End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.25,
|
| 3 |
+
"eval_accuracy": 0.9558620689655173,
|
| 4 |
+
"eval_f1": 0.9562337987051189,
|
| 5 |
+
"eval_loss": 0.14939993619918823,
|
| 6 |
+
"eval_precision": 0.9567608190198278,
|
| 7 |
+
"eval_recall": 0.956465718392097,
|
| 8 |
+
"eval_runtime": 930.9063,
|
| 9 |
+
"eval_samples_per_second": 0.779,
|
| 10 |
+
"eval_steps_per_second": 0.196
|
| 11 |
+
}
|
runs/Apr30_22-59-34_blackhorse/events.out.tfevents.1714517431.blackhorse.5540.3
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:290d37102d1ed7890a79ba8e7c0d287dd45df80c5e7f1a28e13825791fd2c453
|
| 3 |
+
size 1032
|
test_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.25,
|
| 3 |
+
"eval_accuracy": 0.9558620689655173,
|
| 4 |
+
"eval_f1": 0.9562337987051189,
|
| 5 |
+
"eval_loss": 0.14939993619918823,
|
| 6 |
+
"eval_precision": 0.9567608190198278,
|
| 7 |
+
"eval_recall": 0.956465718392097,
|
| 8 |
+
"eval_runtime": 930.9063,
|
| 9 |
+
"eval_samples_per_second": 0.779,
|
| 10 |
+
"eval_steps_per_second": 0.196
|
| 11 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2069 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.9879931389365352,
|
| 3 |
+
"best_model_checkpoint": "videomae-base-finetuned-numbers-augmented/checkpoint-2816",
|
| 4 |
+
"epoch": 3.25,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2816,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0035511363636363635,
|
| 13 |
+
"grad_norm": 11.368436813354492,
|
| 14 |
+
"learning_rate": 1.7730496453900712e-06,
|
| 15 |
+
"loss": 2.4086,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.007102272727272727,
|
| 20 |
+
"grad_norm": 13.567573547363281,
|
| 21 |
+
"learning_rate": 3.5460992907801423e-06,
|
| 22 |
+
"loss": 2.4404,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01065340909090909,
|
| 27 |
+
"grad_norm": 10.261510848999023,
|
| 28 |
+
"learning_rate": 5.319148936170213e-06,
|
| 29 |
+
"loss": 2.4437,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.014204545454545454,
|
| 34 |
+
"grad_norm": 10.864927291870117,
|
| 35 |
+
"learning_rate": 7.092198581560285e-06,
|
| 36 |
+
"loss": 2.4537,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01775568181818182,
|
| 41 |
+
"grad_norm": 13.574213027954102,
|
| 42 |
+
"learning_rate": 8.865248226950355e-06,
|
| 43 |
+
"loss": 2.3604,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.02130681818181818,
|
| 48 |
+
"grad_norm": 12.087525367736816,
|
| 49 |
+
"learning_rate": 1.0638297872340426e-05,
|
| 50 |
+
"loss": 2.4028,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.024857954545454544,
|
| 55 |
+
"grad_norm": 16.888160705566406,
|
| 56 |
+
"learning_rate": 1.2411347517730498e-05,
|
| 57 |
+
"loss": 2.3285,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.028409090909090908,
|
| 62 |
+
"grad_norm": 11.21091365814209,
|
| 63 |
+
"learning_rate": 1.418439716312057e-05,
|
| 64 |
+
"loss": 2.4292,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03196022727272727,
|
| 69 |
+
"grad_norm": 12.686086654663086,
|
| 70 |
+
"learning_rate": 1.595744680851064e-05,
|
| 71 |
+
"loss": 2.3094,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.03551136363636364,
|
| 76 |
+
"grad_norm": 13.735329627990723,
|
| 77 |
+
"learning_rate": 1.773049645390071e-05,
|
| 78 |
+
"loss": 2.2478,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.0390625,
|
| 83 |
+
"grad_norm": 15.364631652832031,
|
| 84 |
+
"learning_rate": 1.950354609929078e-05,
|
| 85 |
+
"loss": 2.2576,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.04261363636363636,
|
| 90 |
+
"grad_norm": 17.786842346191406,
|
| 91 |
+
"learning_rate": 2.1276595744680852e-05,
|
| 92 |
+
"loss": 2.2132,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04616477272727273,
|
| 97 |
+
"grad_norm": 15.442058563232422,
|
| 98 |
+
"learning_rate": 2.3049645390070924e-05,
|
| 99 |
+
"loss": 2.1968,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.04971590909090909,
|
| 104 |
+
"grad_norm": 18.52874755859375,
|
| 105 |
+
"learning_rate": 2.4822695035460995e-05,
|
| 106 |
+
"loss": 1.9613,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.053267045454545456,
|
| 111 |
+
"grad_norm": 16.437816619873047,
|
| 112 |
+
"learning_rate": 2.6595744680851064e-05,
|
| 113 |
+
"loss": 2.055,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.056818181818181816,
|
| 118 |
+
"grad_norm": 16.25989532470703,
|
| 119 |
+
"learning_rate": 2.836879432624114e-05,
|
| 120 |
+
"loss": 2.2668,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.060369318181818184,
|
| 125 |
+
"grad_norm": 16.64414405822754,
|
| 126 |
+
"learning_rate": 3.0141843971631207e-05,
|
| 127 |
+
"loss": 2.0752,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.06392045454545454,
|
| 132 |
+
"grad_norm": 22.03922462463379,
|
| 133 |
+
"learning_rate": 3.191489361702128e-05,
|
| 134 |
+
"loss": 1.9947,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.06747159090909091,
|
| 139 |
+
"grad_norm": 20.077898025512695,
|
| 140 |
+
"learning_rate": 3.3687943262411347e-05,
|
| 141 |
+
"loss": 1.8888,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.07102272727272728,
|
| 146 |
+
"grad_norm": 17.019630432128906,
|
| 147 |
+
"learning_rate": 3.546099290780142e-05,
|
| 148 |
+
"loss": 1.8263,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.07457386363636363,
|
| 153 |
+
"grad_norm": 22.599655151367188,
|
| 154 |
+
"learning_rate": 3.723404255319149e-05,
|
| 155 |
+
"loss": 1.6843,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.078125,
|
| 160 |
+
"grad_norm": 17.443727493286133,
|
| 161 |
+
"learning_rate": 3.900709219858156e-05,
|
| 162 |
+
"loss": 1.7081,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.08167613636363637,
|
| 167 |
+
"grad_norm": 12.519112586975098,
|
| 168 |
+
"learning_rate": 4.078014184397163e-05,
|
| 169 |
+
"loss": 1.4262,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.08522727272727272,
|
| 174 |
+
"grad_norm": 21.003129959106445,
|
| 175 |
+
"learning_rate": 4.2553191489361704e-05,
|
| 176 |
+
"loss": 1.6321,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.08877840909090909,
|
| 181 |
+
"grad_norm": 9.726280212402344,
|
| 182 |
+
"learning_rate": 4.432624113475177e-05,
|
| 183 |
+
"loss": 1.6712,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.09232954545454546,
|
| 188 |
+
"grad_norm": 24.422882080078125,
|
| 189 |
+
"learning_rate": 4.609929078014185e-05,
|
| 190 |
+
"loss": 2.0967,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.09588068181818182,
|
| 195 |
+
"grad_norm": 11.92298698425293,
|
| 196 |
+
"learning_rate": 4.787234042553192e-05,
|
| 197 |
+
"loss": 1.7163,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.09943181818181818,
|
| 202 |
+
"grad_norm": 14.245928764343262,
|
| 203 |
+
"learning_rate": 4.964539007092199e-05,
|
| 204 |
+
"loss": 1.9799,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.10298295454545454,
|
| 209 |
+
"grad_norm": 14.423157691955566,
|
| 210 |
+
"learning_rate": 4.9842146803472775e-05,
|
| 211 |
+
"loss": 1.6602,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.10653409090909091,
|
| 216 |
+
"grad_norm": 11.99425983428955,
|
| 217 |
+
"learning_rate": 4.9644830307813736e-05,
|
| 218 |
+
"loss": 1.6844,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.11008522727272728,
|
| 223 |
+
"grad_norm": 13.40038776397705,
|
| 224 |
+
"learning_rate": 4.94475138121547e-05,
|
| 225 |
+
"loss": 1.6082,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.11363636363636363,
|
| 230 |
+
"grad_norm": 9.363014221191406,
|
| 231 |
+
"learning_rate": 4.925019731649566e-05,
|
| 232 |
+
"loss": 1.7191,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1171875,
|
| 237 |
+
"grad_norm": 16.138887405395508,
|
| 238 |
+
"learning_rate": 4.905288082083662e-05,
|
| 239 |
+
"loss": 1.621,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.12073863636363637,
|
| 244 |
+
"grad_norm": 13.552618980407715,
|
| 245 |
+
"learning_rate": 4.885556432517759e-05,
|
| 246 |
+
"loss": 1.4569,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.12428977272727272,
|
| 251 |
+
"grad_norm": 16.55213737487793,
|
| 252 |
+
"learning_rate": 4.865824782951855e-05,
|
| 253 |
+
"loss": 1.5886,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.1278409090909091,
|
| 258 |
+
"grad_norm": 4.656749248504639,
|
| 259 |
+
"learning_rate": 4.8460931333859514e-05,
|
| 260 |
+
"loss": 1.2231,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.13139204545454544,
|
| 265 |
+
"grad_norm": 9.768326759338379,
|
| 266 |
+
"learning_rate": 4.8263614838200475e-05,
|
| 267 |
+
"loss": 1.65,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.13494318181818182,
|
| 272 |
+
"grad_norm": 20.023557662963867,
|
| 273 |
+
"learning_rate": 4.806629834254144e-05,
|
| 274 |
+
"loss": 1.4166,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.13849431818181818,
|
| 279 |
+
"grad_norm": 18.180370330810547,
|
| 280 |
+
"learning_rate": 4.78689818468824e-05,
|
| 281 |
+
"loss": 1.1376,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.14204545454545456,
|
| 286 |
+
"grad_norm": 16.621734619140625,
|
| 287 |
+
"learning_rate": 4.767166535122337e-05,
|
| 288 |
+
"loss": 1.5926,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.1455965909090909,
|
| 293 |
+
"grad_norm": 17.98001480102539,
|
| 294 |
+
"learning_rate": 4.747434885556433e-05,
|
| 295 |
+
"loss": 1.3632,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.14914772727272727,
|
| 300 |
+
"grad_norm": 20.42987060546875,
|
| 301 |
+
"learning_rate": 4.727703235990529e-05,
|
| 302 |
+
"loss": 1.3245,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.15269886363636365,
|
| 307 |
+
"grad_norm": 11.056243896484375,
|
| 308 |
+
"learning_rate": 4.707971586424625e-05,
|
| 309 |
+
"loss": 1.2668,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.15625,
|
| 314 |
+
"grad_norm": 13.91126537322998,
|
| 315 |
+
"learning_rate": 4.6882399368587215e-05,
|
| 316 |
+
"loss": 1.0435,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.15980113636363635,
|
| 321 |
+
"grad_norm": 19.263099670410156,
|
| 322 |
+
"learning_rate": 4.6685082872928176e-05,
|
| 323 |
+
"loss": 1.3369,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.16335227272727273,
|
| 328 |
+
"grad_norm": 23.511869430541992,
|
| 329 |
+
"learning_rate": 4.6487766377269145e-05,
|
| 330 |
+
"loss": 1.1933,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1669034090909091,
|
| 335 |
+
"grad_norm": 14.228583335876465,
|
| 336 |
+
"learning_rate": 4.629044988161011e-05,
|
| 337 |
+
"loss": 1.591,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.17045454545454544,
|
| 342 |
+
"grad_norm": 19.068967819213867,
|
| 343 |
+
"learning_rate": 4.609313338595106e-05,
|
| 344 |
+
"loss": 0.9233,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.17400568181818182,
|
| 349 |
+
"grad_norm": 11.136780738830566,
|
| 350 |
+
"learning_rate": 4.589581689029203e-05,
|
| 351 |
+
"loss": 1.0833,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.17755681818181818,
|
| 356 |
+
"grad_norm": 10.600316047668457,
|
| 357 |
+
"learning_rate": 4.569850039463299e-05,
|
| 358 |
+
"loss": 1.1883,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.18110795454545456,
|
| 363 |
+
"grad_norm": 14.483977317810059,
|
| 364 |
+
"learning_rate": 4.550118389897396e-05,
|
| 365 |
+
"loss": 1.2224,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.1846590909090909,
|
| 370 |
+
"grad_norm": 20.28068733215332,
|
| 371 |
+
"learning_rate": 4.530386740331492e-05,
|
| 372 |
+
"loss": 0.9271,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.18821022727272727,
|
| 377 |
+
"grad_norm": 8.706443786621094,
|
| 378 |
+
"learning_rate": 4.5106550907655884e-05,
|
| 379 |
+
"loss": 0.876,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.19176136363636365,
|
| 384 |
+
"grad_norm": 24.446735382080078,
|
| 385 |
+
"learning_rate": 4.4909234411996846e-05,
|
| 386 |
+
"loss": 0.9443,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.1953125,
|
| 391 |
+
"grad_norm": 18.790678024291992,
|
| 392 |
+
"learning_rate": 4.471191791633781e-05,
|
| 393 |
+
"loss": 1.5365,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.19886363636363635,
|
| 398 |
+
"grad_norm": 12.385162353515625,
|
| 399 |
+
"learning_rate": 4.451460142067877e-05,
|
| 400 |
+
"loss": 1.1072,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.20241477272727273,
|
| 405 |
+
"grad_norm": 10.875085830688477,
|
| 406 |
+
"learning_rate": 4.431728492501974e-05,
|
| 407 |
+
"loss": 1.3267,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.2059659090909091,
|
| 412 |
+
"grad_norm": 23.155399322509766,
|
| 413 |
+
"learning_rate": 4.41199684293607e-05,
|
| 414 |
+
"loss": 1.0875,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.20951704545454544,
|
| 419 |
+
"grad_norm": 16.206789016723633,
|
| 420 |
+
"learning_rate": 4.392265193370166e-05,
|
| 421 |
+
"loss": 1.532,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.21306818181818182,
|
| 426 |
+
"grad_norm": 10.199226379394531,
|
| 427 |
+
"learning_rate": 4.372533543804262e-05,
|
| 428 |
+
"loss": 1.3542,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.21661931818181818,
|
| 433 |
+
"grad_norm": 24.252439498901367,
|
| 434 |
+
"learning_rate": 4.3528018942383585e-05,
|
| 435 |
+
"loss": 0.9932,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.22017045454545456,
|
| 440 |
+
"grad_norm": 21.493459701538086,
|
| 441 |
+
"learning_rate": 4.333070244672455e-05,
|
| 442 |
+
"loss": 0.9927,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2237215909090909,
|
| 447 |
+
"grad_norm": 14.551782608032227,
|
| 448 |
+
"learning_rate": 4.3133385951065515e-05,
|
| 449 |
+
"loss": 0.86,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.22727272727272727,
|
| 454 |
+
"grad_norm": 36.133609771728516,
|
| 455 |
+
"learning_rate": 4.293606945540648e-05,
|
| 456 |
+
"loss": 0.7792,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.23082386363636365,
|
| 461 |
+
"grad_norm": 13.04199504852295,
|
| 462 |
+
"learning_rate": 4.273875295974744e-05,
|
| 463 |
+
"loss": 0.9251,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.234375,
|
| 468 |
+
"grad_norm": 13.920279502868652,
|
| 469 |
+
"learning_rate": 4.25414364640884e-05,
|
| 470 |
+
"loss": 1.0642,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.23792613636363635,
|
| 475 |
+
"grad_norm": 4.325495719909668,
|
| 476 |
+
"learning_rate": 4.234411996842936e-05,
|
| 477 |
+
"loss": 0.9275,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.24147727272727273,
|
| 482 |
+
"grad_norm": 23.056415557861328,
|
| 483 |
+
"learning_rate": 4.2146803472770324e-05,
|
| 484 |
+
"loss": 1.1907,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.2450284090909091,
|
| 489 |
+
"grad_norm": 11.812889099121094,
|
| 490 |
+
"learning_rate": 4.194948697711129e-05,
|
| 491 |
+
"loss": 1.0329,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.24857954545454544,
|
| 496 |
+
"grad_norm": 11.360099792480469,
|
| 497 |
+
"learning_rate": 4.1752170481452254e-05,
|
| 498 |
+
"loss": 0.8968,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.25,
|
| 503 |
+
"eval_accuracy": 0.6878216123499142,
|
| 504 |
+
"eval_f1": 0.6885242103977541,
|
| 505 |
+
"eval_loss": 0.8689462542533875,
|
| 506 |
+
"eval_precision": 0.742302721464857,
|
| 507 |
+
"eval_recall": 0.688613815159033,
|
| 508 |
+
"eval_runtime": 653.0033,
|
| 509 |
+
"eval_samples_per_second": 0.893,
|
| 510 |
+
"eval_steps_per_second": 0.224,
|
| 511 |
+
"step": 704
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 1.0021306818181819,
|
| 515 |
+
"grad_norm": 5.516427040100098,
|
| 516 |
+
"learning_rate": 4.1554853985793216e-05,
|
| 517 |
+
"loss": 0.9278,
|
| 518 |
+
"step": 710
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.0056818181818181,
|
| 522 |
+
"grad_norm": 8.10287094116211,
|
| 523 |
+
"learning_rate": 4.135753749013418e-05,
|
| 524 |
+
"loss": 0.8833,
|
| 525 |
+
"step": 720
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.0092329545454546,
|
| 529 |
+
"grad_norm": 21.60080337524414,
|
| 530 |
+
"learning_rate": 4.116022099447514e-05,
|
| 531 |
+
"loss": 0.6148,
|
| 532 |
+
"step": 730
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 1.0127840909090908,
|
| 536 |
+
"grad_norm": 17.968292236328125,
|
| 537 |
+
"learning_rate": 4.09629044988161e-05,
|
| 538 |
+
"loss": 0.8788,
|
| 539 |
+
"step": 740
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.0163352272727273,
|
| 543 |
+
"grad_norm": 8.538472175598145,
|
| 544 |
+
"learning_rate": 4.076558800315707e-05,
|
| 545 |
+
"loss": 0.778,
|
| 546 |
+
"step": 750
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 1.0198863636363635,
|
| 550 |
+
"grad_norm": 4.343443870544434,
|
| 551 |
+
"learning_rate": 4.056827150749803e-05,
|
| 552 |
+
"loss": 0.5165,
|
| 553 |
+
"step": 760
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.0234375,
|
| 557 |
+
"grad_norm": 6.1510009765625,
|
| 558 |
+
"learning_rate": 4.037095501183899e-05,
|
| 559 |
+
"loss": 0.8024,
|
| 560 |
+
"step": 770
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.0269886363636365,
|
| 564 |
+
"grad_norm": 23.44464111328125,
|
| 565 |
+
"learning_rate": 4.0173638516179955e-05,
|
| 566 |
+
"loss": 1.0074,
|
| 567 |
+
"step": 780
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 1.0305397727272727,
|
| 571 |
+
"grad_norm": 16.371910095214844,
|
| 572 |
+
"learning_rate": 3.997632202052092e-05,
|
| 573 |
+
"loss": 0.9794,
|
| 574 |
+
"step": 790
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 1.0340909090909092,
|
| 578 |
+
"grad_norm": 8.532549858093262,
|
| 579 |
+
"learning_rate": 3.977900552486188e-05,
|
| 580 |
+
"loss": 0.8435,
|
| 581 |
+
"step": 800
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.0376420454545454,
|
| 585 |
+
"grad_norm": 15.13945484161377,
|
| 586 |
+
"learning_rate": 3.958168902920285e-05,
|
| 587 |
+
"loss": 0.9073,
|
| 588 |
+
"step": 810
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 1.0411931818181819,
|
| 592 |
+
"grad_norm": 12.980558395385742,
|
| 593 |
+
"learning_rate": 3.938437253354381e-05,
|
| 594 |
+
"loss": 0.8077,
|
| 595 |
+
"step": 820
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.0447443181818181,
|
| 599 |
+
"grad_norm": 7.141788005828857,
|
| 600 |
+
"learning_rate": 3.9187056037884764e-05,
|
| 601 |
+
"loss": 0.9853,
|
| 602 |
+
"step": 830
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.0482954545454546,
|
| 606 |
+
"grad_norm": 5.774048328399658,
|
| 607 |
+
"learning_rate": 3.898973954222573e-05,
|
| 608 |
+
"loss": 0.5617,
|
| 609 |
+
"step": 840
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 1.0518465909090908,
|
| 613 |
+
"grad_norm": 30.009641647338867,
|
| 614 |
+
"learning_rate": 3.8792423046566694e-05,
|
| 615 |
+
"loss": 0.8913,
|
| 616 |
+
"step": 850
|
| 617 |
+
},
|
| 618 |
+
{
|
| 619 |
+
"epoch": 1.0553977272727273,
|
| 620 |
+
"grad_norm": 11.07997989654541,
|
| 621 |
+
"learning_rate": 3.8595106550907656e-05,
|
| 622 |
+
"loss": 0.6674,
|
| 623 |
+
"step": 860
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 1.0589488636363635,
|
| 627 |
+
"grad_norm": 13.393827438354492,
|
| 628 |
+
"learning_rate": 3.8397790055248625e-05,
|
| 629 |
+
"loss": 0.5832,
|
| 630 |
+
"step": 870
|
| 631 |
+
},
|
| 632 |
+
{
|
| 633 |
+
"epoch": 1.0625,
|
| 634 |
+
"grad_norm": 10.827837944030762,
|
| 635 |
+
"learning_rate": 3.8200473559589587e-05,
|
| 636 |
+
"loss": 0.6509,
|
| 637 |
+
"step": 880
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.0660511363636365,
|
| 641 |
+
"grad_norm": 27.103227615356445,
|
| 642 |
+
"learning_rate": 3.800315706393054e-05,
|
| 643 |
+
"loss": 0.7135,
|
| 644 |
+
"step": 890
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.0696022727272727,
|
| 648 |
+
"grad_norm": 16.949966430664062,
|
| 649 |
+
"learning_rate": 3.780584056827151e-05,
|
| 650 |
+
"loss": 1.0985,
|
| 651 |
+
"step": 900
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 1.0731534090909092,
|
| 655 |
+
"grad_norm": 11.158037185668945,
|
| 656 |
+
"learning_rate": 3.760852407261247e-05,
|
| 657 |
+
"loss": 0.7335,
|
| 658 |
+
"step": 910
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"epoch": 1.0767045454545454,
|
| 662 |
+
"grad_norm": 26.55095863342285,
|
| 663 |
+
"learning_rate": 3.7411207576953434e-05,
|
| 664 |
+
"loss": 0.7143,
|
| 665 |
+
"step": 920
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.0802556818181819,
|
| 669 |
+
"grad_norm": 7.920712947845459,
|
| 670 |
+
"learning_rate": 3.72138910812944e-05,
|
| 671 |
+
"loss": 0.8197,
|
| 672 |
+
"step": 930
|
| 673 |
+
},
|
| 674 |
+
{
|
| 675 |
+
"epoch": 1.0838068181818181,
|
| 676 |
+
"grad_norm": 2.0704779624938965,
|
| 677 |
+
"learning_rate": 3.7016574585635364e-05,
|
| 678 |
+
"loss": 0.7739,
|
| 679 |
+
"step": 940
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.0873579545454546,
|
| 683 |
+
"grad_norm": 29.551998138427734,
|
| 684 |
+
"learning_rate": 3.681925808997632e-05,
|
| 685 |
+
"loss": 0.9224,
|
| 686 |
+
"step": 950
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.0909090909090908,
|
| 690 |
+
"grad_norm": 9.859967231750488,
|
| 691 |
+
"learning_rate": 3.662194159431729e-05,
|
| 692 |
+
"loss": 0.8151,
|
| 693 |
+
"step": 960
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 1.0944602272727273,
|
| 697 |
+
"grad_norm": 10.744651794433594,
|
| 698 |
+
"learning_rate": 3.642462509865825e-05,
|
| 699 |
+
"loss": 0.4832,
|
| 700 |
+
"step": 970
|
| 701 |
+
},
|
| 702 |
+
{
|
| 703 |
+
"epoch": 1.0980113636363635,
|
| 704 |
+
"grad_norm": 3.085749864578247,
|
| 705 |
+
"learning_rate": 3.622730860299921e-05,
|
| 706 |
+
"loss": 0.5775,
|
| 707 |
+
"step": 980
|
| 708 |
+
},
|
| 709 |
+
{
|
| 710 |
+
"epoch": 1.1015625,
|
| 711 |
+
"grad_norm": 11.848700523376465,
|
| 712 |
+
"learning_rate": 3.602999210734018e-05,
|
| 713 |
+
"loss": 0.4641,
|
| 714 |
+
"step": 990
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 1.1051136363636365,
|
| 718 |
+
"grad_norm": 5.603400230407715,
|
| 719 |
+
"learning_rate": 3.583267561168114e-05,
|
| 720 |
+
"loss": 0.5069,
|
| 721 |
+
"step": 1000
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.1086647727272727,
|
| 725 |
+
"grad_norm": 8.615111351013184,
|
| 726 |
+
"learning_rate": 3.5635359116022096e-05,
|
| 727 |
+
"loss": 0.5031,
|
| 728 |
+
"step": 1010
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.1122159090909092,
|
| 732 |
+
"grad_norm": 23.804479598999023,
|
| 733 |
+
"learning_rate": 3.5438042620363065e-05,
|
| 734 |
+
"loss": 0.7348,
|
| 735 |
+
"step": 1020
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 1.1157670454545454,
|
| 739 |
+
"grad_norm": 1.017351508140564,
|
| 740 |
+
"learning_rate": 3.5240726124704027e-05,
|
| 741 |
+
"loss": 0.7015,
|
| 742 |
+
"step": 1030
|
| 743 |
+
},
|
| 744 |
+
{
|
| 745 |
+
"epoch": 1.1193181818181819,
|
| 746 |
+
"grad_norm": 14.092010498046875,
|
| 747 |
+
"learning_rate": 3.504340962904499e-05,
|
| 748 |
+
"loss": 0.5592,
|
| 749 |
+
"step": 1040
|
| 750 |
+
},
|
| 751 |
+
{
|
| 752 |
+
"epoch": 1.1228693181818181,
|
| 753 |
+
"grad_norm": 20.209379196166992,
|
| 754 |
+
"learning_rate": 3.484609313338596e-05,
|
| 755 |
+
"loss": 0.7704,
|
| 756 |
+
"step": 1050
|
| 757 |
+
},
|
| 758 |
+
{
|
| 759 |
+
"epoch": 1.1264204545454546,
|
| 760 |
+
"grad_norm": 0.6501548290252686,
|
| 761 |
+
"learning_rate": 3.464877663772691e-05,
|
| 762 |
+
"loss": 0.5026,
|
| 763 |
+
"step": 1060
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.1299715909090908,
|
| 767 |
+
"grad_norm": 1.0749989748001099,
|
| 768 |
+
"learning_rate": 3.4451460142067874e-05,
|
| 769 |
+
"loss": 0.5371,
|
| 770 |
+
"step": 1070
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.1335227272727273,
|
| 774 |
+
"grad_norm": 10.37779712677002,
|
| 775 |
+
"learning_rate": 3.425414364640884e-05,
|
| 776 |
+
"loss": 0.4032,
|
| 777 |
+
"step": 1080
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 1.1370738636363638,
|
| 781 |
+
"grad_norm": 30.61161231994629,
|
| 782 |
+
"learning_rate": 3.4056827150749804e-05,
|
| 783 |
+
"loss": 0.6612,
|
| 784 |
+
"step": 1090
|
| 785 |
+
},
|
| 786 |
+
{
|
| 787 |
+
"epoch": 1.140625,
|
| 788 |
+
"grad_norm": 28.0980167388916,
|
| 789 |
+
"learning_rate": 3.3859510655090766e-05,
|
| 790 |
+
"loss": 0.5713,
|
| 791 |
+
"step": 1100
|
| 792 |
+
},
|
| 793 |
+
{
|
| 794 |
+
"epoch": 1.1441761363636362,
|
| 795 |
+
"grad_norm": 25.723461151123047,
|
| 796 |
+
"learning_rate": 3.3662194159431734e-05,
|
| 797 |
+
"loss": 0.6669,
|
| 798 |
+
"step": 1110
|
| 799 |
+
},
|
| 800 |
+
{
|
| 801 |
+
"epoch": 1.1477272727272727,
|
| 802 |
+
"grad_norm": 7.138387203216553,
|
| 803 |
+
"learning_rate": 3.346487766377269e-05,
|
| 804 |
+
"loss": 0.5781,
|
| 805 |
+
"step": 1120
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.1512784090909092,
|
| 809 |
+
"grad_norm": 27.993236541748047,
|
| 810 |
+
"learning_rate": 3.326756116811365e-05,
|
| 811 |
+
"loss": 0.6033,
|
| 812 |
+
"step": 1130
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.1548295454545454,
|
| 816 |
+
"grad_norm": 3.16493558883667,
|
| 817 |
+
"learning_rate": 3.307024467245462e-05,
|
| 818 |
+
"loss": 0.787,
|
| 819 |
+
"step": 1140
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 1.1583806818181819,
|
| 823 |
+
"grad_norm": 11.943394660949707,
|
| 824 |
+
"learning_rate": 3.287292817679558e-05,
|
| 825 |
+
"loss": 0.657,
|
| 826 |
+
"step": 1150
|
| 827 |
+
},
|
| 828 |
+
{
|
| 829 |
+
"epoch": 1.1619318181818181,
|
| 830 |
+
"grad_norm": 3.7604823112487793,
|
| 831 |
+
"learning_rate": 3.267561168113654e-05,
|
| 832 |
+
"loss": 0.4096,
|
| 833 |
+
"step": 1160
|
| 834 |
+
},
|
| 835 |
+
{
|
| 836 |
+
"epoch": 1.1654829545454546,
|
| 837 |
+
"grad_norm": 21.66828727722168,
|
| 838 |
+
"learning_rate": 3.247829518547751e-05,
|
| 839 |
+
"loss": 0.5554,
|
| 840 |
+
"step": 1170
|
| 841 |
+
},
|
| 842 |
+
{
|
| 843 |
+
"epoch": 1.1690340909090908,
|
| 844 |
+
"grad_norm": 3.944326162338257,
|
| 845 |
+
"learning_rate": 3.2280978689818467e-05,
|
| 846 |
+
"loss": 0.8685,
|
| 847 |
+
"step": 1180
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.1725852272727273,
|
| 851 |
+
"grad_norm": 4.541408061981201,
|
| 852 |
+
"learning_rate": 3.208366219415943e-05,
|
| 853 |
+
"loss": 0.651,
|
| 854 |
+
"step": 1190
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.1761363636363638,
|
| 858 |
+
"grad_norm": 16.059410095214844,
|
| 859 |
+
"learning_rate": 3.18863456985004e-05,
|
| 860 |
+
"loss": 0.5313,
|
| 861 |
+
"step": 1200
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 1.1796875,
|
| 865 |
+
"grad_norm": 29.15842056274414,
|
| 866 |
+
"learning_rate": 3.168902920284136e-05,
|
| 867 |
+
"loss": 0.6873,
|
| 868 |
+
"step": 1210
|
| 869 |
+
},
|
| 870 |
+
{
|
| 871 |
+
"epoch": 1.1832386363636362,
|
| 872 |
+
"grad_norm": 27.368764877319336,
|
| 873 |
+
"learning_rate": 3.149171270718232e-05,
|
| 874 |
+
"loss": 0.7953,
|
| 875 |
+
"step": 1220
|
| 876 |
+
},
|
| 877 |
+
{
|
| 878 |
+
"epoch": 1.1867897727272727,
|
| 879 |
+
"grad_norm": 12.318428039550781,
|
| 880 |
+
"learning_rate": 3.129439621152329e-05,
|
| 881 |
+
"loss": 0.316,
|
| 882 |
+
"step": 1230
|
| 883 |
+
},
|
| 884 |
+
{
|
| 885 |
+
"epoch": 1.1903409090909092,
|
| 886 |
+
"grad_norm": 7.381816864013672,
|
| 887 |
+
"learning_rate": 3.1097079715864244e-05,
|
| 888 |
+
"loss": 0.5909,
|
| 889 |
+
"step": 1240
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.1938920454545454,
|
| 893 |
+
"grad_norm": 56.74608612060547,
|
| 894 |
+
"learning_rate": 3.0899763220205206e-05,
|
| 895 |
+
"loss": 0.5496,
|
| 896 |
+
"step": 1250
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.1974431818181819,
|
| 900 |
+
"grad_norm": 5.5123090744018555,
|
| 901 |
+
"learning_rate": 3.0702446724546174e-05,
|
| 902 |
+
"loss": 0.4415,
|
| 903 |
+
"step": 1260
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 1.2009943181818181,
|
| 907 |
+
"grad_norm": 3.0995726585388184,
|
| 908 |
+
"learning_rate": 3.0505130228887136e-05,
|
| 909 |
+
"loss": 0.4572,
|
| 910 |
+
"step": 1270
|
| 911 |
+
},
|
| 912 |
+
{
|
| 913 |
+
"epoch": 1.2045454545454546,
|
| 914 |
+
"grad_norm": 1.3388721942901611,
|
| 915 |
+
"learning_rate": 3.03078137332281e-05,
|
| 916 |
+
"loss": 0.4261,
|
| 917 |
+
"step": 1280
|
| 918 |
+
},
|
| 919 |
+
{
|
| 920 |
+
"epoch": 1.2080965909090908,
|
| 921 |
+
"grad_norm": 3.432847738265991,
|
| 922 |
+
"learning_rate": 3.0110497237569063e-05,
|
| 923 |
+
"loss": 0.5804,
|
| 924 |
+
"step": 1290
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 1.2116477272727273,
|
| 928 |
+
"grad_norm": 5.881500244140625,
|
| 929 |
+
"learning_rate": 2.9913180741910025e-05,
|
| 930 |
+
"loss": 0.3265,
|
| 931 |
+
"step": 1300
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.2151988636363638,
|
| 935 |
+
"grad_norm": 27.702701568603516,
|
| 936 |
+
"learning_rate": 2.971586424625099e-05,
|
| 937 |
+
"loss": 0.8524,
|
| 938 |
+
"step": 1310
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 1.21875,
|
| 942 |
+
"grad_norm": 3.5202479362487793,
|
| 943 |
+
"learning_rate": 2.951854775059195e-05,
|
| 944 |
+
"loss": 0.7208,
|
| 945 |
+
"step": 1320
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 1.2223011363636362,
|
| 949 |
+
"grad_norm": 14.859701156616211,
|
| 950 |
+
"learning_rate": 2.9321231254932913e-05,
|
| 951 |
+
"loss": 0.6345,
|
| 952 |
+
"step": 1330
|
| 953 |
+
},
|
| 954 |
+
{
|
| 955 |
+
"epoch": 1.2258522727272727,
|
| 956 |
+
"grad_norm": 2.571312427520752,
|
| 957 |
+
"learning_rate": 2.912391475927388e-05,
|
| 958 |
+
"loss": 0.3549,
|
| 959 |
+
"step": 1340
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.2294034090909092,
|
| 963 |
+
"grad_norm": 8.97351360321045,
|
| 964 |
+
"learning_rate": 2.892659826361484e-05,
|
| 965 |
+
"loss": 0.507,
|
| 966 |
+
"step": 1350
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 1.2329545454545454,
|
| 970 |
+
"grad_norm": 13.665780067443848,
|
| 971 |
+
"learning_rate": 2.8729281767955802e-05,
|
| 972 |
+
"loss": 0.4561,
|
| 973 |
+
"step": 1360
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.2365056818181819,
|
| 977 |
+
"grad_norm": 13.264731407165527,
|
| 978 |
+
"learning_rate": 2.8531965272296767e-05,
|
| 979 |
+
"loss": 0.539,
|
| 980 |
+
"step": 1370
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 1.2400568181818181,
|
| 984 |
+
"grad_norm": 29.683090209960938,
|
| 985 |
+
"learning_rate": 2.833464877663773e-05,
|
| 986 |
+
"loss": 0.5715,
|
| 987 |
+
"step": 1380
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.2436079545454546,
|
| 991 |
+
"grad_norm": 8.478317260742188,
|
| 992 |
+
"learning_rate": 2.813733228097869e-05,
|
| 993 |
+
"loss": 0.183,
|
| 994 |
+
"step": 1390
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.2471590909090908,
|
| 998 |
+
"grad_norm": 28.569978713989258,
|
| 999 |
+
"learning_rate": 2.7940015785319656e-05,
|
| 1000 |
+
"loss": 0.5002,
|
| 1001 |
+
"step": 1400
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.25,
|
| 1005 |
+
"eval_accuracy": 0.8542024013722127,
|
| 1006 |
+
"eval_f1": 0.8530975388761288,
|
| 1007 |
+
"eval_loss": 0.4373819828033447,
|
| 1008 |
+
"eval_precision": 0.8718298587182326,
|
| 1009 |
+
"eval_recall": 0.853515201868564,
|
| 1010 |
+
"eval_runtime": 695.9457,
|
| 1011 |
+
"eval_samples_per_second": 0.838,
|
| 1012 |
+
"eval_steps_per_second": 0.21,
|
| 1013 |
+
"step": 1408
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 2.000710227272727,
|
| 1017 |
+
"grad_norm": 25.39379119873047,
|
| 1018 |
+
"learning_rate": 2.7742699289660618e-05,
|
| 1019 |
+
"loss": 0.6607,
|
| 1020 |
+
"step": 1410
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 2.0042613636363638,
|
| 1024 |
+
"grad_norm": 3.422687530517578,
|
| 1025 |
+
"learning_rate": 2.754538279400158e-05,
|
| 1026 |
+
"loss": 0.5038,
|
| 1027 |
+
"step": 1420
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 2.0078125,
|
| 1031 |
+
"grad_norm": 1.3007397651672363,
|
| 1032 |
+
"learning_rate": 2.7348066298342545e-05,
|
| 1033 |
+
"loss": 0.3052,
|
| 1034 |
+
"step": 1430
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.0113636363636362,
|
| 1038 |
+
"grad_norm": 15.327472686767578,
|
| 1039 |
+
"learning_rate": 2.7150749802683506e-05,
|
| 1040 |
+
"loss": 0.2933,
|
| 1041 |
+
"step": 1440
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 2.014914772727273,
|
| 1045 |
+
"grad_norm": 23.62551498413086,
|
| 1046 |
+
"learning_rate": 2.6953433307024468e-05,
|
| 1047 |
+
"loss": 0.4165,
|
| 1048 |
+
"step": 1450
|
| 1049 |
+
},
|
| 1050 |
+
{
|
| 1051 |
+
"epoch": 2.018465909090909,
|
| 1052 |
+
"grad_norm": 8.549898147583008,
|
| 1053 |
+
"learning_rate": 2.6756116811365433e-05,
|
| 1054 |
+
"loss": 0.3321,
|
| 1055 |
+
"step": 1460
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 2.0220170454545454,
|
| 1059 |
+
"grad_norm": 25.098161697387695,
|
| 1060 |
+
"learning_rate": 2.6558800315706395e-05,
|
| 1061 |
+
"loss": 0.4109,
|
| 1062 |
+
"step": 1470
|
| 1063 |
+
},
|
| 1064 |
+
{
|
| 1065 |
+
"epoch": 2.0255681818181817,
|
| 1066 |
+
"grad_norm": 9.298599243164062,
|
| 1067 |
+
"learning_rate": 2.6361483820047357e-05,
|
| 1068 |
+
"loss": 0.4261,
|
| 1069 |
+
"step": 1480
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 2.0291193181818183,
|
| 1073 |
+
"grad_norm": 66.2896957397461,
|
| 1074 |
+
"learning_rate": 2.6164167324388322e-05,
|
| 1075 |
+
"loss": 0.3508,
|
| 1076 |
+
"step": 1490
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.0326704545454546,
|
| 1080 |
+
"grad_norm": 1.1833637952804565,
|
| 1081 |
+
"learning_rate": 2.5966850828729284e-05,
|
| 1082 |
+
"loss": 0.1169,
|
| 1083 |
+
"step": 1500
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 2.036221590909091,
|
| 1087 |
+
"grad_norm": 32.40926742553711,
|
| 1088 |
+
"learning_rate": 2.5769534333070246e-05,
|
| 1089 |
+
"loss": 0.7922,
|
| 1090 |
+
"step": 1510
|
| 1091 |
+
},
|
| 1092 |
+
{
|
| 1093 |
+
"epoch": 2.039772727272727,
|
| 1094 |
+
"grad_norm": 31.9284610748291,
|
| 1095 |
+
"learning_rate": 2.557221783741121e-05,
|
| 1096 |
+
"loss": 0.3044,
|
| 1097 |
+
"step": 1520
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.0433238636363638,
|
| 1101 |
+
"grad_norm": 0.32966673374176025,
|
| 1102 |
+
"learning_rate": 2.5374901341752172e-05,
|
| 1103 |
+
"loss": 0.5598,
|
| 1104 |
+
"step": 1530
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 2.046875,
|
| 1108 |
+
"grad_norm": 3.2222678661346436,
|
| 1109 |
+
"learning_rate": 2.5177584846093134e-05,
|
| 1110 |
+
"loss": 0.5305,
|
| 1111 |
+
"step": 1540
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 2.0504261363636362,
|
| 1115 |
+
"grad_norm": 5.157206058502197,
|
| 1116 |
+
"learning_rate": 2.4980268350434096e-05,
|
| 1117 |
+
"loss": 0.3806,
|
| 1118 |
+
"step": 1550
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.053977272727273,
|
| 1122 |
+
"grad_norm": 29.749786376953125,
|
| 1123 |
+
"learning_rate": 2.478295185477506e-05,
|
| 1124 |
+
"loss": 0.6572,
|
| 1125 |
+
"step": 1560
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 2.057528409090909,
|
| 1129 |
+
"grad_norm": 5.4849534034729,
|
| 1130 |
+
"learning_rate": 2.4585635359116023e-05,
|
| 1131 |
+
"loss": 0.2828,
|
| 1132 |
+
"step": 1570
|
| 1133 |
+
},
|
| 1134 |
+
{
|
| 1135 |
+
"epoch": 2.0610795454545454,
|
| 1136 |
+
"grad_norm": 1.8682303428649902,
|
| 1137 |
+
"learning_rate": 2.4388318863456985e-05,
|
| 1138 |
+
"loss": 0.275,
|
| 1139 |
+
"step": 1580
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.0646306818181817,
|
| 1143 |
+
"grad_norm": 31.073657989501953,
|
| 1144 |
+
"learning_rate": 2.419100236779795e-05,
|
| 1145 |
+
"loss": 0.2635,
|
| 1146 |
+
"step": 1590
|
| 1147 |
+
},
|
| 1148 |
+
{
|
| 1149 |
+
"epoch": 2.0681818181818183,
|
| 1150 |
+
"grad_norm": 2.873807668685913,
|
| 1151 |
+
"learning_rate": 2.399368587213891e-05,
|
| 1152 |
+
"loss": 0.2189,
|
| 1153 |
+
"step": 1600
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 2.0717329545454546,
|
| 1157 |
+
"grad_norm": 6.81038761138916,
|
| 1158 |
+
"learning_rate": 2.3796369376479873e-05,
|
| 1159 |
+
"loss": 0.2335,
|
| 1160 |
+
"step": 1610
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.075284090909091,
|
| 1164 |
+
"grad_norm": 5.240432262420654,
|
| 1165 |
+
"learning_rate": 2.359905288082084e-05,
|
| 1166 |
+
"loss": 0.513,
|
| 1167 |
+
"step": 1620
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 2.078835227272727,
|
| 1171 |
+
"grad_norm": 2.8783762454986572,
|
| 1172 |
+
"learning_rate": 2.34017363851618e-05,
|
| 1173 |
+
"loss": 0.2719,
|
| 1174 |
+
"step": 1630
|
| 1175 |
+
},
|
| 1176 |
+
{
|
| 1177 |
+
"epoch": 2.0823863636363638,
|
| 1178 |
+
"grad_norm": 4.599855422973633,
|
| 1179 |
+
"learning_rate": 2.3204419889502762e-05,
|
| 1180 |
+
"loss": 0.2012,
|
| 1181 |
+
"step": 1640
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 2.0859375,
|
| 1185 |
+
"grad_norm": 1.6575523614883423,
|
| 1186 |
+
"learning_rate": 2.3007103393843727e-05,
|
| 1187 |
+
"loss": 0.1299,
|
| 1188 |
+
"step": 1650
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 2.0894886363636362,
|
| 1192 |
+
"grad_norm": 0.3831326961517334,
|
| 1193 |
+
"learning_rate": 2.280978689818469e-05,
|
| 1194 |
+
"loss": 0.1766,
|
| 1195 |
+
"step": 1660
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 2.093039772727273,
|
| 1199 |
+
"grad_norm": 1.1560420989990234,
|
| 1200 |
+
"learning_rate": 2.261247040252565e-05,
|
| 1201 |
+
"loss": 0.3844,
|
| 1202 |
+
"step": 1670
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.096590909090909,
|
| 1206 |
+
"grad_norm": 1.8366397619247437,
|
| 1207 |
+
"learning_rate": 2.2415153906866616e-05,
|
| 1208 |
+
"loss": 0.3113,
|
| 1209 |
+
"step": 1680
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 2.1001420454545454,
|
| 1213 |
+
"grad_norm": 10.303657531738281,
|
| 1214 |
+
"learning_rate": 2.2217837411207578e-05,
|
| 1215 |
+
"loss": 0.1612,
|
| 1216 |
+
"step": 1690
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 2.1036931818181817,
|
| 1220 |
+
"grad_norm": 35.21874237060547,
|
| 1221 |
+
"learning_rate": 2.202052091554854e-05,
|
| 1222 |
+
"loss": 0.454,
|
| 1223 |
+
"step": 1700
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 2.1072443181818183,
|
| 1227 |
+
"grad_norm": 4.324901580810547,
|
| 1228 |
+
"learning_rate": 2.1823204419889505e-05,
|
| 1229 |
+
"loss": 0.5713,
|
| 1230 |
+
"step": 1710
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 2.1107954545454546,
|
| 1234 |
+
"grad_norm": 4.857918739318848,
|
| 1235 |
+
"learning_rate": 2.1625887924230466e-05,
|
| 1236 |
+
"loss": 0.4748,
|
| 1237 |
+
"step": 1720
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 2.114346590909091,
|
| 1241 |
+
"grad_norm": 0.15146464109420776,
|
| 1242 |
+
"learning_rate": 2.1428571428571428e-05,
|
| 1243 |
+
"loss": 0.3261,
|
| 1244 |
+
"step": 1730
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.117897727272727,
|
| 1248 |
+
"grad_norm": 19.752939224243164,
|
| 1249 |
+
"learning_rate": 2.1231254932912393e-05,
|
| 1250 |
+
"loss": 0.2406,
|
| 1251 |
+
"step": 1740
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 2.1214488636363638,
|
| 1255 |
+
"grad_norm": 18.78827476501465,
|
| 1256 |
+
"learning_rate": 2.1033938437253355e-05,
|
| 1257 |
+
"loss": 0.4846,
|
| 1258 |
+
"step": 1750
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 2.125,
|
| 1262 |
+
"grad_norm": 0.27601099014282227,
|
| 1263 |
+
"learning_rate": 2.0836621941594317e-05,
|
| 1264 |
+
"loss": 0.2026,
|
| 1265 |
+
"step": 1760
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 2.1285511363636362,
|
| 1269 |
+
"grad_norm": 0.5743858218193054,
|
| 1270 |
+
"learning_rate": 2.0639305445935282e-05,
|
| 1271 |
+
"loss": 0.2579,
|
| 1272 |
+
"step": 1770
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 2.132102272727273,
|
| 1276 |
+
"grad_norm": 0.10733279585838318,
|
| 1277 |
+
"learning_rate": 2.0441988950276244e-05,
|
| 1278 |
+
"loss": 0.2348,
|
| 1279 |
+
"step": 1780
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 2.135653409090909,
|
| 1283 |
+
"grad_norm": 2.2115983963012695,
|
| 1284 |
+
"learning_rate": 2.0244672454617206e-05,
|
| 1285 |
+
"loss": 0.2632,
|
| 1286 |
+
"step": 1790
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.1392045454545454,
|
| 1290 |
+
"grad_norm": 37.26658248901367,
|
| 1291 |
+
"learning_rate": 2.004735595895817e-05,
|
| 1292 |
+
"loss": 0.4384,
|
| 1293 |
+
"step": 1800
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 2.1427556818181817,
|
| 1297 |
+
"grad_norm": 10.85006332397461,
|
| 1298 |
+
"learning_rate": 1.9850039463299132e-05,
|
| 1299 |
+
"loss": 0.2217,
|
| 1300 |
+
"step": 1810
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 2.1463068181818183,
|
| 1304 |
+
"grad_norm": 4.717857360839844,
|
| 1305 |
+
"learning_rate": 1.9652722967640098e-05,
|
| 1306 |
+
"loss": 0.2748,
|
| 1307 |
+
"step": 1820
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 2.1498579545454546,
|
| 1311 |
+
"grad_norm": 2.3380446434020996,
|
| 1312 |
+
"learning_rate": 1.945540647198106e-05,
|
| 1313 |
+
"loss": 0.1738,
|
| 1314 |
+
"step": 1830
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 2.153409090909091,
|
| 1318 |
+
"grad_norm": 1.6607507467269897,
|
| 1319 |
+
"learning_rate": 1.925808997632202e-05,
|
| 1320 |
+
"loss": 0.2641,
|
| 1321 |
+
"step": 1840
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 2.156960227272727,
|
| 1325 |
+
"grad_norm": 12.202836036682129,
|
| 1326 |
+
"learning_rate": 1.9060773480662986e-05,
|
| 1327 |
+
"loss": 0.2995,
|
| 1328 |
+
"step": 1850
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 2.1605113636363638,
|
| 1332 |
+
"grad_norm": 20.316972732543945,
|
| 1333 |
+
"learning_rate": 1.8863456985003948e-05,
|
| 1334 |
+
"loss": 0.1573,
|
| 1335 |
+
"step": 1860
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 2.1640625,
|
| 1339 |
+
"grad_norm": 34.030296325683594,
|
| 1340 |
+
"learning_rate": 1.866614048934491e-05,
|
| 1341 |
+
"loss": 0.266,
|
| 1342 |
+
"step": 1870
|
| 1343 |
+
},
|
| 1344 |
+
{
|
| 1345 |
+
"epoch": 2.1676136363636362,
|
| 1346 |
+
"grad_norm": 4.90464973449707,
|
| 1347 |
+
"learning_rate": 1.8468823993685875e-05,
|
| 1348 |
+
"loss": 0.1064,
|
| 1349 |
+
"step": 1880
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 2.171164772727273,
|
| 1353 |
+
"grad_norm": 38.46381378173828,
|
| 1354 |
+
"learning_rate": 1.8271507498026837e-05,
|
| 1355 |
+
"loss": 0.2679,
|
| 1356 |
+
"step": 1890
|
| 1357 |
+
},
|
| 1358 |
+
{
|
| 1359 |
+
"epoch": 2.174715909090909,
|
| 1360 |
+
"grad_norm": 52.19645690917969,
|
| 1361 |
+
"learning_rate": 1.80741910023678e-05,
|
| 1362 |
+
"loss": 0.166,
|
| 1363 |
+
"step": 1900
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 2.1782670454545454,
|
| 1367 |
+
"grad_norm": 15.714051246643066,
|
| 1368 |
+
"learning_rate": 1.7876874506708764e-05,
|
| 1369 |
+
"loss": 0.4399,
|
| 1370 |
+
"step": 1910
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 2.1818181818181817,
|
| 1374 |
+
"grad_norm": 50.66694641113281,
|
| 1375 |
+
"learning_rate": 1.7679558011049725e-05,
|
| 1376 |
+
"loss": 0.1816,
|
| 1377 |
+
"step": 1920
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 2.1853693181818183,
|
| 1381 |
+
"grad_norm": 4.8143205642700195,
|
| 1382 |
+
"learning_rate": 1.7482241515390687e-05,
|
| 1383 |
+
"loss": 0.5087,
|
| 1384 |
+
"step": 1930
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 2.1889204545454546,
|
| 1388 |
+
"grad_norm": 1.34380042552948,
|
| 1389 |
+
"learning_rate": 1.7284925019731652e-05,
|
| 1390 |
+
"loss": 0.4257,
|
| 1391 |
+
"step": 1940
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 2.192471590909091,
|
| 1395 |
+
"grad_norm": 6.62140417098999,
|
| 1396 |
+
"learning_rate": 1.7087608524072614e-05,
|
| 1397 |
+
"loss": 0.1767,
|
| 1398 |
+
"step": 1950
|
| 1399 |
+
},
|
| 1400 |
+
{
|
| 1401 |
+
"epoch": 2.196022727272727,
|
| 1402 |
+
"grad_norm": 34.63706588745117,
|
| 1403 |
+
"learning_rate": 1.6890292028413576e-05,
|
| 1404 |
+
"loss": 0.3301,
|
| 1405 |
+
"step": 1960
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 2.1995738636363638,
|
| 1409 |
+
"grad_norm": 3.23504638671875,
|
| 1410 |
+
"learning_rate": 1.669297553275454e-05,
|
| 1411 |
+
"loss": 0.2174,
|
| 1412 |
+
"step": 1970
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 2.203125,
|
| 1416 |
+
"grad_norm": 16.90635871887207,
|
| 1417 |
+
"learning_rate": 1.64956590370955e-05,
|
| 1418 |
+
"loss": 0.4572,
|
| 1419 |
+
"step": 1980
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 2.2066761363636362,
|
| 1423 |
+
"grad_norm": 20.676467895507812,
|
| 1424 |
+
"learning_rate": 1.6298342541436465e-05,
|
| 1425 |
+
"loss": 0.3199,
|
| 1426 |
+
"step": 1990
|
| 1427 |
+
},
|
| 1428 |
+
{
|
| 1429 |
+
"epoch": 2.210227272727273,
|
| 1430 |
+
"grad_norm": 0.31672224402427673,
|
| 1431 |
+
"learning_rate": 1.610102604577743e-05,
|
| 1432 |
+
"loss": 0.2,
|
| 1433 |
+
"step": 2000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 2.213778409090909,
|
| 1437 |
+
"grad_norm": 8.79170036315918,
|
| 1438 |
+
"learning_rate": 1.5903709550118388e-05,
|
| 1439 |
+
"loss": 0.1675,
|
| 1440 |
+
"step": 2010
|
| 1441 |
+
},
|
| 1442 |
+
{
|
| 1443 |
+
"epoch": 2.2173295454545454,
|
| 1444 |
+
"grad_norm": 2.0761942863464355,
|
| 1445 |
+
"learning_rate": 1.5706393054459353e-05,
|
| 1446 |
+
"loss": 0.2614,
|
| 1447 |
+
"step": 2020
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 2.2208806818181817,
|
| 1451 |
+
"grad_norm": 0.06360641866922379,
|
| 1452 |
+
"learning_rate": 1.550907655880032e-05,
|
| 1453 |
+
"loss": 0.2335,
|
| 1454 |
+
"step": 2030
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 2.2244318181818183,
|
| 1458 |
+
"grad_norm": 18.36467170715332,
|
| 1459 |
+
"learning_rate": 1.5311760063141277e-05,
|
| 1460 |
+
"loss": 0.5066,
|
| 1461 |
+
"step": 2040
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 2.2279829545454546,
|
| 1465 |
+
"grad_norm": 9.642960548400879,
|
| 1466 |
+
"learning_rate": 1.5114443567482242e-05,
|
| 1467 |
+
"loss": 0.2035,
|
| 1468 |
+
"step": 2050
|
| 1469 |
+
},
|
| 1470 |
+
{
|
| 1471 |
+
"epoch": 2.231534090909091,
|
| 1472 |
+
"grad_norm": 6.050065040588379,
|
| 1473 |
+
"learning_rate": 1.4917127071823205e-05,
|
| 1474 |
+
"loss": 0.1467,
|
| 1475 |
+
"step": 2060
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.235085227272727,
|
| 1479 |
+
"grad_norm": 0.06556593626737595,
|
| 1480 |
+
"learning_rate": 1.4719810576164167e-05,
|
| 1481 |
+
"loss": 0.1243,
|
| 1482 |
+
"step": 2070
|
| 1483 |
+
},
|
| 1484 |
+
{
|
| 1485 |
+
"epoch": 2.2386363636363638,
|
| 1486 |
+
"grad_norm": 0.29696065187454224,
|
| 1487 |
+
"learning_rate": 1.452249408050513e-05,
|
| 1488 |
+
"loss": 0.2465,
|
| 1489 |
+
"step": 2080
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 2.2421875,
|
| 1493 |
+
"grad_norm": 1.2069284915924072,
|
| 1494 |
+
"learning_rate": 1.4325177584846094e-05,
|
| 1495 |
+
"loss": 0.191,
|
| 1496 |
+
"step": 2090
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 2.2457386363636362,
|
| 1500 |
+
"grad_norm": 2.228632688522339,
|
| 1501 |
+
"learning_rate": 1.4127861089187056e-05,
|
| 1502 |
+
"loss": 0.1718,
|
| 1503 |
+
"step": 2100
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 2.249289772727273,
|
| 1507 |
+
"grad_norm": 1.17184317111969,
|
| 1508 |
+
"learning_rate": 1.393054459352802e-05,
|
| 1509 |
+
"loss": 0.3627,
|
| 1510 |
+
"step": 2110
|
| 1511 |
+
},
|
| 1512 |
+
{
|
| 1513 |
+
"epoch": 2.25,
|
| 1514 |
+
"eval_accuracy": 0.9622641509433962,
|
| 1515 |
+
"eval_f1": 0.9617697305162668,
|
| 1516 |
+
"eval_loss": 0.1109052374958992,
|
| 1517 |
+
"eval_precision": 0.9647490495721144,
|
| 1518 |
+
"eval_recall": 0.9614389495077621,
|
| 1519 |
+
"eval_runtime": 696.0893,
|
| 1520 |
+
"eval_samples_per_second": 0.838,
|
| 1521 |
+
"eval_steps_per_second": 0.21,
|
| 1522 |
+
"step": 2112
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 3.002840909090909,
|
| 1526 |
+
"grad_norm": 0.03763706982135773,
|
| 1527 |
+
"learning_rate": 1.3733228097868983e-05,
|
| 1528 |
+
"loss": 0.0171,
|
| 1529 |
+
"step": 2120
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 3.0063920454545454,
|
| 1533 |
+
"grad_norm": 1.5585871934890747,
|
| 1534 |
+
"learning_rate": 1.3535911602209945e-05,
|
| 1535 |
+
"loss": 0.2023,
|
| 1536 |
+
"step": 2130
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 3.0099431818181817,
|
| 1540 |
+
"grad_norm": 20.643203735351562,
|
| 1541 |
+
"learning_rate": 1.3338595106550908e-05,
|
| 1542 |
+
"loss": 0.1212,
|
| 1543 |
+
"step": 2140
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 3.0134943181818183,
|
| 1547 |
+
"grad_norm": 0.19967862963676453,
|
| 1548 |
+
"learning_rate": 1.3141278610891871e-05,
|
| 1549 |
+
"loss": 0.0201,
|
| 1550 |
+
"step": 2150
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 3.0170454545454546,
|
| 1554 |
+
"grad_norm": 3.6144466400146484,
|
| 1555 |
+
"learning_rate": 1.2943962115232833e-05,
|
| 1556 |
+
"loss": 0.1455,
|
| 1557 |
+
"step": 2160
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 3.020596590909091,
|
| 1561 |
+
"grad_norm": 7.9842610359191895,
|
| 1562 |
+
"learning_rate": 1.2746645619573797e-05,
|
| 1563 |
+
"loss": 0.042,
|
| 1564 |
+
"step": 2170
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 3.024147727272727,
|
| 1568 |
+
"grad_norm": 3.346548080444336,
|
| 1569 |
+
"learning_rate": 1.254932912391476e-05,
|
| 1570 |
+
"loss": 0.0365,
|
| 1571 |
+
"step": 2180
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.0276988636363638,
|
| 1575 |
+
"grad_norm": 0.10210000723600388,
|
| 1576 |
+
"learning_rate": 1.2352012628255722e-05,
|
| 1577 |
+
"loss": 0.118,
|
| 1578 |
+
"step": 2190
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 3.03125,
|
| 1582 |
+
"grad_norm": 2.003911018371582,
|
| 1583 |
+
"learning_rate": 1.2154696132596685e-05,
|
| 1584 |
+
"loss": 0.0461,
|
| 1585 |
+
"step": 2200
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 3.0348011363636362,
|
| 1589 |
+
"grad_norm": 0.06448382884263992,
|
| 1590 |
+
"learning_rate": 1.1957379636937649e-05,
|
| 1591 |
+
"loss": 0.1276,
|
| 1592 |
+
"step": 2210
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 3.038352272727273,
|
| 1596 |
+
"grad_norm": 59.72112274169922,
|
| 1597 |
+
"learning_rate": 1.176006314127861e-05,
|
| 1598 |
+
"loss": 0.1744,
|
| 1599 |
+
"step": 2220
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 3.041903409090909,
|
| 1603 |
+
"grad_norm": 26.68014907836914,
|
| 1604 |
+
"learning_rate": 1.1562746645619574e-05,
|
| 1605 |
+
"loss": 0.0988,
|
| 1606 |
+
"step": 2230
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 3.0454545454545454,
|
| 1610 |
+
"grad_norm": 1.2907896041870117,
|
| 1611 |
+
"learning_rate": 1.1365430149960538e-05,
|
| 1612 |
+
"loss": 0.0432,
|
| 1613 |
+
"step": 2240
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 3.0490056818181817,
|
| 1617 |
+
"grad_norm": 2.807907819747925,
|
| 1618 |
+
"learning_rate": 1.11681136543015e-05,
|
| 1619 |
+
"loss": 0.1671,
|
| 1620 |
+
"step": 2250
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 3.0525568181818183,
|
| 1624 |
+
"grad_norm": 0.22834616899490356,
|
| 1625 |
+
"learning_rate": 1.0970797158642463e-05,
|
| 1626 |
+
"loss": 0.0643,
|
| 1627 |
+
"step": 2260
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 3.0561079545454546,
|
| 1631 |
+
"grad_norm": 0.23886850476264954,
|
| 1632 |
+
"learning_rate": 1.0773480662983426e-05,
|
| 1633 |
+
"loss": 0.1166,
|
| 1634 |
+
"step": 2270
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 3.059659090909091,
|
| 1638 |
+
"grad_norm": 2.190918445587158,
|
| 1639 |
+
"learning_rate": 1.0576164167324388e-05,
|
| 1640 |
+
"loss": 0.0236,
|
| 1641 |
+
"step": 2280
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 3.063210227272727,
|
| 1645 |
+
"grad_norm": 1.075537919998169,
|
| 1646 |
+
"learning_rate": 1.0378847671665353e-05,
|
| 1647 |
+
"loss": 0.1185,
|
| 1648 |
+
"step": 2290
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 3.0667613636363638,
|
| 1652 |
+
"grad_norm": 34.861881256103516,
|
| 1653 |
+
"learning_rate": 1.0181531176006315e-05,
|
| 1654 |
+
"loss": 0.1225,
|
| 1655 |
+
"step": 2300
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 3.0703125,
|
| 1659 |
+
"grad_norm": 0.2916108965873718,
|
| 1660 |
+
"learning_rate": 9.984214680347277e-06,
|
| 1661 |
+
"loss": 0.0877,
|
| 1662 |
+
"step": 2310
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 3.0738636363636362,
|
| 1666 |
+
"grad_norm": 24.542476654052734,
|
| 1667 |
+
"learning_rate": 9.786898184688242e-06,
|
| 1668 |
+
"loss": 0.1007,
|
| 1669 |
+
"step": 2320
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 3.077414772727273,
|
| 1673 |
+
"grad_norm": 1.0566986799240112,
|
| 1674 |
+
"learning_rate": 9.589581689029204e-06,
|
| 1675 |
+
"loss": 0.0563,
|
| 1676 |
+
"step": 2330
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 3.080965909090909,
|
| 1680 |
+
"grad_norm": 20.17405128479004,
|
| 1681 |
+
"learning_rate": 9.392265193370165e-06,
|
| 1682 |
+
"loss": 0.0374,
|
| 1683 |
+
"step": 2340
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 3.0845170454545454,
|
| 1687 |
+
"grad_norm": 0.07401897758245468,
|
| 1688 |
+
"learning_rate": 9.194948697711129e-06,
|
| 1689 |
+
"loss": 0.0495,
|
| 1690 |
+
"step": 2350
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 3.0880681818181817,
|
| 1694 |
+
"grad_norm": 4.210412502288818,
|
| 1695 |
+
"learning_rate": 8.997632202052092e-06,
|
| 1696 |
+
"loss": 0.1265,
|
| 1697 |
+
"step": 2360
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 3.0916193181818183,
|
| 1701 |
+
"grad_norm": 27.83361053466797,
|
| 1702 |
+
"learning_rate": 8.800315706393054e-06,
|
| 1703 |
+
"loss": 0.2213,
|
| 1704 |
+
"step": 2370
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 3.0951704545454546,
|
| 1708 |
+
"grad_norm": 0.04297183081507683,
|
| 1709 |
+
"learning_rate": 8.602999210734018e-06,
|
| 1710 |
+
"loss": 0.1469,
|
| 1711 |
+
"step": 2380
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 3.098721590909091,
|
| 1715 |
+
"grad_norm": 0.7070327401161194,
|
| 1716 |
+
"learning_rate": 8.405682715074981e-06,
|
| 1717 |
+
"loss": 0.0219,
|
| 1718 |
+
"step": 2390
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 3.102272727272727,
|
| 1722 |
+
"grad_norm": 0.22129130363464355,
|
| 1723 |
+
"learning_rate": 8.208366219415943e-06,
|
| 1724 |
+
"loss": 0.0115,
|
| 1725 |
+
"step": 2400
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 3.1058238636363638,
|
| 1729 |
+
"grad_norm": 0.8595547080039978,
|
| 1730 |
+
"learning_rate": 8.011049723756906e-06,
|
| 1731 |
+
"loss": 0.0124,
|
| 1732 |
+
"step": 2410
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 3.109375,
|
| 1736 |
+
"grad_norm": 0.13839711248874664,
|
| 1737 |
+
"learning_rate": 7.81373322809787e-06,
|
| 1738 |
+
"loss": 0.0865,
|
| 1739 |
+
"step": 2420
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 3.1129261363636362,
|
| 1743 |
+
"grad_norm": 0.04325387626886368,
|
| 1744 |
+
"learning_rate": 7.6164167324388314e-06,
|
| 1745 |
+
"loss": 0.0793,
|
| 1746 |
+
"step": 2430
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 3.116477272727273,
|
| 1750 |
+
"grad_norm": 3.90751051902771,
|
| 1751 |
+
"learning_rate": 7.419100236779796e-06,
|
| 1752 |
+
"loss": 0.1025,
|
| 1753 |
+
"step": 2440
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 3.120028409090909,
|
| 1757 |
+
"grad_norm": 0.03920091316103935,
|
| 1758 |
+
"learning_rate": 7.2217837411207575e-06,
|
| 1759 |
+
"loss": 0.2589,
|
| 1760 |
+
"step": 2450
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 3.1235795454545454,
|
| 1764 |
+
"grad_norm": 0.2701418995857239,
|
| 1765 |
+
"learning_rate": 7.02446724546172e-06,
|
| 1766 |
+
"loss": 0.0066,
|
| 1767 |
+
"step": 2460
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 3.1271306818181817,
|
| 1771 |
+
"grad_norm": 0.3685694634914398,
|
| 1772 |
+
"learning_rate": 6.827150749802684e-06,
|
| 1773 |
+
"loss": 0.0537,
|
| 1774 |
+
"step": 2470
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 3.1306818181818183,
|
| 1778 |
+
"grad_norm": 0.03167586028575897,
|
| 1779 |
+
"learning_rate": 6.629834254143646e-06,
|
| 1780 |
+
"loss": 0.2329,
|
| 1781 |
+
"step": 2480
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 3.1342329545454546,
|
| 1785 |
+
"grad_norm": 24.94156265258789,
|
| 1786 |
+
"learning_rate": 6.4325177584846105e-06,
|
| 1787 |
+
"loss": 0.113,
|
| 1788 |
+
"step": 2490
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 3.137784090909091,
|
| 1792 |
+
"grad_norm": 2.3878488540649414,
|
| 1793 |
+
"learning_rate": 6.235201262825572e-06,
|
| 1794 |
+
"loss": 0.0289,
|
| 1795 |
+
"step": 2500
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 3.141335227272727,
|
| 1799 |
+
"grad_norm": 0.1500893086194992,
|
| 1800 |
+
"learning_rate": 6.037884767166536e-06,
|
| 1801 |
+
"loss": 0.0164,
|
| 1802 |
+
"step": 2510
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 3.1448863636363638,
|
| 1806 |
+
"grad_norm": 1.0166605710983276,
|
| 1807 |
+
"learning_rate": 5.840568271507498e-06,
|
| 1808 |
+
"loss": 0.0665,
|
| 1809 |
+
"step": 2520
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 3.1484375,
|
| 1813 |
+
"grad_norm": 8.010841369628906,
|
| 1814 |
+
"learning_rate": 5.643251775848461e-06,
|
| 1815 |
+
"loss": 0.2883,
|
| 1816 |
+
"step": 2530
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 3.1519886363636362,
|
| 1820 |
+
"grad_norm": 11.449795722961426,
|
| 1821 |
+
"learning_rate": 5.445935280189424e-06,
|
| 1822 |
+
"loss": 0.0892,
|
| 1823 |
+
"step": 2540
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 3.155539772727273,
|
| 1827 |
+
"grad_norm": 41.279239654541016,
|
| 1828 |
+
"learning_rate": 5.248618784530387e-06,
|
| 1829 |
+
"loss": 0.0529,
|
| 1830 |
+
"step": 2550
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 3.159090909090909,
|
| 1834 |
+
"grad_norm": 0.06051742658019066,
|
| 1835 |
+
"learning_rate": 5.05130228887135e-06,
|
| 1836 |
+
"loss": 0.1401,
|
| 1837 |
+
"step": 2560
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 3.1626420454545454,
|
| 1841 |
+
"grad_norm": 0.08330044150352478,
|
| 1842 |
+
"learning_rate": 4.853985793212313e-06,
|
| 1843 |
+
"loss": 0.0411,
|
| 1844 |
+
"step": 2570
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 3.1661931818181817,
|
| 1848 |
+
"grad_norm": 44.56931686401367,
|
| 1849 |
+
"learning_rate": 4.656669297553276e-06,
|
| 1850 |
+
"loss": 0.0657,
|
| 1851 |
+
"step": 2580
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 3.1697443181818183,
|
| 1855 |
+
"grad_norm": 1.3760157823562622,
|
| 1856 |
+
"learning_rate": 4.459352801894238e-06,
|
| 1857 |
+
"loss": 0.1188,
|
| 1858 |
+
"step": 2590
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 3.1732954545454546,
|
| 1862 |
+
"grad_norm": 0.1809810996055603,
|
| 1863 |
+
"learning_rate": 4.262036306235202e-06,
|
| 1864 |
+
"loss": 0.0305,
|
| 1865 |
+
"step": 2600
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 3.176846590909091,
|
| 1869 |
+
"grad_norm": 1.571562647819519,
|
| 1870 |
+
"learning_rate": 4.064719810576164e-06,
|
| 1871 |
+
"loss": 0.0448,
|
| 1872 |
+
"step": 2610
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 3.180397727272727,
|
| 1876 |
+
"grad_norm": 0.04679562523961067,
|
| 1877 |
+
"learning_rate": 3.867403314917127e-06,
|
| 1878 |
+
"loss": 0.0876,
|
| 1879 |
+
"step": 2620
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 3.1839488636363638,
|
| 1883 |
+
"grad_norm": 0.40420079231262207,
|
| 1884 |
+
"learning_rate": 3.67008681925809e-06,
|
| 1885 |
+
"loss": 0.0399,
|
| 1886 |
+
"step": 2630
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 3.1875,
|
| 1890 |
+
"grad_norm": 0.8966354727745056,
|
| 1891 |
+
"learning_rate": 3.472770323599053e-06,
|
| 1892 |
+
"loss": 0.0747,
|
| 1893 |
+
"step": 2640
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.1910511363636362,
|
| 1897 |
+
"grad_norm": 0.499515563249588,
|
| 1898 |
+
"learning_rate": 3.2754538279400157e-06,
|
| 1899 |
+
"loss": 0.1019,
|
| 1900 |
+
"step": 2650
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 3.194602272727273,
|
| 1904 |
+
"grad_norm": 0.09785713255405426,
|
| 1905 |
+
"learning_rate": 3.0781373322809787e-06,
|
| 1906 |
+
"loss": 0.032,
|
| 1907 |
+
"step": 2660
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 3.198153409090909,
|
| 1911 |
+
"grad_norm": 0.037965867668390274,
|
| 1912 |
+
"learning_rate": 2.8808208366219414e-06,
|
| 1913 |
+
"loss": 0.0693,
|
| 1914 |
+
"step": 2670
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 3.2017045454545454,
|
| 1918 |
+
"grad_norm": 0.5521709322929382,
|
| 1919 |
+
"learning_rate": 2.683504340962905e-06,
|
| 1920 |
+
"loss": 0.0761,
|
| 1921 |
+
"step": 2680
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 3.2052556818181817,
|
| 1925 |
+
"grad_norm": 0.09314560890197754,
|
| 1926 |
+
"learning_rate": 2.4861878453038674e-06,
|
| 1927 |
+
"loss": 0.1744,
|
| 1928 |
+
"step": 2690
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 3.2088068181818183,
|
| 1932 |
+
"grad_norm": 12.886544227600098,
|
| 1933 |
+
"learning_rate": 2.2888713496448305e-06,
|
| 1934 |
+
"loss": 0.1549,
|
| 1935 |
+
"step": 2700
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.2123579545454546,
|
| 1939 |
+
"grad_norm": 0.18088915944099426,
|
| 1940 |
+
"learning_rate": 2.091554853985793e-06,
|
| 1941 |
+
"loss": 0.186,
|
| 1942 |
+
"step": 2710
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 3.215909090909091,
|
| 1946 |
+
"grad_norm": 0.18304170668125153,
|
| 1947 |
+
"learning_rate": 1.8942383583267563e-06,
|
| 1948 |
+
"loss": 0.0137,
|
| 1949 |
+
"step": 2720
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 3.219460227272727,
|
| 1953 |
+
"grad_norm": 5.272530555725098,
|
| 1954 |
+
"learning_rate": 1.6969218626677192e-06,
|
| 1955 |
+
"loss": 0.0369,
|
| 1956 |
+
"step": 2730
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 3.2230113636363638,
|
| 1960 |
+
"grad_norm": 0.17552484571933746,
|
| 1961 |
+
"learning_rate": 1.499605367008682e-06,
|
| 1962 |
+
"loss": 0.0675,
|
| 1963 |
+
"step": 2740
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 3.2265625,
|
| 1967 |
+
"grad_norm": 0.28992605209350586,
|
| 1968 |
+
"learning_rate": 1.3022888713496448e-06,
|
| 1969 |
+
"loss": 0.0744,
|
| 1970 |
+
"step": 2750
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 3.2301136363636362,
|
| 1974 |
+
"grad_norm": 0.17767734825611115,
|
| 1975 |
+
"learning_rate": 1.1049723756906078e-06,
|
| 1976 |
+
"loss": 0.1551,
|
| 1977 |
+
"step": 2760
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 3.233664772727273,
|
| 1981 |
+
"grad_norm": 0.43689748644828796,
|
| 1982 |
+
"learning_rate": 9.076558800315707e-07,
|
| 1983 |
+
"loss": 0.1754,
|
| 1984 |
+
"step": 2770
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 3.237215909090909,
|
| 1988 |
+
"grad_norm": 0.956597089767456,
|
| 1989 |
+
"learning_rate": 7.103393843725336e-07,
|
| 1990 |
+
"loss": 0.0069,
|
| 1991 |
+
"step": 2780
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 3.2407670454545454,
|
| 1995 |
+
"grad_norm": 25.514036178588867,
|
| 1996 |
+
"learning_rate": 5.130228887134964e-07,
|
| 1997 |
+
"loss": 0.0997,
|
| 1998 |
+
"step": 2790
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 3.2443181818181817,
|
| 2002 |
+
"grad_norm": 3.5862650871276855,
|
| 2003 |
+
"learning_rate": 3.1570639305445935e-07,
|
| 2004 |
+
"loss": 0.1085,
|
| 2005 |
+
"step": 2800
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 3.2478693181818183,
|
| 2009 |
+
"grad_norm": 0.035997312515974045,
|
| 2010 |
+
"learning_rate": 1.1838989739542227e-07,
|
| 2011 |
+
"loss": 0.0289,
|
| 2012 |
+
"step": 2810
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 3.25,
|
| 2016 |
+
"eval_accuracy": 0.9879931389365352,
|
| 2017 |
+
"eval_f1": 0.9879774481091409,
|
| 2018 |
+
"eval_loss": 0.037413839250802994,
|
| 2019 |
+
"eval_precision": 0.9880848996629442,
|
| 2020 |
+
"eval_recall": 0.9879997453526865,
|
| 2021 |
+
"eval_runtime": 692.1711,
|
| 2022 |
+
"eval_samples_per_second": 0.842,
|
| 2023 |
+
"eval_steps_per_second": 0.211,
|
| 2024 |
+
"step": 2816
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 3.25,
|
| 2028 |
+
"step": 2816,
|
| 2029 |
+
"total_flos": 1.4036798084625727e+19,
|
| 2030 |
+
"train_loss": 0.6578278151044453,
|
| 2031 |
+
"train_runtime": 18231.9061,
|
| 2032 |
+
"train_samples_per_second": 0.618,
|
| 2033 |
+
"train_steps_per_second": 0.154
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 3.25,
|
| 2037 |
+
"eval_accuracy": 0.9558620689655173,
|
| 2038 |
+
"eval_f1": 0.9562337987051189,
|
| 2039 |
+
"eval_loss": 0.14939993619918823,
|
| 2040 |
+
"eval_precision": 0.9567608190198278,
|
| 2041 |
+
"eval_recall": 0.956465718392097,
|
| 2042 |
+
"eval_runtime": 911.348,
|
| 2043 |
+
"eval_samples_per_second": 0.796,
|
| 2044 |
+
"eval_steps_per_second": 0.2,
|
| 2045 |
+
"step": 2816
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 3.25,
|
| 2049 |
+
"eval_accuracy": 0.9558620689655173,
|
| 2050 |
+
"eval_f1": 0.9562337987051189,
|
| 2051 |
+
"eval_loss": 0.14939993619918823,
|
| 2052 |
+
"eval_precision": 0.9567608190198278,
|
| 2053 |
+
"eval_recall": 0.956465718392097,
|
| 2054 |
+
"eval_runtime": 930.9063,
|
| 2055 |
+
"eval_samples_per_second": 0.779,
|
| 2056 |
+
"eval_steps_per_second": 0.196,
|
| 2057 |
+
"step": 2816
|
| 2058 |
+
}
|
| 2059 |
+
],
|
| 2060 |
+
"logging_steps": 10,
|
| 2061 |
+
"max_steps": 2816,
|
| 2062 |
+
"num_input_tokens_seen": 0,
|
| 2063 |
+
"num_train_epochs": 9223372036854775807,
|
| 2064 |
+
"save_steps": 500,
|
| 2065 |
+
"total_flos": 1.4036798084625727e+19,
|
| 2066 |
+
"train_batch_size": 4,
|
| 2067 |
+
"trial_name": null,
|
| 2068 |
+
"trial_params": null
|
| 2069 |
+
}
|