End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +1549 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_loss": 1.7273770570755005,
|
| 4 |
+
"eval_runtime": 36.5966,
|
| 5 |
+
"eval_samples": 1999,
|
| 6 |
+
"eval_samples_per_second": 54.623,
|
| 7 |
+
"eval_steps_per_second": 1.721,
|
| 8 |
+
"train_loss": 1.219492632603017,
|
| 9 |
+
"train_runtime": 160361.9384,
|
| 10 |
+
"train_samples": 610320,
|
| 11 |
+
"train_samples_per_second": 38.059,
|
| 12 |
+
"train_steps_per_second": 0.793
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_loss": 1.7273770570755005,
|
| 4 |
+
"eval_runtime": 36.5966,
|
| 5 |
+
"eval_samples": 1999,
|
| 6 |
+
"eval_samples_per_second": 54.623,
|
| 7 |
+
"eval_steps_per_second": 1.721
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 1.219492632603017,
|
| 4 |
+
"train_runtime": 160361.9384,
|
| 5 |
+
"train_samples": 610320,
|
| 6 |
+
"train_samples_per_second": 38.059,
|
| 7 |
+
"train_steps_per_second": 0.793
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1549 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"global_step": 127150,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 4.980338183248132e-05,
|
| 13 |
+
"loss": 8.0518,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.08,
|
| 18 |
+
"learning_rate": 4.9606763664962646e-05,
|
| 19 |
+
"loss": 3.358,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.12,
|
| 24 |
+
"learning_rate": 4.9410145497443964e-05,
|
| 25 |
+
"loss": 2.9014,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.16,
|
| 30 |
+
"learning_rate": 4.921352732992528e-05,
|
| 31 |
+
"loss": 2.6617,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.2,
|
| 36 |
+
"learning_rate": 4.901690916240661e-05,
|
| 37 |
+
"loss": 2.4945,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.24,
|
| 42 |
+
"learning_rate": 4.882029099488793e-05,
|
| 43 |
+
"loss": 2.3686,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.28,
|
| 48 |
+
"learning_rate": 4.862367282736925e-05,
|
| 49 |
+
"loss": 2.2674,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.31,
|
| 54 |
+
"learning_rate": 4.8427054659850575e-05,
|
| 55 |
+
"loss": 2.1889,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.35,
|
| 60 |
+
"learning_rate": 4.8230436492331893e-05,
|
| 61 |
+
"loss": 2.1107,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.39,
|
| 66 |
+
"learning_rate": 4.803381832481321e-05,
|
| 67 |
+
"loss": 2.0427,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.43,
|
| 72 |
+
"learning_rate": 4.783720015729454e-05,
|
| 73 |
+
"loss": 1.9907,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.47,
|
| 78 |
+
"learning_rate": 4.7640581989775855e-05,
|
| 79 |
+
"loss": 1.9401,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.51,
|
| 84 |
+
"learning_rate": 4.744396382225718e-05,
|
| 85 |
+
"loss": 1.9027,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.55,
|
| 90 |
+
"learning_rate": 4.72473456547385e-05,
|
| 91 |
+
"loss": 1.8655,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.59,
|
| 96 |
+
"learning_rate": 4.705072748721982e-05,
|
| 97 |
+
"loss": 1.823,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.63,
|
| 102 |
+
"learning_rate": 4.685410931970115e-05,
|
| 103 |
+
"loss": 1.7995,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.67,
|
| 108 |
+
"learning_rate": 4.6657491152182466e-05,
|
| 109 |
+
"loss": 1.7682,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.71,
|
| 114 |
+
"learning_rate": 4.6460872984663785e-05,
|
| 115 |
+
"loss": 1.7367,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.75,
|
| 120 |
+
"learning_rate": 4.626425481714511e-05,
|
| 121 |
+
"loss": 1.7111,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.79,
|
| 126 |
+
"learning_rate": 4.606763664962643e-05,
|
| 127 |
+
"loss": 1.6903,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.83,
|
| 132 |
+
"learning_rate": 4.5871018482107746e-05,
|
| 133 |
+
"loss": 1.6655,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.87,
|
| 138 |
+
"learning_rate": 4.567440031458907e-05,
|
| 139 |
+
"loss": 1.6468,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.9,
|
| 144 |
+
"learning_rate": 4.547778214707039e-05,
|
| 145 |
+
"loss": 1.6196,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.94,
|
| 150 |
+
"learning_rate": 4.528116397955171e-05,
|
| 151 |
+
"loss": 1.6017,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.98,
|
| 156 |
+
"learning_rate": 4.508454581203303e-05,
|
| 157 |
+
"loss": 1.5889,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.02,
|
| 162 |
+
"learning_rate": 4.488792764451436e-05,
|
| 163 |
+
"loss": 1.5679,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.06,
|
| 168 |
+
"learning_rate": 4.4691309476995676e-05,
|
| 169 |
+
"loss": 1.5419,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.1,
|
| 174 |
+
"learning_rate": 4.4494691309477e-05,
|
| 175 |
+
"loss": 1.5339,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.14,
|
| 180 |
+
"learning_rate": 4.429807314195832e-05,
|
| 181 |
+
"loss": 1.5189,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.18,
|
| 186 |
+
"learning_rate": 4.410145497443964e-05,
|
| 187 |
+
"loss": 1.5071,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.22,
|
| 192 |
+
"learning_rate": 4.390483680692096e-05,
|
| 193 |
+
"loss": 1.4936,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.26,
|
| 198 |
+
"learning_rate": 4.370821863940228e-05,
|
| 199 |
+
"loss": 1.4849,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.3,
|
| 204 |
+
"learning_rate": 4.3511600471883605e-05,
|
| 205 |
+
"loss": 1.4671,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.34,
|
| 210 |
+
"learning_rate": 4.3314982304364923e-05,
|
| 211 |
+
"loss": 1.4519,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.38,
|
| 216 |
+
"learning_rate": 4.311836413684625e-05,
|
| 217 |
+
"loss": 1.4411,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.42,
|
| 222 |
+
"learning_rate": 4.2921745969327573e-05,
|
| 223 |
+
"loss": 1.4355,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.45,
|
| 228 |
+
"learning_rate": 4.272512780180889e-05,
|
| 229 |
+
"loss": 1.4224,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.49,
|
| 234 |
+
"learning_rate": 4.252850963429021e-05,
|
| 235 |
+
"loss": 1.4221,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.53,
|
| 240 |
+
"learning_rate": 4.2331891466771535e-05,
|
| 241 |
+
"loss": 1.4053,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.57,
|
| 246 |
+
"learning_rate": 4.213527329925285e-05,
|
| 247 |
+
"loss": 1.394,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.61,
|
| 252 |
+
"learning_rate": 4.193865513173417e-05,
|
| 253 |
+
"loss": 1.3904,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.65,
|
| 258 |
+
"learning_rate": 4.1742036964215496e-05,
|
| 259 |
+
"loss": 1.377,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.69,
|
| 264 |
+
"learning_rate": 4.1545418796696815e-05,
|
| 265 |
+
"loss": 1.3735,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.73,
|
| 270 |
+
"learning_rate": 4.134880062917813e-05,
|
| 271 |
+
"loss": 1.364,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.77,
|
| 276 |
+
"learning_rate": 4.115218246165946e-05,
|
| 277 |
+
"loss": 1.3605,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.81,
|
| 282 |
+
"learning_rate": 4.095556429414078e-05,
|
| 283 |
+
"loss": 1.3539,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.85,
|
| 288 |
+
"learning_rate": 4.07589461266221e-05,
|
| 289 |
+
"loss": 1.3397,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.89,
|
| 294 |
+
"learning_rate": 4.0562327959103426e-05,
|
| 295 |
+
"loss": 1.3326,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.93,
|
| 300 |
+
"learning_rate": 4.0365709791584744e-05,
|
| 301 |
+
"loss": 1.3295,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.97,
|
| 306 |
+
"learning_rate": 4.016909162406606e-05,
|
| 307 |
+
"loss": 1.3297,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 2.01,
|
| 312 |
+
"learning_rate": 3.997247345654739e-05,
|
| 313 |
+
"loss": 1.312,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 2.04,
|
| 318 |
+
"learning_rate": 3.9775855289028706e-05,
|
| 319 |
+
"loss": 1.2981,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.08,
|
| 324 |
+
"learning_rate": 3.957923712151003e-05,
|
| 325 |
+
"loss": 1.2988,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.12,
|
| 330 |
+
"learning_rate": 3.938261895399135e-05,
|
| 331 |
+
"loss": 1.289,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 2.16,
|
| 336 |
+
"learning_rate": 3.9186000786472674e-05,
|
| 337 |
+
"loss": 1.2861,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 2.2,
|
| 342 |
+
"learning_rate": 3.8989382618954e-05,
|
| 343 |
+
"loss": 1.2823,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 2.24,
|
| 348 |
+
"learning_rate": 3.879276445143532e-05,
|
| 349 |
+
"loss": 1.2718,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 2.28,
|
| 354 |
+
"learning_rate": 3.8596146283916635e-05,
|
| 355 |
+
"loss": 1.2673,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.32,
|
| 360 |
+
"learning_rate": 3.839952811639796e-05,
|
| 361 |
+
"loss": 1.2672,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.36,
|
| 366 |
+
"learning_rate": 3.820290994887928e-05,
|
| 367 |
+
"loss": 1.2538,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.4,
|
| 372 |
+
"learning_rate": 3.80062917813606e-05,
|
| 373 |
+
"loss": 1.2533,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.44,
|
| 378 |
+
"learning_rate": 3.780967361384192e-05,
|
| 379 |
+
"loss": 1.2519,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.48,
|
| 384 |
+
"learning_rate": 3.761305544632324e-05,
|
| 385 |
+
"loss": 1.2437,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.52,
|
| 390 |
+
"learning_rate": 3.741643727880456e-05,
|
| 391 |
+
"loss": 1.2363,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.56,
|
| 396 |
+
"learning_rate": 3.721981911128588e-05,
|
| 397 |
+
"loss": 1.2397,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.6,
|
| 402 |
+
"learning_rate": 3.702320094376721e-05,
|
| 403 |
+
"loss": 1.2328,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.63,
|
| 408 |
+
"learning_rate": 3.6826582776248526e-05,
|
| 409 |
+
"loss": 1.2336,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.67,
|
| 414 |
+
"learning_rate": 3.662996460872985e-05,
|
| 415 |
+
"loss": 1.2271,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.71,
|
| 420 |
+
"learning_rate": 3.643334644121117e-05,
|
| 421 |
+
"loss": 1.2199,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.75,
|
| 426 |
+
"learning_rate": 3.623672827369249e-05,
|
| 427 |
+
"loss": 1.2167,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.79,
|
| 432 |
+
"learning_rate": 3.604011010617381e-05,
|
| 433 |
+
"loss": 1.2115,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.83,
|
| 438 |
+
"learning_rate": 3.584349193865513e-05,
|
| 439 |
+
"loss": 1.2095,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.87,
|
| 444 |
+
"learning_rate": 3.564687377113645e-05,
|
| 445 |
+
"loss": 1.2043,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.91,
|
| 450 |
+
"learning_rate": 3.5450255603617774e-05,
|
| 451 |
+
"loss": 1.2029,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.95,
|
| 456 |
+
"learning_rate": 3.52536374360991e-05,
|
| 457 |
+
"loss": 1.1952,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.99,
|
| 462 |
+
"learning_rate": 3.5057019268580424e-05,
|
| 463 |
+
"loss": 1.1916,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 3.03,
|
| 468 |
+
"learning_rate": 3.486040110106174e-05,
|
| 469 |
+
"loss": 1.183,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 3.07,
|
| 474 |
+
"learning_rate": 3.466378293354306e-05,
|
| 475 |
+
"loss": 1.1823,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 3.11,
|
| 480 |
+
"learning_rate": 3.4467164766024386e-05,
|
| 481 |
+
"loss": 1.1806,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 3.15,
|
| 486 |
+
"learning_rate": 3.4270546598505704e-05,
|
| 487 |
+
"loss": 1.182,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 3.19,
|
| 492 |
+
"learning_rate": 3.407392843098702e-05,
|
| 493 |
+
"loss": 1.1705,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 3.22,
|
| 498 |
+
"learning_rate": 3.387731026346835e-05,
|
| 499 |
+
"loss": 1.174,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 3.26,
|
| 504 |
+
"learning_rate": 3.3680692095949665e-05,
|
| 505 |
+
"loss": 1.1634,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 3.3,
|
| 510 |
+
"learning_rate": 3.3484073928430984e-05,
|
| 511 |
+
"loss": 1.1655,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 3.34,
|
| 516 |
+
"learning_rate": 3.328745576091231e-05,
|
| 517 |
+
"loss": 1.1622,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 3.38,
|
| 522 |
+
"learning_rate": 3.3090837593393633e-05,
|
| 523 |
+
"loss": 1.161,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 3.42,
|
| 528 |
+
"learning_rate": 3.289421942587495e-05,
|
| 529 |
+
"loss": 1.1495,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 3.46,
|
| 534 |
+
"learning_rate": 3.269760125835628e-05,
|
| 535 |
+
"loss": 1.1522,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 3.5,
|
| 540 |
+
"learning_rate": 3.2500983090837595e-05,
|
| 541 |
+
"loss": 1.1483,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 3.54,
|
| 546 |
+
"learning_rate": 3.230436492331891e-05,
|
| 547 |
+
"loss": 1.1516,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 3.58,
|
| 552 |
+
"learning_rate": 3.210774675580024e-05,
|
| 553 |
+
"loss": 1.1475,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 3.62,
|
| 558 |
+
"learning_rate": 3.1911128588281556e-05,
|
| 559 |
+
"loss": 1.1386,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 3.66,
|
| 564 |
+
"learning_rate": 3.1714510420762875e-05,
|
| 565 |
+
"loss": 1.142,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.7,
|
| 570 |
+
"learning_rate": 3.15178922532442e-05,
|
| 571 |
+
"loss": 1.1403,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.74,
|
| 576 |
+
"learning_rate": 3.1321274085725525e-05,
|
| 577 |
+
"loss": 1.1328,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.78,
|
| 582 |
+
"learning_rate": 3.112465591820685e-05,
|
| 583 |
+
"loss": 1.1257,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.81,
|
| 588 |
+
"learning_rate": 3.092803775068817e-05,
|
| 589 |
+
"loss": 1.1259,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.85,
|
| 594 |
+
"learning_rate": 3.0731419583169486e-05,
|
| 595 |
+
"loss": 1.128,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.89,
|
| 600 |
+
"learning_rate": 3.053480141565081e-05,
|
| 601 |
+
"loss": 1.1299,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.93,
|
| 606 |
+
"learning_rate": 3.033818324813213e-05,
|
| 607 |
+
"loss": 1.1285,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.97,
|
| 612 |
+
"learning_rate": 3.0141565080613447e-05,
|
| 613 |
+
"loss": 1.1224,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 4.01,
|
| 618 |
+
"learning_rate": 2.9944946913094772e-05,
|
| 619 |
+
"loss": 1.116,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 4.05,
|
| 624 |
+
"learning_rate": 2.9748328745576094e-05,
|
| 625 |
+
"loss": 1.1125,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 4.09,
|
| 630 |
+
"learning_rate": 2.9551710578057412e-05,
|
| 631 |
+
"loss": 1.1031,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 4.13,
|
| 636 |
+
"learning_rate": 2.9355092410538737e-05,
|
| 637 |
+
"loss": 1.1091,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 4.17,
|
| 642 |
+
"learning_rate": 2.9158474243020055e-05,
|
| 643 |
+
"loss": 1.1058,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 4.21,
|
| 648 |
+
"learning_rate": 2.8961856075501377e-05,
|
| 649 |
+
"loss": 1.1025,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 4.25,
|
| 654 |
+
"learning_rate": 2.8765237907982702e-05,
|
| 655 |
+
"loss": 1.1076,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 4.29,
|
| 660 |
+
"learning_rate": 2.856861974046402e-05,
|
| 661 |
+
"loss": 1.1052,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 4.33,
|
| 666 |
+
"learning_rate": 2.837200157294534e-05,
|
| 667 |
+
"loss": 1.0995,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 4.36,
|
| 672 |
+
"learning_rate": 2.8175383405426664e-05,
|
| 673 |
+
"loss": 1.0948,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 4.4,
|
| 678 |
+
"learning_rate": 2.7978765237907985e-05,
|
| 679 |
+
"loss": 1.0985,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 4.44,
|
| 684 |
+
"learning_rate": 2.7782147070389303e-05,
|
| 685 |
+
"loss": 1.0969,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 4.48,
|
| 690 |
+
"learning_rate": 2.758552890287063e-05,
|
| 691 |
+
"loss": 1.0921,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.52,
|
| 696 |
+
"learning_rate": 2.7388910735351947e-05,
|
| 697 |
+
"loss": 1.0913,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.56,
|
| 702 |
+
"learning_rate": 2.719229256783327e-05,
|
| 703 |
+
"loss": 1.0883,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 4.6,
|
| 708 |
+
"learning_rate": 2.699567440031459e-05,
|
| 709 |
+
"loss": 1.0859,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 4.64,
|
| 714 |
+
"learning_rate": 2.679905623279591e-05,
|
| 715 |
+
"loss": 1.0898,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.68,
|
| 720 |
+
"learning_rate": 2.6602438065277236e-05,
|
| 721 |
+
"loss": 1.0809,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.72,
|
| 726 |
+
"learning_rate": 2.6405819897758555e-05,
|
| 727 |
+
"loss": 1.084,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.76,
|
| 732 |
+
"learning_rate": 2.6209201730239873e-05,
|
| 733 |
+
"loss": 1.0767,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.8,
|
| 738 |
+
"learning_rate": 2.6012583562721198e-05,
|
| 739 |
+
"loss": 1.0709,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.84,
|
| 744 |
+
"learning_rate": 2.581596539520252e-05,
|
| 745 |
+
"loss": 1.0789,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.88,
|
| 750 |
+
"learning_rate": 2.5619347227683838e-05,
|
| 751 |
+
"loss": 1.0779,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.92,
|
| 756 |
+
"learning_rate": 2.5422729060165163e-05,
|
| 757 |
+
"loss": 1.0732,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.95,
|
| 762 |
+
"learning_rate": 2.522611089264648e-05,
|
| 763 |
+
"loss": 1.0718,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.99,
|
| 768 |
+
"learning_rate": 2.5029492725127802e-05,
|
| 769 |
+
"loss": 1.0665,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 5.03,
|
| 774 |
+
"learning_rate": 2.4832874557609124e-05,
|
| 775 |
+
"loss": 1.0673,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 5.07,
|
| 780 |
+
"learning_rate": 2.4636256390090446e-05,
|
| 781 |
+
"loss": 1.0603,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 5.11,
|
| 786 |
+
"learning_rate": 2.4439638222571767e-05,
|
| 787 |
+
"loss": 1.0557,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 5.15,
|
| 792 |
+
"learning_rate": 2.4243020055053085e-05,
|
| 793 |
+
"loss": 1.0563,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 5.19,
|
| 798 |
+
"learning_rate": 2.404640188753441e-05,
|
| 799 |
+
"loss": 1.0557,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 5.23,
|
| 804 |
+
"learning_rate": 2.3849783720015732e-05,
|
| 805 |
+
"loss": 1.0586,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 5.27,
|
| 810 |
+
"learning_rate": 2.3653165552497054e-05,
|
| 811 |
+
"loss": 1.0574,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 5.31,
|
| 816 |
+
"learning_rate": 2.3456547384978372e-05,
|
| 817 |
+
"loss": 1.0578,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 5.35,
|
| 822 |
+
"learning_rate": 2.3259929217459694e-05,
|
| 823 |
+
"loss": 1.0569,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 5.39,
|
| 828 |
+
"learning_rate": 2.3063311049941015e-05,
|
| 829 |
+
"loss": 1.0541,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 5.43,
|
| 834 |
+
"learning_rate": 2.2866692882422337e-05,
|
| 835 |
+
"loss": 1.0532,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 5.47,
|
| 840 |
+
"learning_rate": 2.267007471490366e-05,
|
| 841 |
+
"loss": 1.0538,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 5.51,
|
| 846 |
+
"learning_rate": 2.247345654738498e-05,
|
| 847 |
+
"loss": 1.0467,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 5.54,
|
| 852 |
+
"learning_rate": 2.2276838379866298e-05,
|
| 853 |
+
"loss": 1.0483,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 5.58,
|
| 858 |
+
"learning_rate": 2.2080220212347623e-05,
|
| 859 |
+
"loss": 1.0565,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 5.62,
|
| 864 |
+
"learning_rate": 2.1883602044828945e-05,
|
| 865 |
+
"loss": 1.0494,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 5.66,
|
| 870 |
+
"learning_rate": 2.1686983877310266e-05,
|
| 871 |
+
"loss": 1.0423,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 5.7,
|
| 876 |
+
"learning_rate": 2.1490365709791585e-05,
|
| 877 |
+
"loss": 1.0449,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 5.74,
|
| 882 |
+
"learning_rate": 2.1293747542272906e-05,
|
| 883 |
+
"loss": 1.0447,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 5.78,
|
| 888 |
+
"learning_rate": 2.1097129374754228e-05,
|
| 889 |
+
"loss": 1.0411,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 5.82,
|
| 894 |
+
"learning_rate": 2.090051120723555e-05,
|
| 895 |
+
"loss": 1.0424,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 5.86,
|
| 900 |
+
"learning_rate": 2.070389303971687e-05,
|
| 901 |
+
"loss": 1.0394,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 5.9,
|
| 906 |
+
"learning_rate": 2.0507274872198193e-05,
|
| 907 |
+
"loss": 1.035,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 5.94,
|
| 912 |
+
"learning_rate": 2.031065670467951e-05,
|
| 913 |
+
"loss": 1.0343,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 5.98,
|
| 918 |
+
"learning_rate": 2.0114038537160836e-05,
|
| 919 |
+
"loss": 1.0371,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 6.02,
|
| 924 |
+
"learning_rate": 1.9917420369642157e-05,
|
| 925 |
+
"loss": 1.0269,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 6.06,
|
| 930 |
+
"learning_rate": 1.972080220212348e-05,
|
| 931 |
+
"loss": 1.0361,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 6.1,
|
| 936 |
+
"learning_rate": 1.9524184034604797e-05,
|
| 937 |
+
"loss": 1.0291,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 6.13,
|
| 942 |
+
"learning_rate": 1.932756586708612e-05,
|
| 943 |
+
"loss": 1.0275,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 6.17,
|
| 948 |
+
"learning_rate": 1.913094769956744e-05,
|
| 949 |
+
"loss": 1.0308,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 6.21,
|
| 954 |
+
"learning_rate": 1.8934329532048762e-05,
|
| 955 |
+
"loss": 1.0259,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 6.25,
|
| 960 |
+
"learning_rate": 1.8737711364530084e-05,
|
| 961 |
+
"loss": 1.0238,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 6.29,
|
| 966 |
+
"learning_rate": 1.8541093197011405e-05,
|
| 967 |
+
"loss": 1.0263,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 6.33,
|
| 972 |
+
"learning_rate": 1.8344475029492724e-05,
|
| 973 |
+
"loss": 1.024,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 6.37,
|
| 978 |
+
"learning_rate": 1.814785686197405e-05,
|
| 979 |
+
"loss": 1.0217,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 6.41,
|
| 984 |
+
"learning_rate": 1.795123869445537e-05,
|
| 985 |
+
"loss": 1.0229,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 6.45,
|
| 990 |
+
"learning_rate": 1.7754620526936692e-05,
|
| 991 |
+
"loss": 1.0211,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 6.49,
|
| 996 |
+
"learning_rate": 1.755800235941801e-05,
|
| 997 |
+
"loss": 1.0143,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 6.53,
|
| 1002 |
+
"learning_rate": 1.736138419189933e-05,
|
| 1003 |
+
"loss": 1.0234,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 6.57,
|
| 1008 |
+
"learning_rate": 1.7164766024380653e-05,
|
| 1009 |
+
"loss": 1.0182,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 6.61,
|
| 1014 |
+
"learning_rate": 1.6968147856861975e-05,
|
| 1015 |
+
"loss": 1.0163,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 6.65,
|
| 1020 |
+
"learning_rate": 1.6771529689343296e-05,
|
| 1021 |
+
"loss": 1.0143,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 6.69,
|
| 1026 |
+
"learning_rate": 1.6574911521824618e-05,
|
| 1027 |
+
"loss": 1.0184,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 6.72,
|
| 1032 |
+
"learning_rate": 1.6378293354305936e-05,
|
| 1033 |
+
"loss": 1.0183,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 6.76,
|
| 1038 |
+
"learning_rate": 1.618167518678726e-05,
|
| 1039 |
+
"loss": 1.0174,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 6.8,
|
| 1044 |
+
"learning_rate": 1.5985057019268583e-05,
|
| 1045 |
+
"loss": 1.0151,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 6.84,
|
| 1050 |
+
"learning_rate": 1.5788438851749904e-05,
|
| 1051 |
+
"loss": 1.0116,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 6.88,
|
| 1056 |
+
"learning_rate": 1.5591820684231223e-05,
|
| 1057 |
+
"loss": 1.0114,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 6.92,
|
| 1062 |
+
"learning_rate": 1.5395202516712544e-05,
|
| 1063 |
+
"loss": 1.0144,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 6.96,
|
| 1068 |
+
"learning_rate": 1.5198584349193868e-05,
|
| 1069 |
+
"loss": 1.0106,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 7.0,
|
| 1074 |
+
"learning_rate": 1.5001966181675187e-05,
|
| 1075 |
+
"loss": 1.0196,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 7.04,
|
| 1080 |
+
"learning_rate": 1.4805348014156509e-05,
|
| 1081 |
+
"loss": 1.0041,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 7.08,
|
| 1086 |
+
"learning_rate": 1.460872984663783e-05,
|
| 1087 |
+
"loss": 1.0061,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 7.12,
|
| 1092 |
+
"learning_rate": 1.441211167911915e-05,
|
| 1093 |
+
"loss": 1.0013,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 7.16,
|
| 1098 |
+
"learning_rate": 1.4215493511600472e-05,
|
| 1099 |
+
"loss": 1.0085,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 7.2,
|
| 1104 |
+
"learning_rate": 1.4018875344081794e-05,
|
| 1105 |
+
"loss": 1.0028,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 7.24,
|
| 1110 |
+
"learning_rate": 1.3822257176563117e-05,
|
| 1111 |
+
"loss": 1.0028,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 7.27,
|
| 1116 |
+
"learning_rate": 1.3625639009044435e-05,
|
| 1117 |
+
"loss": 0.9988,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 7.31,
|
| 1122 |
+
"learning_rate": 1.3429020841525759e-05,
|
| 1123 |
+
"loss": 1.0067,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 7.35,
|
| 1128 |
+
"learning_rate": 1.323240267400708e-05,
|
| 1129 |
+
"loss": 1.0072,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 7.39,
|
| 1134 |
+
"learning_rate": 1.30357845064884e-05,
|
| 1135 |
+
"loss": 1.0001,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 7.43,
|
| 1140 |
+
"learning_rate": 1.2839166338969722e-05,
|
| 1141 |
+
"loss": 1.0012,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 7.47,
|
| 1146 |
+
"learning_rate": 1.2642548171451043e-05,
|
| 1147 |
+
"loss": 1.0003,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 7.51,
|
| 1152 |
+
"learning_rate": 1.2445930003932365e-05,
|
| 1153 |
+
"loss": 0.9877,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 7.55,
|
| 1158 |
+
"learning_rate": 1.2249311836413685e-05,
|
| 1159 |
+
"loss": 1.0001,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 7.59,
|
| 1164 |
+
"learning_rate": 1.2052693668895006e-05,
|
| 1165 |
+
"loss": 0.9954,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 7.63,
|
| 1170 |
+
"learning_rate": 1.1856075501376328e-05,
|
| 1171 |
+
"loss": 0.9963,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 7.67,
|
| 1176 |
+
"learning_rate": 1.1659457333857648e-05,
|
| 1177 |
+
"loss": 0.994,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 7.71,
|
| 1182 |
+
"learning_rate": 1.1462839166338971e-05,
|
| 1183 |
+
"loss": 1.003,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 7.75,
|
| 1188 |
+
"learning_rate": 1.1266220998820291e-05,
|
| 1189 |
+
"loss": 0.9931,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 7.79,
|
| 1194 |
+
"learning_rate": 1.1069602831301613e-05,
|
| 1195 |
+
"loss": 0.9983,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 7.83,
|
| 1200 |
+
"learning_rate": 1.0872984663782934e-05,
|
| 1201 |
+
"loss": 0.9922,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 7.86,
|
| 1206 |
+
"learning_rate": 1.0676366496264254e-05,
|
| 1207 |
+
"loss": 0.9968,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 7.9,
|
| 1212 |
+
"learning_rate": 1.0479748328745578e-05,
|
| 1213 |
+
"loss": 0.9956,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 7.94,
|
| 1218 |
+
"learning_rate": 1.0283130161226898e-05,
|
| 1219 |
+
"loss": 0.991,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 7.98,
|
| 1224 |
+
"learning_rate": 1.008651199370822e-05,
|
| 1225 |
+
"loss": 0.9915,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 8.02,
|
| 1230 |
+
"learning_rate": 9.88989382618954e-06,
|
| 1231 |
+
"loss": 0.9859,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 8.06,
|
| 1236 |
+
"learning_rate": 9.69327565867086e-06,
|
| 1237 |
+
"loss": 0.986,
|
| 1238 |
+
"step": 102500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 8.1,
|
| 1242 |
+
"learning_rate": 9.496657491152184e-06,
|
| 1243 |
+
"loss": 0.9937,
|
| 1244 |
+
"step": 103000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 8.14,
|
| 1248 |
+
"learning_rate": 9.300039323633504e-06,
|
| 1249 |
+
"loss": 0.9863,
|
| 1250 |
+
"step": 103500
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 8.18,
|
| 1254 |
+
"learning_rate": 9.103421156114826e-06,
|
| 1255 |
+
"loss": 0.9836,
|
| 1256 |
+
"step": 104000
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 8.22,
|
| 1260 |
+
"learning_rate": 8.906802988596147e-06,
|
| 1261 |
+
"loss": 0.9857,
|
| 1262 |
+
"step": 104500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 8.26,
|
| 1266 |
+
"learning_rate": 8.710184821077467e-06,
|
| 1267 |
+
"loss": 0.989,
|
| 1268 |
+
"step": 105000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 8.3,
|
| 1272 |
+
"learning_rate": 8.51356665355879e-06,
|
| 1273 |
+
"loss": 0.9886,
|
| 1274 |
+
"step": 105500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 8.34,
|
| 1278 |
+
"learning_rate": 8.31694848604011e-06,
|
| 1279 |
+
"loss": 0.9885,
|
| 1280 |
+
"step": 106000
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 8.38,
|
| 1284 |
+
"learning_rate": 8.120330318521432e-06,
|
| 1285 |
+
"loss": 0.9847,
|
| 1286 |
+
"step": 106500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 8.42,
|
| 1290 |
+
"learning_rate": 7.923712151002753e-06,
|
| 1291 |
+
"loss": 0.9846,
|
| 1292 |
+
"step": 107000
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 8.45,
|
| 1296 |
+
"learning_rate": 7.727093983484073e-06,
|
| 1297 |
+
"loss": 0.9882,
|
| 1298 |
+
"step": 107500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 8.49,
|
| 1302 |
+
"learning_rate": 7.530475815965396e-06,
|
| 1303 |
+
"loss": 0.9866,
|
| 1304 |
+
"step": 108000
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 8.53,
|
| 1308 |
+
"learning_rate": 7.333857648446717e-06,
|
| 1309 |
+
"loss": 0.9823,
|
| 1310 |
+
"step": 108500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 8.57,
|
| 1314 |
+
"learning_rate": 7.137239480928037e-06,
|
| 1315 |
+
"loss": 0.9841,
|
| 1316 |
+
"step": 109000
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 8.61,
|
| 1320 |
+
"learning_rate": 6.94062131340936e-06,
|
| 1321 |
+
"loss": 0.9844,
|
| 1322 |
+
"step": 109500
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 8.65,
|
| 1326 |
+
"learning_rate": 6.7440031458906806e-06,
|
| 1327 |
+
"loss": 0.9817,
|
| 1328 |
+
"step": 110000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 8.69,
|
| 1332 |
+
"learning_rate": 6.547384978372002e-06,
|
| 1333 |
+
"loss": 0.9854,
|
| 1334 |
+
"step": 110500
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 8.73,
|
| 1338 |
+
"learning_rate": 6.350766810853323e-06,
|
| 1339 |
+
"loss": 0.9845,
|
| 1340 |
+
"step": 111000
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 8.77,
|
| 1344 |
+
"learning_rate": 6.1541486433346445e-06,
|
| 1345 |
+
"loss": 0.984,
|
| 1346 |
+
"step": 111500
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 8.81,
|
| 1350 |
+
"learning_rate": 5.957530475815966e-06,
|
| 1351 |
+
"loss": 0.9865,
|
| 1352 |
+
"step": 112000
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 8.85,
|
| 1356 |
+
"learning_rate": 5.760912308297287e-06,
|
| 1357 |
+
"loss": 0.9765,
|
| 1358 |
+
"step": 112500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 8.89,
|
| 1362 |
+
"learning_rate": 5.564294140778608e-06,
|
| 1363 |
+
"loss": 0.9867,
|
| 1364 |
+
"step": 113000
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 8.93,
|
| 1368 |
+
"learning_rate": 5.367675973259929e-06,
|
| 1369 |
+
"loss": 0.986,
|
| 1370 |
+
"step": 113500
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 8.97,
|
| 1374 |
+
"learning_rate": 5.171057805741251e-06,
|
| 1375 |
+
"loss": 0.9817,
|
| 1376 |
+
"step": 114000
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 9.01,
|
| 1380 |
+
"learning_rate": 4.9744396382225725e-06,
|
| 1381 |
+
"loss": 0.9774,
|
| 1382 |
+
"step": 114500
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 9.04,
|
| 1386 |
+
"learning_rate": 4.777821470703893e-06,
|
| 1387 |
+
"loss": 0.9811,
|
| 1388 |
+
"step": 115000
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 9.08,
|
| 1392 |
+
"learning_rate": 4.581203303185214e-06,
|
| 1393 |
+
"loss": 0.9759,
|
| 1394 |
+
"step": 115500
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 9.12,
|
| 1398 |
+
"learning_rate": 4.384585135666536e-06,
|
| 1399 |
+
"loss": 0.9769,
|
| 1400 |
+
"step": 116000
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 9.16,
|
| 1404 |
+
"learning_rate": 4.187966968147857e-06,
|
| 1405 |
+
"loss": 0.9807,
|
| 1406 |
+
"step": 116500
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 9.2,
|
| 1410 |
+
"learning_rate": 3.991348800629179e-06,
|
| 1411 |
+
"loss": 0.9822,
|
| 1412 |
+
"step": 117000
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 9.24,
|
| 1416 |
+
"learning_rate": 3.794730633110499e-06,
|
| 1417 |
+
"loss": 0.9843,
|
| 1418 |
+
"step": 117500
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 9.28,
|
| 1422 |
+
"learning_rate": 3.5981124655918208e-06,
|
| 1423 |
+
"loss": 0.9777,
|
| 1424 |
+
"step": 118000
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 9.32,
|
| 1428 |
+
"learning_rate": 3.401494298073142e-06,
|
| 1429 |
+
"loss": 0.977,
|
| 1430 |
+
"step": 118500
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 9.36,
|
| 1434 |
+
"learning_rate": 3.2048761305544636e-06,
|
| 1435 |
+
"loss": 0.9785,
|
| 1436 |
+
"step": 119000
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 9.4,
|
| 1440 |
+
"learning_rate": 3.0082579630357847e-06,
|
| 1441 |
+
"loss": 0.9756,
|
| 1442 |
+
"step": 119500
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 9.44,
|
| 1446 |
+
"learning_rate": 2.811639795517106e-06,
|
| 1447 |
+
"loss": 0.9803,
|
| 1448 |
+
"step": 120000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 9.48,
|
| 1452 |
+
"learning_rate": 2.615021627998427e-06,
|
| 1453 |
+
"loss": 0.9774,
|
| 1454 |
+
"step": 120500
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 9.52,
|
| 1458 |
+
"learning_rate": 2.4184034604797483e-06,
|
| 1459 |
+
"loss": 0.984,
|
| 1460 |
+
"step": 121000
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 9.56,
|
| 1464 |
+
"learning_rate": 2.22178529296107e-06,
|
| 1465 |
+
"loss": 0.9753,
|
| 1466 |
+
"step": 121500
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 9.59,
|
| 1470 |
+
"learning_rate": 2.025167125442391e-06,
|
| 1471 |
+
"loss": 0.9764,
|
| 1472 |
+
"step": 122000
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 9.63,
|
| 1476 |
+
"learning_rate": 1.8285489579237123e-06,
|
| 1477 |
+
"loss": 0.9756,
|
| 1478 |
+
"step": 122500
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 9.67,
|
| 1482 |
+
"learning_rate": 1.6319307904050335e-06,
|
| 1483 |
+
"loss": 0.9733,
|
| 1484 |
+
"step": 123000
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 9.71,
|
| 1488 |
+
"learning_rate": 1.4353126228863549e-06,
|
| 1489 |
+
"loss": 0.9726,
|
| 1490 |
+
"step": 123500
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 9.75,
|
| 1494 |
+
"learning_rate": 1.238694455367676e-06,
|
| 1495 |
+
"loss": 0.9783,
|
| 1496 |
+
"step": 124000
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 9.79,
|
| 1500 |
+
"learning_rate": 1.0420762878489972e-06,
|
| 1501 |
+
"loss": 0.9786,
|
| 1502 |
+
"step": 124500
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 9.83,
|
| 1506 |
+
"learning_rate": 8.454581203303186e-07,
|
| 1507 |
+
"loss": 0.9802,
|
| 1508 |
+
"step": 125000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 9.87,
|
| 1512 |
+
"learning_rate": 6.488399528116398e-07,
|
| 1513 |
+
"loss": 0.979,
|
| 1514 |
+
"step": 125500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 9.91,
|
| 1518 |
+
"learning_rate": 4.522217852929611e-07,
|
| 1519 |
+
"loss": 0.978,
|
| 1520 |
+
"step": 126000
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 9.95,
|
| 1524 |
+
"learning_rate": 2.556036177742823e-07,
|
| 1525 |
+
"loss": 0.9718,
|
| 1526 |
+
"step": 126500
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 9.99,
|
| 1530 |
+
"learning_rate": 5.898545025560362e-08,
|
| 1531 |
+
"loss": 0.9811,
|
| 1532 |
+
"step": 127000
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 10.0,
|
| 1536 |
+
"step": 127150,
|
| 1537 |
+
"total_flos": 5.807966088205763e+17,
|
| 1538 |
+
"train_loss": 1.219492632603017,
|
| 1539 |
+
"train_runtime": 160361.9384,
|
| 1540 |
+
"train_samples_per_second": 38.059,
|
| 1541 |
+
"train_steps_per_second": 0.793
|
| 1542 |
+
}
|
| 1543 |
+
],
|
| 1544 |
+
"max_steps": 127150,
|
| 1545 |
+
"num_train_epochs": 10,
|
| 1546 |
+
"total_flos": 5.807966088205763e+17,
|
| 1547 |
+
"trial_name": null,
|
| 1548 |
+
"trial_params": null
|
| 1549 |
+
}
|