Commit
·
6ab7016
1
Parent(s):
a1504e3
End of training
Browse files- all_results.json +13 -0
- eval_results.json +8 -0
- runs/Feb14_09-58-40_4b8c246c6890/events.out.tfevents.1676377548.4b8c246c6890.455.2 +3 -0
- train_results.json +8 -0
- trainer_state.json +1708 -0
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.970523929063082,
|
| 4 |
+
"eval_loss": 0.07879053056240082,
|
| 5 |
+
"eval_runtime": 168.6571,
|
| 6 |
+
"eval_samples_per_second": 73.22,
|
| 7 |
+
"eval_steps_per_second": 2.289,
|
| 8 |
+
"total_flos": 8.694514267698954e+18,
|
| 9 |
+
"train_loss": 0.3431307647691105,
|
| 10 |
+
"train_runtime": 8584.1671,
|
| 11 |
+
"train_samples_per_second": 40.756,
|
| 12 |
+
"train_steps_per_second": 0.318
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_accuracy": 0.970523929063082,
|
| 4 |
+
"eval_loss": 0.07879053056240082,
|
| 5 |
+
"eval_runtime": 168.6571,
|
| 6 |
+
"eval_samples_per_second": 73.22,
|
| 7 |
+
"eval_steps_per_second": 2.289
|
| 8 |
+
}
|
runs/Feb14_09-58-40_4b8c246c6890/events.out.tfevents.1676377548.4b8c246c6890.455.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5b36c9eeb7f7ade84b830d3131793ce83046d5eeee395a3236b528a3065586a
|
| 3 |
+
size 363
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"total_flos": 8.694514267698954e+18,
|
| 4 |
+
"train_loss": 0.3431307647691105,
|
| 5 |
+
"train_runtime": 8584.1671,
|
| 6 |
+
"train_samples_per_second": 40.756,
|
| 7 |
+
"train_steps_per_second": 0.318
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1708 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.970523929063082,
|
| 3 |
+
"best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-og_dataset_5e/checkpoint-2730",
|
| 4 |
+
"epoch": 4.998628257887518,
|
| 5 |
+
"global_step": 2730,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.02,
|
| 12 |
+
"learning_rate": 1.8315018315018316e-06,
|
| 13 |
+
"loss": 1.3747,
|
| 14 |
+
"step": 10
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.04,
|
| 18 |
+
"learning_rate": 3.663003663003663e-06,
|
| 19 |
+
"loss": 1.3424,
|
| 20 |
+
"step": 20
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.05,
|
| 24 |
+
"learning_rate": 5.494505494505494e-06,
|
| 25 |
+
"loss": 1.2339,
|
| 26 |
+
"step": 30
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.07,
|
| 30 |
+
"learning_rate": 7.326007326007326e-06,
|
| 31 |
+
"loss": 1.113,
|
| 32 |
+
"step": 40
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.09,
|
| 36 |
+
"learning_rate": 9.157509157509158e-06,
|
| 37 |
+
"loss": 0.9805,
|
| 38 |
+
"step": 50
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.11,
|
| 42 |
+
"learning_rate": 1.0989010989010989e-05,
|
| 43 |
+
"loss": 0.8468,
|
| 44 |
+
"step": 60
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.13,
|
| 48 |
+
"learning_rate": 1.282051282051282e-05,
|
| 49 |
+
"loss": 0.8367,
|
| 50 |
+
"step": 70
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.15,
|
| 54 |
+
"learning_rate": 1.4652014652014653e-05,
|
| 55 |
+
"loss": 0.7899,
|
| 56 |
+
"step": 80
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.16,
|
| 60 |
+
"learning_rate": 1.6483516483516486e-05,
|
| 61 |
+
"loss": 0.7616,
|
| 62 |
+
"step": 90
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.18,
|
| 66 |
+
"learning_rate": 1.8315018315018315e-05,
|
| 67 |
+
"loss": 0.7109,
|
| 68 |
+
"step": 100
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.2,
|
| 72 |
+
"learning_rate": 2.0146520146520148e-05,
|
| 73 |
+
"loss": 0.7296,
|
| 74 |
+
"step": 110
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.22,
|
| 78 |
+
"learning_rate": 2.1978021978021977e-05,
|
| 79 |
+
"loss": 0.746,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.24,
|
| 84 |
+
"learning_rate": 2.380952380952381e-05,
|
| 85 |
+
"loss": 0.6468,
|
| 86 |
+
"step": 130
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.26,
|
| 90 |
+
"learning_rate": 2.564102564102564e-05,
|
| 91 |
+
"loss": 0.6482,
|
| 92 |
+
"step": 140
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.27,
|
| 96 |
+
"learning_rate": 2.7472527472527476e-05,
|
| 97 |
+
"loss": 0.6902,
|
| 98 |
+
"step": 150
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.29,
|
| 102 |
+
"learning_rate": 2.9304029304029305e-05,
|
| 103 |
+
"loss": 0.6797,
|
| 104 |
+
"step": 160
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.31,
|
| 108 |
+
"learning_rate": 3.113553113553114e-05,
|
| 109 |
+
"loss": 0.7254,
|
| 110 |
+
"step": 170
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.33,
|
| 114 |
+
"learning_rate": 3.296703296703297e-05,
|
| 115 |
+
"loss": 0.6635,
|
| 116 |
+
"step": 180
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.35,
|
| 120 |
+
"learning_rate": 3.47985347985348e-05,
|
| 121 |
+
"loss": 0.6333,
|
| 122 |
+
"step": 190
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.37,
|
| 126 |
+
"learning_rate": 3.663003663003663e-05,
|
| 127 |
+
"loss": 0.6655,
|
| 128 |
+
"step": 200
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.38,
|
| 132 |
+
"learning_rate": 3.846153846153846e-05,
|
| 133 |
+
"loss": 0.6587,
|
| 134 |
+
"step": 210
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.4,
|
| 138 |
+
"learning_rate": 4.0293040293040296e-05,
|
| 139 |
+
"loss": 0.6673,
|
| 140 |
+
"step": 220
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.42,
|
| 144 |
+
"learning_rate": 4.212454212454213e-05,
|
| 145 |
+
"loss": 0.6875,
|
| 146 |
+
"step": 230
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.44,
|
| 150 |
+
"learning_rate": 4.3956043956043955e-05,
|
| 151 |
+
"loss": 0.6372,
|
| 152 |
+
"step": 240
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.46,
|
| 156 |
+
"learning_rate": 4.578754578754579e-05,
|
| 157 |
+
"loss": 0.6481,
|
| 158 |
+
"step": 250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.48,
|
| 162 |
+
"learning_rate": 4.761904761904762e-05,
|
| 163 |
+
"loss": 0.5944,
|
| 164 |
+
"step": 260
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.49,
|
| 168 |
+
"learning_rate": 4.945054945054945e-05,
|
| 169 |
+
"loss": 0.6148,
|
| 170 |
+
"step": 270
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.51,
|
| 174 |
+
"learning_rate": 4.985754985754986e-05,
|
| 175 |
+
"loss": 0.6211,
|
| 176 |
+
"step": 280
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.53,
|
| 180 |
+
"learning_rate": 4.965404965404965e-05,
|
| 181 |
+
"loss": 0.6081,
|
| 182 |
+
"step": 290
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.55,
|
| 186 |
+
"learning_rate": 4.945054945054945e-05,
|
| 187 |
+
"loss": 0.6008,
|
| 188 |
+
"step": 300
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.57,
|
| 192 |
+
"learning_rate": 4.9247049247049247e-05,
|
| 193 |
+
"loss": 0.6103,
|
| 194 |
+
"step": 310
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.59,
|
| 198 |
+
"learning_rate": 4.904354904354905e-05,
|
| 199 |
+
"loss": 0.5943,
|
| 200 |
+
"step": 320
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.6,
|
| 204 |
+
"learning_rate": 4.884004884004884e-05,
|
| 205 |
+
"loss": 0.6131,
|
| 206 |
+
"step": 330
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.62,
|
| 210 |
+
"learning_rate": 4.863654863654864e-05,
|
| 211 |
+
"loss": 0.5773,
|
| 212 |
+
"step": 340
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.64,
|
| 216 |
+
"learning_rate": 4.8433048433048433e-05,
|
| 217 |
+
"loss": 0.5885,
|
| 218 |
+
"step": 350
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.66,
|
| 222 |
+
"learning_rate": 4.8229548229548234e-05,
|
| 223 |
+
"loss": 0.5702,
|
| 224 |
+
"step": 360
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.68,
|
| 228 |
+
"learning_rate": 4.802604802604803e-05,
|
| 229 |
+
"loss": 0.5774,
|
| 230 |
+
"step": 370
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.7,
|
| 234 |
+
"learning_rate": 4.782254782254783e-05,
|
| 235 |
+
"loss": 0.5542,
|
| 236 |
+
"step": 380
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.71,
|
| 240 |
+
"learning_rate": 4.761904761904762e-05,
|
| 241 |
+
"loss": 0.5746,
|
| 242 |
+
"step": 390
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.73,
|
| 246 |
+
"learning_rate": 4.741554741554742e-05,
|
| 247 |
+
"loss": 0.5137,
|
| 248 |
+
"step": 400
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.75,
|
| 252 |
+
"learning_rate": 4.7212047212047214e-05,
|
| 253 |
+
"loss": 0.4981,
|
| 254 |
+
"step": 410
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.77,
|
| 258 |
+
"learning_rate": 4.700854700854701e-05,
|
| 259 |
+
"loss": 0.4805,
|
| 260 |
+
"step": 420
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.79,
|
| 264 |
+
"learning_rate": 4.680504680504681e-05,
|
| 265 |
+
"loss": 0.5002,
|
| 266 |
+
"step": 430
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.8,
|
| 270 |
+
"learning_rate": 4.66015466015466e-05,
|
| 271 |
+
"loss": 0.5002,
|
| 272 |
+
"step": 440
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.82,
|
| 276 |
+
"learning_rate": 4.63980463980464e-05,
|
| 277 |
+
"loss": 0.5202,
|
| 278 |
+
"step": 450
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.84,
|
| 282 |
+
"learning_rate": 4.6194546194546194e-05,
|
| 283 |
+
"loss": 0.5185,
|
| 284 |
+
"step": 460
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.86,
|
| 288 |
+
"learning_rate": 4.5991045991045994e-05,
|
| 289 |
+
"loss": 0.4858,
|
| 290 |
+
"step": 470
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.88,
|
| 294 |
+
"learning_rate": 4.578754578754579e-05,
|
| 295 |
+
"loss": 0.5254,
|
| 296 |
+
"step": 480
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.9,
|
| 300 |
+
"learning_rate": 4.558404558404559e-05,
|
| 301 |
+
"loss": 0.5179,
|
| 302 |
+
"step": 490
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.91,
|
| 306 |
+
"learning_rate": 4.538054538054538e-05,
|
| 307 |
+
"loss": 0.4956,
|
| 308 |
+
"step": 500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.93,
|
| 312 |
+
"learning_rate": 4.517704517704518e-05,
|
| 313 |
+
"loss": 0.4152,
|
| 314 |
+
"step": 510
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.95,
|
| 318 |
+
"learning_rate": 4.4973544973544974e-05,
|
| 319 |
+
"loss": 0.4272,
|
| 320 |
+
"step": 520
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.97,
|
| 324 |
+
"learning_rate": 4.4770044770044775e-05,
|
| 325 |
+
"loss": 0.4864,
|
| 326 |
+
"step": 530
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.99,
|
| 330 |
+
"learning_rate": 4.4566544566544575e-05,
|
| 331 |
+
"loss": 0.4626,
|
| 332 |
+
"step": 540
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.0,
|
| 336 |
+
"eval_accuracy": 0.8578022511944287,
|
| 337 |
+
"eval_loss": 0.34676697850227356,
|
| 338 |
+
"eval_runtime": 181.8589,
|
| 339 |
+
"eval_samples_per_second": 67.904,
|
| 340 |
+
"eval_steps_per_second": 2.123,
|
| 341 |
+
"step": 546
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 1.01,
|
| 345 |
+
"learning_rate": 4.436304436304437e-05,
|
| 346 |
+
"loss": 0.4772,
|
| 347 |
+
"step": 550
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 1.03,
|
| 351 |
+
"learning_rate": 4.415954415954416e-05,
|
| 352 |
+
"loss": 0.4954,
|
| 353 |
+
"step": 560
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.04,
|
| 357 |
+
"learning_rate": 4.3956043956043955e-05,
|
| 358 |
+
"loss": 0.4469,
|
| 359 |
+
"step": 570
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.06,
|
| 363 |
+
"learning_rate": 4.3752543752543755e-05,
|
| 364 |
+
"loss": 0.4596,
|
| 365 |
+
"step": 580
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 1.08,
|
| 369 |
+
"learning_rate": 4.354904354904355e-05,
|
| 370 |
+
"loss": 0.4428,
|
| 371 |
+
"step": 590
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 1.1,
|
| 375 |
+
"learning_rate": 4.334554334554335e-05,
|
| 376 |
+
"loss": 0.4451,
|
| 377 |
+
"step": 600
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 1.12,
|
| 381 |
+
"learning_rate": 4.314204314204314e-05,
|
| 382 |
+
"loss": 0.4503,
|
| 383 |
+
"step": 610
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 1.14,
|
| 387 |
+
"learning_rate": 4.293854293854294e-05,
|
| 388 |
+
"loss": 0.4325,
|
| 389 |
+
"step": 620
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 1.15,
|
| 393 |
+
"learning_rate": 4.2735042735042735e-05,
|
| 394 |
+
"loss": 0.4398,
|
| 395 |
+
"step": 630
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.17,
|
| 399 |
+
"learning_rate": 4.2531542531542535e-05,
|
| 400 |
+
"loss": 0.4212,
|
| 401 |
+
"step": 640
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.19,
|
| 405 |
+
"learning_rate": 4.232804232804233e-05,
|
| 406 |
+
"loss": 0.4225,
|
| 407 |
+
"step": 650
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 1.21,
|
| 411 |
+
"learning_rate": 4.212454212454213e-05,
|
| 412 |
+
"loss": 0.4071,
|
| 413 |
+
"step": 660
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 1.23,
|
| 417 |
+
"learning_rate": 4.192104192104192e-05,
|
| 418 |
+
"loss": 0.3837,
|
| 419 |
+
"step": 670
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 1.25,
|
| 423 |
+
"learning_rate": 4.171754171754172e-05,
|
| 424 |
+
"loss": 0.3924,
|
| 425 |
+
"step": 680
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.26,
|
| 429 |
+
"learning_rate": 4.1514041514041515e-05,
|
| 430 |
+
"loss": 0.4217,
|
| 431 |
+
"step": 690
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 1.28,
|
| 435 |
+
"learning_rate": 4.131054131054131e-05,
|
| 436 |
+
"loss": 0.4186,
|
| 437 |
+
"step": 700
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.3,
|
| 441 |
+
"learning_rate": 4.110704110704111e-05,
|
| 442 |
+
"loss": 0.3931,
|
| 443 |
+
"step": 710
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.32,
|
| 447 |
+
"learning_rate": 4.09035409035409e-05,
|
| 448 |
+
"loss": 0.4057,
|
| 449 |
+
"step": 720
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 1.34,
|
| 453 |
+
"learning_rate": 4.07000407000407e-05,
|
| 454 |
+
"loss": 0.3512,
|
| 455 |
+
"step": 730
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 1.35,
|
| 459 |
+
"learning_rate": 4.0496540496540496e-05,
|
| 460 |
+
"loss": 0.3781,
|
| 461 |
+
"step": 740
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 1.37,
|
| 465 |
+
"learning_rate": 4.0293040293040296e-05,
|
| 466 |
+
"loss": 0.3737,
|
| 467 |
+
"step": 750
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.39,
|
| 471 |
+
"learning_rate": 4.008954008954009e-05,
|
| 472 |
+
"loss": 0.3852,
|
| 473 |
+
"step": 760
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 1.41,
|
| 477 |
+
"learning_rate": 3.988603988603989e-05,
|
| 478 |
+
"loss": 0.37,
|
| 479 |
+
"step": 770
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.43,
|
| 483 |
+
"learning_rate": 3.968253968253968e-05,
|
| 484 |
+
"loss": 0.3596,
|
| 485 |
+
"step": 780
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.45,
|
| 489 |
+
"learning_rate": 3.947903947903948e-05,
|
| 490 |
+
"loss": 0.434,
|
| 491 |
+
"step": 790
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 1.46,
|
| 495 |
+
"learning_rate": 3.9275539275539276e-05,
|
| 496 |
+
"loss": 0.3793,
|
| 497 |
+
"step": 800
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 1.48,
|
| 501 |
+
"learning_rate": 3.9072039072039076e-05,
|
| 502 |
+
"loss": 0.3407,
|
| 503 |
+
"step": 810
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.5,
|
| 507 |
+
"learning_rate": 3.8868538868538876e-05,
|
| 508 |
+
"loss": 0.3758,
|
| 509 |
+
"step": 820
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 1.52,
|
| 513 |
+
"learning_rate": 3.866503866503867e-05,
|
| 514 |
+
"loss": 0.3451,
|
| 515 |
+
"step": 830
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 1.54,
|
| 519 |
+
"learning_rate": 3.846153846153846e-05,
|
| 520 |
+
"loss": 0.355,
|
| 521 |
+
"step": 840
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 1.56,
|
| 525 |
+
"learning_rate": 3.8258038258038256e-05,
|
| 526 |
+
"loss": 0.3811,
|
| 527 |
+
"step": 850
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 1.57,
|
| 531 |
+
"learning_rate": 3.8054538054538056e-05,
|
| 532 |
+
"loss": 0.3691,
|
| 533 |
+
"step": 860
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 1.59,
|
| 537 |
+
"learning_rate": 3.785103785103785e-05,
|
| 538 |
+
"loss": 0.3389,
|
| 539 |
+
"step": 870
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 1.61,
|
| 543 |
+
"learning_rate": 3.764753764753765e-05,
|
| 544 |
+
"loss": 0.3486,
|
| 545 |
+
"step": 880
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.63,
|
| 549 |
+
"learning_rate": 3.744403744403744e-05,
|
| 550 |
+
"loss": 0.393,
|
| 551 |
+
"step": 890
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 1.65,
|
| 555 |
+
"learning_rate": 3.724053724053724e-05,
|
| 556 |
+
"loss": 0.3611,
|
| 557 |
+
"step": 900
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 1.67,
|
| 561 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 562 |
+
"loss": 0.3613,
|
| 563 |
+
"step": 910
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 1.68,
|
| 567 |
+
"learning_rate": 3.683353683353684e-05,
|
| 568 |
+
"loss": 0.3248,
|
| 569 |
+
"step": 920
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 1.7,
|
| 573 |
+
"learning_rate": 3.663003663003663e-05,
|
| 574 |
+
"loss": 0.3462,
|
| 575 |
+
"step": 930
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 1.72,
|
| 579 |
+
"learning_rate": 3.642653642653643e-05,
|
| 580 |
+
"loss": 0.3602,
|
| 581 |
+
"step": 940
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 1.74,
|
| 585 |
+
"learning_rate": 3.622303622303623e-05,
|
| 586 |
+
"loss": 0.3113,
|
| 587 |
+
"step": 950
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.76,
|
| 591 |
+
"learning_rate": 3.6019536019536024e-05,
|
| 592 |
+
"loss": 0.3538,
|
| 593 |
+
"step": 960
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 1.78,
|
| 597 |
+
"learning_rate": 3.581603581603582e-05,
|
| 598 |
+
"loss": 0.3354,
|
| 599 |
+
"step": 970
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 1.79,
|
| 603 |
+
"learning_rate": 3.561253561253561e-05,
|
| 604 |
+
"loss": 0.3265,
|
| 605 |
+
"step": 980
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.81,
|
| 609 |
+
"learning_rate": 3.540903540903541e-05,
|
| 610 |
+
"loss": 0.316,
|
| 611 |
+
"step": 990
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 1.83,
|
| 615 |
+
"learning_rate": 3.5205535205535204e-05,
|
| 616 |
+
"loss": 0.3425,
|
| 617 |
+
"step": 1000
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 1.85,
|
| 621 |
+
"learning_rate": 3.5002035002035004e-05,
|
| 622 |
+
"loss": 0.3139,
|
| 623 |
+
"step": 1010
|
| 624 |
+
},
|
| 625 |
+
{
|
| 626 |
+
"epoch": 1.87,
|
| 627 |
+
"learning_rate": 3.47985347985348e-05,
|
| 628 |
+
"loss": 0.3099,
|
| 629 |
+
"step": 1020
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.89,
|
| 633 |
+
"learning_rate": 3.45950345950346e-05,
|
| 634 |
+
"loss": 0.3213,
|
| 635 |
+
"step": 1030
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 1.9,
|
| 639 |
+
"learning_rate": 3.439153439153439e-05,
|
| 640 |
+
"loss": 0.348,
|
| 641 |
+
"step": 1040
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 1.92,
|
| 645 |
+
"learning_rate": 3.418803418803419e-05,
|
| 646 |
+
"loss": 0.2939,
|
| 647 |
+
"step": 1050
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.94,
|
| 651 |
+
"learning_rate": 3.3984533984533984e-05,
|
| 652 |
+
"loss": 0.3144,
|
| 653 |
+
"step": 1060
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 1.96,
|
| 657 |
+
"learning_rate": 3.3781033781033784e-05,
|
| 658 |
+
"loss": 0.2839,
|
| 659 |
+
"step": 1070
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 1.98,
|
| 663 |
+
"learning_rate": 3.357753357753358e-05,
|
| 664 |
+
"loss": 0.3333,
|
| 665 |
+
"step": 1080
|
| 666 |
+
},
|
| 667 |
+
{
|
| 668 |
+
"epoch": 1.99,
|
| 669 |
+
"learning_rate": 3.337403337403338e-05,
|
| 670 |
+
"loss": 0.2915,
|
| 671 |
+
"step": 1090
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 2.0,
|
| 675 |
+
"eval_accuracy": 0.9199935217426513,
|
| 676 |
+
"eval_loss": 0.19980975985527039,
|
| 677 |
+
"eval_runtime": 180.6917,
|
| 678 |
+
"eval_samples_per_second": 68.343,
|
| 679 |
+
"eval_steps_per_second": 2.136,
|
| 680 |
+
"step": 1092
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 2.01,
|
| 684 |
+
"learning_rate": 3.317053317053318e-05,
|
| 685 |
+
"loss": 0.3417,
|
| 686 |
+
"step": 1100
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 2.03,
|
| 690 |
+
"learning_rate": 3.296703296703297e-05,
|
| 691 |
+
"loss": 0.3149,
|
| 692 |
+
"step": 1110
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 2.05,
|
| 696 |
+
"learning_rate": 3.2763532763532764e-05,
|
| 697 |
+
"loss": 0.3164,
|
| 698 |
+
"step": 1120
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 2.07,
|
| 702 |
+
"learning_rate": 3.256003256003256e-05,
|
| 703 |
+
"loss": 0.3036,
|
| 704 |
+
"step": 1130
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.09,
|
| 708 |
+
"learning_rate": 3.235653235653236e-05,
|
| 709 |
+
"loss": 0.3119,
|
| 710 |
+
"step": 1140
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.11,
|
| 714 |
+
"learning_rate": 3.215303215303215e-05,
|
| 715 |
+
"loss": 0.3086,
|
| 716 |
+
"step": 1150
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.12,
|
| 720 |
+
"learning_rate": 3.194953194953195e-05,
|
| 721 |
+
"loss": 0.2903,
|
| 722 |
+
"step": 1160
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 2.14,
|
| 726 |
+
"learning_rate": 3.1746031746031745e-05,
|
| 727 |
+
"loss": 0.3125,
|
| 728 |
+
"step": 1170
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 2.16,
|
| 732 |
+
"learning_rate": 3.1542531542531545e-05,
|
| 733 |
+
"loss": 0.3052,
|
| 734 |
+
"step": 1180
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.18,
|
| 738 |
+
"learning_rate": 3.133903133903134e-05,
|
| 739 |
+
"loss": 0.2671,
|
| 740 |
+
"step": 1190
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 2.2,
|
| 744 |
+
"learning_rate": 3.113553113553114e-05,
|
| 745 |
+
"loss": 0.3126,
|
| 746 |
+
"step": 1200
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 2.22,
|
| 750 |
+
"learning_rate": 3.093203093203093e-05,
|
| 751 |
+
"loss": 0.2785,
|
| 752 |
+
"step": 1210
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.23,
|
| 756 |
+
"learning_rate": 3.072853072853073e-05,
|
| 757 |
+
"loss": 0.2877,
|
| 758 |
+
"step": 1220
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.25,
|
| 762 |
+
"learning_rate": 3.052503052503053e-05,
|
| 763 |
+
"loss": 0.2819,
|
| 764 |
+
"step": 1230
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 2.27,
|
| 768 |
+
"learning_rate": 3.0321530321530322e-05,
|
| 769 |
+
"loss": 0.2669,
|
| 770 |
+
"step": 1240
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 2.29,
|
| 774 |
+
"learning_rate": 3.0118030118030122e-05,
|
| 775 |
+
"loss": 0.3046,
|
| 776 |
+
"step": 1250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 2.31,
|
| 780 |
+
"learning_rate": 2.9914529914529915e-05,
|
| 781 |
+
"loss": 0.2825,
|
| 782 |
+
"step": 1260
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 2.33,
|
| 786 |
+
"learning_rate": 2.9711029711029715e-05,
|
| 787 |
+
"loss": 0.2749,
|
| 788 |
+
"step": 1270
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 2.34,
|
| 792 |
+
"learning_rate": 2.950752950752951e-05,
|
| 793 |
+
"loss": 0.2673,
|
| 794 |
+
"step": 1280
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 2.36,
|
| 798 |
+
"learning_rate": 2.9304029304029305e-05,
|
| 799 |
+
"loss": 0.2823,
|
| 800 |
+
"step": 1290
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 2.38,
|
| 804 |
+
"learning_rate": 2.91005291005291e-05,
|
| 805 |
+
"loss": 0.2895,
|
| 806 |
+
"step": 1300
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 2.4,
|
| 810 |
+
"learning_rate": 2.88970288970289e-05,
|
| 811 |
+
"loss": 0.2611,
|
| 812 |
+
"step": 1310
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 2.42,
|
| 816 |
+
"learning_rate": 2.8693528693528692e-05,
|
| 817 |
+
"loss": 0.2632,
|
| 818 |
+
"step": 1320
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 2.44,
|
| 822 |
+
"learning_rate": 2.8490028490028492e-05,
|
| 823 |
+
"loss": 0.2639,
|
| 824 |
+
"step": 1330
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 2.45,
|
| 828 |
+
"learning_rate": 2.8286528286528286e-05,
|
| 829 |
+
"loss": 0.263,
|
| 830 |
+
"step": 1340
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 2.47,
|
| 834 |
+
"learning_rate": 2.8083028083028086e-05,
|
| 835 |
+
"loss": 0.2685,
|
| 836 |
+
"step": 1350
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 2.49,
|
| 840 |
+
"learning_rate": 2.7879527879527882e-05,
|
| 841 |
+
"loss": 0.2882,
|
| 842 |
+
"step": 1360
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 2.51,
|
| 846 |
+
"learning_rate": 2.7676027676027676e-05,
|
| 847 |
+
"loss": 0.2422,
|
| 848 |
+
"step": 1370
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 2.53,
|
| 852 |
+
"learning_rate": 2.7472527472527476e-05,
|
| 853 |
+
"loss": 0.2834,
|
| 854 |
+
"step": 1380
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 2.55,
|
| 858 |
+
"learning_rate": 2.726902726902727e-05,
|
| 859 |
+
"loss": 0.2568,
|
| 860 |
+
"step": 1390
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 2.56,
|
| 864 |
+
"learning_rate": 2.706552706552707e-05,
|
| 865 |
+
"loss": 0.2672,
|
| 866 |
+
"step": 1400
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 2.58,
|
| 870 |
+
"learning_rate": 2.6862026862026863e-05,
|
| 871 |
+
"loss": 0.2288,
|
| 872 |
+
"step": 1410
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.6,
|
| 876 |
+
"learning_rate": 2.6658526658526663e-05,
|
| 877 |
+
"loss": 0.265,
|
| 878 |
+
"step": 1420
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 2.62,
|
| 882 |
+
"learning_rate": 2.6455026455026456e-05,
|
| 883 |
+
"loss": 0.2578,
|
| 884 |
+
"step": 1430
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.64,
|
| 888 |
+
"learning_rate": 2.6251526251526253e-05,
|
| 889 |
+
"loss": 0.2503,
|
| 890 |
+
"step": 1440
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 2.65,
|
| 894 |
+
"learning_rate": 2.6048026048026046e-05,
|
| 895 |
+
"loss": 0.2676,
|
| 896 |
+
"step": 1450
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 2.67,
|
| 900 |
+
"learning_rate": 2.5844525844525846e-05,
|
| 901 |
+
"loss": 0.2336,
|
| 902 |
+
"step": 1460
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 2.69,
|
| 906 |
+
"learning_rate": 2.564102564102564e-05,
|
| 907 |
+
"loss": 0.246,
|
| 908 |
+
"step": 1470
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.71,
|
| 912 |
+
"learning_rate": 2.543752543752544e-05,
|
| 913 |
+
"loss": 0.2511,
|
| 914 |
+
"step": 1480
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.73,
|
| 918 |
+
"learning_rate": 2.5234025234025233e-05,
|
| 919 |
+
"loss": 0.2704,
|
| 920 |
+
"step": 1490
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 2.75,
|
| 924 |
+
"learning_rate": 2.5030525030525033e-05,
|
| 925 |
+
"loss": 0.3,
|
| 926 |
+
"step": 1500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.76,
|
| 930 |
+
"learning_rate": 2.4827024827024827e-05,
|
| 931 |
+
"loss": 0.2671,
|
| 932 |
+
"step": 1510
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.78,
|
| 936 |
+
"learning_rate": 2.4623524623524623e-05,
|
| 937 |
+
"loss": 0.2473,
|
| 938 |
+
"step": 1520
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.8,
|
| 942 |
+
"learning_rate": 2.442002442002442e-05,
|
| 943 |
+
"loss": 0.2289,
|
| 944 |
+
"step": 1530
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.82,
|
| 948 |
+
"learning_rate": 2.4216524216524217e-05,
|
| 949 |
+
"loss": 0.246,
|
| 950 |
+
"step": 1540
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.84,
|
| 954 |
+
"learning_rate": 2.4013024013024013e-05,
|
| 955 |
+
"loss": 0.2828,
|
| 956 |
+
"step": 1550
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.86,
|
| 960 |
+
"learning_rate": 2.380952380952381e-05,
|
| 961 |
+
"loss": 0.2726,
|
| 962 |
+
"step": 1560
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 2.87,
|
| 966 |
+
"learning_rate": 2.3606023606023607e-05,
|
| 967 |
+
"loss": 0.2543,
|
| 968 |
+
"step": 1570
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.89,
|
| 972 |
+
"learning_rate": 2.3402523402523404e-05,
|
| 973 |
+
"loss": 0.2714,
|
| 974 |
+
"step": 1580
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.91,
|
| 978 |
+
"learning_rate": 2.31990231990232e-05,
|
| 979 |
+
"loss": 0.2416,
|
| 980 |
+
"step": 1590
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 2.93,
|
| 984 |
+
"learning_rate": 2.2995522995522997e-05,
|
| 985 |
+
"loss": 0.2495,
|
| 986 |
+
"step": 1600
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.95,
|
| 990 |
+
"learning_rate": 2.2792022792022794e-05,
|
| 991 |
+
"loss": 0.2569,
|
| 992 |
+
"step": 1610
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 2.97,
|
| 996 |
+
"learning_rate": 2.258852258852259e-05,
|
| 997 |
+
"loss": 0.2437,
|
| 998 |
+
"step": 1620
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.98,
|
| 1002 |
+
"learning_rate": 2.2385022385022387e-05,
|
| 1003 |
+
"loss": 0.2333,
|
| 1004 |
+
"step": 1630
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 3.0,
|
| 1008 |
+
"eval_accuracy": 0.9565956757632197,
|
| 1009 |
+
"eval_loss": 0.11550568044185638,
|
| 1010 |
+
"eval_runtime": 172.257,
|
| 1011 |
+
"eval_samples_per_second": 71.689,
|
| 1012 |
+
"eval_steps_per_second": 2.241,
|
| 1013 |
+
"step": 1638
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 3.0,
|
| 1017 |
+
"learning_rate": 2.2181522181522184e-05,
|
| 1018 |
+
"loss": 0.2693,
|
| 1019 |
+
"step": 1640
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 3.02,
|
| 1023 |
+
"learning_rate": 2.1978021978021977e-05,
|
| 1024 |
+
"loss": 0.2483,
|
| 1025 |
+
"step": 1650
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 3.04,
|
| 1029 |
+
"learning_rate": 2.1774521774521774e-05,
|
| 1030 |
+
"loss": 0.2182,
|
| 1031 |
+
"step": 1660
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 3.06,
|
| 1035 |
+
"learning_rate": 2.157102157102157e-05,
|
| 1036 |
+
"loss": 0.243,
|
| 1037 |
+
"step": 1670
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 3.08,
|
| 1041 |
+
"learning_rate": 2.1367521367521368e-05,
|
| 1042 |
+
"loss": 0.2373,
|
| 1043 |
+
"step": 1680
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 3.1,
|
| 1047 |
+
"learning_rate": 2.1164021164021164e-05,
|
| 1048 |
+
"loss": 0.2434,
|
| 1049 |
+
"step": 1690
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 3.11,
|
| 1053 |
+
"learning_rate": 2.096052096052096e-05,
|
| 1054 |
+
"loss": 0.2494,
|
| 1055 |
+
"step": 1700
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 3.13,
|
| 1059 |
+
"learning_rate": 2.0757020757020758e-05,
|
| 1060 |
+
"loss": 0.2529,
|
| 1061 |
+
"step": 1710
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 3.15,
|
| 1065 |
+
"learning_rate": 2.0553520553520554e-05,
|
| 1066 |
+
"loss": 0.2363,
|
| 1067 |
+
"step": 1720
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 3.17,
|
| 1071 |
+
"learning_rate": 2.035002035002035e-05,
|
| 1072 |
+
"loss": 0.2397,
|
| 1073 |
+
"step": 1730
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 3.19,
|
| 1077 |
+
"learning_rate": 2.0146520146520148e-05,
|
| 1078 |
+
"loss": 0.2121,
|
| 1079 |
+
"step": 1740
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 3.2,
|
| 1083 |
+
"learning_rate": 1.9943019943019945e-05,
|
| 1084 |
+
"loss": 0.2478,
|
| 1085 |
+
"step": 1750
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 3.22,
|
| 1089 |
+
"learning_rate": 1.973951973951974e-05,
|
| 1090 |
+
"loss": 0.2239,
|
| 1091 |
+
"step": 1760
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 3.24,
|
| 1095 |
+
"learning_rate": 1.9536019536019538e-05,
|
| 1096 |
+
"loss": 0.209,
|
| 1097 |
+
"step": 1770
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 3.26,
|
| 1101 |
+
"learning_rate": 1.9332519332519335e-05,
|
| 1102 |
+
"loss": 0.2458,
|
| 1103 |
+
"step": 1780
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 3.28,
|
| 1107 |
+
"learning_rate": 1.9129019129019128e-05,
|
| 1108 |
+
"loss": 0.2203,
|
| 1109 |
+
"step": 1790
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 3.3,
|
| 1113 |
+
"learning_rate": 1.8925518925518925e-05,
|
| 1114 |
+
"loss": 0.2039,
|
| 1115 |
+
"step": 1800
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 3.31,
|
| 1119 |
+
"learning_rate": 1.872201872201872e-05,
|
| 1120 |
+
"loss": 0.2239,
|
| 1121 |
+
"step": 1810
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 3.33,
|
| 1125 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 1126 |
+
"loss": 0.2587,
|
| 1127 |
+
"step": 1820
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 3.35,
|
| 1131 |
+
"learning_rate": 1.8315018315018315e-05,
|
| 1132 |
+
"loss": 0.2194,
|
| 1133 |
+
"step": 1830
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 3.37,
|
| 1137 |
+
"learning_rate": 1.8111518111518115e-05,
|
| 1138 |
+
"loss": 0.2218,
|
| 1139 |
+
"step": 1840
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 3.39,
|
| 1143 |
+
"learning_rate": 1.790801790801791e-05,
|
| 1144 |
+
"loss": 0.2305,
|
| 1145 |
+
"step": 1850
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 3.41,
|
| 1149 |
+
"learning_rate": 1.7704517704517705e-05,
|
| 1150 |
+
"loss": 0.2053,
|
| 1151 |
+
"step": 1860
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 3.42,
|
| 1155 |
+
"learning_rate": 1.7501017501017502e-05,
|
| 1156 |
+
"loss": 0.2416,
|
| 1157 |
+
"step": 1870
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 3.44,
|
| 1161 |
+
"learning_rate": 1.72975172975173e-05,
|
| 1162 |
+
"loss": 0.1992,
|
| 1163 |
+
"step": 1880
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 3.46,
|
| 1167 |
+
"learning_rate": 1.7094017094017095e-05,
|
| 1168 |
+
"loss": 0.22,
|
| 1169 |
+
"step": 1890
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 3.48,
|
| 1173 |
+
"learning_rate": 1.6890516890516892e-05,
|
| 1174 |
+
"loss": 0.1947,
|
| 1175 |
+
"step": 1900
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 3.5,
|
| 1179 |
+
"learning_rate": 1.668701668701669e-05,
|
| 1180 |
+
"loss": 0.1955,
|
| 1181 |
+
"step": 1910
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 3.52,
|
| 1185 |
+
"learning_rate": 1.6483516483516486e-05,
|
| 1186 |
+
"loss": 0.2087,
|
| 1187 |
+
"step": 1920
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 3.53,
|
| 1191 |
+
"learning_rate": 1.628001628001628e-05,
|
| 1192 |
+
"loss": 0.2331,
|
| 1193 |
+
"step": 1930
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 3.55,
|
| 1197 |
+
"learning_rate": 1.6076516076516076e-05,
|
| 1198 |
+
"loss": 0.2065,
|
| 1199 |
+
"step": 1940
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 3.57,
|
| 1203 |
+
"learning_rate": 1.5873015873015872e-05,
|
| 1204 |
+
"loss": 0.1832,
|
| 1205 |
+
"step": 1950
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 3.59,
|
| 1209 |
+
"learning_rate": 1.566951566951567e-05,
|
| 1210 |
+
"loss": 0.2141,
|
| 1211 |
+
"step": 1960
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 3.61,
|
| 1215 |
+
"learning_rate": 1.5466015466015466e-05,
|
| 1216 |
+
"loss": 0.2224,
|
| 1217 |
+
"step": 1970
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 3.63,
|
| 1221 |
+
"learning_rate": 1.5262515262515266e-05,
|
| 1222 |
+
"loss": 0.211,
|
| 1223 |
+
"step": 1980
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 3.64,
|
| 1227 |
+
"learning_rate": 1.5059015059015061e-05,
|
| 1228 |
+
"loss": 0.1843,
|
| 1229 |
+
"step": 1990
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 3.66,
|
| 1233 |
+
"learning_rate": 1.4855514855514858e-05,
|
| 1234 |
+
"loss": 0.2199,
|
| 1235 |
+
"step": 2000
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 3.68,
|
| 1239 |
+
"learning_rate": 1.4652014652014653e-05,
|
| 1240 |
+
"loss": 0.1834,
|
| 1241 |
+
"step": 2010
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 3.7,
|
| 1245 |
+
"learning_rate": 1.444851444851445e-05,
|
| 1246 |
+
"loss": 0.1961,
|
| 1247 |
+
"step": 2020
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 3.72,
|
| 1251 |
+
"learning_rate": 1.4245014245014246e-05,
|
| 1252 |
+
"loss": 0.2311,
|
| 1253 |
+
"step": 2030
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 3.74,
|
| 1257 |
+
"learning_rate": 1.4041514041514043e-05,
|
| 1258 |
+
"loss": 0.2076,
|
| 1259 |
+
"step": 2040
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 3.75,
|
| 1263 |
+
"learning_rate": 1.3838013838013838e-05,
|
| 1264 |
+
"loss": 0.2132,
|
| 1265 |
+
"step": 2050
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 3.77,
|
| 1269 |
+
"learning_rate": 1.3634513634513635e-05,
|
| 1270 |
+
"loss": 0.2069,
|
| 1271 |
+
"step": 2060
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 3.79,
|
| 1275 |
+
"learning_rate": 1.3431013431013431e-05,
|
| 1276 |
+
"loss": 0.2255,
|
| 1277 |
+
"step": 2070
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 3.81,
|
| 1281 |
+
"learning_rate": 1.3227513227513228e-05,
|
| 1282 |
+
"loss": 0.2241,
|
| 1283 |
+
"step": 2080
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 3.83,
|
| 1287 |
+
"learning_rate": 1.3024013024013023e-05,
|
| 1288 |
+
"loss": 0.1935,
|
| 1289 |
+
"step": 2090
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 3.84,
|
| 1293 |
+
"learning_rate": 1.282051282051282e-05,
|
| 1294 |
+
"loss": 0.2236,
|
| 1295 |
+
"step": 2100
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 3.86,
|
| 1299 |
+
"learning_rate": 1.2617012617012617e-05,
|
| 1300 |
+
"loss": 0.2063,
|
| 1301 |
+
"step": 2110
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 3.88,
|
| 1305 |
+
"learning_rate": 1.2413512413512413e-05,
|
| 1306 |
+
"loss": 0.194,
|
| 1307 |
+
"step": 2120
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 3.9,
|
| 1311 |
+
"learning_rate": 1.221001221001221e-05,
|
| 1312 |
+
"loss": 0.2024,
|
| 1313 |
+
"step": 2130
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 3.92,
|
| 1317 |
+
"learning_rate": 1.2006512006512007e-05,
|
| 1318 |
+
"loss": 0.1805,
|
| 1319 |
+
"step": 2140
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 3.94,
|
| 1323 |
+
"learning_rate": 1.1803011803011803e-05,
|
| 1324 |
+
"loss": 0.1947,
|
| 1325 |
+
"step": 2150
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 3.95,
|
| 1329 |
+
"learning_rate": 1.15995115995116e-05,
|
| 1330 |
+
"loss": 0.2187,
|
| 1331 |
+
"step": 2160
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 3.97,
|
| 1335 |
+
"learning_rate": 1.1396011396011397e-05,
|
| 1336 |
+
"loss": 0.2181,
|
| 1337 |
+
"step": 2170
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 3.99,
|
| 1341 |
+
"learning_rate": 1.1192511192511194e-05,
|
| 1342 |
+
"loss": 0.2019,
|
| 1343 |
+
"step": 2180
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 4.0,
|
| 1347 |
+
"eval_accuracy": 0.9633978459794316,
|
| 1348 |
+
"eval_loss": 0.09769698977470398,
|
| 1349 |
+
"eval_runtime": 171.1977,
|
| 1350 |
+
"eval_samples_per_second": 72.133,
|
| 1351 |
+
"eval_steps_per_second": 2.255,
|
| 1352 |
+
"step": 2184
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 4.01,
|
| 1356 |
+
"learning_rate": 1.0989010989010989e-05,
|
| 1357 |
+
"loss": 0.2576,
|
| 1358 |
+
"step": 2190
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 4.03,
|
| 1362 |
+
"learning_rate": 1.0785510785510785e-05,
|
| 1363 |
+
"loss": 0.2054,
|
| 1364 |
+
"step": 2200
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 4.05,
|
| 1368 |
+
"learning_rate": 1.0582010582010582e-05,
|
| 1369 |
+
"loss": 0.2028,
|
| 1370 |
+
"step": 2210
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 4.07,
|
| 1374 |
+
"learning_rate": 1.0378510378510379e-05,
|
| 1375 |
+
"loss": 0.1979,
|
| 1376 |
+
"step": 2220
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 4.08,
|
| 1380 |
+
"learning_rate": 1.0175010175010176e-05,
|
| 1381 |
+
"loss": 0.2037,
|
| 1382 |
+
"step": 2230
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 4.1,
|
| 1386 |
+
"learning_rate": 9.971509971509972e-06,
|
| 1387 |
+
"loss": 0.2049,
|
| 1388 |
+
"step": 2240
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 4.12,
|
| 1392 |
+
"learning_rate": 9.768009768009769e-06,
|
| 1393 |
+
"loss": 0.2185,
|
| 1394 |
+
"step": 2250
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 4.14,
|
| 1398 |
+
"learning_rate": 9.564509564509564e-06,
|
| 1399 |
+
"loss": 0.1796,
|
| 1400 |
+
"step": 2260
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 4.16,
|
| 1404 |
+
"learning_rate": 9.36100936100936e-06,
|
| 1405 |
+
"loss": 0.2016,
|
| 1406 |
+
"step": 2270
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 4.18,
|
| 1410 |
+
"learning_rate": 9.157509157509158e-06,
|
| 1411 |
+
"loss": 0.1976,
|
| 1412 |
+
"step": 2280
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 4.19,
|
| 1416 |
+
"learning_rate": 8.954008954008954e-06,
|
| 1417 |
+
"loss": 0.1821,
|
| 1418 |
+
"step": 2290
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 4.21,
|
| 1422 |
+
"learning_rate": 8.750508750508751e-06,
|
| 1423 |
+
"loss": 0.1943,
|
| 1424 |
+
"step": 2300
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 4.23,
|
| 1428 |
+
"learning_rate": 8.547008547008548e-06,
|
| 1429 |
+
"loss": 0.1965,
|
| 1430 |
+
"step": 2310
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 4.25,
|
| 1434 |
+
"learning_rate": 8.343508343508344e-06,
|
| 1435 |
+
"loss": 0.1822,
|
| 1436 |
+
"step": 2320
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 4.27,
|
| 1440 |
+
"learning_rate": 8.14000814000814e-06,
|
| 1441 |
+
"loss": 0.1936,
|
| 1442 |
+
"step": 2330
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 4.29,
|
| 1446 |
+
"learning_rate": 7.936507936507936e-06,
|
| 1447 |
+
"loss": 0.1981,
|
| 1448 |
+
"step": 2340
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 4.3,
|
| 1452 |
+
"learning_rate": 7.733007733007733e-06,
|
| 1453 |
+
"loss": 0.1931,
|
| 1454 |
+
"step": 2350
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 4.32,
|
| 1458 |
+
"learning_rate": 7.5295075295075305e-06,
|
| 1459 |
+
"loss": 0.2053,
|
| 1460 |
+
"step": 2360
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 4.34,
|
| 1464 |
+
"learning_rate": 7.326007326007326e-06,
|
| 1465 |
+
"loss": 0.1934,
|
| 1466 |
+
"step": 2370
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 4.36,
|
| 1470 |
+
"learning_rate": 7.122507122507123e-06,
|
| 1471 |
+
"loss": 0.1906,
|
| 1472 |
+
"step": 2380
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 4.38,
|
| 1476 |
+
"learning_rate": 6.919006919006919e-06,
|
| 1477 |
+
"loss": 0.1878,
|
| 1478 |
+
"step": 2390
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 4.4,
|
| 1482 |
+
"learning_rate": 6.715506715506716e-06,
|
| 1483 |
+
"loss": 0.1953,
|
| 1484 |
+
"step": 2400
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 4.41,
|
| 1488 |
+
"learning_rate": 6.5120065120065115e-06,
|
| 1489 |
+
"loss": 0.2172,
|
| 1490 |
+
"step": 2410
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 4.43,
|
| 1494 |
+
"learning_rate": 6.308506308506308e-06,
|
| 1495 |
+
"loss": 0.2084,
|
| 1496 |
+
"step": 2420
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 4.45,
|
| 1500 |
+
"learning_rate": 6.105006105006105e-06,
|
| 1501 |
+
"loss": 0.1821,
|
| 1502 |
+
"step": 2430
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 4.47,
|
| 1506 |
+
"learning_rate": 5.901505901505902e-06,
|
| 1507 |
+
"loss": 0.2012,
|
| 1508 |
+
"step": 2440
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 4.49,
|
| 1512 |
+
"learning_rate": 5.6980056980056985e-06,
|
| 1513 |
+
"loss": 0.2058,
|
| 1514 |
+
"step": 2450
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 4.5,
|
| 1518 |
+
"learning_rate": 5.494505494505494e-06,
|
| 1519 |
+
"loss": 0.1918,
|
| 1520 |
+
"step": 2460
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 4.52,
|
| 1524 |
+
"learning_rate": 5.291005291005291e-06,
|
| 1525 |
+
"loss": 0.1953,
|
| 1526 |
+
"step": 2470
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 4.54,
|
| 1530 |
+
"learning_rate": 5.087505087505088e-06,
|
| 1531 |
+
"loss": 0.2129,
|
| 1532 |
+
"step": 2480
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 4.56,
|
| 1536 |
+
"learning_rate": 4.8840048840048845e-06,
|
| 1537 |
+
"loss": 0.1748,
|
| 1538 |
+
"step": 2490
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 4.58,
|
| 1542 |
+
"learning_rate": 4.68050468050468e-06,
|
| 1543 |
+
"loss": 0.2034,
|
| 1544 |
+
"step": 2500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 4.6,
|
| 1548 |
+
"learning_rate": 4.477004477004477e-06,
|
| 1549 |
+
"loss": 0.1731,
|
| 1550 |
+
"step": 2510
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 4.61,
|
| 1554 |
+
"learning_rate": 4.273504273504274e-06,
|
| 1555 |
+
"loss": 0.166,
|
| 1556 |
+
"step": 2520
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 4.63,
|
| 1560 |
+
"learning_rate": 4.07000407000407e-06,
|
| 1561 |
+
"loss": 0.188,
|
| 1562 |
+
"step": 2530
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 4.65,
|
| 1566 |
+
"learning_rate": 3.8665038665038664e-06,
|
| 1567 |
+
"loss": 0.1726,
|
| 1568 |
+
"step": 2540
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 4.67,
|
| 1572 |
+
"learning_rate": 3.663003663003663e-06,
|
| 1573 |
+
"loss": 0.181,
|
| 1574 |
+
"step": 2550
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 4.69,
|
| 1578 |
+
"learning_rate": 3.4595034595034595e-06,
|
| 1579 |
+
"loss": 0.1683,
|
| 1580 |
+
"step": 2560
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 4.71,
|
| 1584 |
+
"learning_rate": 3.2560032560032558e-06,
|
| 1585 |
+
"loss": 0.186,
|
| 1586 |
+
"step": 2570
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 4.72,
|
| 1590 |
+
"learning_rate": 3.0525030525030525e-06,
|
| 1591 |
+
"loss": 0.2046,
|
| 1592 |
+
"step": 2580
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 4.74,
|
| 1596 |
+
"learning_rate": 2.8490028490028492e-06,
|
| 1597 |
+
"loss": 0.1942,
|
| 1598 |
+
"step": 2590
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 4.76,
|
| 1602 |
+
"learning_rate": 2.6455026455026455e-06,
|
| 1603 |
+
"loss": 0.1819,
|
| 1604 |
+
"step": 2600
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 4.78,
|
| 1608 |
+
"learning_rate": 2.4420024420024423e-06,
|
| 1609 |
+
"loss": 0.1931,
|
| 1610 |
+
"step": 2610
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 4.8,
|
| 1614 |
+
"learning_rate": 2.2385022385022386e-06,
|
| 1615 |
+
"loss": 0.2086,
|
| 1616 |
+
"step": 2620
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 4.82,
|
| 1620 |
+
"learning_rate": 2.035002035002035e-06,
|
| 1621 |
+
"loss": 0.1911,
|
| 1622 |
+
"step": 2630
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 4.83,
|
| 1626 |
+
"learning_rate": 1.8315018315018316e-06,
|
| 1627 |
+
"loss": 0.186,
|
| 1628 |
+
"step": 2640
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 4.85,
|
| 1632 |
+
"learning_rate": 1.6280016280016279e-06,
|
| 1633 |
+
"loss": 0.1697,
|
| 1634 |
+
"step": 2650
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 4.87,
|
| 1638 |
+
"learning_rate": 1.4245014245014246e-06,
|
| 1639 |
+
"loss": 0.1676,
|
| 1640 |
+
"step": 2660
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 4.89,
|
| 1644 |
+
"learning_rate": 1.2210012210012211e-06,
|
| 1645 |
+
"loss": 0.2149,
|
| 1646 |
+
"step": 2670
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 4.91,
|
| 1650 |
+
"learning_rate": 1.0175010175010174e-06,
|
| 1651 |
+
"loss": 0.1704,
|
| 1652 |
+
"step": 2680
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 4.93,
|
| 1656 |
+
"learning_rate": 8.140008140008139e-07,
|
| 1657 |
+
"loss": 0.184,
|
| 1658 |
+
"step": 2690
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 4.94,
|
| 1662 |
+
"learning_rate": 6.105006105006106e-07,
|
| 1663 |
+
"loss": 0.1887,
|
| 1664 |
+
"step": 2700
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 4.96,
|
| 1668 |
+
"learning_rate": 4.0700040700040697e-07,
|
| 1669 |
+
"loss": 0.1865,
|
| 1670 |
+
"step": 2710
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 4.98,
|
| 1674 |
+
"learning_rate": 2.0350020350020349e-07,
|
| 1675 |
+
"loss": 0.1652,
|
| 1676 |
+
"step": 2720
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 5.0,
|
| 1680 |
+
"learning_rate": 0.0,
|
| 1681 |
+
"loss": 0.1713,
|
| 1682 |
+
"step": 2730
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 5.0,
|
| 1686 |
+
"eval_accuracy": 0.970523929063082,
|
| 1687 |
+
"eval_loss": 0.07879053056240082,
|
| 1688 |
+
"eval_runtime": 169.4659,
|
| 1689 |
+
"eval_samples_per_second": 72.87,
|
| 1690 |
+
"eval_steps_per_second": 2.278,
|
| 1691 |
+
"step": 2730
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 5.0,
|
| 1695 |
+
"step": 2730,
|
| 1696 |
+
"total_flos": 8.694514267698954e+18,
|
| 1697 |
+
"train_loss": 0.3431307647691105,
|
| 1698 |
+
"train_runtime": 8584.1671,
|
| 1699 |
+
"train_samples_per_second": 40.756,
|
| 1700 |
+
"train_steps_per_second": 0.318
|
| 1701 |
+
}
|
| 1702 |
+
],
|
| 1703 |
+
"max_steps": 2730,
|
| 1704 |
+
"num_train_epochs": 5,
|
| 1705 |
+
"total_flos": 8.694514267698954e+18,
|
| 1706 |
+
"trial_name": null,
|
| 1707 |
+
"trial_params": null
|
| 1708 |
+
}
|