"loss": 0.2467,
|
| 1798 |
+
"step": 2440
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 9.096326530612245,
|
| 1802 |
+
"grad_norm": 5.503404140472412,
|
| 1803 |
+
"learning_rate": 0.0,
|
| 1804 |
+
"loss": 0.1596,
|
| 1805 |
+
"step": 2450
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 9.096326530612245,
|
| 1809 |
+
"eval_accuracy": 0.7795321637426901,
|
| 1810 |
+
"eval_loss": 1.1184571981430054,
|
| 1811 |
+
"eval_runtime": 767.3021,
|
| 1812 |
+
"eval_samples_per_second": 2.229,
|
| 1813 |
+
"eval_steps_per_second": 0.094,
|
| 1814 |
+
"step": 2450
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 9.096326530612245,
|
| 1818 |
+
"step": 2450,
|
| 1819 |
+
"total_flos": 7.308381545357289e+19,
|
| 1820 |
+
"train_loss": 0.6407903413869896,
|
| 1821 |
+
"train_runtime": 33040.9588,
|
| 1822 |
+
"train_samples_per_second": 1.78,
|
| 1823 |
+
"train_steps_per_second": 0.074
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 9.096326530612245,
|
| 1827 |
+
"eval_accuracy": 0.8550922761449077,
|
| 1828 |
+
"eval_loss": 0.5230759382247925,
|
| 1829 |
+
"eval_runtime": 638.0216,
|
| 1830 |
+
"eval_samples_per_second": 2.293,
|
| 1831 |
+
"eval_steps_per_second": 0.096,
|
| 1832 |
+
"step": 2450
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 9.096326530612245,
|
| 1836 |
+
"eval_accuracy": 0.8550922761449077,
|
| 1837 |
+
"eval_loss": 0.5230759382247925,
|
| 1838 |
+
"eval_runtime": 633.2216,
|
| 1839 |
+
"eval_samples_per_second": 2.31,
|
| 1840 |
+
"eval_steps_per_second": 0.096,
|
| 1841 |
+
"step": 2450
|
| 1842 |
+
}
|
| 1843 |
+
],
|
| 1844 |
+
"logging_steps": 10,
|
| 1845 |
+
"max_steps": 2450,
|
| 1846 |
+
"num_input_tokens_seen": 0,
|
| 1847 |
+
"num_train_epochs": 9223372036854775807,
|
| 1848 |
+
"save_steps": 500,
|
| 1849 |
+
"stateful_callbacks": {
|
| 1850 |
+
"TrainerControl": {
|
| 1851 |
+
"args": {
|
| 1852 |
+
"should_epoch_stop": false,
|
| 1853 |
+
"should_evaluate": false,
|
| 1854 |
+
"should_log": false,
|
| 1855 |
+
"should_save": true,
|
| 1856 |
+
"should_training_stop": true
|
| 1857 |
+
},
|
| 1858 |
+
"attributes": {}
|
| 1859 |
+
}
|
| 1860 |
+
},
|
| 1861 |
+
"total_flos": 7.308381545357289e+19,
|
| 1862 |
+
"train_batch_size": 24,
|
| 1863 |
+
"trial_name": null,
|
| 1864 |
+
"trial_params": null
|
| 1865 |
+
}
|