End of training
Browse files- README.md +1 -0
- all_results.json +13 -0
- eval_results.json +8 -0
- train_results.json +8 -0
- trainer_state.json +1480 -0
README.md
CHANGED
|
@@ -2,6 +2,7 @@
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/wav2vec2-base
|
| 4 |
tags:
|
|
|
|
| 5 |
- generated_from_trainer
|
| 6 |
datasets:
|
| 7 |
- superb
|
|
|
|
| 2 |
license: apache-2.0
|
| 3 |
base_model: facebook/wav2vec2-base
|
| 4 |
tags:
|
| 5 |
+
- audio-classification
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- superb
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.996869129618034,
|
| 3 |
+
"eval_accuracy": 0.9832303618711385,
|
| 4 |
+
"eval_loss": 0.07819115370512009,
|
| 5 |
+
"eval_runtime": 10.6152,
|
| 6 |
+
"eval_samples_per_second": 640.403,
|
| 7 |
+
"eval_steps_per_second": 20.066,
|
| 8 |
+
"total_flos": 2.3180194781952e+18,
|
| 9 |
+
"train_loss": 0.40026830736557045,
|
| 10 |
+
"train_runtime": 1087.6247,
|
| 11 |
+
"train_samples_per_second": 234.888,
|
| 12 |
+
"train_steps_per_second": 1.834
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.996869129618034,
|
| 3 |
+
"eval_accuracy": 0.9832303618711385,
|
| 4 |
+
"eval_loss": 0.07819115370512009,
|
| 5 |
+
"eval_runtime": 10.6152,
|
| 6 |
+
"eval_samples_per_second": 640.403,
|
| 7 |
+
"eval_steps_per_second": 20.066
|
| 8 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.996869129618034,
|
| 3 |
+
"total_flos": 2.3180194781952e+18,
|
| 4 |
+
"train_loss": 0.40026830736557045,
|
| 5 |
+
"train_runtime": 1087.6247,
|
| 6 |
+
"train_samples_per_second": 234.888,
|
| 7 |
+
"train_steps_per_second": 1.834
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1480 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9832303618711385,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-base-ft-keyword-spotting/checkpoint-1995",
|
| 4 |
+
"epoch": 4.996869129618034,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1995,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.025046963055729492,
|
| 13 |
+
"grad_norm": 1.278334379196167,
|
| 14 |
+
"learning_rate": 1.5e-06,
|
| 15 |
+
"loss": 2.4699,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.050093926111458985,
|
| 20 |
+
"grad_norm": 1.3655153512954712,
|
| 21 |
+
"learning_rate": 3e-06,
|
| 22 |
+
"loss": 2.4539,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.07514088916718847,
|
| 27 |
+
"grad_norm": 1.3241969347000122,
|
| 28 |
+
"learning_rate": 4.5e-06,
|
| 29 |
+
"loss": 2.4044,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.10018785222291797,
|
| 34 |
+
"grad_norm": 1.7512290477752686,
|
| 35 |
+
"learning_rate": 6e-06,
|
| 36 |
+
"loss": 2.3261,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.12523481527864747,
|
| 41 |
+
"grad_norm": 2.1144192218780518,
|
| 42 |
+
"learning_rate": 7.5e-06,
|
| 43 |
+
"loss": 2.2156,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.15028177833437695,
|
| 48 |
+
"grad_norm": 2.416064739227295,
|
| 49 |
+
"learning_rate": 9e-06,
|
| 50 |
+
"loss": 2.0569,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.17532874139010646,
|
| 55 |
+
"grad_norm": 2.3332056999206543,
|
| 56 |
+
"learning_rate": 1.05e-05,
|
| 57 |
+
"loss": 1.8437,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.20037570444583594,
|
| 62 |
+
"grad_norm": 2.037410259246826,
|
| 63 |
+
"learning_rate": 1.2e-05,
|
| 64 |
+
"loss": 1.7693,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.22542266750156542,
|
| 69 |
+
"grad_norm": 2.00655198097229,
|
| 70 |
+
"learning_rate": 1.3500000000000001e-05,
|
| 71 |
+
"loss": 1.6565,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.25046963055729493,
|
| 76 |
+
"grad_norm": 1.5333060026168823,
|
| 77 |
+
"learning_rate": 1.5e-05,
|
| 78 |
+
"loss": 1.6115,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.27551659361302444,
|
| 83 |
+
"grad_norm": 0.9409968852996826,
|
| 84 |
+
"learning_rate": 1.65e-05,
|
| 85 |
+
"loss": 1.6253,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.3005635566687539,
|
| 90 |
+
"grad_norm": 1.4992973804473877,
|
| 91 |
+
"learning_rate": 1.8e-05,
|
| 92 |
+
"loss": 1.5118,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.3256105197244834,
|
| 97 |
+
"grad_norm": 0.5201308727264404,
|
| 98 |
+
"learning_rate": 1.95e-05,
|
| 99 |
+
"loss": 1.503,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.3506574827802129,
|
| 104 |
+
"grad_norm": 1.1067086458206177,
|
| 105 |
+
"learning_rate": 2.1e-05,
|
| 106 |
+
"loss": 1.5415,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.37570444583594237,
|
| 111 |
+
"grad_norm": 1.8773168325424194,
|
| 112 |
+
"learning_rate": 2.25e-05,
|
| 113 |
+
"loss": 1.4933,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.4007514088916719,
|
| 118 |
+
"grad_norm": 2.180652618408203,
|
| 119 |
+
"learning_rate": 2.4e-05,
|
| 120 |
+
"loss": 1.449,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.4257983719474014,
|
| 125 |
+
"grad_norm": 1.6153920888900757,
|
| 126 |
+
"learning_rate": 2.55e-05,
|
| 127 |
+
"loss": 1.4281,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.45084533500313084,
|
| 132 |
+
"grad_norm": 3.9366486072540283,
|
| 133 |
+
"learning_rate": 2.7000000000000002e-05,
|
| 134 |
+
"loss": 1.3706,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.47589229805886035,
|
| 139 |
+
"grad_norm": 8.680933952331543,
|
| 140 |
+
"learning_rate": 2.8499999999999998e-05,
|
| 141 |
+
"loss": 1.3632,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5009392611145899,
|
| 146 |
+
"grad_norm": 1.6787041425704956,
|
| 147 |
+
"learning_rate": 3e-05,
|
| 148 |
+
"loss": 1.3052,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.5259862241703194,
|
| 153 |
+
"grad_norm": 3.0554661750793457,
|
| 154 |
+
"learning_rate": 2.9832869080779945e-05,
|
| 155 |
+
"loss": 1.2246,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.5510331872260489,
|
| 160 |
+
"grad_norm": 2.5610270500183105,
|
| 161 |
+
"learning_rate": 2.9665738161559886e-05,
|
| 162 |
+
"loss": 1.1641,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.5760801502817783,
|
| 167 |
+
"grad_norm": 4.368018627166748,
|
| 168 |
+
"learning_rate": 2.9498607242339834e-05,
|
| 169 |
+
"loss": 1.1112,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.6011271133375078,
|
| 174 |
+
"grad_norm": 3.1300201416015625,
|
| 175 |
+
"learning_rate": 2.933147632311978e-05,
|
| 176 |
+
"loss": 1.1034,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.6261740763932373,
|
| 181 |
+
"grad_norm": 3.441697597503662,
|
| 182 |
+
"learning_rate": 2.916434540389972e-05,
|
| 183 |
+
"loss": 0.9748,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.6512210394489668,
|
| 188 |
+
"grad_norm": 2.930171251296997,
|
| 189 |
+
"learning_rate": 2.8997214484679665e-05,
|
| 190 |
+
"loss": 0.9908,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.6762680025046963,
|
| 195 |
+
"grad_norm": 4.085419178009033,
|
| 196 |
+
"learning_rate": 2.8830083565459613e-05,
|
| 197 |
+
"loss": 0.8302,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.7013149655604258,
|
| 202 |
+
"grad_norm": 3.5097601413726807,
|
| 203 |
+
"learning_rate": 2.8662952646239554e-05,
|
| 204 |
+
"loss": 0.788,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.7263619286161553,
|
| 209 |
+
"grad_norm": 4.0369062423706055,
|
| 210 |
+
"learning_rate": 2.84958217270195e-05,
|
| 211 |
+
"loss": 0.766,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.7514088916718847,
|
| 216 |
+
"grad_norm": 2.472437858581543,
|
| 217 |
+
"learning_rate": 2.8328690807799443e-05,
|
| 218 |
+
"loss": 0.742,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.7764558547276142,
|
| 223 |
+
"grad_norm": 3.6972556114196777,
|
| 224 |
+
"learning_rate": 2.8161559888579388e-05,
|
| 225 |
+
"loss": 0.7595,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.8015028177833438,
|
| 230 |
+
"grad_norm": 3.38409161567688,
|
| 231 |
+
"learning_rate": 2.7994428969359332e-05,
|
| 232 |
+
"loss": 0.6941,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.8265497808390733,
|
| 237 |
+
"grad_norm": 2.243342399597168,
|
| 238 |
+
"learning_rate": 2.7827298050139277e-05,
|
| 239 |
+
"loss": 0.6347,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.8515967438948028,
|
| 244 |
+
"grad_norm": 4.12843132019043,
|
| 245 |
+
"learning_rate": 2.7660167130919218e-05,
|
| 246 |
+
"loss": 0.6031,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.8766437069505323,
|
| 251 |
+
"grad_norm": 2.2556419372558594,
|
| 252 |
+
"learning_rate": 2.7493036211699166e-05,
|
| 253 |
+
"loss": 0.5767,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.9016906700062617,
|
| 258 |
+
"grad_norm": 3.9663896560668945,
|
| 259 |
+
"learning_rate": 2.732590529247911e-05,
|
| 260 |
+
"loss": 0.5565,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.9267376330619912,
|
| 265 |
+
"grad_norm": 3.594534397125244,
|
| 266 |
+
"learning_rate": 2.7158774373259055e-05,
|
| 267 |
+
"loss": 0.5529,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.9517845961177207,
|
| 272 |
+
"grad_norm": 2.7648637294769287,
|
| 273 |
+
"learning_rate": 2.6991643454038996e-05,
|
| 274 |
+
"loss": 0.4842,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.9768315591734502,
|
| 279 |
+
"grad_norm": 3.66634202003479,
|
| 280 |
+
"learning_rate": 2.6824512534818944e-05,
|
| 281 |
+
"loss": 0.5065,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.9993738259236068,
|
| 286 |
+
"eval_accuracy": 0.9713150926743159,
|
| 287 |
+
"eval_loss": 0.3426418602466583,
|
| 288 |
+
"eval_runtime": 10.5903,
|
| 289 |
+
"eval_samples_per_second": 641.906,
|
| 290 |
+
"eval_steps_per_second": 20.113,
|
| 291 |
+
"step": 399
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.0018785222291797,
|
| 295 |
+
"grad_norm": 2.2331535816192627,
|
| 296 |
+
"learning_rate": 2.665738161559889e-05,
|
| 297 |
+
"loss": 0.4929,
|
| 298 |
+
"step": 400
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.0269254852849092,
|
| 302 |
+
"grad_norm": 2.8309195041656494,
|
| 303 |
+
"learning_rate": 2.649025069637883e-05,
|
| 304 |
+
"loss": 0.413,
|
| 305 |
+
"step": 410
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.0519724483406387,
|
| 309 |
+
"grad_norm": 3.615835428237915,
|
| 310 |
+
"learning_rate": 2.6323119777158774e-05,
|
| 311 |
+
"loss": 0.4205,
|
| 312 |
+
"step": 420
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.0770194113963683,
|
| 316 |
+
"grad_norm": 4.138455867767334,
|
| 317 |
+
"learning_rate": 2.6155988857938722e-05,
|
| 318 |
+
"loss": 0.3937,
|
| 319 |
+
"step": 430
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 1.1020663744520978,
|
| 323 |
+
"grad_norm": 3.2305445671081543,
|
| 324 |
+
"learning_rate": 2.5988857938718663e-05,
|
| 325 |
+
"loss": 0.4097,
|
| 326 |
+
"step": 440
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.127113337507827,
|
| 330 |
+
"grad_norm": 3.1868467330932617,
|
| 331 |
+
"learning_rate": 2.5821727019498608e-05,
|
| 332 |
+
"loss": 0.3661,
|
| 333 |
+
"step": 450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.1521603005635566,
|
| 337 |
+
"grad_norm": 1.8689473867416382,
|
| 338 |
+
"learning_rate": 2.5654596100278553e-05,
|
| 339 |
+
"loss": 0.3327,
|
| 340 |
+
"step": 460
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.177207263619286,
|
| 344 |
+
"grad_norm": 1.8282161951065063,
|
| 345 |
+
"learning_rate": 2.5487465181058497e-05,
|
| 346 |
+
"loss": 0.3211,
|
| 347 |
+
"step": 470
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.2022542266750156,
|
| 351 |
+
"grad_norm": 3.445858955383301,
|
| 352 |
+
"learning_rate": 2.532033426183844e-05,
|
| 353 |
+
"loss": 0.3468,
|
| 354 |
+
"step": 480
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.227301189730745,
|
| 358 |
+
"grad_norm": 3.0723907947540283,
|
| 359 |
+
"learning_rate": 2.5153203342618386e-05,
|
| 360 |
+
"loss": 0.3533,
|
| 361 |
+
"step": 490
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.2523481527864746,
|
| 365 |
+
"grad_norm": 4.13520622253418,
|
| 366 |
+
"learning_rate": 2.4986072423398327e-05,
|
| 367 |
+
"loss": 0.2817,
|
| 368 |
+
"step": 500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.277395115842204,
|
| 372 |
+
"grad_norm": 6.625748634338379,
|
| 373 |
+
"learning_rate": 2.4818941504178275e-05,
|
| 374 |
+
"loss": 0.3295,
|
| 375 |
+
"step": 510
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.3024420788979336,
|
| 379 |
+
"grad_norm": 1.7438695430755615,
|
| 380 |
+
"learning_rate": 2.465181058495822e-05,
|
| 381 |
+
"loss": 0.2915,
|
| 382 |
+
"step": 520
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.3274890419536631,
|
| 386 |
+
"grad_norm": 4.603183746337891,
|
| 387 |
+
"learning_rate": 2.448467966573816e-05,
|
| 388 |
+
"loss": 0.3447,
|
| 389 |
+
"step": 530
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.3525360050093926,
|
| 393 |
+
"grad_norm": 2.3780734539031982,
|
| 394 |
+
"learning_rate": 2.4317548746518106e-05,
|
| 395 |
+
"loss": 0.2678,
|
| 396 |
+
"step": 540
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.3775829680651221,
|
| 400 |
+
"grad_norm": 2.449040174484253,
|
| 401 |
+
"learning_rate": 2.415041782729805e-05,
|
| 402 |
+
"loss": 0.3029,
|
| 403 |
+
"step": 550
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 1.4026299311208517,
|
| 407 |
+
"grad_norm": 3.899897336959839,
|
| 408 |
+
"learning_rate": 2.3983286908077995e-05,
|
| 409 |
+
"loss": 0.2609,
|
| 410 |
+
"step": 560
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.4276768941765812,
|
| 414 |
+
"grad_norm": 4.0847978591918945,
|
| 415 |
+
"learning_rate": 2.381615598885794e-05,
|
| 416 |
+
"loss": 0.2979,
|
| 417 |
+
"step": 570
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.4527238572323107,
|
| 421 |
+
"grad_norm": 3.0735533237457275,
|
| 422 |
+
"learning_rate": 2.3649025069637884e-05,
|
| 423 |
+
"loss": 0.2612,
|
| 424 |
+
"step": 580
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.4777708202880402,
|
| 428 |
+
"grad_norm": 4.23598051071167,
|
| 429 |
+
"learning_rate": 2.3481894150417825e-05,
|
| 430 |
+
"loss": 0.2657,
|
| 431 |
+
"step": 590
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.5028177833437697,
|
| 435 |
+
"grad_norm": 2.6922895908355713,
|
| 436 |
+
"learning_rate": 2.3314763231197773e-05,
|
| 437 |
+
"loss": 0.2332,
|
| 438 |
+
"step": 600
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 1.5278647463994992,
|
| 442 |
+
"grad_norm": 2.444784641265869,
|
| 443 |
+
"learning_rate": 2.3147632311977718e-05,
|
| 444 |
+
"loss": 0.2646,
|
| 445 |
+
"step": 610
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 1.5529117094552287,
|
| 449 |
+
"grad_norm": 3.595010757446289,
|
| 450 |
+
"learning_rate": 2.298050139275766e-05,
|
| 451 |
+
"loss": 0.2258,
|
| 452 |
+
"step": 620
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.577958672510958,
|
| 456 |
+
"grad_norm": 4.990175724029541,
|
| 457 |
+
"learning_rate": 2.2813370473537603e-05,
|
| 458 |
+
"loss": 0.2477,
|
| 459 |
+
"step": 630
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.6030056355666875,
|
| 463 |
+
"grad_norm": 2.4451816082000732,
|
| 464 |
+
"learning_rate": 2.264623955431755e-05,
|
| 465 |
+
"loss": 0.2268,
|
| 466 |
+
"step": 640
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 1.628052598622417,
|
| 470 |
+
"grad_norm": 5.3694233894348145,
|
| 471 |
+
"learning_rate": 2.2479108635097492e-05,
|
| 472 |
+
"loss": 0.226,
|
| 473 |
+
"step": 650
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.6530995616781465,
|
| 477 |
+
"grad_norm": 4.013739585876465,
|
| 478 |
+
"learning_rate": 2.2311977715877437e-05,
|
| 479 |
+
"loss": 0.2519,
|
| 480 |
+
"step": 660
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.678146524733876,
|
| 484 |
+
"grad_norm": 2.913640022277832,
|
| 485 |
+
"learning_rate": 2.214484679665738e-05,
|
| 486 |
+
"loss": 0.2723,
|
| 487 |
+
"step": 670
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 1.7031934877896056,
|
| 491 |
+
"grad_norm": 3.3302664756774902,
|
| 492 |
+
"learning_rate": 2.1977715877437326e-05,
|
| 493 |
+
"loss": 0.2412,
|
| 494 |
+
"step": 680
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.7282404508453348,
|
| 498 |
+
"grad_norm": 3.4415056705474854,
|
| 499 |
+
"learning_rate": 2.181058495821727e-05,
|
| 500 |
+
"loss": 0.2463,
|
| 501 |
+
"step": 690
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.7532874139010644,
|
| 505 |
+
"grad_norm": 2.424220085144043,
|
| 506 |
+
"learning_rate": 2.1643454038997215e-05,
|
| 507 |
+
"loss": 0.198,
|
| 508 |
+
"step": 700
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 1.7783343769567939,
|
| 512 |
+
"grad_norm": 3.404729127883911,
|
| 513 |
+
"learning_rate": 2.147632311977716e-05,
|
| 514 |
+
"loss": 0.2263,
|
| 515 |
+
"step": 710
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.8033813400125234,
|
| 519 |
+
"grad_norm": 2.495832681655884,
|
| 520 |
+
"learning_rate": 2.1309192200557104e-05,
|
| 521 |
+
"loss": 0.2146,
|
| 522 |
+
"step": 720
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.8284283030682529,
|
| 526 |
+
"grad_norm": 3.8898746967315674,
|
| 527 |
+
"learning_rate": 2.114206128133705e-05,
|
| 528 |
+
"loss": 0.2289,
|
| 529 |
+
"step": 730
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 1.8534752661239824,
|
| 533 |
+
"grad_norm": 2.768136978149414,
|
| 534 |
+
"learning_rate": 2.0974930362116993e-05,
|
| 535 |
+
"loss": 0.2144,
|
| 536 |
+
"step": 740
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.878522229179712,
|
| 540 |
+
"grad_norm": 2.6077802181243896,
|
| 541 |
+
"learning_rate": 2.0807799442896935e-05,
|
| 542 |
+
"loss": 0.2269,
|
| 543 |
+
"step": 750
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.9035691922354414,
|
| 547 |
+
"grad_norm": 2.223156452178955,
|
| 548 |
+
"learning_rate": 2.0640668523676883e-05,
|
| 549 |
+
"loss": 0.1809,
|
| 550 |
+
"step": 760
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.928616155291171,
|
| 554 |
+
"grad_norm": 2.7288057804107666,
|
| 555 |
+
"learning_rate": 2.0473537604456827e-05,
|
| 556 |
+
"loss": 0.1833,
|
| 557 |
+
"step": 770
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.9536631183469004,
|
| 561 |
+
"grad_norm": 2.93996000289917,
|
| 562 |
+
"learning_rate": 2.0306406685236768e-05,
|
| 563 |
+
"loss": 0.2276,
|
| 564 |
+
"step": 780
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.97871008140263,
|
| 568 |
+
"grad_norm": 1.5655947923660278,
|
| 569 |
+
"learning_rate": 2.0139275766016713e-05,
|
| 570 |
+
"loss": 0.2089,
|
| 571 |
+
"step": 790
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.9987476518472134,
|
| 575 |
+
"eval_accuracy": 0.9780817887614004,
|
| 576 |
+
"eval_loss": 0.12669295072555542,
|
| 577 |
+
"eval_runtime": 10.6875,
|
| 578 |
+
"eval_samples_per_second": 636.068,
|
| 579 |
+
"eval_steps_per_second": 19.93,
|
| 580 |
+
"step": 798
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 2.0037570444583594,
|
| 584 |
+
"grad_norm": 2.9249160289764404,
|
| 585 |
+
"learning_rate": 1.997214484679666e-05,
|
| 586 |
+
"loss": 0.201,
|
| 587 |
+
"step": 800
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 2.028804007514089,
|
| 591 |
+
"grad_norm": 2.023646116256714,
|
| 592 |
+
"learning_rate": 1.9805013927576602e-05,
|
| 593 |
+
"loss": 0.1868,
|
| 594 |
+
"step": 810
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.0538509705698185,
|
| 598 |
+
"grad_norm": 1.4999943971633911,
|
| 599 |
+
"learning_rate": 1.9637883008356546e-05,
|
| 600 |
+
"loss": 0.1885,
|
| 601 |
+
"step": 820
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 2.078897933625548,
|
| 605 |
+
"grad_norm": 3.2979843616485596,
|
| 606 |
+
"learning_rate": 1.947075208913649e-05,
|
| 607 |
+
"loss": 0.1904,
|
| 608 |
+
"step": 830
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.1039448966812775,
|
| 612 |
+
"grad_norm": 1.8831233978271484,
|
| 613 |
+
"learning_rate": 1.9303621169916436e-05,
|
| 614 |
+
"loss": 0.185,
|
| 615 |
+
"step": 840
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 2.128991859737007,
|
| 619 |
+
"grad_norm": 1.3918356895446777,
|
| 620 |
+
"learning_rate": 1.913649025069638e-05,
|
| 621 |
+
"loss": 0.1829,
|
| 622 |
+
"step": 850
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 2.1540388227927365,
|
| 626 |
+
"grad_norm": 1.2814128398895264,
|
| 627 |
+
"learning_rate": 1.8969359331476325e-05,
|
| 628 |
+
"loss": 0.1667,
|
| 629 |
+
"step": 860
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 2.179085785848466,
|
| 633 |
+
"grad_norm": 3.2199411392211914,
|
| 634 |
+
"learning_rate": 1.8802228412256266e-05,
|
| 635 |
+
"loss": 0.179,
|
| 636 |
+
"step": 870
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.2041327489041955,
|
| 640 |
+
"grad_norm": 3.293250560760498,
|
| 641 |
+
"learning_rate": 1.863509749303621e-05,
|
| 642 |
+
"loss": 0.213,
|
| 643 |
+
"step": 880
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 2.229179711959925,
|
| 647 |
+
"grad_norm": 2.887782335281372,
|
| 648 |
+
"learning_rate": 1.846796657381616e-05,
|
| 649 |
+
"loss": 0.1872,
|
| 650 |
+
"step": 890
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.254226675015654,
|
| 654 |
+
"grad_norm": 3.306884288787842,
|
| 655 |
+
"learning_rate": 1.83008356545961e-05,
|
| 656 |
+
"loss": 0.1762,
|
| 657 |
+
"step": 900
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.279273638071384,
|
| 661 |
+
"grad_norm": 3.3350651264190674,
|
| 662 |
+
"learning_rate": 1.8133704735376044e-05,
|
| 663 |
+
"loss": 0.2153,
|
| 664 |
+
"step": 910
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 2.304320601127113,
|
| 668 |
+
"grad_norm": 1.8800674676895142,
|
| 669 |
+
"learning_rate": 1.796657381615599e-05,
|
| 670 |
+
"loss": 0.2069,
|
| 671 |
+
"step": 920
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 2.329367564182843,
|
| 675 |
+
"grad_norm": 3.2760367393493652,
|
| 676 |
+
"learning_rate": 1.7799442896935933e-05,
|
| 677 |
+
"loss": 0.1756,
|
| 678 |
+
"step": 930
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 2.354414527238572,
|
| 682 |
+
"grad_norm": 4.827606678009033,
|
| 683 |
+
"learning_rate": 1.7632311977715878e-05,
|
| 684 |
+
"loss": 0.1673,
|
| 685 |
+
"step": 940
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 2.3794614902943017,
|
| 689 |
+
"grad_norm": 4.115577697753906,
|
| 690 |
+
"learning_rate": 1.7465181058495822e-05,
|
| 691 |
+
"loss": 0.1503,
|
| 692 |
+
"step": 950
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.404508453350031,
|
| 696 |
+
"grad_norm": 2.163761854171753,
|
| 697 |
+
"learning_rate": 1.7298050139275764e-05,
|
| 698 |
+
"loss": 0.2029,
|
| 699 |
+
"step": 960
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.4295554164057607,
|
| 703 |
+
"grad_norm": 3.1453959941864014,
|
| 704 |
+
"learning_rate": 1.713091922005571e-05,
|
| 705 |
+
"loss": 0.1748,
|
| 706 |
+
"step": 970
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 2.45460237946149,
|
| 710 |
+
"grad_norm": 0.9899351596832275,
|
| 711 |
+
"learning_rate": 1.6963788300835656e-05,
|
| 712 |
+
"loss": 0.1589,
|
| 713 |
+
"step": 980
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.4796493425172197,
|
| 717 |
+
"grad_norm": 2.109621524810791,
|
| 718 |
+
"learning_rate": 1.6796657381615597e-05,
|
| 719 |
+
"loss": 0.1709,
|
| 720 |
+
"step": 990
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.504696305572949,
|
| 724 |
+
"grad_norm": 4.260733604431152,
|
| 725 |
+
"learning_rate": 1.6629526462395542e-05,
|
| 726 |
+
"loss": 0.1776,
|
| 727 |
+
"step": 1000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 2.5297432686286787,
|
| 731 |
+
"grad_norm": 1.1494958400726318,
|
| 732 |
+
"learning_rate": 1.646239554317549e-05,
|
| 733 |
+
"loss": 0.1651,
|
| 734 |
+
"step": 1010
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.554790231684408,
|
| 738 |
+
"grad_norm": 4.246566295623779,
|
| 739 |
+
"learning_rate": 1.6295264623955434e-05,
|
| 740 |
+
"loss": 0.1941,
|
| 741 |
+
"step": 1020
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.5798371947401377,
|
| 745 |
+
"grad_norm": 2.06291127204895,
|
| 746 |
+
"learning_rate": 1.6128133704735375e-05,
|
| 747 |
+
"loss": 0.1276,
|
| 748 |
+
"step": 1030
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 2.6048841577958672,
|
| 752 |
+
"grad_norm": 2.7648720741271973,
|
| 753 |
+
"learning_rate": 1.596100278551532e-05,
|
| 754 |
+
"loss": 0.1725,
|
| 755 |
+
"step": 1040
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 2.6299311208515967,
|
| 759 |
+
"grad_norm": 1.355832815170288,
|
| 760 |
+
"learning_rate": 1.5793871866295268e-05,
|
| 761 |
+
"loss": 0.1585,
|
| 762 |
+
"step": 1050
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 2.6549780839073263,
|
| 766 |
+
"grad_norm": 3.2558505535125732,
|
| 767 |
+
"learning_rate": 1.562674094707521e-05,
|
| 768 |
+
"loss": 0.1778,
|
| 769 |
+
"step": 1060
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 2.6800250469630558,
|
| 773 |
+
"grad_norm": 6.209780693054199,
|
| 774 |
+
"learning_rate": 1.5459610027855154e-05,
|
| 775 |
+
"loss": 0.1789,
|
| 776 |
+
"step": 1070
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.7050720100187853,
|
| 780 |
+
"grad_norm": 2.586533784866333,
|
| 781 |
+
"learning_rate": 1.5292479108635098e-05,
|
| 782 |
+
"loss": 0.1648,
|
| 783 |
+
"step": 1080
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.730118973074515,
|
| 787 |
+
"grad_norm": 3.024934768676758,
|
| 788 |
+
"learning_rate": 1.5125348189415043e-05,
|
| 789 |
+
"loss": 0.1428,
|
| 790 |
+
"step": 1090
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 2.7551659361302443,
|
| 794 |
+
"grad_norm": 3.368821382522583,
|
| 795 |
+
"learning_rate": 1.4958217270194987e-05,
|
| 796 |
+
"loss": 0.1499,
|
| 797 |
+
"step": 1100
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 2.780212899185974,
|
| 801 |
+
"grad_norm": 2.3457772731781006,
|
| 802 |
+
"learning_rate": 1.479108635097493e-05,
|
| 803 |
+
"loss": 0.1733,
|
| 804 |
+
"step": 1110
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.8052598622417033,
|
| 808 |
+
"grad_norm": 1.2713210582733154,
|
| 809 |
+
"learning_rate": 1.4623955431754876e-05,
|
| 810 |
+
"loss": 0.1533,
|
| 811 |
+
"step": 1120
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 2.830306825297433,
|
| 815 |
+
"grad_norm": 2.4018795490264893,
|
| 816 |
+
"learning_rate": 1.445682451253482e-05,
|
| 817 |
+
"loss": 0.1713,
|
| 818 |
+
"step": 1130
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.8553537883531623,
|
| 822 |
+
"grad_norm": 2.1352932453155518,
|
| 823 |
+
"learning_rate": 1.4289693593314764e-05,
|
| 824 |
+
"loss": 0.1723,
|
| 825 |
+
"step": 1140
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.8804007514088914,
|
| 829 |
+
"grad_norm": 1.5749989748001099,
|
| 830 |
+
"learning_rate": 1.4122562674094708e-05,
|
| 831 |
+
"loss": 0.1965,
|
| 832 |
+
"step": 1150
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 2.9054477144646214,
|
| 836 |
+
"grad_norm": 3.0269646644592285,
|
| 837 |
+
"learning_rate": 1.3955431754874653e-05,
|
| 838 |
+
"loss": 0.1965,
|
| 839 |
+
"step": 1160
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 2.9304946775203504,
|
| 843 |
+
"grad_norm": 2.5006277561187744,
|
| 844 |
+
"learning_rate": 1.3788300835654596e-05,
|
| 845 |
+
"loss": 0.163,
|
| 846 |
+
"step": 1170
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 2.9555416405760804,
|
| 850 |
+
"grad_norm": 1.5534733533859253,
|
| 851 |
+
"learning_rate": 1.362116991643454e-05,
|
| 852 |
+
"loss": 0.1569,
|
| 853 |
+
"step": 1180
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 2.9805886036318094,
|
| 857 |
+
"grad_norm": 2.774616241455078,
|
| 858 |
+
"learning_rate": 1.3454038997214485e-05,
|
| 859 |
+
"loss": 0.1834,
|
| 860 |
+
"step": 1190
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 2.9981214777708205,
|
| 864 |
+
"eval_accuracy": 0.980435422182995,
|
| 865 |
+
"eval_loss": 0.08890607953071594,
|
| 866 |
+
"eval_runtime": 10.7458,
|
| 867 |
+
"eval_samples_per_second": 632.617,
|
| 868 |
+
"eval_steps_per_second": 19.822,
|
| 869 |
+
"step": 1197
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 3.005635566687539,
|
| 873 |
+
"grad_norm": 3.6320059299468994,
|
| 874 |
+
"learning_rate": 1.3286908077994428e-05,
|
| 875 |
+
"loss": 0.1771,
|
| 876 |
+
"step": 1200
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 3.0306825297432685,
|
| 880 |
+
"grad_norm": 1.4830687046051025,
|
| 881 |
+
"learning_rate": 1.3119777158774374e-05,
|
| 882 |
+
"loss": 0.171,
|
| 883 |
+
"step": 1210
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 3.055729492798998,
|
| 887 |
+
"grad_norm": 1.7326056957244873,
|
| 888 |
+
"learning_rate": 1.2952646239554317e-05,
|
| 889 |
+
"loss": 0.1576,
|
| 890 |
+
"step": 1220
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.0807764558547275,
|
| 894 |
+
"grad_norm": 2.301881790161133,
|
| 895 |
+
"learning_rate": 1.2785515320334262e-05,
|
| 896 |
+
"loss": 0.1675,
|
| 897 |
+
"step": 1230
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 3.105823418910457,
|
| 901 |
+
"grad_norm": 3.6545441150665283,
|
| 902 |
+
"learning_rate": 1.2618384401114206e-05,
|
| 903 |
+
"loss": 0.1707,
|
| 904 |
+
"step": 1240
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 3.1308703819661865,
|
| 908 |
+
"grad_norm": 1.9802820682525635,
|
| 909 |
+
"learning_rate": 1.245125348189415e-05,
|
| 910 |
+
"loss": 0.1744,
|
| 911 |
+
"step": 1250
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 3.155917345021916,
|
| 915 |
+
"grad_norm": 2.116532564163208,
|
| 916 |
+
"learning_rate": 1.2284122562674095e-05,
|
| 917 |
+
"loss": 0.1592,
|
| 918 |
+
"step": 1260
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 3.1809643080776455,
|
| 922 |
+
"grad_norm": 2.7397706508636475,
|
| 923 |
+
"learning_rate": 1.211699164345404e-05,
|
| 924 |
+
"loss": 0.1567,
|
| 925 |
+
"step": 1270
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 3.206011271133375,
|
| 929 |
+
"grad_norm": 2.8573620319366455,
|
| 930 |
+
"learning_rate": 1.1949860724233983e-05,
|
| 931 |
+
"loss": 0.1673,
|
| 932 |
+
"step": 1280
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 3.2310582341891045,
|
| 936 |
+
"grad_norm": 3.157711982727051,
|
| 937 |
+
"learning_rate": 1.1782729805013929e-05,
|
| 938 |
+
"loss": 0.1345,
|
| 939 |
+
"step": 1290
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.256105197244834,
|
| 943 |
+
"grad_norm": 4.402627944946289,
|
| 944 |
+
"learning_rate": 1.1615598885793872e-05,
|
| 945 |
+
"loss": 0.1689,
|
| 946 |
+
"step": 1300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 3.2811521603005636,
|
| 950 |
+
"grad_norm": 1.3651645183563232,
|
| 951 |
+
"learning_rate": 1.1448467966573816e-05,
|
| 952 |
+
"loss": 0.0912,
|
| 953 |
+
"step": 1310
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 3.306199123356293,
|
| 957 |
+
"grad_norm": 3.8698039054870605,
|
| 958 |
+
"learning_rate": 1.1281337047353761e-05,
|
| 959 |
+
"loss": 0.1594,
|
| 960 |
+
"step": 1320
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 3.3312460864120226,
|
| 964 |
+
"grad_norm": 3.035409450531006,
|
| 965 |
+
"learning_rate": 1.1114206128133705e-05,
|
| 966 |
+
"loss": 0.1601,
|
| 967 |
+
"step": 1330
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 3.356293049467752,
|
| 971 |
+
"grad_norm": 1.8605338335037231,
|
| 972 |
+
"learning_rate": 1.0947075208913648e-05,
|
| 973 |
+
"loss": 0.1713,
|
| 974 |
+
"step": 1340
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 3.3813400125234816,
|
| 978 |
+
"grad_norm": 3.120457887649536,
|
| 979 |
+
"learning_rate": 1.0779944289693595e-05,
|
| 980 |
+
"loss": 0.1542,
|
| 981 |
+
"step": 1350
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.406386975579211,
|
| 985 |
+
"grad_norm": 3.9332990646362305,
|
| 986 |
+
"learning_rate": 1.0612813370473537e-05,
|
| 987 |
+
"loss": 0.1451,
|
| 988 |
+
"step": 1360
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 3.4314339386349406,
|
| 992 |
+
"grad_norm": 3.2169320583343506,
|
| 993 |
+
"learning_rate": 1.0445682451253482e-05,
|
| 994 |
+
"loss": 0.1417,
|
| 995 |
+
"step": 1370
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 3.45648090169067,
|
| 999 |
+
"grad_norm": 1.290043592453003,
|
| 1000 |
+
"learning_rate": 1.0278551532033427e-05,
|
| 1001 |
+
"loss": 0.1672,
|
| 1002 |
+
"step": 1380
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 3.4815278647463996,
|
| 1006 |
+
"grad_norm": 1.9304981231689453,
|
| 1007 |
+
"learning_rate": 1.0111420612813371e-05,
|
| 1008 |
+
"loss": 0.1558,
|
| 1009 |
+
"step": 1390
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 3.506574827802129,
|
| 1013 |
+
"grad_norm": 1.4201843738555908,
|
| 1014 |
+
"learning_rate": 9.944289693593314e-06,
|
| 1015 |
+
"loss": 0.1535,
|
| 1016 |
+
"step": 1400
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 3.5316217908578587,
|
| 1020 |
+
"grad_norm": 1.5534597635269165,
|
| 1021 |
+
"learning_rate": 9.77715877437326e-06,
|
| 1022 |
+
"loss": 0.1275,
|
| 1023 |
+
"step": 1410
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.5566687539135877,
|
| 1027 |
+
"grad_norm": 3.183288097381592,
|
| 1028 |
+
"learning_rate": 9.610027855153203e-06,
|
| 1029 |
+
"loss": 0.1588,
|
| 1030 |
+
"step": 1420
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 3.5817157169693177,
|
| 1034 |
+
"grad_norm": 3.252284288406372,
|
| 1035 |
+
"learning_rate": 9.44289693593315e-06,
|
| 1036 |
+
"loss": 0.1266,
|
| 1037 |
+
"step": 1430
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 3.6067626800250467,
|
| 1041 |
+
"grad_norm": 3.5988216400146484,
|
| 1042 |
+
"learning_rate": 9.275766016713092e-06,
|
| 1043 |
+
"loss": 0.1484,
|
| 1044 |
+
"step": 1440
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 3.6318096430807767,
|
| 1048 |
+
"grad_norm": 0.8129429221153259,
|
| 1049 |
+
"learning_rate": 9.108635097493037e-06,
|
| 1050 |
+
"loss": 0.1249,
|
| 1051 |
+
"step": 1450
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 3.6568566061365058,
|
| 1055 |
+
"grad_norm": 1.1750004291534424,
|
| 1056 |
+
"learning_rate": 8.941504178272981e-06,
|
| 1057 |
+
"loss": 0.1552,
|
| 1058 |
+
"step": 1460
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 3.6819035691922357,
|
| 1062 |
+
"grad_norm": 2.3609495162963867,
|
| 1063 |
+
"learning_rate": 8.774373259052926e-06,
|
| 1064 |
+
"loss": 0.1506,
|
| 1065 |
+
"step": 1470
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.706950532247965,
|
| 1069 |
+
"grad_norm": 3.020759344100952,
|
| 1070 |
+
"learning_rate": 8.607242339832869e-06,
|
| 1071 |
+
"loss": 0.1492,
|
| 1072 |
+
"step": 1480
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 3.7319974953036943,
|
| 1076 |
+
"grad_norm": 2.7013983726501465,
|
| 1077 |
+
"learning_rate": 8.440111420612815e-06,
|
| 1078 |
+
"loss": 0.1045,
|
| 1079 |
+
"step": 1490
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.757044458359424,
|
| 1083 |
+
"grad_norm": 3.6517138481140137,
|
| 1084 |
+
"learning_rate": 8.272980501392758e-06,
|
| 1085 |
+
"loss": 0.1354,
|
| 1086 |
+
"step": 1500
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 3.7820914214151533,
|
| 1090 |
+
"grad_norm": 1.283496618270874,
|
| 1091 |
+
"learning_rate": 8.1058495821727e-06,
|
| 1092 |
+
"loss": 0.1376,
|
| 1093 |
+
"step": 1510
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 3.807138384470883,
|
| 1097 |
+
"grad_norm": 4.068831443786621,
|
| 1098 |
+
"learning_rate": 7.938718662952647e-06,
|
| 1099 |
+
"loss": 0.147,
|
| 1100 |
+
"step": 1520
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 3.8321853475266123,
|
| 1104 |
+
"grad_norm": 1.8731778860092163,
|
| 1105 |
+
"learning_rate": 7.77158774373259e-06,
|
| 1106 |
+
"loss": 0.134,
|
| 1107 |
+
"step": 1530
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.857232310582342,
|
| 1111 |
+
"grad_norm": 2.750255823135376,
|
| 1112 |
+
"learning_rate": 7.604456824512535e-06,
|
| 1113 |
+
"loss": 0.1398,
|
| 1114 |
+
"step": 1540
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 3.8822792736380713,
|
| 1118 |
+
"grad_norm": 2.020031213760376,
|
| 1119 |
+
"learning_rate": 7.43732590529248e-06,
|
| 1120 |
+
"loss": 0.1236,
|
| 1121 |
+
"step": 1550
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 3.907326236693801,
|
| 1125 |
+
"grad_norm": 3.4288525581359863,
|
| 1126 |
+
"learning_rate": 7.2701949860724235e-06,
|
| 1127 |
+
"loss": 0.1223,
|
| 1128 |
+
"step": 1560
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 3.9323731997495304,
|
| 1132 |
+
"grad_norm": 2.974663019180298,
|
| 1133 |
+
"learning_rate": 7.103064066852368e-06,
|
| 1134 |
+
"loss": 0.1596,
|
| 1135 |
+
"step": 1570
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 3.95742016280526,
|
| 1139 |
+
"grad_norm": 2.157737970352173,
|
| 1140 |
+
"learning_rate": 6.935933147632313e-06,
|
| 1141 |
+
"loss": 0.1277,
|
| 1142 |
+
"step": 1580
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 3.9824671258609894,
|
| 1146 |
+
"grad_norm": 2.2225987911224365,
|
| 1147 |
+
"learning_rate": 6.768802228412256e-06,
|
| 1148 |
+
"loss": 0.142,
|
| 1149 |
+
"step": 1590
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 4.0,
|
| 1153 |
+
"eval_accuracy": 0.981318034716093,
|
| 1154 |
+
"eval_loss": 0.08539879322052002,
|
| 1155 |
+
"eval_runtime": 10.6636,
|
| 1156 |
+
"eval_samples_per_second": 637.494,
|
| 1157 |
+
"eval_steps_per_second": 19.974,
|
| 1158 |
+
"step": 1597
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 4.007514088916719,
|
| 1162 |
+
"grad_norm": 1.7441836595535278,
|
| 1163 |
+
"learning_rate": 6.601671309192201e-06,
|
| 1164 |
+
"loss": 0.1491,
|
| 1165 |
+
"step": 1600
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 4.032561051972448,
|
| 1169 |
+
"grad_norm": 2.0195930004119873,
|
| 1170 |
+
"learning_rate": 6.4345403899721455e-06,
|
| 1171 |
+
"loss": 0.1357,
|
| 1172 |
+
"step": 1610
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 4.057608015028178,
|
| 1176 |
+
"grad_norm": 2.4423089027404785,
|
| 1177 |
+
"learning_rate": 6.267409470752089e-06,
|
| 1178 |
+
"loss": 0.1438,
|
| 1179 |
+
"step": 1620
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 4.082654978083907,
|
| 1183 |
+
"grad_norm": 2.6411936283111572,
|
| 1184 |
+
"learning_rate": 6.100278551532034e-06,
|
| 1185 |
+
"loss": 0.131,
|
| 1186 |
+
"step": 1630
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 4.107701941139637,
|
| 1190 |
+
"grad_norm": 0.8893925547599792,
|
| 1191 |
+
"learning_rate": 5.933147632311978e-06,
|
| 1192 |
+
"loss": 0.1249,
|
| 1193 |
+
"step": 1640
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 4.132748904195366,
|
| 1197 |
+
"grad_norm": 2.2715697288513184,
|
| 1198 |
+
"learning_rate": 5.766016713091923e-06,
|
| 1199 |
+
"loss": 0.1642,
|
| 1200 |
+
"step": 1650
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 4.157795867251096,
|
| 1204 |
+
"grad_norm": 1.903985619544983,
|
| 1205 |
+
"learning_rate": 5.598885793871867e-06,
|
| 1206 |
+
"loss": 0.1418,
|
| 1207 |
+
"step": 1660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 4.182842830306825,
|
| 1211 |
+
"grad_norm": 2.41510272026062,
|
| 1212 |
+
"learning_rate": 5.43175487465181e-06,
|
| 1213 |
+
"loss": 0.1535,
|
| 1214 |
+
"step": 1670
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 4.207889793362555,
|
| 1218 |
+
"grad_norm": 3.837315320968628,
|
| 1219 |
+
"learning_rate": 5.264623955431755e-06,
|
| 1220 |
+
"loss": 0.1433,
|
| 1221 |
+
"step": 1680
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 4.232936756418284,
|
| 1225 |
+
"grad_norm": 2.4811267852783203,
|
| 1226 |
+
"learning_rate": 5.0974930362116986e-06,
|
| 1227 |
+
"loss": 0.121,
|
| 1228 |
+
"step": 1690
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 4.257983719474014,
|
| 1232 |
+
"grad_norm": 4.462907791137695,
|
| 1233 |
+
"learning_rate": 4.930362116991643e-06,
|
| 1234 |
+
"loss": 0.1267,
|
| 1235 |
+
"step": 1700
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 4.283030682529743,
|
| 1239 |
+
"grad_norm": 3.019132137298584,
|
| 1240 |
+
"learning_rate": 4.763231197771588e-06,
|
| 1241 |
+
"loss": 0.1151,
|
| 1242 |
+
"step": 1710
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 4.308077645585473,
|
| 1246 |
+
"grad_norm": 3.2697532176971436,
|
| 1247 |
+
"learning_rate": 4.596100278551532e-06,
|
| 1248 |
+
"loss": 0.1068,
|
| 1249 |
+
"step": 1720
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 4.333124608641202,
|
| 1253 |
+
"grad_norm": 3.3036513328552246,
|
| 1254 |
+
"learning_rate": 4.428969359331476e-06,
|
| 1255 |
+
"loss": 0.1009,
|
| 1256 |
+
"step": 1730
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 4.358171571696932,
|
| 1260 |
+
"grad_norm": 2.5975863933563232,
|
| 1261 |
+
"learning_rate": 4.2618384401114205e-06,
|
| 1262 |
+
"loss": 0.1168,
|
| 1263 |
+
"step": 1740
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.383218534752661,
|
| 1267 |
+
"grad_norm": 1.6975170373916626,
|
| 1268 |
+
"learning_rate": 4.094707520891365e-06,
|
| 1269 |
+
"loss": 0.1549,
|
| 1270 |
+
"step": 1750
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 4.408265497808391,
|
| 1274 |
+
"grad_norm": 2.671894073486328,
|
| 1275 |
+
"learning_rate": 3.927576601671309e-06,
|
| 1276 |
+
"loss": 0.1211,
|
| 1277 |
+
"step": 1760
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 4.43331246086412,
|
| 1281 |
+
"grad_norm": 4.581605911254883,
|
| 1282 |
+
"learning_rate": 3.7604456824512533e-06,
|
| 1283 |
+
"loss": 0.1406,
|
| 1284 |
+
"step": 1770
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 4.45835942391985,
|
| 1288 |
+
"grad_norm": 3.0454249382019043,
|
| 1289 |
+
"learning_rate": 3.593314763231198e-06,
|
| 1290 |
+
"loss": 0.137,
|
| 1291 |
+
"step": 1780
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 4.483406386975579,
|
| 1295 |
+
"grad_norm": 2.0320184230804443,
|
| 1296 |
+
"learning_rate": 3.426183844011142e-06,
|
| 1297 |
+
"loss": 0.1105,
|
| 1298 |
+
"step": 1790
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 4.508453350031308,
|
| 1302 |
+
"grad_norm": 1.5015639066696167,
|
| 1303 |
+
"learning_rate": 3.259052924791086e-06,
|
| 1304 |
+
"loss": 0.1282,
|
| 1305 |
+
"step": 1800
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.533500313087038,
|
| 1309 |
+
"grad_norm": 2.1987569332122803,
|
| 1310 |
+
"learning_rate": 3.0919220055710307e-06,
|
| 1311 |
+
"loss": 0.1111,
|
| 1312 |
+
"step": 1810
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 4.558547276142768,
|
| 1316 |
+
"grad_norm": 1.3748631477355957,
|
| 1317 |
+
"learning_rate": 2.924791086350975e-06,
|
| 1318 |
+
"loss": 0.1326,
|
| 1319 |
+
"step": 1820
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 4.583594239198497,
|
| 1323 |
+
"grad_norm": 1.8067643642425537,
|
| 1324 |
+
"learning_rate": 2.7576601671309194e-06,
|
| 1325 |
+
"loss": 0.1112,
|
| 1326 |
+
"step": 1830
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 4.608641202254226,
|
| 1330 |
+
"grad_norm": 2.4993462562561035,
|
| 1331 |
+
"learning_rate": 2.5905292479108636e-06,
|
| 1332 |
+
"loss": 0.1014,
|
| 1333 |
+
"step": 1840
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 4.633688165309956,
|
| 1337 |
+
"grad_norm": 3.3623006343841553,
|
| 1338 |
+
"learning_rate": 2.4233983286908077e-06,
|
| 1339 |
+
"loss": 0.1319,
|
| 1340 |
+
"step": 1850
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 4.658735128365686,
|
| 1344 |
+
"grad_norm": 2.678271770477295,
|
| 1345 |
+
"learning_rate": 2.2562674094707523e-06,
|
| 1346 |
+
"loss": 0.1391,
|
| 1347 |
+
"step": 1860
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.683782091421415,
|
| 1351 |
+
"grad_norm": 1.3857640027999878,
|
| 1352 |
+
"learning_rate": 2.0891364902506964e-06,
|
| 1353 |
+
"loss": 0.1149,
|
| 1354 |
+
"step": 1870
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 4.708829054477144,
|
| 1358 |
+
"grad_norm": 2.552626371383667,
|
| 1359 |
+
"learning_rate": 1.922005571030641e-06,
|
| 1360 |
+
"loss": 0.119,
|
| 1361 |
+
"step": 1880
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 4.733876017532874,
|
| 1365 |
+
"grad_norm": 1.3172489404678345,
|
| 1366 |
+
"learning_rate": 1.7548746518105849e-06,
|
| 1367 |
+
"loss": 0.1226,
|
| 1368 |
+
"step": 1890
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 4.758922980588603,
|
| 1372 |
+
"grad_norm": 4.114231586456299,
|
| 1373 |
+
"learning_rate": 1.5877437325905292e-06,
|
| 1374 |
+
"loss": 0.1191,
|
| 1375 |
+
"step": 1900
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 4.783969943644333,
|
| 1379 |
+
"grad_norm": 3.1994917392730713,
|
| 1380 |
+
"learning_rate": 1.4206128133704736e-06,
|
| 1381 |
+
"loss": 0.1064,
|
| 1382 |
+
"step": 1910
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 4.809016906700062,
|
| 1386 |
+
"grad_norm": 2.785364866256714,
|
| 1387 |
+
"learning_rate": 1.253481894150418e-06,
|
| 1388 |
+
"loss": 0.1092,
|
| 1389 |
+
"step": 1920
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.834063869755792,
|
| 1393 |
+
"grad_norm": 1.6309309005737305,
|
| 1394 |
+
"learning_rate": 1.0863509749303623e-06,
|
| 1395 |
+
"loss": 0.1355,
|
| 1396 |
+
"step": 1930
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 4.859110832811521,
|
| 1400 |
+
"grad_norm": 3.174809694290161,
|
| 1401 |
+
"learning_rate": 9.192200557103064e-07,
|
| 1402 |
+
"loss": 0.1211,
|
| 1403 |
+
"step": 1940
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 4.884157795867251,
|
| 1407 |
+
"grad_norm": 2.4154794216156006,
|
| 1408 |
+
"learning_rate": 7.520891364902508e-07,
|
| 1409 |
+
"loss": 0.1135,
|
| 1410 |
+
"step": 1950
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 4.90920475892298,
|
| 1414 |
+
"grad_norm": 2.3144145011901855,
|
| 1415 |
+
"learning_rate": 5.84958217270195e-07,
|
| 1416 |
+
"loss": 0.1121,
|
| 1417 |
+
"step": 1960
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 4.93425172197871,
|
| 1421 |
+
"grad_norm": 2.8474748134613037,
|
| 1422 |
+
"learning_rate": 4.178272980501393e-07,
|
| 1423 |
+
"loss": 0.143,
|
| 1424 |
+
"step": 1970
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 4.959298685034439,
|
| 1428 |
+
"grad_norm": 2.4842255115509033,
|
| 1429 |
+
"learning_rate": 2.506963788300836e-07,
|
| 1430 |
+
"loss": 0.116,
|
| 1431 |
+
"step": 1980
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.984345648090169,
|
| 1435 |
+
"grad_norm": 2.9249684810638428,
|
| 1436 |
+
"learning_rate": 8.356545961002785e-08,
|
| 1437 |
+
"loss": 0.1472,
|
| 1438 |
+
"step": 1990
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 4.996869129618034,
|
| 1442 |
+
"eval_accuracy": 0.9832303618711385,
|
| 1443 |
+
"eval_loss": 0.07819115370512009,
|
| 1444 |
+
"eval_runtime": 10.7075,
|
| 1445 |
+
"eval_samples_per_second": 634.883,
|
| 1446 |
+
"eval_steps_per_second": 19.893,
|
| 1447 |
+
"step": 1995
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 4.996869129618034,
|
| 1451 |
+
"step": 1995,
|
| 1452 |
+
"total_flos": 2.3180194781952e+18,
|
| 1453 |
+
"train_loss": 0.40026830736557045,
|
| 1454 |
+
"train_runtime": 1087.6247,
|
| 1455 |
+
"train_samples_per_second": 234.888,
|
| 1456 |
+
"train_steps_per_second": 1.834
|
| 1457 |
+
}
|
| 1458 |
+
],
|
| 1459 |
+
"logging_steps": 10,
|
| 1460 |
+
"max_steps": 1995,
|
| 1461 |
+
"num_input_tokens_seen": 0,
|
| 1462 |
+
"num_train_epochs": 5,
|
| 1463 |
+
"save_steps": 500,
|
| 1464 |
+
"stateful_callbacks": {
|
| 1465 |
+
"TrainerControl": {
|
| 1466 |
+
"args": {
|
| 1467 |
+
"should_epoch_stop": false,
|
| 1468 |
+
"should_evaluate": false,
|
| 1469 |
+
"should_log": false,
|
| 1470 |
+
"should_save": true,
|
| 1471 |
+
"should_training_stop": true
|
| 1472 |
+
},
|
| 1473 |
+
"attributes": {}
|
| 1474 |
+
}
|
| 1475 |
+
},
|
| 1476 |
+
"total_flos": 2.3180194781952e+18,
|
| 1477 |
+
"train_batch_size": 32,
|
| 1478 |
+
"trial_name": null,
|
| 1479 |
+
"trial_params": null
|
| 1480 |
+
}
|