End of training
Browse files- all_results.json +16 -0
- eval_results.json +11 -0
- runs/Aug31_14-32-13_bad3846a2b09/events.out.tfevents.1725122684.bad3846a2b09.1416.4 +3 -0
- train_results.json +8 -0
- trainer_state.json +1359 -0
all_results.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.6991409248766222,
|
| 4 |
+
"eval_f1": 0.6766571542539821,
|
| 5 |
+
"eval_loss": 0.7419390082359314,
|
| 6 |
+
"eval_precision": 0.6830062110035815,
|
| 7 |
+
"eval_recall": 0.6991409248766222,
|
| 8 |
+
"eval_runtime": 80.0564,
|
| 9 |
+
"eval_samples_per_second": 136.679,
|
| 10 |
+
"eval_steps_per_second": 2.136,
|
| 11 |
+
"total_flos": 3.3914202248568177e+19,
|
| 12 |
+
"train_loss": 0.790359598014787,
|
| 13 |
+
"train_runtime": 7828.8109,
|
| 14 |
+
"train_samples_per_second": 55.901,
|
| 15 |
+
"train_steps_per_second": 0.218
|
| 16 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.6991409248766222,
|
| 4 |
+
"eval_f1": 0.6766571542539821,
|
| 5 |
+
"eval_loss": 0.7419390082359314,
|
| 6 |
+
"eval_precision": 0.6830062110035815,
|
| 7 |
+
"eval_recall": 0.6991409248766222,
|
| 8 |
+
"eval_runtime": 80.0564,
|
| 9 |
+
"eval_samples_per_second": 136.679,
|
| 10 |
+
"eval_steps_per_second": 2.136
|
| 11 |
+
}
|
runs/Aug31_14-32-13_bad3846a2b09/events.out.tfevents.1725122684.bad3846a2b09.1416.4
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0da12348d4574c75322deaf8b4d24e9e33f0f9a22759a7366d25f7b52f7a05b2
|
| 3 |
+
size 560
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"total_flos": 3.3914202248568177e+19,
|
| 4 |
+
"train_loss": 0.790359598014787,
|
| 5 |
+
"train_runtime": 7828.8109,
|
| 6 |
+
"train_samples_per_second": 55.901,
|
| 7 |
+
"train_steps_per_second": 0.218
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1359 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6991409248766222,
|
| 3 |
+
"best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-hongrui_mammogram_v_1/checkpoint-1710",
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1710,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.05847953216374269,
|
| 13 |
+
"grad_norm": 1.5846091508865356,
|
| 14 |
+
"learning_rate": 2.9239766081871343e-06,
|
| 15 |
+
"loss": 1.3844,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.11695906432748537,
|
| 20 |
+
"grad_norm": 1.26529061794281,
|
| 21 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 22 |
+
"loss": 1.3401,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.17543859649122806,
|
| 27 |
+
"grad_norm": 1.0668294429779053,
|
| 28 |
+
"learning_rate": 8.771929824561403e-06,
|
| 29 |
+
"loss": 1.2478,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.23391812865497075,
|
| 34 |
+
"grad_norm": 0.7071924805641174,
|
| 35 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 36 |
+
"loss": 1.1693,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.29239766081871343,
|
| 41 |
+
"grad_norm": 0.7040536403656006,
|
| 42 |
+
"learning_rate": 1.4619883040935673e-05,
|
| 43 |
+
"loss": 1.0874,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.3508771929824561,
|
| 48 |
+
"grad_norm": 0.5056448578834534,
|
| 49 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 50 |
+
"loss": 1.0551,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.4093567251461988,
|
| 55 |
+
"grad_norm": 0.6976526975631714,
|
| 56 |
+
"learning_rate": 2.046783625730994e-05,
|
| 57 |
+
"loss": 1.0349,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.4678362573099415,
|
| 62 |
+
"grad_norm": 0.6909885406494141,
|
| 63 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 64 |
+
"loss": 0.9862,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.5263157894736842,
|
| 69 |
+
"grad_norm": 1.2905045747756958,
|
| 70 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 71 |
+
"loss": 0.9897,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.5847953216374269,
|
| 76 |
+
"grad_norm": 1.0667847394943237,
|
| 77 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 78 |
+
"loss": 0.9316,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.6432748538011696,
|
| 83 |
+
"grad_norm": 0.7333235144615173,
|
| 84 |
+
"learning_rate": 3.216374269005848e-05,
|
| 85 |
+
"loss": 0.9224,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.7017543859649122,
|
| 90 |
+
"grad_norm": 0.8830112814903259,
|
| 91 |
+
"learning_rate": 3.508771929824561e-05,
|
| 92 |
+
"loss": 0.9127,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.7602339181286549,
|
| 97 |
+
"grad_norm": 0.7408013343811035,
|
| 98 |
+
"learning_rate": 3.8011695906432746e-05,
|
| 99 |
+
"loss": 0.9349,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.8187134502923976,
|
| 104 |
+
"grad_norm": 0.6464580297470093,
|
| 105 |
+
"learning_rate": 4.093567251461988e-05,
|
| 106 |
+
"loss": 0.9046,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.8771929824561403,
|
| 111 |
+
"grad_norm": 0.9568632245063782,
|
| 112 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 113 |
+
"loss": 0.8981,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.935672514619883,
|
| 118 |
+
"grad_norm": 0.755339503288269,
|
| 119 |
+
"learning_rate": 4.678362573099415e-05,
|
| 120 |
+
"loss": 0.9012,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.9941520467836257,
|
| 125 |
+
"grad_norm": 0.8921021819114685,
|
| 126 |
+
"learning_rate": 4.970760233918128e-05,
|
| 127 |
+
"loss": 0.8576,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.0,
|
| 132 |
+
"eval_accuracy": 0.6677938219703893,
|
| 133 |
+
"eval_f1": 0.6066566978834349,
|
| 134 |
+
"eval_loss": 0.8430724143981934,
|
| 135 |
+
"eval_precision": 0.7751300124185229,
|
| 136 |
+
"eval_recall": 0.6677938219703893,
|
| 137 |
+
"eval_runtime": 78.2201,
|
| 138 |
+
"eval_samples_per_second": 139.887,
|
| 139 |
+
"eval_steps_per_second": 2.186,
|
| 140 |
+
"step": 171
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.0526315789473684,
|
| 144 |
+
"grad_norm": 0.7332887053489685,
|
| 145 |
+
"learning_rate": 4.970760233918128e-05,
|
| 146 |
+
"loss": 0.8527,
|
| 147 |
+
"step": 180
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.1111111111111112,
|
| 151 |
+
"grad_norm": 0.776443362236023,
|
| 152 |
+
"learning_rate": 4.938271604938271e-05,
|
| 153 |
+
"loss": 0.8561,
|
| 154 |
+
"step": 190
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.1695906432748537,
|
| 158 |
+
"grad_norm": 0.7149679660797119,
|
| 159 |
+
"learning_rate": 4.9057829759584143e-05,
|
| 160 |
+
"loss": 0.8435,
|
| 161 |
+
"step": 200
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.2280701754385965,
|
| 165 |
+
"grad_norm": 0.8708255290985107,
|
| 166 |
+
"learning_rate": 4.8732943469785574e-05,
|
| 167 |
+
"loss": 0.8332,
|
| 168 |
+
"step": 210
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.286549707602339,
|
| 172 |
+
"grad_norm": 0.8141400814056396,
|
| 173 |
+
"learning_rate": 4.8408057179987004e-05,
|
| 174 |
+
"loss": 0.8377,
|
| 175 |
+
"step": 220
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.345029239766082,
|
| 179 |
+
"grad_norm": 1.188214659690857,
|
| 180 |
+
"learning_rate": 4.8083170890188434e-05,
|
| 181 |
+
"loss": 0.8456,
|
| 182 |
+
"step": 230
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.4035087719298245,
|
| 186 |
+
"grad_norm": 0.6589232087135315,
|
| 187 |
+
"learning_rate": 4.7758284600389865e-05,
|
| 188 |
+
"loss": 0.8387,
|
| 189 |
+
"step": 240
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.4619883040935673,
|
| 193 |
+
"grad_norm": 0.9304301142692566,
|
| 194 |
+
"learning_rate": 4.7433398310591295e-05,
|
| 195 |
+
"loss": 0.8521,
|
| 196 |
+
"step": 250
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.52046783625731,
|
| 200 |
+
"grad_norm": 0.8288267254829407,
|
| 201 |
+
"learning_rate": 4.7108512020792725e-05,
|
| 202 |
+
"loss": 0.8453,
|
| 203 |
+
"step": 260
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.5789473684210527,
|
| 207 |
+
"grad_norm": 0.8100181221961975,
|
| 208 |
+
"learning_rate": 4.678362573099415e-05,
|
| 209 |
+
"loss": 0.8398,
|
| 210 |
+
"step": 270
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.6374269005847952,
|
| 214 |
+
"grad_norm": 1.2519994974136353,
|
| 215 |
+
"learning_rate": 4.645873944119558e-05,
|
| 216 |
+
"loss": 0.835,
|
| 217 |
+
"step": 280
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.695906432748538,
|
| 221 |
+
"grad_norm": 0.868988573551178,
|
| 222 |
+
"learning_rate": 4.613385315139701e-05,
|
| 223 |
+
"loss": 0.8,
|
| 224 |
+
"step": 290
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.7543859649122808,
|
| 228 |
+
"grad_norm": 0.7534909844398499,
|
| 229 |
+
"learning_rate": 4.580896686159844e-05,
|
| 230 |
+
"loss": 0.8103,
|
| 231 |
+
"step": 300
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.8128654970760234,
|
| 235 |
+
"grad_norm": 0.9954193234443665,
|
| 236 |
+
"learning_rate": 4.548408057179987e-05,
|
| 237 |
+
"loss": 0.8065,
|
| 238 |
+
"step": 310
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.871345029239766,
|
| 242 |
+
"grad_norm": 0.6459550857543945,
|
| 243 |
+
"learning_rate": 4.51591942820013e-05,
|
| 244 |
+
"loss": 0.8196,
|
| 245 |
+
"step": 320
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.9298245614035088,
|
| 249 |
+
"grad_norm": 0.9698415398597717,
|
| 250 |
+
"learning_rate": 4.483430799220273e-05,
|
| 251 |
+
"loss": 0.8464,
|
| 252 |
+
"step": 330
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.9883040935672516,
|
| 256 |
+
"grad_norm": 0.9862537980079651,
|
| 257 |
+
"learning_rate": 4.450942170240416e-05,
|
| 258 |
+
"loss": 0.8297,
|
| 259 |
+
"step": 340
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 2.0,
|
| 263 |
+
"eval_accuracy": 0.6791263023213306,
|
| 264 |
+
"eval_f1": 0.6182249859592962,
|
| 265 |
+
"eval_loss": 0.796485424041748,
|
| 266 |
+
"eval_precision": 0.6757921489303887,
|
| 267 |
+
"eval_recall": 0.6791263023213306,
|
| 268 |
+
"eval_runtime": 78.3952,
|
| 269 |
+
"eval_samples_per_second": 139.575,
|
| 270 |
+
"eval_steps_per_second": 2.181,
|
| 271 |
+
"step": 342
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 2.046783625730994,
|
| 275 |
+
"grad_norm": 0.7553840279579163,
|
| 276 |
+
"learning_rate": 4.418453541260559e-05,
|
| 277 |
+
"loss": 0.8262,
|
| 278 |
+
"step": 350
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.1052631578947367,
|
| 282 |
+
"grad_norm": 0.8275452852249146,
|
| 283 |
+
"learning_rate": 4.3859649122807014e-05,
|
| 284 |
+
"loss": 0.809,
|
| 285 |
+
"step": 360
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 2.1637426900584797,
|
| 289 |
+
"grad_norm": 0.6450644731521606,
|
| 290 |
+
"learning_rate": 4.3534762833008445e-05,
|
| 291 |
+
"loss": 0.7939,
|
| 292 |
+
"step": 370
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 2.2222222222222223,
|
| 296 |
+
"grad_norm": 0.6809207201004028,
|
| 297 |
+
"learning_rate": 4.3209876543209875e-05,
|
| 298 |
+
"loss": 0.8157,
|
| 299 |
+
"step": 380
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 2.280701754385965,
|
| 303 |
+
"grad_norm": 0.9741197228431702,
|
| 304 |
+
"learning_rate": 4.2884990253411305e-05,
|
| 305 |
+
"loss": 0.8126,
|
| 306 |
+
"step": 390
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 2.3391812865497075,
|
| 310 |
+
"grad_norm": 0.6929029822349548,
|
| 311 |
+
"learning_rate": 4.2560103963612735e-05,
|
| 312 |
+
"loss": 0.8155,
|
| 313 |
+
"step": 400
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 2.39766081871345,
|
| 317 |
+
"grad_norm": 0.836373507976532,
|
| 318 |
+
"learning_rate": 4.2235217673814166e-05,
|
| 319 |
+
"loss": 0.8145,
|
| 320 |
+
"step": 410
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 2.456140350877193,
|
| 324 |
+
"grad_norm": 0.7719972729682922,
|
| 325 |
+
"learning_rate": 4.1910331384015596e-05,
|
| 326 |
+
"loss": 0.7963,
|
| 327 |
+
"step": 420
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 2.5146198830409356,
|
| 331 |
+
"grad_norm": 0.5631088018417358,
|
| 332 |
+
"learning_rate": 4.1585445094217026e-05,
|
| 333 |
+
"loss": 0.7839,
|
| 334 |
+
"step": 430
|
| 335 |
+
},
|
| 336 |
+
{
|
| 337 |
+
"epoch": 2.573099415204678,
|
| 338 |
+
"grad_norm": 0.7167489528656006,
|
| 339 |
+
"learning_rate": 4.1260558804418457e-05,
|
| 340 |
+
"loss": 0.7837,
|
| 341 |
+
"step": 440
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 2.6315789473684212,
|
| 345 |
+
"grad_norm": 0.8713414072990417,
|
| 346 |
+
"learning_rate": 4.093567251461988e-05,
|
| 347 |
+
"loss": 0.8046,
|
| 348 |
+
"step": 450
|
| 349 |
+
},
|
| 350 |
+
{
|
| 351 |
+
"epoch": 2.690058479532164,
|
| 352 |
+
"grad_norm": 1.0731910467147827,
|
| 353 |
+
"learning_rate": 4.061078622482131e-05,
|
| 354 |
+
"loss": 0.7813,
|
| 355 |
+
"step": 460
|
| 356 |
+
},
|
| 357 |
+
{
|
| 358 |
+
"epoch": 2.7485380116959064,
|
| 359 |
+
"grad_norm": 0.6702953577041626,
|
| 360 |
+
"learning_rate": 4.028589993502274e-05,
|
| 361 |
+
"loss": 0.8076,
|
| 362 |
+
"step": 470
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.807017543859649,
|
| 366 |
+
"grad_norm": 0.6061651706695557,
|
| 367 |
+
"learning_rate": 3.996101364522417e-05,
|
| 368 |
+
"loss": 0.786,
|
| 369 |
+
"step": 480
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.8654970760233915,
|
| 373 |
+
"grad_norm": 0.9429291486740112,
|
| 374 |
+
"learning_rate": 3.96361273554256e-05,
|
| 375 |
+
"loss": 0.817,
|
| 376 |
+
"step": 490
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 2.9239766081871346,
|
| 380 |
+
"grad_norm": 0.7229118347167969,
|
| 381 |
+
"learning_rate": 3.931124106562703e-05,
|
| 382 |
+
"loss": 0.8003,
|
| 383 |
+
"step": 500
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 2.982456140350877,
|
| 387 |
+
"grad_norm": 0.702900230884552,
|
| 388 |
+
"learning_rate": 3.898635477582846e-05,
|
| 389 |
+
"loss": 0.8303,
|
| 390 |
+
"step": 510
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 3.0,
|
| 394 |
+
"eval_accuracy": 0.6842441966733687,
|
| 395 |
+
"eval_f1": 0.6360238176585905,
|
| 396 |
+
"eval_loss": 0.7872248888015747,
|
| 397 |
+
"eval_precision": 0.6704334160874683,
|
| 398 |
+
"eval_recall": 0.6842441966733687,
|
| 399 |
+
"eval_runtime": 77.468,
|
| 400 |
+
"eval_samples_per_second": 141.245,
|
| 401 |
+
"eval_steps_per_second": 2.207,
|
| 402 |
+
"step": 513
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 3.0409356725146197,
|
| 406 |
+
"grad_norm": 0.6098015308380127,
|
| 407 |
+
"learning_rate": 3.866146848602989e-05,
|
| 408 |
+
"loss": 0.7717,
|
| 409 |
+
"step": 520
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 3.0994152046783627,
|
| 413 |
+
"grad_norm": 0.7494928240776062,
|
| 414 |
+
"learning_rate": 3.8336582196231315e-05,
|
| 415 |
+
"loss": 0.7807,
|
| 416 |
+
"step": 530
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 3.1578947368421053,
|
| 420 |
+
"grad_norm": 0.8039376735687256,
|
| 421 |
+
"learning_rate": 3.8011695906432746e-05,
|
| 422 |
+
"loss": 0.79,
|
| 423 |
+
"step": 540
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 3.216374269005848,
|
| 427 |
+
"grad_norm": 0.687044084072113,
|
| 428 |
+
"learning_rate": 3.7686809616634176e-05,
|
| 429 |
+
"loss": 0.8004,
|
| 430 |
+
"step": 550
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 3.2748538011695905,
|
| 434 |
+
"grad_norm": 0.9368821978569031,
|
| 435 |
+
"learning_rate": 3.7361923326835606e-05,
|
| 436 |
+
"loss": 0.8002,
|
| 437 |
+
"step": 560
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 3.3333333333333335,
|
| 441 |
+
"grad_norm": 1.0392132997512817,
|
| 442 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 443 |
+
"loss": 0.8042,
|
| 444 |
+
"step": 570
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 3.391812865497076,
|
| 448 |
+
"grad_norm": 0.5632928013801575,
|
| 449 |
+
"learning_rate": 3.671215074723847e-05,
|
| 450 |
+
"loss": 0.7746,
|
| 451 |
+
"step": 580
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 3.4502923976608186,
|
| 455 |
+
"grad_norm": 0.6281052231788635,
|
| 456 |
+
"learning_rate": 3.63872644574399e-05,
|
| 457 |
+
"loss": 0.7691,
|
| 458 |
+
"step": 590
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 3.5087719298245617,
|
| 462 |
+
"grad_norm": 0.6072255969047546,
|
| 463 |
+
"learning_rate": 3.606237816764133e-05,
|
| 464 |
+
"loss": 0.7552,
|
| 465 |
+
"step": 600
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 3.5672514619883042,
|
| 469 |
+
"grad_norm": 0.6243124604225159,
|
| 470 |
+
"learning_rate": 3.573749187784276e-05,
|
| 471 |
+
"loss": 0.7905,
|
| 472 |
+
"step": 610
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 3.625730994152047,
|
| 476 |
+
"grad_norm": 0.8322011828422546,
|
| 477 |
+
"learning_rate": 3.541260558804418e-05,
|
| 478 |
+
"loss": 0.7772,
|
| 479 |
+
"step": 620
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 3.6842105263157894,
|
| 483 |
+
"grad_norm": 0.9370966553688049,
|
| 484 |
+
"learning_rate": 3.508771929824561e-05,
|
| 485 |
+
"loss": 0.7683,
|
| 486 |
+
"step": 630
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 3.742690058479532,
|
| 490 |
+
"grad_norm": 0.6632276177406311,
|
| 491 |
+
"learning_rate": 3.476283300844704e-05,
|
| 492 |
+
"loss": 0.7853,
|
| 493 |
+
"step": 640
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 3.801169590643275,
|
| 497 |
+
"grad_norm": 0.678115963935852,
|
| 498 |
+
"learning_rate": 3.443794671864847e-05,
|
| 499 |
+
"loss": 0.7691,
|
| 500 |
+
"step": 650
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 3.8596491228070176,
|
| 504 |
+
"grad_norm": 0.6135697960853577,
|
| 505 |
+
"learning_rate": 3.41130604288499e-05,
|
| 506 |
+
"loss": 0.7778,
|
| 507 |
+
"step": 660
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 3.91812865497076,
|
| 511 |
+
"grad_norm": 0.6042221188545227,
|
| 512 |
+
"learning_rate": 3.378817413905133e-05,
|
| 513 |
+
"loss": 0.7983,
|
| 514 |
+
"step": 670
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 3.976608187134503,
|
| 518 |
+
"grad_norm": 0.8058825135231018,
|
| 519 |
+
"learning_rate": 3.346328784925276e-05,
|
| 520 |
+
"loss": 0.7814,
|
| 521 |
+
"step": 680
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 4.0,
|
| 525 |
+
"eval_accuracy": 0.6843355876439408,
|
| 526 |
+
"eval_f1": 0.659727957333552,
|
| 527 |
+
"eval_loss": 0.7717081308364868,
|
| 528 |
+
"eval_precision": 0.6601378633948723,
|
| 529 |
+
"eval_recall": 0.6843355876439408,
|
| 530 |
+
"eval_runtime": 78.1449,
|
| 531 |
+
"eval_samples_per_second": 140.022,
|
| 532 |
+
"eval_steps_per_second": 2.188,
|
| 533 |
+
"step": 684
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 4.035087719298246,
|
| 537 |
+
"grad_norm": 0.966820478439331,
|
| 538 |
+
"learning_rate": 3.313840155945419e-05,
|
| 539 |
+
"loss": 0.7736,
|
| 540 |
+
"step": 690
|
| 541 |
+
},
|
| 542 |
+
{
|
| 543 |
+
"epoch": 4.093567251461988,
|
| 544 |
+
"grad_norm": 1.1176632642745972,
|
| 545 |
+
"learning_rate": 3.281351526965562e-05,
|
| 546 |
+
"loss": 0.7529,
|
| 547 |
+
"step": 700
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 4.152046783625731,
|
| 551 |
+
"grad_norm": 0.6780201196670532,
|
| 552 |
+
"learning_rate": 3.248862897985705e-05,
|
| 553 |
+
"loss": 0.7722,
|
| 554 |
+
"step": 710
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 4.2105263157894735,
|
| 558 |
+
"grad_norm": 0.8257865905761719,
|
| 559 |
+
"learning_rate": 3.216374269005848e-05,
|
| 560 |
+
"loss": 0.771,
|
| 561 |
+
"step": 720
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 4.269005847953216,
|
| 565 |
+
"grad_norm": 0.6105393767356873,
|
| 566 |
+
"learning_rate": 3.183885640025991e-05,
|
| 567 |
+
"loss": 0.7883,
|
| 568 |
+
"step": 730
|
| 569 |
+
},
|
| 570 |
+
{
|
| 571 |
+
"epoch": 4.3274853801169595,
|
| 572 |
+
"grad_norm": 0.8537980914115906,
|
| 573 |
+
"learning_rate": 3.151397011046134e-05,
|
| 574 |
+
"loss": 0.7698,
|
| 575 |
+
"step": 740
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 4.385964912280702,
|
| 579 |
+
"grad_norm": 0.8124959468841553,
|
| 580 |
+
"learning_rate": 3.118908382066277e-05,
|
| 581 |
+
"loss": 0.7737,
|
| 582 |
+
"step": 750
|
| 583 |
+
},
|
| 584 |
+
{
|
| 585 |
+
"epoch": 4.444444444444445,
|
| 586 |
+
"grad_norm": 0.8992810845375061,
|
| 587 |
+
"learning_rate": 3.08641975308642e-05,
|
| 588 |
+
"loss": 0.7809,
|
| 589 |
+
"step": 760
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 4.502923976608187,
|
| 593 |
+
"grad_norm": 0.6706241965293884,
|
| 594 |
+
"learning_rate": 3.053931124106563e-05,
|
| 595 |
+
"loss": 0.7629,
|
| 596 |
+
"step": 770
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 4.56140350877193,
|
| 600 |
+
"grad_norm": 0.8107186555862427,
|
| 601 |
+
"learning_rate": 3.0214424951267055e-05,
|
| 602 |
+
"loss": 0.7683,
|
| 603 |
+
"step": 780
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 4.619883040935672,
|
| 607 |
+
"grad_norm": 0.7054151296615601,
|
| 608 |
+
"learning_rate": 2.9889538661468486e-05,
|
| 609 |
+
"loss": 0.7367,
|
| 610 |
+
"step": 790
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 4.678362573099415,
|
| 614 |
+
"grad_norm": 0.7284806966781616,
|
| 615 |
+
"learning_rate": 2.9564652371669916e-05,
|
| 616 |
+
"loss": 0.7476,
|
| 617 |
+
"step": 800
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 4.7368421052631575,
|
| 621 |
+
"grad_norm": 1.1839812994003296,
|
| 622 |
+
"learning_rate": 2.9239766081871346e-05,
|
| 623 |
+
"loss": 0.7565,
|
| 624 |
+
"step": 810
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 4.7953216374269,
|
| 628 |
+
"grad_norm": 0.7781530618667603,
|
| 629 |
+
"learning_rate": 2.8914879792072773e-05,
|
| 630 |
+
"loss": 0.7737,
|
| 631 |
+
"step": 820
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 4.853801169590644,
|
| 635 |
+
"grad_norm": 0.7338679432868958,
|
| 636 |
+
"learning_rate": 2.8589993502274203e-05,
|
| 637 |
+
"loss": 0.7519,
|
| 638 |
+
"step": 830
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 4.912280701754386,
|
| 642 |
+
"grad_norm": 1.015286922454834,
|
| 643 |
+
"learning_rate": 2.8265107212475634e-05,
|
| 644 |
+
"loss": 0.8023,
|
| 645 |
+
"step": 840
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 4.970760233918129,
|
| 649 |
+
"grad_norm": 0.8456217050552368,
|
| 650 |
+
"learning_rate": 2.7940220922677064e-05,
|
| 651 |
+
"loss": 0.7768,
|
| 652 |
+
"step": 850
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 5.0,
|
| 656 |
+
"eval_accuracy": 0.6905501736428441,
|
| 657 |
+
"eval_f1": 0.6543647035652124,
|
| 658 |
+
"eval_loss": 0.7693981528282166,
|
| 659 |
+
"eval_precision": 0.6774590515406133,
|
| 660 |
+
"eval_recall": 0.6905501736428441,
|
| 661 |
+
"eval_runtime": 77.1215,
|
| 662 |
+
"eval_samples_per_second": 141.88,
|
| 663 |
+
"eval_steps_per_second": 2.217,
|
| 664 |
+
"step": 855
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 5.029239766081871,
|
| 668 |
+
"grad_norm": 0.624717116355896,
|
| 669 |
+
"learning_rate": 2.761533463287849e-05,
|
| 670 |
+
"loss": 0.7482,
|
| 671 |
+
"step": 860
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 5.087719298245614,
|
| 675 |
+
"grad_norm": 0.752734363079071,
|
| 676 |
+
"learning_rate": 2.729044834307992e-05,
|
| 677 |
+
"loss": 0.7259,
|
| 678 |
+
"step": 870
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 5.146198830409356,
|
| 682 |
+
"grad_norm": 0.6503344178199768,
|
| 683 |
+
"learning_rate": 2.696556205328135e-05,
|
| 684 |
+
"loss": 0.7488,
|
| 685 |
+
"step": 880
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 5.204678362573099,
|
| 689 |
+
"grad_norm": 0.8620956540107727,
|
| 690 |
+
"learning_rate": 2.664067576348278e-05,
|
| 691 |
+
"loss": 0.7704,
|
| 692 |
+
"step": 890
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 5.2631578947368425,
|
| 696 |
+
"grad_norm": 0.6938666701316833,
|
| 697 |
+
"learning_rate": 2.6315789473684212e-05,
|
| 698 |
+
"loss": 0.76,
|
| 699 |
+
"step": 900
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 5.321637426900585,
|
| 703 |
+
"grad_norm": 0.8206263184547424,
|
| 704 |
+
"learning_rate": 2.599090318388564e-05,
|
| 705 |
+
"loss": 0.7685,
|
| 706 |
+
"step": 910
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 5.380116959064328,
|
| 710 |
+
"grad_norm": 0.8919401168823242,
|
| 711 |
+
"learning_rate": 2.566601689408707e-05,
|
| 712 |
+
"loss": 0.7673,
|
| 713 |
+
"step": 920
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 5.43859649122807,
|
| 717 |
+
"grad_norm": 0.9412862062454224,
|
| 718 |
+
"learning_rate": 2.53411306042885e-05,
|
| 719 |
+
"loss": 0.7403,
|
| 720 |
+
"step": 930
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 5.497076023391813,
|
| 724 |
+
"grad_norm": 1.1093353033065796,
|
| 725 |
+
"learning_rate": 2.501624431448993e-05,
|
| 726 |
+
"loss": 0.743,
|
| 727 |
+
"step": 940
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 5.555555555555555,
|
| 731 |
+
"grad_norm": 0.6838064193725586,
|
| 732 |
+
"learning_rate": 2.4691358024691357e-05,
|
| 733 |
+
"loss": 0.7641,
|
| 734 |
+
"step": 950
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 5.614035087719298,
|
| 738 |
+
"grad_norm": 0.7546567320823669,
|
| 739 |
+
"learning_rate": 2.4366471734892787e-05,
|
| 740 |
+
"loss": 0.7648,
|
| 741 |
+
"step": 960
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 5.6725146198830405,
|
| 745 |
+
"grad_norm": 0.7909204363822937,
|
| 746 |
+
"learning_rate": 2.4041585445094217e-05,
|
| 747 |
+
"loss": 0.7872,
|
| 748 |
+
"step": 970
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 5.730994152046784,
|
| 752 |
+
"grad_norm": 0.6969336867332458,
|
| 753 |
+
"learning_rate": 2.3716699155295647e-05,
|
| 754 |
+
"loss": 0.7638,
|
| 755 |
+
"step": 980
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 5.7894736842105265,
|
| 759 |
+
"grad_norm": 0.7838913202285767,
|
| 760 |
+
"learning_rate": 2.3391812865497074e-05,
|
| 761 |
+
"loss": 0.747,
|
| 762 |
+
"step": 990
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 5.847953216374269,
|
| 766 |
+
"grad_norm": 0.8347417712211609,
|
| 767 |
+
"learning_rate": 2.3066926575698505e-05,
|
| 768 |
+
"loss": 0.7749,
|
| 769 |
+
"step": 1000
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 5.906432748538012,
|
| 773 |
+
"grad_norm": 0.9646545052528381,
|
| 774 |
+
"learning_rate": 2.2742040285899935e-05,
|
| 775 |
+
"loss": 0.737,
|
| 776 |
+
"step": 1010
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 5.964912280701754,
|
| 780 |
+
"grad_norm": 0.6134990453720093,
|
| 781 |
+
"learning_rate": 2.2417153996101365e-05,
|
| 782 |
+
"loss": 0.7415,
|
| 783 |
+
"step": 1020
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 6.0,
|
| 787 |
+
"eval_accuracy": 0.6962164138183148,
|
| 788 |
+
"eval_f1": 0.671796652878038,
|
| 789 |
+
"eval_loss": 0.7572136521339417,
|
| 790 |
+
"eval_precision": 0.6763761787994358,
|
| 791 |
+
"eval_recall": 0.6962164138183148,
|
| 792 |
+
"eval_runtime": 77.9563,
|
| 793 |
+
"eval_samples_per_second": 140.361,
|
| 794 |
+
"eval_steps_per_second": 2.194,
|
| 795 |
+
"step": 1026
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 6.023391812865497,
|
| 799 |
+
"grad_norm": 0.7049497961997986,
|
| 800 |
+
"learning_rate": 2.2092267706302795e-05,
|
| 801 |
+
"loss": 0.7598,
|
| 802 |
+
"step": 1030
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 6.081871345029239,
|
| 806 |
+
"grad_norm": 0.9780289530754089,
|
| 807 |
+
"learning_rate": 2.1767381416504222e-05,
|
| 808 |
+
"loss": 0.7472,
|
| 809 |
+
"step": 1040
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 6.140350877192983,
|
| 813 |
+
"grad_norm": 0.7058891654014587,
|
| 814 |
+
"learning_rate": 2.1442495126705653e-05,
|
| 815 |
+
"loss": 0.742,
|
| 816 |
+
"step": 1050
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 6.1988304093567255,
|
| 820 |
+
"grad_norm": 0.8734349012374878,
|
| 821 |
+
"learning_rate": 2.1117608836907083e-05,
|
| 822 |
+
"loss": 0.7581,
|
| 823 |
+
"step": 1060
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 6.257309941520468,
|
| 827 |
+
"grad_norm": 0.8839743733406067,
|
| 828 |
+
"learning_rate": 2.0792722547108513e-05,
|
| 829 |
+
"loss": 0.7516,
|
| 830 |
+
"step": 1070
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 6.315789473684211,
|
| 834 |
+
"grad_norm": 0.6963735222816467,
|
| 835 |
+
"learning_rate": 2.046783625730994e-05,
|
| 836 |
+
"loss": 0.7412,
|
| 837 |
+
"step": 1080
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 6.374269005847953,
|
| 841 |
+
"grad_norm": 0.9337784647941589,
|
| 842 |
+
"learning_rate": 2.014294996751137e-05,
|
| 843 |
+
"loss": 0.7402,
|
| 844 |
+
"step": 1090
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 6.432748538011696,
|
| 848 |
+
"grad_norm": 0.6648013591766357,
|
| 849 |
+
"learning_rate": 1.98180636777128e-05,
|
| 850 |
+
"loss": 0.7513,
|
| 851 |
+
"step": 1100
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 6.491228070175438,
|
| 855 |
+
"grad_norm": 1.072342872619629,
|
| 856 |
+
"learning_rate": 1.949317738791423e-05,
|
| 857 |
+
"loss": 0.7406,
|
| 858 |
+
"step": 1110
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 6.549707602339181,
|
| 862 |
+
"grad_norm": 1.0100135803222656,
|
| 863 |
+
"learning_rate": 1.9168291098115658e-05,
|
| 864 |
+
"loss": 0.7643,
|
| 865 |
+
"step": 1120
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 6.6081871345029235,
|
| 869 |
+
"grad_norm": 0.7687884569168091,
|
| 870 |
+
"learning_rate": 1.8843404808317088e-05,
|
| 871 |
+
"loss": 0.7404,
|
| 872 |
+
"step": 1130
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 6.666666666666667,
|
| 876 |
+
"grad_norm": 0.750688374042511,
|
| 877 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 878 |
+
"loss": 0.7391,
|
| 879 |
+
"step": 1140
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 6.7251461988304095,
|
| 883 |
+
"grad_norm": 0.7010438442230225,
|
| 884 |
+
"learning_rate": 1.819363222871995e-05,
|
| 885 |
+
"loss": 0.7534,
|
| 886 |
+
"step": 1150
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 6.783625730994152,
|
| 890 |
+
"grad_norm": 1.0012060403823853,
|
| 891 |
+
"learning_rate": 1.786874593892138e-05,
|
| 892 |
+
"loss": 0.7485,
|
| 893 |
+
"step": 1160
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 6.842105263157895,
|
| 897 |
+
"grad_norm": 0.8860548734664917,
|
| 898 |
+
"learning_rate": 1.7543859649122806e-05,
|
| 899 |
+
"loss": 0.7606,
|
| 900 |
+
"step": 1170
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 6.900584795321637,
|
| 904 |
+
"grad_norm": 0.969633936882019,
|
| 905 |
+
"learning_rate": 1.7218973359324236e-05,
|
| 906 |
+
"loss": 0.7592,
|
| 907 |
+
"step": 1180
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 6.95906432748538,
|
| 911 |
+
"grad_norm": 0.8473331928253174,
|
| 912 |
+
"learning_rate": 1.6894087069525666e-05,
|
| 913 |
+
"loss": 0.7351,
|
| 914 |
+
"step": 1190
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 7.0,
|
| 918 |
+
"eval_accuracy": 0.692195211113142,
|
| 919 |
+
"eval_f1": 0.6568623393542826,
|
| 920 |
+
"eval_loss": 0.754936695098877,
|
| 921 |
+
"eval_precision": 0.6648205901494669,
|
| 922 |
+
"eval_recall": 0.692195211113142,
|
| 923 |
+
"eval_runtime": 77.9138,
|
| 924 |
+
"eval_samples_per_second": 140.437,
|
| 925 |
+
"eval_steps_per_second": 2.195,
|
| 926 |
+
"step": 1197
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 7.017543859649122,
|
| 930 |
+
"grad_norm": 0.7504809498786926,
|
| 931 |
+
"learning_rate": 1.6569200779727097e-05,
|
| 932 |
+
"loss": 0.7102,
|
| 933 |
+
"step": 1200
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 7.076023391812866,
|
| 937 |
+
"grad_norm": 1.306260585784912,
|
| 938 |
+
"learning_rate": 1.6244314489928523e-05,
|
| 939 |
+
"loss": 0.7574,
|
| 940 |
+
"step": 1210
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 7.1345029239766085,
|
| 944 |
+
"grad_norm": 0.7214799523353577,
|
| 945 |
+
"learning_rate": 1.5919428200129954e-05,
|
| 946 |
+
"loss": 0.7355,
|
| 947 |
+
"step": 1220
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 7.192982456140351,
|
| 951 |
+
"grad_norm": 0.8254335522651672,
|
| 952 |
+
"learning_rate": 1.5594541910331384e-05,
|
| 953 |
+
"loss": 0.758,
|
| 954 |
+
"step": 1230
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 7.251461988304094,
|
| 958 |
+
"grad_norm": 0.8644353747367859,
|
| 959 |
+
"learning_rate": 1.5269655620532814e-05,
|
| 960 |
+
"loss": 0.7131,
|
| 961 |
+
"step": 1240
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 7.309941520467836,
|
| 965 |
+
"grad_norm": 0.7876085638999939,
|
| 966 |
+
"learning_rate": 1.4944769330734243e-05,
|
| 967 |
+
"loss": 0.7437,
|
| 968 |
+
"step": 1250
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 7.368421052631579,
|
| 972 |
+
"grad_norm": 0.904586136341095,
|
| 973 |
+
"learning_rate": 1.4619883040935673e-05,
|
| 974 |
+
"loss": 0.7464,
|
| 975 |
+
"step": 1260
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 7.426900584795321,
|
| 979 |
+
"grad_norm": 1.033260464668274,
|
| 980 |
+
"learning_rate": 1.4294996751137102e-05,
|
| 981 |
+
"loss": 0.7226,
|
| 982 |
+
"step": 1270
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 7.485380116959064,
|
| 986 |
+
"grad_norm": 0.9057112336158752,
|
| 987 |
+
"learning_rate": 1.3970110461338532e-05,
|
| 988 |
+
"loss": 0.7425,
|
| 989 |
+
"step": 1280
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 7.543859649122807,
|
| 993 |
+
"grad_norm": 0.8631776571273804,
|
| 994 |
+
"learning_rate": 1.364522417153996e-05,
|
| 995 |
+
"loss": 0.7178,
|
| 996 |
+
"step": 1290
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 7.60233918128655,
|
| 1000 |
+
"grad_norm": 0.8566320538520813,
|
| 1001 |
+
"learning_rate": 1.332033788174139e-05,
|
| 1002 |
+
"loss": 0.7151,
|
| 1003 |
+
"step": 1300
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 7.6608187134502925,
|
| 1007 |
+
"grad_norm": 1.056127905845642,
|
| 1008 |
+
"learning_rate": 1.299545159194282e-05,
|
| 1009 |
+
"loss": 0.7574,
|
| 1010 |
+
"step": 1310
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 7.719298245614035,
|
| 1014 |
+
"grad_norm": 1.0582066774368286,
|
| 1015 |
+
"learning_rate": 1.267056530214425e-05,
|
| 1016 |
+
"loss": 0.7462,
|
| 1017 |
+
"step": 1320
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 7.777777777777778,
|
| 1021 |
+
"grad_norm": 1.0808275938034058,
|
| 1022 |
+
"learning_rate": 1.2345679012345678e-05,
|
| 1023 |
+
"loss": 0.7181,
|
| 1024 |
+
"step": 1330
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 7.83625730994152,
|
| 1028 |
+
"grad_norm": 0.8452061414718628,
|
| 1029 |
+
"learning_rate": 1.2020792722547109e-05,
|
| 1030 |
+
"loss": 0.7686,
|
| 1031 |
+
"step": 1340
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 7.894736842105263,
|
| 1035 |
+
"grad_norm": 0.7253689765930176,
|
| 1036 |
+
"learning_rate": 1.1695906432748537e-05,
|
| 1037 |
+
"loss": 0.7174,
|
| 1038 |
+
"step": 1350
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 7.953216374269006,
|
| 1042 |
+
"grad_norm": 0.9176128506660461,
|
| 1043 |
+
"learning_rate": 1.1371020142949967e-05,
|
| 1044 |
+
"loss": 0.7197,
|
| 1045 |
+
"step": 1360
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 8.0,
|
| 1049 |
+
"eval_accuracy": 0.6985925790531895,
|
| 1050 |
+
"eval_f1": 0.6855055863067254,
|
| 1051 |
+
"eval_loss": 0.7478834390640259,
|
| 1052 |
+
"eval_precision": 0.6925926647987316,
|
| 1053 |
+
"eval_recall": 0.6985925790531895,
|
| 1054 |
+
"eval_runtime": 77.8555,
|
| 1055 |
+
"eval_samples_per_second": 140.542,
|
| 1056 |
+
"eval_steps_per_second": 2.196,
|
| 1057 |
+
"step": 1368
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 8.011695906432749,
|
| 1061 |
+
"grad_norm": 0.7897553443908691,
|
| 1062 |
+
"learning_rate": 1.1046133853151398e-05,
|
| 1063 |
+
"loss": 0.755,
|
| 1064 |
+
"step": 1370
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 8.070175438596491,
|
| 1068 |
+
"grad_norm": 0.7324469685554504,
|
| 1069 |
+
"learning_rate": 1.0721247563352826e-05,
|
| 1070 |
+
"loss": 0.7243,
|
| 1071 |
+
"step": 1380
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 8.128654970760234,
|
| 1075 |
+
"grad_norm": 0.7983306646347046,
|
| 1076 |
+
"learning_rate": 1.0396361273554257e-05,
|
| 1077 |
+
"loss": 0.7294,
|
| 1078 |
+
"step": 1390
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 8.187134502923977,
|
| 1082 |
+
"grad_norm": 0.9110460877418518,
|
| 1083 |
+
"learning_rate": 1.0071474983755685e-05,
|
| 1084 |
+
"loss": 0.7027,
|
| 1085 |
+
"step": 1400
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 8.24561403508772,
|
| 1089 |
+
"grad_norm": 0.9574342966079712,
|
| 1090 |
+
"learning_rate": 9.746588693957115e-06,
|
| 1091 |
+
"loss": 0.7131,
|
| 1092 |
+
"step": 1410
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 8.304093567251462,
|
| 1096 |
+
"grad_norm": 0.7169631719589233,
|
| 1097 |
+
"learning_rate": 9.421702404158544e-06,
|
| 1098 |
+
"loss": 0.7365,
|
| 1099 |
+
"step": 1420
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 8.362573099415204,
|
| 1103 |
+
"grad_norm": 0.9551491737365723,
|
| 1104 |
+
"learning_rate": 9.096816114359974e-06,
|
| 1105 |
+
"loss": 0.7313,
|
| 1106 |
+
"step": 1430
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 8.421052631578947,
|
| 1110 |
+
"grad_norm": 1.159575343132019,
|
| 1111 |
+
"learning_rate": 8.771929824561403e-06,
|
| 1112 |
+
"loss": 0.7438,
|
| 1113 |
+
"step": 1440
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 8.47953216374269,
|
| 1117 |
+
"grad_norm": 0.8166360259056091,
|
| 1118 |
+
"learning_rate": 8.447043534762833e-06,
|
| 1119 |
+
"loss": 0.7355,
|
| 1120 |
+
"step": 1450
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 8.538011695906432,
|
| 1124 |
+
"grad_norm": 0.8369165062904358,
|
| 1125 |
+
"learning_rate": 8.122157244964262e-06,
|
| 1126 |
+
"loss": 0.7183,
|
| 1127 |
+
"step": 1460
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 8.596491228070175,
|
| 1131 |
+
"grad_norm": 0.7923627495765686,
|
| 1132 |
+
"learning_rate": 7.797270955165692e-06,
|
| 1133 |
+
"loss": 0.711,
|
| 1134 |
+
"step": 1470
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 8.654970760233919,
|
| 1138 |
+
"grad_norm": 0.7623910903930664,
|
| 1139 |
+
"learning_rate": 7.4723846653671214e-06,
|
| 1140 |
+
"loss": 0.7377,
|
| 1141 |
+
"step": 1480
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 8.713450292397662,
|
| 1145 |
+
"grad_norm": 1.088745355606079,
|
| 1146 |
+
"learning_rate": 7.147498375568551e-06,
|
| 1147 |
+
"loss": 0.7199,
|
| 1148 |
+
"step": 1490
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 8.771929824561404,
|
| 1152 |
+
"grad_norm": 0.8672430515289307,
|
| 1153 |
+
"learning_rate": 6.82261208576998e-06,
|
| 1154 |
+
"loss": 0.7234,
|
| 1155 |
+
"step": 1500
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 8.830409356725147,
|
| 1159 |
+
"grad_norm": 0.77957683801651,
|
| 1160 |
+
"learning_rate": 6.49772579597141e-06,
|
| 1161 |
+
"loss": 0.7348,
|
| 1162 |
+
"step": 1510
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 8.88888888888889,
|
| 1166 |
+
"grad_norm": 0.982523500919342,
|
| 1167 |
+
"learning_rate": 6.172839506172839e-06,
|
| 1168 |
+
"loss": 0.7348,
|
| 1169 |
+
"step": 1520
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 8.947368421052632,
|
| 1173 |
+
"grad_norm": 0.8758224844932556,
|
| 1174 |
+
"learning_rate": 5.8479532163742686e-06,
|
| 1175 |
+
"loss": 0.7087,
|
| 1176 |
+
"step": 1530
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 9.0,
|
| 1180 |
+
"eval_accuracy": 0.6978614512886127,
|
| 1181 |
+
"eval_f1": 0.6697434161663234,
|
| 1182 |
+
"eval_loss": 0.744518518447876,
|
| 1183 |
+
"eval_precision": 0.6792260519903555,
|
| 1184 |
+
"eval_recall": 0.6978614512886127,
|
| 1185 |
+
"eval_runtime": 77.9794,
|
| 1186 |
+
"eval_samples_per_second": 140.319,
|
| 1187 |
+
"eval_steps_per_second": 2.193,
|
| 1188 |
+
"step": 1539
|
| 1189 |
+
},
|
| 1190 |
+
{
|
| 1191 |
+
"epoch": 9.005847953216374,
|
| 1192 |
+
"grad_norm": 0.7864174246788025,
|
| 1193 |
+
"learning_rate": 5.523066926575699e-06,
|
| 1194 |
+
"loss": 0.713,
|
| 1195 |
+
"step": 1540
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 9.064327485380117,
|
| 1199 |
+
"grad_norm": 1.0881294012069702,
|
| 1200 |
+
"learning_rate": 5.198180636777128e-06,
|
| 1201 |
+
"loss": 0.7092,
|
| 1202 |
+
"step": 1550
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 9.12280701754386,
|
| 1206 |
+
"grad_norm": 1.0221022367477417,
|
| 1207 |
+
"learning_rate": 4.873294346978558e-06,
|
| 1208 |
+
"loss": 0.7463,
|
| 1209 |
+
"step": 1560
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 9.181286549707602,
|
| 1213 |
+
"grad_norm": 0.8976357579231262,
|
| 1214 |
+
"learning_rate": 4.548408057179987e-06,
|
| 1215 |
+
"loss": 0.7392,
|
| 1216 |
+
"step": 1570
|
| 1217 |
+
},
|
| 1218 |
+
{
|
| 1219 |
+
"epoch": 9.239766081871345,
|
| 1220 |
+
"grad_norm": 0.8547308444976807,
|
| 1221 |
+
"learning_rate": 4.2235217673814166e-06,
|
| 1222 |
+
"loss": 0.7275,
|
| 1223 |
+
"step": 1580
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 9.298245614035087,
|
| 1227 |
+
"grad_norm": 0.9341883063316345,
|
| 1228 |
+
"learning_rate": 3.898635477582846e-06,
|
| 1229 |
+
"loss": 0.7183,
|
| 1230 |
+
"step": 1590
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"epoch": 9.35672514619883,
|
| 1234 |
+
"grad_norm": 0.9447769522666931,
|
| 1235 |
+
"learning_rate": 3.5737491877842754e-06,
|
| 1236 |
+
"loss": 0.6966,
|
| 1237 |
+
"step": 1600
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 9.415204678362572,
|
| 1241 |
+
"grad_norm": 1.0918306112289429,
|
| 1242 |
+
"learning_rate": 3.248862897985705e-06,
|
| 1243 |
+
"loss": 0.7161,
|
| 1244 |
+
"step": 1610
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 9.473684210526315,
|
| 1248 |
+
"grad_norm": 0.9330850839614868,
|
| 1249 |
+
"learning_rate": 2.9239766081871343e-06,
|
| 1250 |
+
"loss": 0.7093,
|
| 1251 |
+
"step": 1620
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 9.53216374269006,
|
| 1255 |
+
"grad_norm": 0.8420782685279846,
|
| 1256 |
+
"learning_rate": 2.599090318388564e-06,
|
| 1257 |
+
"loss": 0.7259,
|
| 1258 |
+
"step": 1630
|
| 1259 |
+
},
|
| 1260 |
+
{
|
| 1261 |
+
"epoch": 9.590643274853802,
|
| 1262 |
+
"grad_norm": 0.9159696698188782,
|
| 1263 |
+
"learning_rate": 2.2742040285899936e-06,
|
| 1264 |
+
"loss": 0.7265,
|
| 1265 |
+
"step": 1640
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 9.649122807017545,
|
| 1269 |
+
"grad_norm": 1.0164194107055664,
|
| 1270 |
+
"learning_rate": 1.949317738791423e-06,
|
| 1271 |
+
"loss": 0.7458,
|
| 1272 |
+
"step": 1650
|
| 1273 |
+
},
|
| 1274 |
+
{
|
| 1275 |
+
"epoch": 9.707602339181287,
|
| 1276 |
+
"grad_norm": 0.8425694704055786,
|
| 1277 |
+
"learning_rate": 1.6244314489928524e-06,
|
| 1278 |
+
"loss": 0.7149,
|
| 1279 |
+
"step": 1660
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 9.76608187134503,
|
| 1283 |
+
"grad_norm": 1.1017402410507202,
|
| 1284 |
+
"learning_rate": 1.299545159194282e-06,
|
| 1285 |
+
"loss": 0.7101,
|
| 1286 |
+
"step": 1670
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 9.824561403508772,
|
| 1290 |
+
"grad_norm": 0.8220164179801941,
|
| 1291 |
+
"learning_rate": 9.746588693957115e-07,
|
| 1292 |
+
"loss": 0.7174,
|
| 1293 |
+
"step": 1680
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 9.883040935672515,
|
| 1297 |
+
"grad_norm": 0.7968518733978271,
|
| 1298 |
+
"learning_rate": 6.49772579597141e-07,
|
| 1299 |
+
"loss": 0.7236,
|
| 1300 |
+
"step": 1690
|
| 1301 |
+
},
|
| 1302 |
+
{
|
| 1303 |
+
"epoch": 9.941520467836257,
|
| 1304 |
+
"grad_norm": 0.7491603493690491,
|
| 1305 |
+
"learning_rate": 3.248862897985705e-07,
|
| 1306 |
+
"loss": 0.6978,
|
| 1307 |
+
"step": 1700
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 10.0,
|
| 1311 |
+
"grad_norm": 0.8862149119377136,
|
| 1312 |
+
"learning_rate": 0.0,
|
| 1313 |
+
"loss": 0.6977,
|
| 1314 |
+
"step": 1710
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 10.0,
|
| 1318 |
+
"eval_accuracy": 0.6991409248766222,
|
| 1319 |
+
"eval_f1": 0.6766571542539821,
|
| 1320 |
+
"eval_loss": 0.7419390082359314,
|
| 1321 |
+
"eval_precision": 0.6830062110035815,
|
| 1322 |
+
"eval_recall": 0.6991409248766222,
|
| 1323 |
+
"eval_runtime": 78.4178,
|
| 1324 |
+
"eval_samples_per_second": 139.535,
|
| 1325 |
+
"eval_steps_per_second": 2.181,
|
| 1326 |
+
"step": 1710
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 10.0,
|
| 1330 |
+
"step": 1710,
|
| 1331 |
+
"total_flos": 3.3914202248568177e+19,
|
| 1332 |
+
"train_loss": 0.790359598014787,
|
| 1333 |
+
"train_runtime": 7828.8109,
|
| 1334 |
+
"train_samples_per_second": 55.901,
|
| 1335 |
+
"train_steps_per_second": 0.218
|
| 1336 |
+
}
|
| 1337 |
+
],
|
| 1338 |
+
"logging_steps": 10,
|
| 1339 |
+
"max_steps": 1710,
|
| 1340 |
+
"num_input_tokens_seen": 0,
|
| 1341 |
+
"num_train_epochs": 10,
|
| 1342 |
+
"save_steps": 500,
|
| 1343 |
+
"stateful_callbacks": {
|
| 1344 |
+
"TrainerControl": {
|
| 1345 |
+
"args": {
|
| 1346 |
+
"should_epoch_stop": false,
|
| 1347 |
+
"should_evaluate": false,
|
| 1348 |
+
"should_log": false,
|
| 1349 |
+
"should_save": true,
|
| 1350 |
+
"should_training_stop": true
|
| 1351 |
+
},
|
| 1352 |
+
"attributes": {}
|
| 1353 |
+
}
|
| 1354 |
+
},
|
| 1355 |
+
"total_flos": 3.3914202248568177e+19,
|
| 1356 |
+
"train_batch_size": 64,
|
| 1357 |
+
"trial_name": null,
|
| 1358 |
+
"trial_params": null
|
| 1359 |
+
}
|