End of training
Browse files- all_results.json +18 -0
- eval_results.json +13 -0
- pytorch_model.bin +1 -1
- runs/Dec26_15-04-41_46cffe085132/events.out.tfevents.1640531576.46cffe085132.2699.0 +2 -2
- runs/Dec26_15-04-41_46cffe085132/events.out.tfevents.1640596685.46cffe085132.2699.2 +3 -0
- train_results.json +8 -0
- trainer_state.json +1345 -0
all_results.json
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_gen_len": 63.1723,
|
| 4 |
+
"eval_loss": 1.215167760848999,
|
| 5 |
+
"eval_rouge1": 41.698,
|
| 6 |
+
"eval_rouge2": 31.3548,
|
| 7 |
+
"eval_rougeL": 38.2817,
|
| 8 |
+
"eval_rougeLsum": 39.6349,
|
| 9 |
+
"eval_runtime": 3094.5108,
|
| 10 |
+
"eval_samples": 11394,
|
| 11 |
+
"eval_samples_per_second": 3.682,
|
| 12 |
+
"eval_steps_per_second": 0.614,
|
| 13 |
+
"train_loss": 0.744481670575304,
|
| 14 |
+
"train_runtime": 62011.6805,
|
| 15 |
+
"train_samples": 220887,
|
| 16 |
+
"train_samples_per_second": 10.686,
|
| 17 |
+
"train_steps_per_second": 1.781
|
| 18 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"eval_gen_len": 63.1723,
|
| 4 |
+
"eval_loss": 1.215167760848999,
|
| 5 |
+
"eval_rouge1": 41.698,
|
| 6 |
+
"eval_rouge2": 31.3548,
|
| 7 |
+
"eval_rougeL": 38.2817,
|
| 8 |
+
"eval_rougeLsum": 39.6349,
|
| 9 |
+
"eval_runtime": 3094.5108,
|
| 10 |
+
"eval_samples": 11394,
|
| 11 |
+
"eval_samples_per_second": 3.682,
|
| 12 |
+
"eval_steps_per_second": 0.614
|
| 13 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 557979193
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:60664a4c271c80b0cc8e7fec6b817425de79f9bbc8bae6369ccfb8eb0dda59bb
|
| 3 |
size 557979193
|
runs/Dec26_15-04-41_46cffe085132/events.out.tfevents.1640531576.46cffe085132.2699.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80642ebb8ce1fcefdc7bad2cbc40c94269029f44ba225651bd3aa4ae38761ff8
|
| 3 |
+
size 39975
|
runs/Dec26_15-04-41_46cffe085132/events.out.tfevents.1640596685.46cffe085132.2699.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1fa0dccbecc97898a7ad8bc011c4cb6e918c362c4aa6de4528f9625b1778bece
|
| 3 |
+
size 575
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"train_loss": 0.744481670575304,
|
| 4 |
+
"train_runtime": 62011.6805,
|
| 5 |
+
"train_samples": 220887,
|
| 6 |
+
"train_samples_per_second": 10.686,
|
| 7 |
+
"train_steps_per_second": 1.781
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1345 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 110445,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.9773642989723395e-05,
|
| 13 |
+
"loss": 1.1879,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 4.954728597944679e-05,
|
| 19 |
+
"loss": 1.0573,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 4.932092896917018e-05,
|
| 25 |
+
"loss": 1.0167,
|
| 26 |
+
"step": 1500
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.05,
|
| 30 |
+
"learning_rate": 4.909457195889357e-05,
|
| 31 |
+
"loss": 1.0412,
|
| 32 |
+
"step": 2000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.07,
|
| 36 |
+
"learning_rate": 4.886821494861696e-05,
|
| 37 |
+
"loss": 1.0694,
|
| 38 |
+
"step": 2500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.08,
|
| 42 |
+
"learning_rate": 4.864185793834035e-05,
|
| 43 |
+
"loss": 0.9758,
|
| 44 |
+
"step": 3000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1,
|
| 48 |
+
"learning_rate": 4.841550092806374e-05,
|
| 49 |
+
"loss": 0.9976,
|
| 50 |
+
"step": 3500
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 4.818914391778714e-05,
|
| 55 |
+
"loss": 0.984,
|
| 56 |
+
"step": 4000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.12,
|
| 60 |
+
"learning_rate": 4.7962786907510526e-05,
|
| 61 |
+
"loss": 0.9322,
|
| 62 |
+
"step": 4500
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.14,
|
| 66 |
+
"learning_rate": 4.773642989723392e-05,
|
| 67 |
+
"loss": 0.9718,
|
| 68 |
+
"step": 5000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.15,
|
| 72 |
+
"learning_rate": 4.751007288695731e-05,
|
| 73 |
+
"loss": 0.9602,
|
| 74 |
+
"step": 5500
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.16,
|
| 78 |
+
"learning_rate": 4.72837158766807e-05,
|
| 79 |
+
"loss": 0.9755,
|
| 80 |
+
"step": 6000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.18,
|
| 84 |
+
"learning_rate": 4.7057358866404096e-05,
|
| 85 |
+
"loss": 0.9505,
|
| 86 |
+
"step": 6500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.19,
|
| 90 |
+
"learning_rate": 4.683100185612749e-05,
|
| 91 |
+
"loss": 0.9638,
|
| 92 |
+
"step": 7000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.2,
|
| 96 |
+
"learning_rate": 4.6604644845850873e-05,
|
| 97 |
+
"loss": 0.9089,
|
| 98 |
+
"step": 7500
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.22,
|
| 102 |
+
"learning_rate": 4.6378287835574266e-05,
|
| 103 |
+
"loss": 0.883,
|
| 104 |
+
"step": 8000
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.23,
|
| 108 |
+
"learning_rate": 4.6151930825297665e-05,
|
| 109 |
+
"loss": 0.899,
|
| 110 |
+
"step": 8500
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.24,
|
| 114 |
+
"learning_rate": 4.592557381502106e-05,
|
| 115 |
+
"loss": 0.8908,
|
| 116 |
+
"step": 9000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.26,
|
| 120 |
+
"learning_rate": 4.569921680474444e-05,
|
| 121 |
+
"loss": 0.9031,
|
| 122 |
+
"step": 9500
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.27,
|
| 126 |
+
"learning_rate": 4.5472859794467835e-05,
|
| 127 |
+
"loss": 0.8931,
|
| 128 |
+
"step": 10000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.29,
|
| 132 |
+
"learning_rate": 4.524650278419123e-05,
|
| 133 |
+
"loss": 0.9001,
|
| 134 |
+
"step": 10500
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.3,
|
| 138 |
+
"learning_rate": 4.502014577391462e-05,
|
| 139 |
+
"loss": 0.8865,
|
| 140 |
+
"step": 11000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.31,
|
| 144 |
+
"learning_rate": 4.479378876363801e-05,
|
| 145 |
+
"loss": 0.8692,
|
| 146 |
+
"step": 11500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.33,
|
| 150 |
+
"learning_rate": 4.4567431753361404e-05,
|
| 151 |
+
"loss": 0.8683,
|
| 152 |
+
"step": 12000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.34,
|
| 156 |
+
"learning_rate": 4.434107474308479e-05,
|
| 157 |
+
"loss": 0.8443,
|
| 158 |
+
"step": 12500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.35,
|
| 162 |
+
"learning_rate": 4.411471773280819e-05,
|
| 163 |
+
"loss": 0.8733,
|
| 164 |
+
"step": 13000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.37,
|
| 168 |
+
"learning_rate": 4.388836072253158e-05,
|
| 169 |
+
"loss": 0.8967,
|
| 170 |
+
"step": 13500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.38,
|
| 174 |
+
"learning_rate": 4.3662003712254973e-05,
|
| 175 |
+
"loss": 0.8516,
|
| 176 |
+
"step": 14000
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.39,
|
| 180 |
+
"learning_rate": 4.343564670197836e-05,
|
| 181 |
+
"loss": 0.8582,
|
| 182 |
+
"step": 14500
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.41,
|
| 186 |
+
"learning_rate": 4.320928969170175e-05,
|
| 187 |
+
"loss": 0.8722,
|
| 188 |
+
"step": 15000
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.42,
|
| 192 |
+
"learning_rate": 4.298293268142515e-05,
|
| 193 |
+
"loss": 0.8478,
|
| 194 |
+
"step": 15500
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.43,
|
| 198 |
+
"learning_rate": 4.2756575671148536e-05,
|
| 199 |
+
"loss": 0.8789,
|
| 200 |
+
"step": 16000
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.45,
|
| 204 |
+
"learning_rate": 4.253021866087193e-05,
|
| 205 |
+
"loss": 0.8153,
|
| 206 |
+
"step": 16500
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.46,
|
| 210 |
+
"learning_rate": 4.230386165059532e-05,
|
| 211 |
+
"loss": 0.8486,
|
| 212 |
+
"step": 17000
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.48,
|
| 216 |
+
"learning_rate": 4.207750464031871e-05,
|
| 217 |
+
"loss": 0.8443,
|
| 218 |
+
"step": 17500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.49,
|
| 222 |
+
"learning_rate": 4.1851147630042105e-05,
|
| 223 |
+
"loss": 0.8261,
|
| 224 |
+
"step": 18000
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.5,
|
| 228 |
+
"learning_rate": 4.16247906197655e-05,
|
| 229 |
+
"loss": 0.8534,
|
| 230 |
+
"step": 18500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.52,
|
| 234 |
+
"learning_rate": 4.139843360948889e-05,
|
| 235 |
+
"loss": 0.8342,
|
| 236 |
+
"step": 19000
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.53,
|
| 240 |
+
"learning_rate": 4.1172076599212275e-05,
|
| 241 |
+
"loss": 0.8622,
|
| 242 |
+
"step": 19500
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.54,
|
| 246 |
+
"learning_rate": 4.0945719588935674e-05,
|
| 247 |
+
"loss": 0.8703,
|
| 248 |
+
"step": 20000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.56,
|
| 252 |
+
"learning_rate": 4.071936257865907e-05,
|
| 253 |
+
"loss": 0.8315,
|
| 254 |
+
"step": 20500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.57,
|
| 258 |
+
"learning_rate": 4.049300556838245e-05,
|
| 259 |
+
"loss": 0.8126,
|
| 260 |
+
"step": 21000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.58,
|
| 264 |
+
"learning_rate": 4.0266648558105845e-05,
|
| 265 |
+
"loss": 0.8236,
|
| 266 |
+
"step": 21500
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.6,
|
| 270 |
+
"learning_rate": 4.004029154782924e-05,
|
| 271 |
+
"loss": 0.854,
|
| 272 |
+
"step": 22000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.61,
|
| 276 |
+
"learning_rate": 3.981393453755263e-05,
|
| 277 |
+
"loss": 0.8196,
|
| 278 |
+
"step": 22500
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.62,
|
| 282 |
+
"learning_rate": 3.958757752727602e-05,
|
| 283 |
+
"loss": 0.7705,
|
| 284 |
+
"step": 23000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.64,
|
| 288 |
+
"learning_rate": 3.9361220516999414e-05,
|
| 289 |
+
"loss": 0.7824,
|
| 290 |
+
"step": 23500
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.65,
|
| 294 |
+
"learning_rate": 3.9134863506722806e-05,
|
| 295 |
+
"loss": 0.7825,
|
| 296 |
+
"step": 24000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.67,
|
| 300 |
+
"learning_rate": 3.89085064964462e-05,
|
| 301 |
+
"loss": 0.7789,
|
| 302 |
+
"step": 24500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.68,
|
| 306 |
+
"learning_rate": 3.868214948616959e-05,
|
| 307 |
+
"loss": 0.8073,
|
| 308 |
+
"step": 25000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.69,
|
| 312 |
+
"learning_rate": 3.845579247589298e-05,
|
| 313 |
+
"loss": 0.8173,
|
| 314 |
+
"step": 25500
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.71,
|
| 318 |
+
"learning_rate": 3.822943546561637e-05,
|
| 319 |
+
"loss": 0.8299,
|
| 320 |
+
"step": 26000
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.72,
|
| 324 |
+
"learning_rate": 3.800307845533976e-05,
|
| 325 |
+
"loss": 0.8018,
|
| 326 |
+
"step": 26500
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.73,
|
| 330 |
+
"learning_rate": 3.777672144506316e-05,
|
| 331 |
+
"loss": 0.8511,
|
| 332 |
+
"step": 27000
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.75,
|
| 336 |
+
"learning_rate": 3.7550364434786546e-05,
|
| 337 |
+
"loss": 0.7932,
|
| 338 |
+
"step": 27500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.76,
|
| 342 |
+
"learning_rate": 3.732400742450994e-05,
|
| 343 |
+
"loss": 0.793,
|
| 344 |
+
"step": 28000
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.77,
|
| 348 |
+
"learning_rate": 3.709765041423333e-05,
|
| 349 |
+
"loss": 0.7984,
|
| 350 |
+
"step": 28500
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.79,
|
| 354 |
+
"learning_rate": 3.687129340395672e-05,
|
| 355 |
+
"loss": 0.814,
|
| 356 |
+
"step": 29000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.8,
|
| 360 |
+
"learning_rate": 3.6644936393680115e-05,
|
| 361 |
+
"loss": 0.7919,
|
| 362 |
+
"step": 29500
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.81,
|
| 366 |
+
"learning_rate": 3.641857938340351e-05,
|
| 367 |
+
"loss": 0.8149,
|
| 368 |
+
"step": 30000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.83,
|
| 372 |
+
"learning_rate": 3.61922223731269e-05,
|
| 373 |
+
"loss": 0.7847,
|
| 374 |
+
"step": 30500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.84,
|
| 378 |
+
"learning_rate": 3.5965865362850285e-05,
|
| 379 |
+
"loss": 0.783,
|
| 380 |
+
"step": 31000
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.86,
|
| 384 |
+
"learning_rate": 3.5739508352573684e-05,
|
| 385 |
+
"loss": 0.7998,
|
| 386 |
+
"step": 31500
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.87,
|
| 390 |
+
"learning_rate": 3.5513151342297076e-05,
|
| 391 |
+
"loss": 0.7888,
|
| 392 |
+
"step": 32000
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.88,
|
| 396 |
+
"learning_rate": 3.528679433202046e-05,
|
| 397 |
+
"loss": 0.7895,
|
| 398 |
+
"step": 32500
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.9,
|
| 402 |
+
"learning_rate": 3.5060437321743854e-05,
|
| 403 |
+
"loss": 0.796,
|
| 404 |
+
"step": 33000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.91,
|
| 408 |
+
"learning_rate": 3.4834080311467246e-05,
|
| 409 |
+
"loss": 0.7896,
|
| 410 |
+
"step": 33500
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.92,
|
| 414 |
+
"learning_rate": 3.460772330119064e-05,
|
| 415 |
+
"loss": 0.8016,
|
| 416 |
+
"step": 34000
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.94,
|
| 420 |
+
"learning_rate": 3.438136629091403e-05,
|
| 421 |
+
"loss": 0.7789,
|
| 422 |
+
"step": 34500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.95,
|
| 426 |
+
"learning_rate": 3.415500928063742e-05,
|
| 427 |
+
"loss": 0.7974,
|
| 428 |
+
"step": 35000
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.96,
|
| 432 |
+
"learning_rate": 3.3928652270360816e-05,
|
| 433 |
+
"loss": 0.7701,
|
| 434 |
+
"step": 35500
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.98,
|
| 438 |
+
"learning_rate": 3.37022952600842e-05,
|
| 439 |
+
"loss": 0.753,
|
| 440 |
+
"step": 36000
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.99,
|
| 444 |
+
"learning_rate": 3.34759382498076e-05,
|
| 445 |
+
"loss": 0.7681,
|
| 446 |
+
"step": 36500
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.01,
|
| 450 |
+
"learning_rate": 3.324958123953099e-05,
|
| 451 |
+
"loss": 0.7854,
|
| 452 |
+
"step": 37000
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.02,
|
| 456 |
+
"learning_rate": 3.302322422925438e-05,
|
| 457 |
+
"loss": 0.7462,
|
| 458 |
+
"step": 37500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.03,
|
| 462 |
+
"learning_rate": 3.279686721897777e-05,
|
| 463 |
+
"loss": 0.7028,
|
| 464 |
+
"step": 38000
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.05,
|
| 468 |
+
"learning_rate": 3.257051020870116e-05,
|
| 469 |
+
"loss": 0.761,
|
| 470 |
+
"step": 38500
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.06,
|
| 474 |
+
"learning_rate": 3.234415319842456e-05,
|
| 475 |
+
"loss": 0.7391,
|
| 476 |
+
"step": 39000
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.07,
|
| 480 |
+
"learning_rate": 3.211779618814795e-05,
|
| 481 |
+
"loss": 0.7345,
|
| 482 |
+
"step": 39500
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.09,
|
| 486 |
+
"learning_rate": 3.189143917787134e-05,
|
| 487 |
+
"loss": 0.7191,
|
| 488 |
+
"step": 40000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.1,
|
| 492 |
+
"learning_rate": 3.166508216759473e-05,
|
| 493 |
+
"loss": 0.7528,
|
| 494 |
+
"step": 40500
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.11,
|
| 498 |
+
"learning_rate": 3.1438725157318124e-05,
|
| 499 |
+
"loss": 0.7449,
|
| 500 |
+
"step": 41000
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.13,
|
| 504 |
+
"learning_rate": 3.1212368147041517e-05,
|
| 505 |
+
"loss": 0.7769,
|
| 506 |
+
"step": 41500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.14,
|
| 510 |
+
"learning_rate": 3.098601113676491e-05,
|
| 511 |
+
"loss": 0.7451,
|
| 512 |
+
"step": 42000
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.15,
|
| 516 |
+
"learning_rate": 3.0759654126488294e-05,
|
| 517 |
+
"loss": 0.7561,
|
| 518 |
+
"step": 42500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.17,
|
| 522 |
+
"learning_rate": 3.053329711621169e-05,
|
| 523 |
+
"loss": 0.7299,
|
| 524 |
+
"step": 43000
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.18,
|
| 528 |
+
"learning_rate": 3.0306940105935082e-05,
|
| 529 |
+
"loss": 0.7381,
|
| 530 |
+
"step": 43500
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 1.2,
|
| 534 |
+
"learning_rate": 3.0080583095658478e-05,
|
| 535 |
+
"loss": 0.7397,
|
| 536 |
+
"step": 44000
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.21,
|
| 540 |
+
"learning_rate": 2.9854226085381864e-05,
|
| 541 |
+
"loss": 0.691,
|
| 542 |
+
"step": 44500
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.22,
|
| 546 |
+
"learning_rate": 2.962786907510526e-05,
|
| 547 |
+
"loss": 0.73,
|
| 548 |
+
"step": 45000
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.24,
|
| 552 |
+
"learning_rate": 2.940151206482865e-05,
|
| 553 |
+
"loss": 0.7078,
|
| 554 |
+
"step": 45500
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.25,
|
| 558 |
+
"learning_rate": 2.917515505455204e-05,
|
| 559 |
+
"loss": 0.7254,
|
| 560 |
+
"step": 46000
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.26,
|
| 564 |
+
"learning_rate": 2.8948798044275433e-05,
|
| 565 |
+
"loss": 0.7223,
|
| 566 |
+
"step": 46500
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.28,
|
| 570 |
+
"learning_rate": 2.8722441033998825e-05,
|
| 571 |
+
"loss": 0.7113,
|
| 572 |
+
"step": 47000
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.29,
|
| 576 |
+
"learning_rate": 2.8496084023722214e-05,
|
| 577 |
+
"loss": 0.7237,
|
| 578 |
+
"step": 47500
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.3,
|
| 582 |
+
"learning_rate": 2.8269727013445606e-05,
|
| 583 |
+
"loss": 0.7221,
|
| 584 |
+
"step": 48000
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.32,
|
| 588 |
+
"learning_rate": 2.8043370003169002e-05,
|
| 589 |
+
"loss": 0.7242,
|
| 590 |
+
"step": 48500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.33,
|
| 594 |
+
"learning_rate": 2.7817012992892394e-05,
|
| 595 |
+
"loss": 0.7214,
|
| 596 |
+
"step": 49000
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.34,
|
| 600 |
+
"learning_rate": 2.7590655982615783e-05,
|
| 601 |
+
"loss": 0.7345,
|
| 602 |
+
"step": 49500
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.36,
|
| 606 |
+
"learning_rate": 2.7364298972339176e-05,
|
| 607 |
+
"loss": 0.7302,
|
| 608 |
+
"step": 50000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.37,
|
| 612 |
+
"learning_rate": 2.7137941962062568e-05,
|
| 613 |
+
"loss": 0.7494,
|
| 614 |
+
"step": 50500
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.39,
|
| 618 |
+
"learning_rate": 2.6911584951785957e-05,
|
| 619 |
+
"loss": 0.7391,
|
| 620 |
+
"step": 51000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.4,
|
| 624 |
+
"learning_rate": 2.668522794150935e-05,
|
| 625 |
+
"loss": 0.6945,
|
| 626 |
+
"step": 51500
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.41,
|
| 630 |
+
"learning_rate": 2.645887093123274e-05,
|
| 631 |
+
"loss": 0.7372,
|
| 632 |
+
"step": 52000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.43,
|
| 636 |
+
"learning_rate": 2.623251392095613e-05,
|
| 637 |
+
"loss": 0.707,
|
| 638 |
+
"step": 52500
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.44,
|
| 642 |
+
"learning_rate": 2.6006156910679523e-05,
|
| 643 |
+
"loss": 0.7253,
|
| 644 |
+
"step": 53000
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.45,
|
| 648 |
+
"learning_rate": 2.577979990040292e-05,
|
| 649 |
+
"loss": 0.7346,
|
| 650 |
+
"step": 53500
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.47,
|
| 654 |
+
"learning_rate": 2.555344289012631e-05,
|
| 655 |
+
"loss": 0.7237,
|
| 656 |
+
"step": 54000
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.48,
|
| 660 |
+
"learning_rate": 2.53270858798497e-05,
|
| 661 |
+
"loss": 0.7106,
|
| 662 |
+
"step": 54500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.49,
|
| 666 |
+
"learning_rate": 2.5100728869573092e-05,
|
| 667 |
+
"loss": 0.7043,
|
| 668 |
+
"step": 55000
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.51,
|
| 672 |
+
"learning_rate": 2.4874371859296484e-05,
|
| 673 |
+
"loss": 0.7339,
|
| 674 |
+
"step": 55500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.52,
|
| 678 |
+
"learning_rate": 2.4648014849019877e-05,
|
| 679 |
+
"loss": 0.7156,
|
| 680 |
+
"step": 56000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.53,
|
| 684 |
+
"learning_rate": 2.4421657838743266e-05,
|
| 685 |
+
"loss": 0.6913,
|
| 686 |
+
"step": 56500
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.55,
|
| 690 |
+
"learning_rate": 2.4195300828466658e-05,
|
| 691 |
+
"loss": 0.6981,
|
| 692 |
+
"step": 57000
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.56,
|
| 696 |
+
"learning_rate": 2.396894381819005e-05,
|
| 697 |
+
"loss": 0.7133,
|
| 698 |
+
"step": 57500
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.58,
|
| 702 |
+
"learning_rate": 2.3742586807913442e-05,
|
| 703 |
+
"loss": 0.7359,
|
| 704 |
+
"step": 58000
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.59,
|
| 708 |
+
"learning_rate": 2.3516229797636835e-05,
|
| 709 |
+
"loss": 0.7033,
|
| 710 |
+
"step": 58500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.6,
|
| 714 |
+
"learning_rate": 2.3289872787360227e-05,
|
| 715 |
+
"loss": 0.7246,
|
| 716 |
+
"step": 59000
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.62,
|
| 720 |
+
"learning_rate": 2.3063515777083616e-05,
|
| 721 |
+
"loss": 0.7254,
|
| 722 |
+
"step": 59500
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.63,
|
| 726 |
+
"learning_rate": 2.2837158766807008e-05,
|
| 727 |
+
"loss": 0.7335,
|
| 728 |
+
"step": 60000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.64,
|
| 732 |
+
"learning_rate": 2.26108017565304e-05,
|
| 733 |
+
"loss": 0.735,
|
| 734 |
+
"step": 60500
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.66,
|
| 738 |
+
"learning_rate": 2.2384444746253793e-05,
|
| 739 |
+
"loss": 0.6991,
|
| 740 |
+
"step": 61000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.67,
|
| 744 |
+
"learning_rate": 2.2158087735977185e-05,
|
| 745 |
+
"loss": 0.702,
|
| 746 |
+
"step": 61500
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.68,
|
| 750 |
+
"learning_rate": 2.1931730725700574e-05,
|
| 751 |
+
"loss": 0.7185,
|
| 752 |
+
"step": 62000
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.7,
|
| 756 |
+
"learning_rate": 2.170537371542397e-05,
|
| 757 |
+
"loss": 0.7132,
|
| 758 |
+
"step": 62500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.71,
|
| 762 |
+
"learning_rate": 2.147901670514736e-05,
|
| 763 |
+
"loss": 0.6873,
|
| 764 |
+
"step": 63000
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.72,
|
| 768 |
+
"learning_rate": 2.125265969487075e-05,
|
| 769 |
+
"loss": 0.7491,
|
| 770 |
+
"step": 63500
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.74,
|
| 774 |
+
"learning_rate": 2.1026302684594143e-05,
|
| 775 |
+
"loss": 0.7064,
|
| 776 |
+
"step": 64000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.75,
|
| 780 |
+
"learning_rate": 2.0799945674317532e-05,
|
| 781 |
+
"loss": 0.723,
|
| 782 |
+
"step": 64500
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.77,
|
| 786 |
+
"learning_rate": 2.0573588664040928e-05,
|
| 787 |
+
"loss": 0.6922,
|
| 788 |
+
"step": 65000
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.78,
|
| 792 |
+
"learning_rate": 2.0347231653764317e-05,
|
| 793 |
+
"loss": 0.6727,
|
| 794 |
+
"step": 65500
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.79,
|
| 798 |
+
"learning_rate": 2.0120874643487713e-05,
|
| 799 |
+
"loss": 0.6876,
|
| 800 |
+
"step": 66000
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.81,
|
| 804 |
+
"learning_rate": 1.98945176332111e-05,
|
| 805 |
+
"loss": 0.7399,
|
| 806 |
+
"step": 66500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.82,
|
| 810 |
+
"learning_rate": 1.9668160622934494e-05,
|
| 811 |
+
"loss": 0.7107,
|
| 812 |
+
"step": 67000
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.83,
|
| 816 |
+
"learning_rate": 1.9441803612657886e-05,
|
| 817 |
+
"loss": 0.6767,
|
| 818 |
+
"step": 67500
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.85,
|
| 822 |
+
"learning_rate": 1.9215446602381275e-05,
|
| 823 |
+
"loss": 0.7102,
|
| 824 |
+
"step": 68000
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.86,
|
| 828 |
+
"learning_rate": 1.898908959210467e-05,
|
| 829 |
+
"loss": 0.7183,
|
| 830 |
+
"step": 68500
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.87,
|
| 834 |
+
"learning_rate": 1.876273258182806e-05,
|
| 835 |
+
"loss": 0.6808,
|
| 836 |
+
"step": 69000
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.89,
|
| 840 |
+
"learning_rate": 1.8536375571551452e-05,
|
| 841 |
+
"loss": 0.7165,
|
| 842 |
+
"step": 69500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.9,
|
| 846 |
+
"learning_rate": 1.8310018561274844e-05,
|
| 847 |
+
"loss": 0.7035,
|
| 848 |
+
"step": 70000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.91,
|
| 852 |
+
"learning_rate": 1.8083661550998233e-05,
|
| 853 |
+
"loss": 0.6831,
|
| 854 |
+
"step": 70500
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.93,
|
| 858 |
+
"learning_rate": 1.785730454072163e-05,
|
| 859 |
+
"loss": 0.7026,
|
| 860 |
+
"step": 71000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.94,
|
| 864 |
+
"learning_rate": 1.7630947530445018e-05,
|
| 865 |
+
"loss": 0.7105,
|
| 866 |
+
"step": 71500
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 1.96,
|
| 870 |
+
"learning_rate": 1.740459052016841e-05,
|
| 871 |
+
"loss": 0.7007,
|
| 872 |
+
"step": 72000
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.97,
|
| 876 |
+
"learning_rate": 1.7178233509891802e-05,
|
| 877 |
+
"loss": 0.6993,
|
| 878 |
+
"step": 72500
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.98,
|
| 882 |
+
"learning_rate": 1.6951876499615195e-05,
|
| 883 |
+
"loss": 0.7108,
|
| 884 |
+
"step": 73000
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.0,
|
| 888 |
+
"learning_rate": 1.6725519489338587e-05,
|
| 889 |
+
"loss": 0.71,
|
| 890 |
+
"step": 73500
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 2.01,
|
| 894 |
+
"learning_rate": 1.6499162479061976e-05,
|
| 895 |
+
"loss": 0.6825,
|
| 896 |
+
"step": 74000
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 2.02,
|
| 900 |
+
"learning_rate": 1.627280546878537e-05,
|
| 901 |
+
"loss": 0.6683,
|
| 902 |
+
"step": 74500
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 2.04,
|
| 906 |
+
"learning_rate": 1.604644845850876e-05,
|
| 907 |
+
"loss": 0.6797,
|
| 908 |
+
"step": 75000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.05,
|
| 912 |
+
"learning_rate": 1.5820091448232153e-05,
|
| 913 |
+
"loss": 0.6507,
|
| 914 |
+
"step": 75500
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.06,
|
| 918 |
+
"learning_rate": 1.5593734437955545e-05,
|
| 919 |
+
"loss": 0.6438,
|
| 920 |
+
"step": 76000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 2.08,
|
| 924 |
+
"learning_rate": 1.5367377427678938e-05,
|
| 925 |
+
"loss": 0.6587,
|
| 926 |
+
"step": 76500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.09,
|
| 930 |
+
"learning_rate": 1.5141020417402326e-05,
|
| 931 |
+
"loss": 0.6977,
|
| 932 |
+
"step": 77000
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.11,
|
| 936 |
+
"learning_rate": 1.491466340712572e-05,
|
| 937 |
+
"loss": 0.6654,
|
| 938 |
+
"step": 77500
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.12,
|
| 942 |
+
"learning_rate": 1.4688306396849111e-05,
|
| 943 |
+
"loss": 0.6654,
|
| 944 |
+
"step": 78000
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.13,
|
| 948 |
+
"learning_rate": 1.4461949386572503e-05,
|
| 949 |
+
"loss": 0.6217,
|
| 950 |
+
"step": 78500
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.15,
|
| 954 |
+
"learning_rate": 1.4235592376295894e-05,
|
| 955 |
+
"loss": 0.638,
|
| 956 |
+
"step": 79000
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.16,
|
| 960 |
+
"learning_rate": 1.4009235366019285e-05,
|
| 961 |
+
"loss": 0.6636,
|
| 962 |
+
"step": 79500
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 2.17,
|
| 966 |
+
"learning_rate": 1.3782878355742679e-05,
|
| 967 |
+
"loss": 0.6635,
|
| 968 |
+
"step": 80000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.19,
|
| 972 |
+
"learning_rate": 1.355652134546607e-05,
|
| 973 |
+
"loss": 0.6802,
|
| 974 |
+
"step": 80500
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.2,
|
| 978 |
+
"learning_rate": 1.3330164335189463e-05,
|
| 979 |
+
"loss": 0.6387,
|
| 980 |
+
"step": 81000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 2.21,
|
| 984 |
+
"learning_rate": 1.3103807324912854e-05,
|
| 985 |
+
"loss": 0.6476,
|
| 986 |
+
"step": 81500
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.23,
|
| 990 |
+
"learning_rate": 1.2877450314636244e-05,
|
| 991 |
+
"loss": 0.6495,
|
| 992 |
+
"step": 82000
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 2.24,
|
| 996 |
+
"learning_rate": 1.2651093304359637e-05,
|
| 997 |
+
"loss": 0.6372,
|
| 998 |
+
"step": 82500
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.25,
|
| 1002 |
+
"learning_rate": 1.2424736294083027e-05,
|
| 1003 |
+
"loss": 0.6304,
|
| 1004 |
+
"step": 83000
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.27,
|
| 1008 |
+
"learning_rate": 1.219837928380642e-05,
|
| 1009 |
+
"loss": 0.6643,
|
| 1010 |
+
"step": 83500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.28,
|
| 1014 |
+
"learning_rate": 1.1972022273529812e-05,
|
| 1015 |
+
"loss": 0.6454,
|
| 1016 |
+
"step": 84000
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 2.3,
|
| 1020 |
+
"learning_rate": 1.1745665263253204e-05,
|
| 1021 |
+
"loss": 0.6847,
|
| 1022 |
+
"step": 84500
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.31,
|
| 1026 |
+
"learning_rate": 1.1519308252976595e-05,
|
| 1027 |
+
"loss": 0.6396,
|
| 1028 |
+
"step": 85000
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 2.32,
|
| 1032 |
+
"learning_rate": 1.1292951242699987e-05,
|
| 1033 |
+
"loss": 0.638,
|
| 1034 |
+
"step": 85500
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.34,
|
| 1038 |
+
"learning_rate": 1.106659423242338e-05,
|
| 1039 |
+
"loss": 0.662,
|
| 1040 |
+
"step": 86000
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.35,
|
| 1044 |
+
"learning_rate": 1.084023722214677e-05,
|
| 1045 |
+
"loss": 0.6449,
|
| 1046 |
+
"step": 86500
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.36,
|
| 1050 |
+
"learning_rate": 1.0613880211870162e-05,
|
| 1051 |
+
"loss": 0.6452,
|
| 1052 |
+
"step": 87000
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.38,
|
| 1056 |
+
"learning_rate": 1.0387523201593553e-05,
|
| 1057 |
+
"loss": 0.6658,
|
| 1058 |
+
"step": 87500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 2.39,
|
| 1062 |
+
"learning_rate": 1.0161166191316945e-05,
|
| 1063 |
+
"loss": 0.6677,
|
| 1064 |
+
"step": 88000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.4,
|
| 1068 |
+
"learning_rate": 9.934809181040338e-06,
|
| 1069 |
+
"loss": 0.6543,
|
| 1070 |
+
"step": 88500
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.42,
|
| 1074 |
+
"learning_rate": 9.70845217076373e-06,
|
| 1075 |
+
"loss": 0.6425,
|
| 1076 |
+
"step": 89000
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.43,
|
| 1080 |
+
"learning_rate": 9.48209516048712e-06,
|
| 1081 |
+
"loss": 0.6593,
|
| 1082 |
+
"step": 89500
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.44,
|
| 1086 |
+
"learning_rate": 9.255738150210511e-06,
|
| 1087 |
+
"loss": 0.641,
|
| 1088 |
+
"step": 90000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.46,
|
| 1092 |
+
"learning_rate": 9.029381139933904e-06,
|
| 1093 |
+
"loss": 0.6167,
|
| 1094 |
+
"step": 90500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.47,
|
| 1098 |
+
"learning_rate": 8.803024129657296e-06,
|
| 1099 |
+
"loss": 0.6498,
|
| 1100 |
+
"step": 91000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 2.49,
|
| 1104 |
+
"learning_rate": 8.576667119380688e-06,
|
| 1105 |
+
"loss": 0.667,
|
| 1106 |
+
"step": 91500
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.5,
|
| 1110 |
+
"learning_rate": 8.35031010910408e-06,
|
| 1111 |
+
"loss": 0.6352,
|
| 1112 |
+
"step": 92000
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.51,
|
| 1116 |
+
"learning_rate": 8.123953098827471e-06,
|
| 1117 |
+
"loss": 0.6837,
|
| 1118 |
+
"step": 92500
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.53,
|
| 1122 |
+
"learning_rate": 7.897596088550863e-06,
|
| 1123 |
+
"loss": 0.6471,
|
| 1124 |
+
"step": 93000
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.54,
|
| 1128 |
+
"learning_rate": 7.671239078274254e-06,
|
| 1129 |
+
"loss": 0.6601,
|
| 1130 |
+
"step": 93500
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.55,
|
| 1134 |
+
"learning_rate": 7.444882067997646e-06,
|
| 1135 |
+
"loss": 0.653,
|
| 1136 |
+
"step": 94000
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.57,
|
| 1140 |
+
"learning_rate": 7.218525057721039e-06,
|
| 1141 |
+
"loss": 0.6495,
|
| 1142 |
+
"step": 94500
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 2.58,
|
| 1146 |
+
"learning_rate": 6.992168047444429e-06,
|
| 1147 |
+
"loss": 0.63,
|
| 1148 |
+
"step": 95000
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.59,
|
| 1152 |
+
"learning_rate": 6.7658110371678216e-06,
|
| 1153 |
+
"loss": 0.6371,
|
| 1154 |
+
"step": 95500
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.61,
|
| 1158 |
+
"learning_rate": 6.539454026891213e-06,
|
| 1159 |
+
"loss": 0.6551,
|
| 1160 |
+
"step": 96000
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.62,
|
| 1164 |
+
"learning_rate": 6.313097016614605e-06,
|
| 1165 |
+
"loss": 0.6621,
|
| 1166 |
+
"step": 96500
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 2.63,
|
| 1170 |
+
"learning_rate": 6.086740006337997e-06,
|
| 1171 |
+
"loss": 0.6169,
|
| 1172 |
+
"step": 97000
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.65,
|
| 1176 |
+
"learning_rate": 5.860382996061388e-06,
|
| 1177 |
+
"loss": 0.622,
|
| 1178 |
+
"step": 97500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.66,
|
| 1182 |
+
"learning_rate": 5.63402598578478e-06,
|
| 1183 |
+
"loss": 0.6535,
|
| 1184 |
+
"step": 98000
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 2.68,
|
| 1188 |
+
"learning_rate": 5.407668975508172e-06,
|
| 1189 |
+
"loss": 0.6599,
|
| 1190 |
+
"step": 98500
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.69,
|
| 1194 |
+
"learning_rate": 5.1813119652315635e-06,
|
| 1195 |
+
"loss": 0.6369,
|
| 1196 |
+
"step": 99000
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.7,
|
| 1200 |
+
"learning_rate": 4.954954954954955e-06,
|
| 1201 |
+
"loss": 0.6215,
|
| 1202 |
+
"step": 99500
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.72,
|
| 1206 |
+
"learning_rate": 4.728597944678347e-06,
|
| 1207 |
+
"loss": 0.6548,
|
| 1208 |
+
"step": 100000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 2.73,
|
| 1212 |
+
"learning_rate": 4.502240934401739e-06,
|
| 1213 |
+
"loss": 0.6472,
|
| 1214 |
+
"step": 100500
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.74,
|
| 1218 |
+
"learning_rate": 4.27588392412513e-06,
|
| 1219 |
+
"loss": 0.6335,
|
| 1220 |
+
"step": 101000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.76,
|
| 1224 |
+
"learning_rate": 4.049526913848522e-06,
|
| 1225 |
+
"loss": 0.6288,
|
| 1226 |
+
"step": 101500
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 2.77,
|
| 1230 |
+
"learning_rate": 3.823169903571914e-06,
|
| 1231 |
+
"loss": 0.6547,
|
| 1232 |
+
"step": 102000
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.78,
|
| 1236 |
+
"learning_rate": 3.596812893295305e-06,
|
| 1237 |
+
"loss": 0.6538,
|
| 1238 |
+
"step": 102500
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.8,
|
| 1242 |
+
"learning_rate": 3.3704558830186973e-06,
|
| 1243 |
+
"loss": 0.6365,
|
| 1244 |
+
"step": 103000
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.81,
|
| 1248 |
+
"learning_rate": 3.144098872742089e-06,
|
| 1249 |
+
"loss": 0.6731,
|
| 1250 |
+
"step": 103500
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 2.82,
|
| 1254 |
+
"learning_rate": 2.9177418624654807e-06,
|
| 1255 |
+
"loss": 0.6494,
|
| 1256 |
+
"step": 104000
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.84,
|
| 1260 |
+
"learning_rate": 2.6913848521888725e-06,
|
| 1261 |
+
"loss": 0.6325,
|
| 1262 |
+
"step": 104500
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.85,
|
| 1266 |
+
"learning_rate": 2.465027841912264e-06,
|
| 1267 |
+
"loss": 0.6237,
|
| 1268 |
+
"step": 105000
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 2.87,
|
| 1272 |
+
"learning_rate": 2.238670831635656e-06,
|
| 1273 |
+
"loss": 0.649,
|
| 1274 |
+
"step": 105500
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.88,
|
| 1278 |
+
"learning_rate": 2.0123138213590474e-06,
|
| 1279 |
+
"loss": 0.6477,
|
| 1280 |
+
"step": 106000
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.89,
|
| 1284 |
+
"learning_rate": 1.7859568110824392e-06,
|
| 1285 |
+
"loss": 0.6642,
|
| 1286 |
+
"step": 106500
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.91,
|
| 1290 |
+
"learning_rate": 1.559599800805831e-06,
|
| 1291 |
+
"loss": 0.6515,
|
| 1292 |
+
"step": 107000
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 2.92,
|
| 1296 |
+
"learning_rate": 1.3332427905292228e-06,
|
| 1297 |
+
"loss": 0.655,
|
| 1298 |
+
"step": 107500
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 2.93,
|
| 1302 |
+
"learning_rate": 1.1068857802526145e-06,
|
| 1303 |
+
"loss": 0.6645,
|
| 1304 |
+
"step": 108000
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 2.95,
|
| 1308 |
+
"learning_rate": 8.805287699760061e-07,
|
| 1309 |
+
"loss": 0.6654,
|
| 1310 |
+
"step": 108500
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 2.96,
|
| 1314 |
+
"learning_rate": 6.541717596993979e-07,
|
| 1315 |
+
"loss": 0.6457,
|
| 1316 |
+
"step": 109000
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 2.97,
|
| 1320 |
+
"learning_rate": 4.278147494227896e-07,
|
| 1321 |
+
"loss": 0.669,
|
| 1322 |
+
"step": 109500
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 2.99,
|
| 1326 |
+
"learning_rate": 2.0145773914618135e-07,
|
| 1327 |
+
"loss": 0.6577,
|
| 1328 |
+
"step": 110000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 3.0,
|
| 1332 |
+
"step": 110445,
|
| 1333 |
+
"total_flos": 4.0396527424641024e+17,
|
| 1334 |
+
"train_loss": 0.744481670575304,
|
| 1335 |
+
"train_runtime": 62011.6805,
|
| 1336 |
+
"train_samples_per_second": 10.686,
|
| 1337 |
+
"train_steps_per_second": 1.781
|
| 1338 |
+
}
|
| 1339 |
+
],
|
| 1340 |
+
"max_steps": 110445,
|
| 1341 |
+
"num_train_epochs": 3,
|
| 1342 |
+
"total_flos": 4.0396527424641024e+17,
|
| 1343 |
+
"trial_name": null,
|
| 1344 |
+
"trial_params": null
|
| 1345 |
+
}
|