Zheng Li
commited on
End of training
Browse files
README.md
CHANGED
|
@@ -3,6 +3,7 @@ library_name: transformers
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: facebook/wav2vec2-base
|
| 5 |
tags:
|
|
|
|
| 6 |
- generated_from_trainer
|
| 7 |
datasets:
|
| 8 |
- superb
|
|
|
|
| 3 |
license: apache-2.0
|
| 4 |
base_model: facebook/wav2vec2-base
|
| 5 |
tags:
|
| 6 |
+
- audio-classification
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- superb
|
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.989355040701315,
|
| 3 |
+
"eval_accuracy": 0.9810238305383936,
|
| 4 |
+
"eval_loss": 0.08548342436552048,
|
| 5 |
+
"eval_runtime": 4.5788,
|
| 6 |
+
"eval_samples_per_second": 1484.67,
|
| 7 |
+
"eval_steps_per_second": 46.519,
|
| 8 |
+
"total_flos": 2.31453318852096e+18,
|
| 9 |
+
"train_loss": 0.40246317323885467,
|
| 10 |
+
"train_runtime": 627.2253,
|
| 11 |
+
"train_samples_per_second": 407.302,
|
| 12 |
+
"train_steps_per_second": 3.181
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.989355040701315,
|
| 3 |
+
"eval_accuracy": 0.9810238305383936,
|
| 4 |
+
"eval_loss": 0.08548342436552048,
|
| 5 |
+
"eval_runtime": 4.5788,
|
| 6 |
+
"eval_samples_per_second": 1484.67,
|
| 7 |
+
"eval_steps_per_second": 46.519
|
| 8 |
+
}
|
runs/Apr08_15-13-01_cs-Precision-7960-Tower/events.out.tfevents.1744140353.cs-Precision-7960-Tower.23127.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8d6d6e018405a7b8fe7c007d0ec50a5761ceae1c8286a5297dd890668d69344
|
| 3 |
+
size 411
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 4.989355040701315,
|
| 3 |
+
"total_flos": 2.31453318852096e+18,
|
| 4 |
+
"train_loss": 0.40246317323885467,
|
| 5 |
+
"train_runtime": 627.2253,
|
| 6 |
+
"train_samples_per_second": 407.302,
|
| 7 |
+
"train_steps_per_second": 3.181
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1480 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9810238305383936,
|
| 3 |
+
"best_model_checkpoint": "wav2vec2-base-ft-keyword-spotting/checkpoint-1995",
|
| 4 |
+
"epoch": 4.989355040701315,
|
| 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": 1.3846150636672974,
|
| 14 |
+
"learning_rate": 1.5e-06,
|
| 15 |
+
"loss": 2.4917,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.050093926111458985,
|
| 20 |
+
"grad_norm": 1.5032284259796143,
|
| 21 |
+
"learning_rate": 3e-06,
|
| 22 |
+
"loss": 2.4701,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.07514088916718847,
|
| 27 |
+
"grad_norm": 1.4602274894714355,
|
| 28 |
+
"learning_rate": 4.5e-06,
|
| 29 |
+
"loss": 2.415,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.10018785222291797,
|
| 34 |
+
"grad_norm": 2.0218751430511475,
|
| 35 |
+
"learning_rate": 6e-06,
|
| 36 |
+
"loss": 2.3156,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.12523481527864747,
|
| 41 |
+
"grad_norm": 2.49809193611145,
|
| 42 |
+
"learning_rate": 7.5e-06,
|
| 43 |
+
"loss": 2.1799,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.15028177833437695,
|
| 48 |
+
"grad_norm": 2.539499282836914,
|
| 49 |
+
"learning_rate": 9e-06,
|
| 50 |
+
"loss": 1.9986,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.17532874139010646,
|
| 55 |
+
"grad_norm": 2.2324798107147217,
|
| 56 |
+
"learning_rate": 1.05e-05,
|
| 57 |
+
"loss": 1.7848,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.20037570444583594,
|
| 62 |
+
"grad_norm": 1.8337464332580566,
|
| 63 |
+
"learning_rate": 1.2e-05,
|
| 64 |
+
"loss": 1.7187,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.22542266750156542,
|
| 69 |
+
"grad_norm": 1.7367889881134033,
|
| 70 |
+
"learning_rate": 1.3500000000000001e-05,
|
| 71 |
+
"loss": 1.6049,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.25046963055729493,
|
| 76 |
+
"grad_norm": 1.1910614967346191,
|
| 77 |
+
"learning_rate": 1.5e-05,
|
| 78 |
+
"loss": 1.5746,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.27551659361302444,
|
| 83 |
+
"grad_norm": 0.6088196635246277,
|
| 84 |
+
"learning_rate": 1.65e-05,
|
| 85 |
+
"loss": 1.6007,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.3005635566687539,
|
| 90 |
+
"grad_norm": 1.1780052185058594,
|
| 91 |
+
"learning_rate": 1.8e-05,
|
| 92 |
+
"loss": 1.4912,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.3256105197244834,
|
| 97 |
+
"grad_norm": 0.6295515894889832,
|
| 98 |
+
"learning_rate": 1.95e-05,
|
| 99 |
+
"loss": 1.4884,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.3506574827802129,
|
| 104 |
+
"grad_norm": 1.570447325706482,
|
| 105 |
+
"learning_rate": 2.1e-05,
|
| 106 |
+
"loss": 1.5284,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.37570444583594237,
|
| 111 |
+
"grad_norm": 1.6432406902313232,
|
| 112 |
+
"learning_rate": 2.25e-05,
|
| 113 |
+
"loss": 1.4759,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.4007514088916719,
|
| 118 |
+
"grad_norm": 1.9464057683944702,
|
| 119 |
+
"learning_rate": 2.4e-05,
|
| 120 |
+
"loss": 1.4358,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.4257983719474014,
|
| 125 |
+
"grad_norm": 3.1392250061035156,
|
| 126 |
+
"learning_rate": 2.55e-05,
|
| 127 |
+
"loss": 1.4103,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.45084533500313084,
|
| 132 |
+
"grad_norm": 2.4613516330718994,
|
| 133 |
+
"learning_rate": 2.7000000000000002e-05,
|
| 134 |
+
"loss": 1.3404,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.47589229805886035,
|
| 139 |
+
"grad_norm": 2.962592363357544,
|
| 140 |
+
"learning_rate": 2.8499999999999998e-05,
|
| 141 |
+
"loss": 1.3514,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.5009392611145899,
|
| 146 |
+
"grad_norm": 3.589179039001465,
|
| 147 |
+
"learning_rate": 3e-05,
|
| 148 |
+
"loss": 1.32,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.5259862241703194,
|
| 153 |
+
"grad_norm": 7.765381813049316,
|
| 154 |
+
"learning_rate": 2.9832869080779945e-05,
|
| 155 |
+
"loss": 1.261,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.5510331872260489,
|
| 160 |
+
"grad_norm": 4.343158721923828,
|
| 161 |
+
"learning_rate": 2.9665738161559886e-05,
|
| 162 |
+
"loss": 1.1817,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.5760801502817783,
|
| 167 |
+
"grad_norm": 2.7387561798095703,
|
| 168 |
+
"learning_rate": 2.9498607242339834e-05,
|
| 169 |
+
"loss": 1.0597,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.6011271133375078,
|
| 174 |
+
"grad_norm": 3.2945845127105713,
|
| 175 |
+
"learning_rate": 2.933147632311978e-05,
|
| 176 |
+
"loss": 1.0775,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.6261740763932373,
|
| 181 |
+
"grad_norm": 3.404000759124756,
|
| 182 |
+
"learning_rate": 2.916434540389972e-05,
|
| 183 |
+
"loss": 0.944,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.6512210394489668,
|
| 188 |
+
"grad_norm": 2.9037885665893555,
|
| 189 |
+
"learning_rate": 2.8997214484679665e-05,
|
| 190 |
+
"loss": 0.965,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.6762680025046963,
|
| 195 |
+
"grad_norm": 3.4317057132720947,
|
| 196 |
+
"learning_rate": 2.8830083565459613e-05,
|
| 197 |
+
"loss": 0.8292,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.7013149655604258,
|
| 202 |
+
"grad_norm": 3.8530466556549072,
|
| 203 |
+
"learning_rate": 2.8662952646239554e-05,
|
| 204 |
+
"loss": 0.7791,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.7263619286161553,
|
| 209 |
+
"grad_norm": 3.3247628211975098,
|
| 210 |
+
"learning_rate": 2.84958217270195e-05,
|
| 211 |
+
"loss": 0.7834,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.7514088916718847,
|
| 216 |
+
"grad_norm": 3.122251510620117,
|
| 217 |
+
"learning_rate": 2.8328690807799443e-05,
|
| 218 |
+
"loss": 0.7576,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.7764558547276142,
|
| 223 |
+
"grad_norm": 5.345429420471191,
|
| 224 |
+
"learning_rate": 2.8161559888579388e-05,
|
| 225 |
+
"loss": 0.7777,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.8015028177833438,
|
| 230 |
+
"grad_norm": 3.185194253921509,
|
| 231 |
+
"learning_rate": 2.7994428969359332e-05,
|
| 232 |
+
"loss": 0.7278,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.8265497808390733,
|
| 237 |
+
"grad_norm": 2.6544058322906494,
|
| 238 |
+
"learning_rate": 2.7827298050139277e-05,
|
| 239 |
+
"loss": 0.6651,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.8515967438948028,
|
| 244 |
+
"grad_norm": 3.633167028427124,
|
| 245 |
+
"learning_rate": 2.7660167130919218e-05,
|
| 246 |
+
"loss": 0.6521,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.8766437069505323,
|
| 251 |
+
"grad_norm": 1.6489408016204834,
|
| 252 |
+
"learning_rate": 2.7493036211699166e-05,
|
| 253 |
+
"loss": 0.6365,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.9016906700062617,
|
| 258 |
+
"grad_norm": 3.1760778427124023,
|
| 259 |
+
"learning_rate": 2.732590529247911e-05,
|
| 260 |
+
"loss": 0.5956,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.9267376330619912,
|
| 265 |
+
"grad_norm": 2.8132238388061523,
|
| 266 |
+
"learning_rate": 2.7158774373259055e-05,
|
| 267 |
+
"loss": 0.5725,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.9517845961177207,
|
| 272 |
+
"grad_norm": 3.009735107421875,
|
| 273 |
+
"learning_rate": 2.6991643454038996e-05,
|
| 274 |
+
"loss": 0.5173,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.9768315591734502,
|
| 279 |
+
"grad_norm": 2.352182388305664,
|
| 280 |
+
"learning_rate": 2.6824512534818944e-05,
|
| 281 |
+
"loss": 0.5199,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.0,
|
| 286 |
+
"grad_norm": 0.45235970616340637,
|
| 287 |
+
"learning_rate": 2.665738161559889e-05,
|
| 288 |
+
"loss": 0.487,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.0,
|
| 293 |
+
"eval_accuracy": 0.9701382759635186,
|
| 294 |
+
"eval_loss": 0.3712795078754425,
|
| 295 |
+
"eval_runtime": 4.5332,
|
| 296 |
+
"eval_samples_per_second": 1499.6,
|
| 297 |
+
"eval_steps_per_second": 46.987,
|
| 298 |
+
"step": 400
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 1.0250469630557295,
|
| 302 |
+
"grad_norm": 2.322962999343872,
|
| 303 |
+
"learning_rate": 2.649025069637883e-05,
|
| 304 |
+
"loss": 0.4671,
|
| 305 |
+
"step": 410
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 1.050093926111459,
|
| 309 |
+
"grad_norm": 3.166677236557007,
|
| 310 |
+
"learning_rate": 2.6323119777158774e-05,
|
| 311 |
+
"loss": 0.4417,
|
| 312 |
+
"step": 420
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 1.0751408891671885,
|
| 316 |
+
"grad_norm": 2.0986714363098145,
|
| 317 |
+
"learning_rate": 2.6155988857938722e-05,
|
| 318 |
+
"loss": 0.4148,
|
| 319 |
+
"step": 430
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 1.100187852222918,
|
| 323 |
+
"grad_norm": 3.847125291824341,
|
| 324 |
+
"learning_rate": 2.5988857938718663e-05,
|
| 325 |
+
"loss": 0.4359,
|
| 326 |
+
"step": 440
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.1252348152786475,
|
| 330 |
+
"grad_norm": 2.775067090988159,
|
| 331 |
+
"learning_rate": 2.5821727019498608e-05,
|
| 332 |
+
"loss": 0.3778,
|
| 333 |
+
"step": 450
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.150281778334377,
|
| 337 |
+
"grad_norm": 4.5280914306640625,
|
| 338 |
+
"learning_rate": 2.5654596100278553e-05,
|
| 339 |
+
"loss": 0.3776,
|
| 340 |
+
"step": 460
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 1.1753287413901066,
|
| 344 |
+
"grad_norm": 2.7972805500030518,
|
| 345 |
+
"learning_rate": 2.5487465181058497e-05,
|
| 346 |
+
"loss": 0.3381,
|
| 347 |
+
"step": 470
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.2003757044458359,
|
| 351 |
+
"grad_norm": 3.291175603866577,
|
| 352 |
+
"learning_rate": 2.532033426183844e-05,
|
| 353 |
+
"loss": 0.3599,
|
| 354 |
+
"step": 480
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 1.2254226675015654,
|
| 358 |
+
"grad_norm": 3.2598955631256104,
|
| 359 |
+
"learning_rate": 2.5153203342618386e-05,
|
| 360 |
+
"loss": 0.3365,
|
| 361 |
+
"step": 490
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 1.2504696305572949,
|
| 365 |
+
"grad_norm": 1.4448655843734741,
|
| 366 |
+
"learning_rate": 2.4986072423398327e-05,
|
| 367 |
+
"loss": 0.319,
|
| 368 |
+
"step": 500
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 1.2755165936130244,
|
| 372 |
+
"grad_norm": 2.443610191345215,
|
| 373 |
+
"learning_rate": 2.4818941504178275e-05,
|
| 374 |
+
"loss": 0.3468,
|
| 375 |
+
"step": 510
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 1.300563556668754,
|
| 379 |
+
"grad_norm": 4.200746536254883,
|
| 380 |
+
"learning_rate": 2.465181058495822e-05,
|
| 381 |
+
"loss": 0.3075,
|
| 382 |
+
"step": 520
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 1.3256105197244834,
|
| 386 |
+
"grad_norm": 2.271791458129883,
|
| 387 |
+
"learning_rate": 2.448467966573816e-05,
|
| 388 |
+
"loss": 0.3326,
|
| 389 |
+
"step": 530
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.350657482780213,
|
| 393 |
+
"grad_norm": 2.7881224155426025,
|
| 394 |
+
"learning_rate": 2.4317548746518106e-05,
|
| 395 |
+
"loss": 0.2874,
|
| 396 |
+
"step": 540
|
| 397 |
+
},
|
| 398 |
+
{
|
| 399 |
+
"epoch": 1.3757044458359424,
|
| 400 |
+
"grad_norm": 1.9204065799713135,
|
| 401 |
+
"learning_rate": 2.415041782729805e-05,
|
| 402 |
+
"loss": 0.2942,
|
| 403 |
+
"step": 550
|
| 404 |
+
},
|
| 405 |
+
{
|
| 406 |
+
"epoch": 1.400751408891672,
|
| 407 |
+
"grad_norm": 3.521867275238037,
|
| 408 |
+
"learning_rate": 2.3983286908077995e-05,
|
| 409 |
+
"loss": 0.2733,
|
| 410 |
+
"step": 560
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 1.4257983719474014,
|
| 414 |
+
"grad_norm": 5.327916622161865,
|
| 415 |
+
"learning_rate": 2.381615598885794e-05,
|
| 416 |
+
"loss": 0.2843,
|
| 417 |
+
"step": 570
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 1.4508453350031307,
|
| 421 |
+
"grad_norm": 1.7485443353652954,
|
| 422 |
+
"learning_rate": 2.3649025069637884e-05,
|
| 423 |
+
"loss": 0.2635,
|
| 424 |
+
"step": 580
|
| 425 |
+
},
|
| 426 |
+
{
|
| 427 |
+
"epoch": 1.4758922980588602,
|
| 428 |
+
"grad_norm": 1.733283281326294,
|
| 429 |
+
"learning_rate": 2.3481894150417825e-05,
|
| 430 |
+
"loss": 0.2598,
|
| 431 |
+
"step": 590
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.5009392611145898,
|
| 435 |
+
"grad_norm": 4.292211532592773,
|
| 436 |
+
"learning_rate": 2.3314763231197773e-05,
|
| 437 |
+
"loss": 0.2626,
|
| 438 |
+
"step": 600
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 1.5259862241703193,
|
| 442 |
+
"grad_norm": 3.2708067893981934,
|
| 443 |
+
"learning_rate": 2.3147632311977718e-05,
|
| 444 |
+
"loss": 0.2783,
|
| 445 |
+
"step": 610
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 1.5510331872260488,
|
| 449 |
+
"grad_norm": 1.9687224626541138,
|
| 450 |
+
"learning_rate": 2.298050139275766e-05,
|
| 451 |
+
"loss": 0.2292,
|
| 452 |
+
"step": 620
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.5760801502817783,
|
| 456 |
+
"grad_norm": 4.437740802764893,
|
| 457 |
+
"learning_rate": 2.2813370473537603e-05,
|
| 458 |
+
"loss": 0.2475,
|
| 459 |
+
"step": 630
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 1.6011271133375078,
|
| 463 |
+
"grad_norm": 1.799039363861084,
|
| 464 |
+
"learning_rate": 2.264623955431755e-05,
|
| 465 |
+
"loss": 0.2312,
|
| 466 |
+
"step": 640
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 1.6261740763932373,
|
| 470 |
+
"grad_norm": 2.9007081985473633,
|
| 471 |
+
"learning_rate": 2.2479108635097492e-05,
|
| 472 |
+
"loss": 0.2475,
|
| 473 |
+
"step": 650
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.6512210394489668,
|
| 477 |
+
"grad_norm": 5.133575439453125,
|
| 478 |
+
"learning_rate": 2.2311977715877437e-05,
|
| 479 |
+
"loss": 0.2499,
|
| 480 |
+
"step": 660
|
| 481 |
+
},
|
| 482 |
+
{
|
| 483 |
+
"epoch": 1.6762680025046963,
|
| 484 |
+
"grad_norm": 4.255336761474609,
|
| 485 |
+
"learning_rate": 2.214484679665738e-05,
|
| 486 |
+
"loss": 0.2798,
|
| 487 |
+
"step": 670
|
| 488 |
+
},
|
| 489 |
+
{
|
| 490 |
+
"epoch": 1.7013149655604258,
|
| 491 |
+
"grad_norm": 2.9616687297821045,
|
| 492 |
+
"learning_rate": 2.1977715877437326e-05,
|
| 493 |
+
"loss": 0.2532,
|
| 494 |
+
"step": 680
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.7263619286161553,
|
| 498 |
+
"grad_norm": 2.9893698692321777,
|
| 499 |
+
"learning_rate": 2.181058495821727e-05,
|
| 500 |
+
"loss": 0.2411,
|
| 501 |
+
"step": 690
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 1.7514088916718848,
|
| 505 |
+
"grad_norm": 1.4978973865509033,
|
| 506 |
+
"learning_rate": 2.1643454038997215e-05,
|
| 507 |
+
"loss": 0.2401,
|
| 508 |
+
"step": 700
|
| 509 |
+
},
|
| 510 |
+
{
|
| 511 |
+
"epoch": 1.7764558547276144,
|
| 512 |
+
"grad_norm": 4.000032901763916,
|
| 513 |
+
"learning_rate": 2.147632311977716e-05,
|
| 514 |
+
"loss": 0.2187,
|
| 515 |
+
"step": 710
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.8015028177833439,
|
| 519 |
+
"grad_norm": 2.2018449306488037,
|
| 520 |
+
"learning_rate": 2.1309192200557104e-05,
|
| 521 |
+
"loss": 0.2366,
|
| 522 |
+
"step": 720
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 1.8265497808390734,
|
| 526 |
+
"grad_norm": 2.023028612136841,
|
| 527 |
+
"learning_rate": 2.114206128133705e-05,
|
| 528 |
+
"loss": 0.2389,
|
| 529 |
+
"step": 730
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 1.8515967438948029,
|
| 533 |
+
"grad_norm": 2.1953046321868896,
|
| 534 |
+
"learning_rate": 2.0974930362116993e-05,
|
| 535 |
+
"loss": 0.2316,
|
| 536 |
+
"step": 740
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.8766437069505324,
|
| 540 |
+
"grad_norm": 2.016090154647827,
|
| 541 |
+
"learning_rate": 2.0807799442896935e-05,
|
| 542 |
+
"loss": 0.2276,
|
| 543 |
+
"step": 750
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 1.9016906700062617,
|
| 547 |
+
"grad_norm": 4.887736797332764,
|
| 548 |
+
"learning_rate": 2.0640668523676883e-05,
|
| 549 |
+
"loss": 0.1759,
|
| 550 |
+
"step": 760
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 1.9267376330619912,
|
| 554 |
+
"grad_norm": 3.064952850341797,
|
| 555 |
+
"learning_rate": 2.0473537604456827e-05,
|
| 556 |
+
"loss": 0.1821,
|
| 557 |
+
"step": 770
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.9517845961177207,
|
| 561 |
+
"grad_norm": 3.6720850467681885,
|
| 562 |
+
"learning_rate": 2.0306406685236768e-05,
|
| 563 |
+
"loss": 0.2252,
|
| 564 |
+
"step": 780
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 1.9768315591734502,
|
| 568 |
+
"grad_norm": 4.621647357940674,
|
| 569 |
+
"learning_rate": 2.0139275766016713e-05,
|
| 570 |
+
"loss": 0.2204,
|
| 571 |
+
"step": 790
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 2.0,
|
| 575 |
+
"grad_norm": 0.6409714221954346,
|
| 576 |
+
"learning_rate": 1.997214484679666e-05,
|
| 577 |
+
"loss": 0.1701,
|
| 578 |
+
"step": 800
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 2.0,
|
| 582 |
+
"eval_accuracy": 0.9748455428067079,
|
| 583 |
+
"eval_loss": 0.1344912201166153,
|
| 584 |
+
"eval_runtime": 4.4272,
|
| 585 |
+
"eval_samples_per_second": 1535.524,
|
| 586 |
+
"eval_steps_per_second": 48.112,
|
| 587 |
+
"step": 800
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 2.0250469630557295,
|
| 591 |
+
"grad_norm": 2.7449936866760254,
|
| 592 |
+
"learning_rate": 1.9805013927576602e-05,
|
| 593 |
+
"loss": 0.2049,
|
| 594 |
+
"step": 810
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 2.050093926111459,
|
| 598 |
+
"grad_norm": 4.835960388183594,
|
| 599 |
+
"learning_rate": 1.9637883008356546e-05,
|
| 600 |
+
"loss": 0.1921,
|
| 601 |
+
"step": 820
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 2.0751408891671885,
|
| 605 |
+
"grad_norm": 3.658618211746216,
|
| 606 |
+
"learning_rate": 1.947075208913649e-05,
|
| 607 |
+
"loss": 0.1837,
|
| 608 |
+
"step": 830
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 2.100187852222918,
|
| 612 |
+
"grad_norm": 3.6431124210357666,
|
| 613 |
+
"learning_rate": 1.9303621169916436e-05,
|
| 614 |
+
"loss": 0.1865,
|
| 615 |
+
"step": 840
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 2.1252348152786475,
|
| 619 |
+
"grad_norm": 4.338460922241211,
|
| 620 |
+
"learning_rate": 1.913649025069638e-05,
|
| 621 |
+
"loss": 0.192,
|
| 622 |
+
"step": 850
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 2.150281778334377,
|
| 626 |
+
"grad_norm": 2.3245649337768555,
|
| 627 |
+
"learning_rate": 1.8969359331476325e-05,
|
| 628 |
+
"loss": 0.1676,
|
| 629 |
+
"step": 860
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 2.1753287413901066,
|
| 633 |
+
"grad_norm": 4.586086750030518,
|
| 634 |
+
"learning_rate": 1.8802228412256266e-05,
|
| 635 |
+
"loss": 0.1865,
|
| 636 |
+
"step": 870
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 2.200375704445836,
|
| 640 |
+
"grad_norm": 3.0299103260040283,
|
| 641 |
+
"learning_rate": 1.863509749303621e-05,
|
| 642 |
+
"loss": 0.1936,
|
| 643 |
+
"step": 880
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 2.2254226675015656,
|
| 647 |
+
"grad_norm": 1.2470777034759521,
|
| 648 |
+
"learning_rate": 1.846796657381616e-05,
|
| 649 |
+
"loss": 0.1919,
|
| 650 |
+
"step": 890
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 2.250469630557295,
|
| 654 |
+
"grad_norm": 2.023383378982544,
|
| 655 |
+
"learning_rate": 1.83008356545961e-05,
|
| 656 |
+
"loss": 0.1863,
|
| 657 |
+
"step": 900
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 2.2755165936130246,
|
| 661 |
+
"grad_norm": 3.2981998920440674,
|
| 662 |
+
"learning_rate": 1.8133704735376044e-05,
|
| 663 |
+
"loss": 0.2287,
|
| 664 |
+
"step": 910
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 2.300563556668754,
|
| 668 |
+
"grad_norm": 1.7662588357925415,
|
| 669 |
+
"learning_rate": 1.796657381615599e-05,
|
| 670 |
+
"loss": 0.2036,
|
| 671 |
+
"step": 920
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 2.325610519724483,
|
| 675 |
+
"grad_norm": 2.323029041290283,
|
| 676 |
+
"learning_rate": 1.7799442896935933e-05,
|
| 677 |
+
"loss": 0.1698,
|
| 678 |
+
"step": 930
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 2.350657482780213,
|
| 682 |
+
"grad_norm": 3.8338754177093506,
|
| 683 |
+
"learning_rate": 1.7632311977715878e-05,
|
| 684 |
+
"loss": 0.1699,
|
| 685 |
+
"step": 940
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 2.375704445835942,
|
| 689 |
+
"grad_norm": 1.8302233219146729,
|
| 690 |
+
"learning_rate": 1.7465181058495822e-05,
|
| 691 |
+
"loss": 0.1718,
|
| 692 |
+
"step": 950
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.4007514088916717,
|
| 696 |
+
"grad_norm": 3.018829584121704,
|
| 697 |
+
"learning_rate": 1.7298050139275764e-05,
|
| 698 |
+
"loss": 0.18,
|
| 699 |
+
"step": 960
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 2.425798371947401,
|
| 703 |
+
"grad_norm": 1.8487874269485474,
|
| 704 |
+
"learning_rate": 1.713091922005571e-05,
|
| 705 |
+
"loss": 0.1827,
|
| 706 |
+
"step": 970
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 2.4508453350031307,
|
| 710 |
+
"grad_norm": 3.3210318088531494,
|
| 711 |
+
"learning_rate": 1.6963788300835656e-05,
|
| 712 |
+
"loss": 0.1662,
|
| 713 |
+
"step": 980
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 2.4758922980588602,
|
| 717 |
+
"grad_norm": 4.057735919952393,
|
| 718 |
+
"learning_rate": 1.6796657381615597e-05,
|
| 719 |
+
"loss": 0.1576,
|
| 720 |
+
"step": 990
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.5009392611145898,
|
| 724 |
+
"grad_norm": 2.257659435272217,
|
| 725 |
+
"learning_rate": 1.6629526462395542e-05,
|
| 726 |
+
"loss": 0.1733,
|
| 727 |
+
"step": 1000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 2.5259862241703193,
|
| 731 |
+
"grad_norm": 4.195764064788818,
|
| 732 |
+
"learning_rate": 1.646239554317549e-05,
|
| 733 |
+
"loss": 0.1787,
|
| 734 |
+
"step": 1010
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.5510331872260488,
|
| 738 |
+
"grad_norm": 2.0992043018341064,
|
| 739 |
+
"learning_rate": 1.6295264623955434e-05,
|
| 740 |
+
"loss": 0.1893,
|
| 741 |
+
"step": 1020
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 2.5760801502817783,
|
| 745 |
+
"grad_norm": 4.466301441192627,
|
| 746 |
+
"learning_rate": 1.6128133704735375e-05,
|
| 747 |
+
"loss": 0.148,
|
| 748 |
+
"step": 1030
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 2.601127113337508,
|
| 752 |
+
"grad_norm": 4.416186809539795,
|
| 753 |
+
"learning_rate": 1.596100278551532e-05,
|
| 754 |
+
"loss": 0.1642,
|
| 755 |
+
"step": 1040
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 2.6261740763932373,
|
| 759 |
+
"grad_norm": 1.4752304553985596,
|
| 760 |
+
"learning_rate": 1.5793871866295268e-05,
|
| 761 |
+
"loss": 0.165,
|
| 762 |
+
"step": 1050
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 2.651221039448967,
|
| 766 |
+
"grad_norm": 3.157634735107422,
|
| 767 |
+
"learning_rate": 1.562674094707521e-05,
|
| 768 |
+
"loss": 0.1645,
|
| 769 |
+
"step": 1060
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 2.6762680025046963,
|
| 773 |
+
"grad_norm": 2.568713426589966,
|
| 774 |
+
"learning_rate": 1.5459610027855154e-05,
|
| 775 |
+
"loss": 0.1806,
|
| 776 |
+
"step": 1070
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.701314965560426,
|
| 780 |
+
"grad_norm": 3.1504478454589844,
|
| 781 |
+
"learning_rate": 1.5292479108635098e-05,
|
| 782 |
+
"loss": 0.1768,
|
| 783 |
+
"step": 1080
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 2.7263619286161553,
|
| 787 |
+
"grad_norm": 1.673834204673767,
|
| 788 |
+
"learning_rate": 1.5125348189415043e-05,
|
| 789 |
+
"loss": 0.1526,
|
| 790 |
+
"step": 1090
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 2.751408891671885,
|
| 794 |
+
"grad_norm": 2.3160338401794434,
|
| 795 |
+
"learning_rate": 1.4958217270194987e-05,
|
| 796 |
+
"loss": 0.1309,
|
| 797 |
+
"step": 1100
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 2.7764558547276144,
|
| 801 |
+
"grad_norm": 4.241952419281006,
|
| 802 |
+
"learning_rate": 1.479108635097493e-05,
|
| 803 |
+
"loss": 0.1992,
|
| 804 |
+
"step": 1110
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 2.801502817783344,
|
| 808 |
+
"grad_norm": 2.4908902645111084,
|
| 809 |
+
"learning_rate": 1.4623955431754876e-05,
|
| 810 |
+
"loss": 0.1632,
|
| 811 |
+
"step": 1120
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 2.8265497808390734,
|
| 815 |
+
"grad_norm": 4.086286544799805,
|
| 816 |
+
"learning_rate": 1.445682451253482e-05,
|
| 817 |
+
"loss": 0.1805,
|
| 818 |
+
"step": 1130
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.851596743894803,
|
| 822 |
+
"grad_norm": 1.9868454933166504,
|
| 823 |
+
"learning_rate": 1.4289693593314764e-05,
|
| 824 |
+
"loss": 0.1579,
|
| 825 |
+
"step": 1140
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 2.8766437069505324,
|
| 829 |
+
"grad_norm": 4.5509934425354,
|
| 830 |
+
"learning_rate": 1.4122562674094708e-05,
|
| 831 |
+
"loss": 0.1892,
|
| 832 |
+
"step": 1150
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 2.9016906700062615,
|
| 836 |
+
"grad_norm": 2.554227828979492,
|
| 837 |
+
"learning_rate": 1.3955431754874653e-05,
|
| 838 |
+
"loss": 0.185,
|
| 839 |
+
"step": 1160
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 2.9267376330619914,
|
| 843 |
+
"grad_norm": 3.8383281230926514,
|
| 844 |
+
"learning_rate": 1.3788300835654596e-05,
|
| 845 |
+
"loss": 0.1869,
|
| 846 |
+
"step": 1170
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 2.9517845961177205,
|
| 850 |
+
"grad_norm": 2.4846155643463135,
|
| 851 |
+
"learning_rate": 1.362116991643454e-05,
|
| 852 |
+
"loss": 0.1728,
|
| 853 |
+
"step": 1180
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 2.9768315591734504,
|
| 857 |
+
"grad_norm": 2.692121982574463,
|
| 858 |
+
"learning_rate": 1.3454038997214485e-05,
|
| 859 |
+
"loss": 0.1492,
|
| 860 |
+
"step": 1190
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 3.0,
|
| 864 |
+
"grad_norm": 1.1535016298294067,
|
| 865 |
+
"learning_rate": 1.3286908077994428e-05,
|
| 866 |
+
"loss": 0.1905,
|
| 867 |
+
"step": 1200
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 3.0,
|
| 871 |
+
"eval_accuracy": 0.9785230950279494,
|
| 872 |
+
"eval_loss": 0.100662961602211,
|
| 873 |
+
"eval_runtime": 4.2957,
|
| 874 |
+
"eval_samples_per_second": 1582.527,
|
| 875 |
+
"eval_steps_per_second": 49.585,
|
| 876 |
+
"step": 1200
|
| 877 |
+
},
|
| 878 |
+
{
|
| 879 |
+
"epoch": 3.0250469630557295,
|
| 880 |
+
"grad_norm": 2.339324474334717,
|
| 881 |
+
"learning_rate": 1.3119777158774374e-05,
|
| 882 |
+
"loss": 0.1671,
|
| 883 |
+
"step": 1210
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 3.050093926111459,
|
| 887 |
+
"grad_norm": 1.8304985761642456,
|
| 888 |
+
"learning_rate": 1.2952646239554317e-05,
|
| 889 |
+
"loss": 0.1595,
|
| 890 |
+
"step": 1220
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 3.0751408891671885,
|
| 894 |
+
"grad_norm": 1.7035090923309326,
|
| 895 |
+
"learning_rate": 1.2785515320334262e-05,
|
| 896 |
+
"loss": 0.1464,
|
| 897 |
+
"step": 1230
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 3.100187852222918,
|
| 901 |
+
"grad_norm": 2.2386112213134766,
|
| 902 |
+
"learning_rate": 1.2618384401114206e-05,
|
| 903 |
+
"loss": 0.1828,
|
| 904 |
+
"step": 1240
|
| 905 |
+
},
|
| 906 |
+
{
|
| 907 |
+
"epoch": 3.1252348152786475,
|
| 908 |
+
"grad_norm": 2.6203246116638184,
|
| 909 |
+
"learning_rate": 1.245125348189415e-05,
|
| 910 |
+
"loss": 0.1709,
|
| 911 |
+
"step": 1250
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"epoch": 3.150281778334377,
|
| 915 |
+
"grad_norm": 3.4411773681640625,
|
| 916 |
+
"learning_rate": 1.2284122562674095e-05,
|
| 917 |
+
"loss": 0.1696,
|
| 918 |
+
"step": 1260
|
| 919 |
+
},
|
| 920 |
+
{
|
| 921 |
+
"epoch": 3.1753287413901066,
|
| 922 |
+
"grad_norm": 1.443657636642456,
|
| 923 |
+
"learning_rate": 1.211699164345404e-05,
|
| 924 |
+
"loss": 0.1694,
|
| 925 |
+
"step": 1270
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 3.200375704445836,
|
| 929 |
+
"grad_norm": 3.1852123737335205,
|
| 930 |
+
"learning_rate": 1.1949860724233983e-05,
|
| 931 |
+
"loss": 0.1472,
|
| 932 |
+
"step": 1280
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 3.2254226675015656,
|
| 936 |
+
"grad_norm": 3.547802686691284,
|
| 937 |
+
"learning_rate": 1.1782729805013929e-05,
|
| 938 |
+
"loss": 0.1367,
|
| 939 |
+
"step": 1290
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 3.250469630557295,
|
| 943 |
+
"grad_norm": 2.098524332046509,
|
| 944 |
+
"learning_rate": 1.1615598885793872e-05,
|
| 945 |
+
"loss": 0.1674,
|
| 946 |
+
"step": 1300
|
| 947 |
+
},
|
| 948 |
+
{
|
| 949 |
+
"epoch": 3.2755165936130246,
|
| 950 |
+
"grad_norm": 2.3405473232269287,
|
| 951 |
+
"learning_rate": 1.1448467966573816e-05,
|
| 952 |
+
"loss": 0.1233,
|
| 953 |
+
"step": 1310
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 3.300563556668754,
|
| 957 |
+
"grad_norm": 2.582063913345337,
|
| 958 |
+
"learning_rate": 1.1281337047353761e-05,
|
| 959 |
+
"loss": 0.1542,
|
| 960 |
+
"step": 1320
|
| 961 |
+
},
|
| 962 |
+
{
|
| 963 |
+
"epoch": 3.325610519724483,
|
| 964 |
+
"grad_norm": 2.138692855834961,
|
| 965 |
+
"learning_rate": 1.1114206128133705e-05,
|
| 966 |
+
"loss": 0.1583,
|
| 967 |
+
"step": 1330
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 3.350657482780213,
|
| 971 |
+
"grad_norm": 2.5060031414031982,
|
| 972 |
+
"learning_rate": 1.0947075208913648e-05,
|
| 973 |
+
"loss": 0.1309,
|
| 974 |
+
"step": 1340
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 3.375704445835942,
|
| 978 |
+
"grad_norm": 5.018398761749268,
|
| 979 |
+
"learning_rate": 1.0779944289693595e-05,
|
| 980 |
+
"loss": 0.158,
|
| 981 |
+
"step": 1350
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 3.4007514088916717,
|
| 985 |
+
"grad_norm": 3.4569902420043945,
|
| 986 |
+
"learning_rate": 1.0612813370473537e-05,
|
| 987 |
+
"loss": 0.1576,
|
| 988 |
+
"step": 1360
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"epoch": 3.425798371947401,
|
| 992 |
+
"grad_norm": 2.4191341400146484,
|
| 993 |
+
"learning_rate": 1.0445682451253482e-05,
|
| 994 |
+
"loss": 0.1363,
|
| 995 |
+
"step": 1370
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 3.4508453350031307,
|
| 999 |
+
"grad_norm": 1.7096883058547974,
|
| 1000 |
+
"learning_rate": 1.0278551532033427e-05,
|
| 1001 |
+
"loss": 0.1741,
|
| 1002 |
+
"step": 1380
|
| 1003 |
+
},
|
| 1004 |
+
{
|
| 1005 |
+
"epoch": 3.4758922980588602,
|
| 1006 |
+
"grad_norm": 1.6938859224319458,
|
| 1007 |
+
"learning_rate": 1.0111420612813371e-05,
|
| 1008 |
+
"loss": 0.1454,
|
| 1009 |
+
"step": 1390
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 3.5009392611145898,
|
| 1013 |
+
"grad_norm": 0.8238410353660583,
|
| 1014 |
+
"learning_rate": 9.944289693593314e-06,
|
| 1015 |
+
"loss": 0.1603,
|
| 1016 |
+
"step": 1400
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 3.5259862241703193,
|
| 1020 |
+
"grad_norm": 2.195200204849243,
|
| 1021 |
+
"learning_rate": 9.77715877437326e-06,
|
| 1022 |
+
"loss": 0.1471,
|
| 1023 |
+
"step": 1410
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 3.5510331872260488,
|
| 1027 |
+
"grad_norm": 1.5268114805221558,
|
| 1028 |
+
"learning_rate": 9.610027855153203e-06,
|
| 1029 |
+
"loss": 0.1308,
|
| 1030 |
+
"step": 1420
|
| 1031 |
+
},
|
| 1032 |
+
{
|
| 1033 |
+
"epoch": 3.5760801502817783,
|
| 1034 |
+
"grad_norm": 3.5758233070373535,
|
| 1035 |
+
"learning_rate": 9.44289693593315e-06,
|
| 1036 |
+
"loss": 0.1713,
|
| 1037 |
+
"step": 1430
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 3.601127113337508,
|
| 1041 |
+
"grad_norm": 2.7455337047576904,
|
| 1042 |
+
"learning_rate": 9.275766016713092e-06,
|
| 1043 |
+
"loss": 0.13,
|
| 1044 |
+
"step": 1440
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 3.6261740763932373,
|
| 1048 |
+
"grad_norm": 2.4125545024871826,
|
| 1049 |
+
"learning_rate": 9.108635097493037e-06,
|
| 1050 |
+
"loss": 0.1444,
|
| 1051 |
+
"step": 1450
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 3.651221039448967,
|
| 1055 |
+
"grad_norm": 4.1105852127075195,
|
| 1056 |
+
"learning_rate": 8.941504178272981e-06,
|
| 1057 |
+
"loss": 0.1581,
|
| 1058 |
+
"step": 1460
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 3.6762680025046963,
|
| 1062 |
+
"grad_norm": 1.781172513961792,
|
| 1063 |
+
"learning_rate": 8.774373259052926e-06,
|
| 1064 |
+
"loss": 0.1413,
|
| 1065 |
+
"step": 1470
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 3.701314965560426,
|
| 1069 |
+
"grad_norm": 0.9671861529350281,
|
| 1070 |
+
"learning_rate": 8.607242339832869e-06,
|
| 1071 |
+
"loss": 0.155,
|
| 1072 |
+
"step": 1480
|
| 1073 |
+
},
|
| 1074 |
+
{
|
| 1075 |
+
"epoch": 3.7263619286161553,
|
| 1076 |
+
"grad_norm": 3.277911424636841,
|
| 1077 |
+
"learning_rate": 8.440111420612815e-06,
|
| 1078 |
+
"loss": 0.1244,
|
| 1079 |
+
"step": 1490
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.751408891671885,
|
| 1083 |
+
"grad_norm": 2.6499757766723633,
|
| 1084 |
+
"learning_rate": 8.272980501392758e-06,
|
| 1085 |
+
"loss": 0.1462,
|
| 1086 |
+
"step": 1500
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 3.7764558547276144,
|
| 1090 |
+
"grad_norm": 1.484277367591858,
|
| 1091 |
+
"learning_rate": 8.1058495821727e-06,
|
| 1092 |
+
"loss": 0.1522,
|
| 1093 |
+
"step": 1510
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 3.801502817783344,
|
| 1097 |
+
"grad_norm": 3.2849831581115723,
|
| 1098 |
+
"learning_rate": 7.938718662952647e-06,
|
| 1099 |
+
"loss": 0.1485,
|
| 1100 |
+
"step": 1520
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 3.8265497808390734,
|
| 1104 |
+
"grad_norm": 2.0737528800964355,
|
| 1105 |
+
"learning_rate": 7.77158774373259e-06,
|
| 1106 |
+
"loss": 0.1243,
|
| 1107 |
+
"step": 1530
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 3.851596743894803,
|
| 1111 |
+
"grad_norm": 2.977562665939331,
|
| 1112 |
+
"learning_rate": 7.604456824512535e-06,
|
| 1113 |
+
"loss": 0.1266,
|
| 1114 |
+
"step": 1540
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 3.8766437069505324,
|
| 1118 |
+
"grad_norm": 2.5057878494262695,
|
| 1119 |
+
"learning_rate": 7.43732590529248e-06,
|
| 1120 |
+
"loss": 0.1349,
|
| 1121 |
+
"step": 1550
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 3.9016906700062615,
|
| 1125 |
+
"grad_norm": 2.0009357929229736,
|
| 1126 |
+
"learning_rate": 7.2701949860724235e-06,
|
| 1127 |
+
"loss": 0.1296,
|
| 1128 |
+
"step": 1560
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 3.9267376330619914,
|
| 1132 |
+
"grad_norm": 3.119765043258667,
|
| 1133 |
+
"learning_rate": 7.103064066852368e-06,
|
| 1134 |
+
"loss": 0.1308,
|
| 1135 |
+
"step": 1570
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 3.9517845961177205,
|
| 1139 |
+
"grad_norm": 2.181718111038208,
|
| 1140 |
+
"learning_rate": 6.935933147632313e-06,
|
| 1141 |
+
"loss": 0.1654,
|
| 1142 |
+
"step": 1580
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 3.9768315591734504,
|
| 1146 |
+
"grad_norm": 3.2763113975524902,
|
| 1147 |
+
"learning_rate": 6.768802228412256e-06,
|
| 1148 |
+
"loss": 0.1541,
|
| 1149 |
+
"step": 1590
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 4.0,
|
| 1153 |
+
"grad_norm": 3.0005788803100586,
|
| 1154 |
+
"learning_rate": 6.601671309192201e-06,
|
| 1155 |
+
"loss": 0.1365,
|
| 1156 |
+
"step": 1600
|
| 1157 |
+
},
|
| 1158 |
+
{
|
| 1159 |
+
"epoch": 4.0,
|
| 1160 |
+
"eval_accuracy": 0.980876728449544,
|
| 1161 |
+
"eval_loss": 0.08496350795030594,
|
| 1162 |
+
"eval_runtime": 4.414,
|
| 1163 |
+
"eval_samples_per_second": 1540.114,
|
| 1164 |
+
"eval_steps_per_second": 48.256,
|
| 1165 |
+
"step": 1600
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 4.025046963055729,
|
| 1169 |
+
"grad_norm": 3.285486936569214,
|
| 1170 |
+
"learning_rate": 6.4345403899721455e-06,
|
| 1171 |
+
"loss": 0.128,
|
| 1172 |
+
"step": 1610
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 4.050093926111459,
|
| 1176 |
+
"grad_norm": 2.414086103439331,
|
| 1177 |
+
"learning_rate": 6.267409470752089e-06,
|
| 1178 |
+
"loss": 0.1321,
|
| 1179 |
+
"step": 1620
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 4.075140889167188,
|
| 1183 |
+
"grad_norm": 2.2890965938568115,
|
| 1184 |
+
"learning_rate": 6.100278551532034e-06,
|
| 1185 |
+
"loss": 0.1298,
|
| 1186 |
+
"step": 1630
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 4.100187852222918,
|
| 1190 |
+
"grad_norm": 1.8132153749465942,
|
| 1191 |
+
"learning_rate": 5.933147632311978e-06,
|
| 1192 |
+
"loss": 0.1539,
|
| 1193 |
+
"step": 1640
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 4.125234815278647,
|
| 1197 |
+
"grad_norm": 2.8671135902404785,
|
| 1198 |
+
"learning_rate": 5.766016713091923e-06,
|
| 1199 |
+
"loss": 0.1302,
|
| 1200 |
+
"step": 1650
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 4.150281778334377,
|
| 1204 |
+
"grad_norm": 2.470564365386963,
|
| 1205 |
+
"learning_rate": 5.598885793871867e-06,
|
| 1206 |
+
"loss": 0.1612,
|
| 1207 |
+
"step": 1660
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 4.175328741390106,
|
| 1211 |
+
"grad_norm": 2.1133806705474854,
|
| 1212 |
+
"learning_rate": 5.43175487465181e-06,
|
| 1213 |
+
"loss": 0.148,
|
| 1214 |
+
"step": 1670
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 4.200375704445836,
|
| 1218 |
+
"grad_norm": 1.391147255897522,
|
| 1219 |
+
"learning_rate": 5.264623955431755e-06,
|
| 1220 |
+
"loss": 0.1455,
|
| 1221 |
+
"step": 1680
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 4.225422667501565,
|
| 1225 |
+
"grad_norm": 2.964405059814453,
|
| 1226 |
+
"learning_rate": 5.0974930362116986e-06,
|
| 1227 |
+
"loss": 0.1244,
|
| 1228 |
+
"step": 1690
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 4.250469630557295,
|
| 1232 |
+
"grad_norm": 3.051886796951294,
|
| 1233 |
+
"learning_rate": 4.930362116991643e-06,
|
| 1234 |
+
"loss": 0.1544,
|
| 1235 |
+
"step": 1700
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 4.275516593613024,
|
| 1239 |
+
"grad_norm": 2.3174211978912354,
|
| 1240 |
+
"learning_rate": 4.763231197771588e-06,
|
| 1241 |
+
"loss": 0.1179,
|
| 1242 |
+
"step": 1710
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 4.300563556668754,
|
| 1246 |
+
"grad_norm": 2.169579267501831,
|
| 1247 |
+
"learning_rate": 4.596100278551532e-06,
|
| 1248 |
+
"loss": 0.1197,
|
| 1249 |
+
"step": 1720
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 4.325610519724483,
|
| 1253 |
+
"grad_norm": 4.180757999420166,
|
| 1254 |
+
"learning_rate": 4.428969359331476e-06,
|
| 1255 |
+
"loss": 0.1147,
|
| 1256 |
+
"step": 1730
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 4.350657482780213,
|
| 1260 |
+
"grad_norm": 2.835193634033203,
|
| 1261 |
+
"learning_rate": 4.2618384401114205e-06,
|
| 1262 |
+
"loss": 0.1199,
|
| 1263 |
+
"step": 1740
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 4.375704445835942,
|
| 1267 |
+
"grad_norm": 2.561140298843384,
|
| 1268 |
+
"learning_rate": 4.094707520891365e-06,
|
| 1269 |
+
"loss": 0.1369,
|
| 1270 |
+
"step": 1750
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 4.400751408891672,
|
| 1274 |
+
"grad_norm": 1.964847445487976,
|
| 1275 |
+
"learning_rate": 3.927576601671309e-06,
|
| 1276 |
+
"loss": 0.1422,
|
| 1277 |
+
"step": 1760
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 4.425798371947401,
|
| 1281 |
+
"grad_norm": 3.416602611541748,
|
| 1282 |
+
"learning_rate": 3.7604456824512533e-06,
|
| 1283 |
+
"loss": 0.1258,
|
| 1284 |
+
"step": 1770
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 4.450845335003131,
|
| 1288 |
+
"grad_norm": 2.517287254333496,
|
| 1289 |
+
"learning_rate": 3.593314763231198e-06,
|
| 1290 |
+
"loss": 0.1334,
|
| 1291 |
+
"step": 1780
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 4.47589229805886,
|
| 1295 |
+
"grad_norm": 1.725051760673523,
|
| 1296 |
+
"learning_rate": 3.426183844011142e-06,
|
| 1297 |
+
"loss": 0.116,
|
| 1298 |
+
"step": 1790
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 4.50093926111459,
|
| 1302 |
+
"grad_norm": 3.366607666015625,
|
| 1303 |
+
"learning_rate": 3.259052924791086e-06,
|
| 1304 |
+
"loss": 0.1166,
|
| 1305 |
+
"step": 1800
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 4.525986224170319,
|
| 1309 |
+
"grad_norm": 4.129775524139404,
|
| 1310 |
+
"learning_rate": 3.0919220055710307e-06,
|
| 1311 |
+
"loss": 0.1238,
|
| 1312 |
+
"step": 1810
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 4.551033187226049,
|
| 1316 |
+
"grad_norm": 1.5678294897079468,
|
| 1317 |
+
"learning_rate": 2.924791086350975e-06,
|
| 1318 |
+
"loss": 0.1409,
|
| 1319 |
+
"step": 1820
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 4.576080150281778,
|
| 1323 |
+
"grad_norm": 2.8961803913116455,
|
| 1324 |
+
"learning_rate": 2.7576601671309194e-06,
|
| 1325 |
+
"loss": 0.1002,
|
| 1326 |
+
"step": 1830
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 4.601127113337508,
|
| 1330 |
+
"grad_norm": 1.3110554218292236,
|
| 1331 |
+
"learning_rate": 2.5905292479108636e-06,
|
| 1332 |
+
"loss": 0.0937,
|
| 1333 |
+
"step": 1840
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 4.626174076393237,
|
| 1337 |
+
"grad_norm": 3.0743277072906494,
|
| 1338 |
+
"learning_rate": 2.4233983286908077e-06,
|
| 1339 |
+
"loss": 0.1362,
|
| 1340 |
+
"step": 1850
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 4.651221039448966,
|
| 1344 |
+
"grad_norm": 3.078073263168335,
|
| 1345 |
+
"learning_rate": 2.2562674094707523e-06,
|
| 1346 |
+
"loss": 0.1542,
|
| 1347 |
+
"step": 1860
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 4.676268002504696,
|
| 1351 |
+
"grad_norm": 1.4623353481292725,
|
| 1352 |
+
"learning_rate": 2.0891364902506964e-06,
|
| 1353 |
+
"loss": 0.1279,
|
| 1354 |
+
"step": 1870
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 4.701314965560426,
|
| 1358 |
+
"grad_norm": 0.6655158996582031,
|
| 1359 |
+
"learning_rate": 1.922005571030641e-06,
|
| 1360 |
+
"loss": 0.0937,
|
| 1361 |
+
"step": 1880
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 4.726361928616155,
|
| 1365 |
+
"grad_norm": 2.3055014610290527,
|
| 1366 |
+
"learning_rate": 1.7548746518105849e-06,
|
| 1367 |
+
"loss": 0.1278,
|
| 1368 |
+
"step": 1890
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 4.751408891671884,
|
| 1372 |
+
"grad_norm": 2.902233839035034,
|
| 1373 |
+
"learning_rate": 1.5877437325905292e-06,
|
| 1374 |
+
"loss": 0.1067,
|
| 1375 |
+
"step": 1900
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 4.776455854727614,
|
| 1379 |
+
"grad_norm": 1.5290515422821045,
|
| 1380 |
+
"learning_rate": 1.4206128133704736e-06,
|
| 1381 |
+
"loss": 0.1299,
|
| 1382 |
+
"step": 1910
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 4.801502817783343,
|
| 1386 |
+
"grad_norm": 2.1846730709075928,
|
| 1387 |
+
"learning_rate": 1.253481894150418e-06,
|
| 1388 |
+
"loss": 0.0939,
|
| 1389 |
+
"step": 1920
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 4.826549780839073,
|
| 1393 |
+
"grad_norm": 1.173969030380249,
|
| 1394 |
+
"learning_rate": 1.0863509749303623e-06,
|
| 1395 |
+
"loss": 0.1181,
|
| 1396 |
+
"step": 1930
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 4.851596743894802,
|
| 1400 |
+
"grad_norm": 4.177393913269043,
|
| 1401 |
+
"learning_rate": 9.192200557103064e-07,
|
| 1402 |
+
"loss": 0.1373,
|
| 1403 |
+
"step": 1940
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 4.876643706950532,
|
| 1407 |
+
"grad_norm": 1.459058403968811,
|
| 1408 |
+
"learning_rate": 7.520891364902508e-07,
|
| 1409 |
+
"loss": 0.1161,
|
| 1410 |
+
"step": 1950
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 4.9016906700062615,
|
| 1414 |
+
"grad_norm": 0.865409255027771,
|
| 1415 |
+
"learning_rate": 5.84958217270195e-07,
|
| 1416 |
+
"loss": 0.1233,
|
| 1417 |
+
"step": 1960
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 4.926737633061991,
|
| 1421 |
+
"grad_norm": 1.5550708770751953,
|
| 1422 |
+
"learning_rate": 4.178272980501393e-07,
|
| 1423 |
+
"loss": 0.1172,
|
| 1424 |
+
"step": 1970
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 4.9517845961177205,
|
| 1428 |
+
"grad_norm": 1.4279484748840332,
|
| 1429 |
+
"learning_rate": 2.506963788300836e-07,
|
| 1430 |
+
"loss": 0.1313,
|
| 1431 |
+
"step": 1980
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 4.97683155917345,
|
| 1435 |
+
"grad_norm": 1.5015976428985596,
|
| 1436 |
+
"learning_rate": 8.356545961002785e-08,
|
| 1437 |
+
"loss": 0.1317,
|
| 1438 |
+
"step": 1990
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 4.989355040701315,
|
| 1442 |
+
"eval_accuracy": 0.9810238305383936,
|
| 1443 |
+
"eval_loss": 0.08548342436552048,
|
| 1444 |
+
"eval_runtime": 4.864,
|
| 1445 |
+
"eval_samples_per_second": 1397.623,
|
| 1446 |
+
"eval_steps_per_second": 43.791,
|
| 1447 |
+
"step": 1995
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 4.989355040701315,
|
| 1451 |
+
"step": 1995,
|
| 1452 |
+
"total_flos": 2.31453318852096e+18,
|
| 1453 |
+
"train_loss": 0.40246317323885467,
|
| 1454 |
+
"train_runtime": 627.2253,
|
| 1455 |
+
"train_samples_per_second": 407.302,
|
| 1456 |
+
"train_steps_per_second": 3.181
|
| 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": 2.31453318852096e+18,
|
| 1477 |
+
"train_batch_size": 32,
|
| 1478 |
+
"trial_name": null,
|
| 1479 |
+
"trial_params": null
|
| 1480 |
+
}
|