Training in progress, step 603, checkpoint
Browse files
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 101184122
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:816c4df0d8b8e012e0f09ffc6fdf9ccdccf3e7710fcab9b3c9e32dc9253071c0
|
| 3 |
size 101184122
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14512
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be613d3fe8e245e7abd0f538f5bdef3d1876d27be56d06ed45314f9b72f85fde
|
| 3 |
size 14512
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14512
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a47fc2ac671fb35eea71ae222e402f44674cb96ad093aaf574572f31279bbbfa
|
| 3 |
size 14512
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1064
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7664ab897ed124bbeb09cf4530c0e76a35f6774388ff7703c1e60c754fa9fc97
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch": 0.
|
| 5 |
"eval_steps": 201,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2845,6 +2845,1421 @@
|
|
| 2845 |
"eval_samples_per_second": 22.312,
|
| 2846 |
"eval_steps_per_second": 5.582,
|
| 2847 |
"step": 402
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2848 |
}
|
| 2849 |
],
|
| 2850 |
"logging_steps": 1,
|
|
@@ -2864,7 +4279,7 @@
|
|
| 2864 |
"attributes": {}
|
| 2865 |
}
|
| 2866 |
},
|
| 2867 |
-
"total_flos":
|
| 2868 |
"train_batch_size": 2,
|
| 2869 |
"trial_name": null,
|
| 2870 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7521047708138447,
|
| 5 |
"eval_steps": 201,
|
| 6 |
+
"global_step": 603,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2845 |
"eval_samples_per_second": 22.312,
|
| 2846 |
"eval_steps_per_second": 5.582,
|
| 2847 |
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.5026504521359526,
|
| 2851 |
+
"grad_norm": NaN,
|
| 2852 |
+
"learning_rate": 0.00010939833731574967,
|
| 2853 |
+
"loss": 0.0,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.503897723729342,
|
| 2858 |
+
"grad_norm": NaN,
|
| 2859 |
+
"learning_rate": 0.00010898178662706471,
|
| 2860 |
+
"loss": 0.0,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.5051449953227315,
|
| 2865 |
+
"grad_norm": NaN,
|
| 2866 |
+
"learning_rate": 0.00010856507876392166,
|
| 2867 |
+
"loss": 0.0,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.5063922669161209,
|
| 2872 |
+
"grad_norm": NaN,
|
| 2873 |
+
"learning_rate": 0.00010814822101839224,
|
| 2874 |
+
"loss": 0.0,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.5076395385095105,
|
| 2879 |
+
"grad_norm": NaN,
|
| 2880 |
+
"learning_rate": 0.00010773122068517103,
|
| 2881 |
+
"loss": 0.0,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.5088868101028999,
|
| 2886 |
+
"grad_norm": NaN,
|
| 2887 |
+
"learning_rate": 0.00010731408506144782,
|
| 2888 |
+
"loss": 0.0,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.5101340816962894,
|
| 2893 |
+
"grad_norm": NaN,
|
| 2894 |
+
"learning_rate": 0.00010689682144677983,
|
| 2895 |
+
"loss": 0.0,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.5113813532896788,
|
| 2900 |
+
"grad_norm": NaN,
|
| 2901 |
+
"learning_rate": 0.00010647943714296405,
|
| 2902 |
+
"loss": 0.0,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.5126286248830683,
|
| 2907 |
+
"grad_norm": NaN,
|
| 2908 |
+
"learning_rate": 0.00010606193945390943,
|
| 2909 |
+
"loss": 0.0,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.5138758964764577,
|
| 2914 |
+
"grad_norm": NaN,
|
| 2915 |
+
"learning_rate": 0.00010564433568550909,
|
| 2916 |
+
"loss": 0.0,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.5151231680698473,
|
| 2921 |
+
"grad_norm": NaN,
|
| 2922 |
+
"learning_rate": 0.00010522663314551247,
|
| 2923 |
+
"loss": 0.0,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.5163704396632367,
|
| 2928 |
+
"grad_norm": NaN,
|
| 2929 |
+
"learning_rate": 0.00010480883914339736,
|
| 2930 |
+
"loss": 0.0,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.5176177112566261,
|
| 2935 |
+
"grad_norm": NaN,
|
| 2936 |
+
"learning_rate": 0.0001043909609902422,
|
| 2937 |
+
"loss": 0.0,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.5188649828500156,
|
| 2942 |
+
"grad_norm": NaN,
|
| 2943 |
+
"learning_rate": 0.00010397300599859785,
|
| 2944 |
+
"loss": 0.0,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.520112254443405,
|
| 2949 |
+
"grad_norm": NaN,
|
| 2950 |
+
"learning_rate": 0.00010355498148235996,
|
| 2951 |
+
"loss": 0.0,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.5213595260367945,
|
| 2956 |
+
"grad_norm": NaN,
|
| 2957 |
+
"learning_rate": 0.00010313689475664063,
|
| 2958 |
+
"loss": 0.0,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.522606797630184,
|
| 2963 |
+
"grad_norm": NaN,
|
| 2964 |
+
"learning_rate": 0.0001027187531376407,
|
| 2965 |
+
"loss": 0.0,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.5238540692235735,
|
| 2970 |
+
"grad_norm": NaN,
|
| 2971 |
+
"learning_rate": 0.00010230056394252161,
|
| 2972 |
+
"loss": 0.0,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.5251013408169629,
|
| 2977 |
+
"grad_norm": NaN,
|
| 2978 |
+
"learning_rate": 0.00010188233448927724,
|
| 2979 |
+
"loss": 0.0,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.5263486124103524,
|
| 2984 |
+
"grad_norm": NaN,
|
| 2985 |
+
"learning_rate": 0.00010146407209660607,
|
| 2986 |
+
"loss": 0.0,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.5275958840037418,
|
| 2991 |
+
"grad_norm": NaN,
|
| 2992 |
+
"learning_rate": 0.00010104578408378289,
|
| 2993 |
+
"loss": 0.0,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.5288431555971312,
|
| 2998 |
+
"grad_norm": NaN,
|
| 2999 |
+
"learning_rate": 0.00010062747777053094,
|
| 3000 |
+
"loss": 0.0,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.5300904271905207,
|
| 3005 |
+
"grad_norm": NaN,
|
| 3006 |
+
"learning_rate": 0.00010020916047689358,
|
| 3007 |
+
"loss": 0.0,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.5313376987839102,
|
| 3012 |
+
"grad_norm": NaN,
|
| 3013 |
+
"learning_rate": 9.979083952310643e-05,
|
| 3014 |
+
"loss": 0.0,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.5325849703772997,
|
| 3019 |
+
"grad_norm": NaN,
|
| 3020 |
+
"learning_rate": 9.937252222946908e-05,
|
| 3021 |
+
"loss": 0.0,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.5338322419706891,
|
| 3026 |
+
"grad_norm": NaN,
|
| 3027 |
+
"learning_rate": 9.895421591621712e-05,
|
| 3028 |
+
"loss": 0.0,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.5350795135640786,
|
| 3033 |
+
"grad_norm": NaN,
|
| 3034 |
+
"learning_rate": 9.853592790339396e-05,
|
| 3035 |
+
"loss": 0.0,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.536326785157468,
|
| 3040 |
+
"grad_norm": NaN,
|
| 3041 |
+
"learning_rate": 9.811766551072278e-05,
|
| 3042 |
+
"loss": 0.0,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.5375740567508575,
|
| 3047 |
+
"grad_norm": NaN,
|
| 3048 |
+
"learning_rate": 9.769943605747844e-05,
|
| 3049 |
+
"loss": 0.0,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.538821328344247,
|
| 3054 |
+
"grad_norm": NaN,
|
| 3055 |
+
"learning_rate": 9.72812468623593e-05,
|
| 3056 |
+
"loss": 0.0,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.5400685999376365,
|
| 3061 |
+
"grad_norm": NaN,
|
| 3062 |
+
"learning_rate": 9.686310524335938e-05,
|
| 3063 |
+
"loss": 0.0,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.5413158715310259,
|
| 3068 |
+
"grad_norm": NaN,
|
| 3069 |
+
"learning_rate": 9.644501851764007e-05,
|
| 3070 |
+
"loss": 0.0,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.5425631431244153,
|
| 3075 |
+
"grad_norm": NaN,
|
| 3076 |
+
"learning_rate": 9.602699400140218e-05,
|
| 3077 |
+
"loss": 0.0,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.5438104147178048,
|
| 3082 |
+
"grad_norm": NaN,
|
| 3083 |
+
"learning_rate": 9.560903900975785e-05,
|
| 3084 |
+
"loss": 0.0,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.5450576863111942,
|
| 3089 |
+
"grad_norm": NaN,
|
| 3090 |
+
"learning_rate": 9.519116085660267e-05,
|
| 3091 |
+
"loss": 0.0,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.5463049579045838,
|
| 3096 |
+
"grad_norm": NaN,
|
| 3097 |
+
"learning_rate": 9.477336685448754e-05,
|
| 3098 |
+
"loss": 0.0,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.5475522294979732,
|
| 3103 |
+
"grad_norm": NaN,
|
| 3104 |
+
"learning_rate": 9.435566431449092e-05,
|
| 3105 |
+
"loss": 0.0,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.5487995010913627,
|
| 3110 |
+
"grad_norm": NaN,
|
| 3111 |
+
"learning_rate": 9.39380605460906e-05,
|
| 3112 |
+
"loss": 0.0,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.5500467726847521,
|
| 3117 |
+
"grad_norm": NaN,
|
| 3118 |
+
"learning_rate": 9.352056285703599e-05,
|
| 3119 |
+
"loss": 0.0,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.5512940442781415,
|
| 3124 |
+
"grad_norm": NaN,
|
| 3125 |
+
"learning_rate": 9.31031785532202e-05,
|
| 3126 |
+
"loss": 0.0,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.552541315871531,
|
| 3131 |
+
"grad_norm": NaN,
|
| 3132 |
+
"learning_rate": 9.268591493855222e-05,
|
| 3133 |
+
"loss": 0.0,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.5537885874649204,
|
| 3138 |
+
"grad_norm": NaN,
|
| 3139 |
+
"learning_rate": 9.226877931482898e-05,
|
| 3140 |
+
"loss": 0.0,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.55503585905831,
|
| 3145 |
+
"grad_norm": NaN,
|
| 3146 |
+
"learning_rate": 9.18517789816078e-05,
|
| 3147 |
+
"loss": 0.0,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.5562831306516994,
|
| 3152 |
+
"grad_norm": NaN,
|
| 3153 |
+
"learning_rate": 9.143492123607838e-05,
|
| 3154 |
+
"loss": 0.0,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.5575304022450889,
|
| 3159 |
+
"grad_norm": NaN,
|
| 3160 |
+
"learning_rate": 9.101821337293532e-05,
|
| 3161 |
+
"loss": 0.0,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.5587776738384783,
|
| 3166 |
+
"grad_norm": NaN,
|
| 3167 |
+
"learning_rate": 9.060166268425038e-05,
|
| 3168 |
+
"loss": 0.0,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.5600249454318678,
|
| 3173 |
+
"grad_norm": NaN,
|
| 3174 |
+
"learning_rate": 9.018527645934488e-05,
|
| 3175 |
+
"loss": 0.0,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.5612722170252572,
|
| 3180 |
+
"grad_norm": NaN,
|
| 3181 |
+
"learning_rate": 8.976906198466213e-05,
|
| 3182 |
+
"loss": 0.0,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.5625194886186468,
|
| 3187 |
+
"grad_norm": NaN,
|
| 3188 |
+
"learning_rate": 8.935302654364e-05,
|
| 3189 |
+
"loss": 0.0,
|
| 3190 |
+
"step": 451
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 0.5637667602120362,
|
| 3194 |
+
"grad_norm": NaN,
|
| 3195 |
+
"learning_rate": 8.893717741658336e-05,
|
| 3196 |
+
"loss": 0.0,
|
| 3197 |
+
"step": 452
|
| 3198 |
+
},
|
| 3199 |
+
{
|
| 3200 |
+
"epoch": 0.5650140318054256,
|
| 3201 |
+
"grad_norm": NaN,
|
| 3202 |
+
"learning_rate": 8.852152188053674e-05,
|
| 3203 |
+
"loss": 0.0,
|
| 3204 |
+
"step": 453
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 0.5662613033988151,
|
| 3208 |
+
"grad_norm": NaN,
|
| 3209 |
+
"learning_rate": 8.810606720915697e-05,
|
| 3210 |
+
"loss": 0.0,
|
| 3211 |
+
"step": 454
|
| 3212 |
+
},
|
| 3213 |
+
{
|
| 3214 |
+
"epoch": 0.5675085749922045,
|
| 3215 |
+
"grad_norm": NaN,
|
| 3216 |
+
"learning_rate": 8.769082067258585e-05,
|
| 3217 |
+
"loss": 0.0,
|
| 3218 |
+
"step": 455
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 0.568755846585594,
|
| 3222 |
+
"grad_norm": NaN,
|
| 3223 |
+
"learning_rate": 8.727578953732303e-05,
|
| 3224 |
+
"loss": 0.0,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 0.5700031181789835,
|
| 3229 |
+
"grad_norm": NaN,
|
| 3230 |
+
"learning_rate": 8.686098106609889e-05,
|
| 3231 |
+
"loss": 0.0,
|
| 3232 |
+
"step": 457
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 0.571250389772373,
|
| 3236 |
+
"grad_norm": NaN,
|
| 3237 |
+
"learning_rate": 8.644640251774722e-05,
|
| 3238 |
+
"loss": 0.0,
|
| 3239 |
+
"step": 458
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.5724976613657624,
|
| 3243 |
+
"grad_norm": NaN,
|
| 3244 |
+
"learning_rate": 8.603206114707837e-05,
|
| 3245 |
+
"loss": 0.0,
|
| 3246 |
+
"step": 459
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.5737449329591519,
|
| 3250 |
+
"grad_norm": NaN,
|
| 3251 |
+
"learning_rate": 8.561796420475227e-05,
|
| 3252 |
+
"loss": 0.0,
|
| 3253 |
+
"step": 460
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.5749922045525413,
|
| 3257 |
+
"grad_norm": NaN,
|
| 3258 |
+
"learning_rate": 8.52041189371515e-05,
|
| 3259 |
+
"loss": 0.0,
|
| 3260 |
+
"step": 461
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.5762394761459307,
|
| 3264 |
+
"grad_norm": NaN,
|
| 3265 |
+
"learning_rate": 8.479053258625467e-05,
|
| 3266 |
+
"loss": 0.0,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.5774867477393202,
|
| 3271 |
+
"grad_norm": NaN,
|
| 3272 |
+
"learning_rate": 8.437721238950938e-05,
|
| 3273 |
+
"loss": 0.0,
|
| 3274 |
+
"step": 463
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.5787340193327097,
|
| 3278 |
+
"grad_norm": NaN,
|
| 3279 |
+
"learning_rate": 8.396416557970576e-05,
|
| 3280 |
+
"loss": 0.0,
|
| 3281 |
+
"step": 464
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.5799812909260992,
|
| 3285 |
+
"grad_norm": NaN,
|
| 3286 |
+
"learning_rate": 8.355139938484995e-05,
|
| 3287 |
+
"loss": 0.0,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.5812285625194886,
|
| 3292 |
+
"grad_norm": NaN,
|
| 3293 |
+
"learning_rate": 8.313892102803749e-05,
|
| 3294 |
+
"loss": 0.0,
|
| 3295 |
+
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.5824758341128781,
|
| 3299 |
+
"grad_norm": NaN,
|
| 3300 |
+
"learning_rate": 8.272673772732695e-05,
|
| 3301 |
+
"loss": 0.0,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.5837231057062675,
|
| 3306 |
+
"grad_norm": NaN,
|
| 3307 |
+
"learning_rate": 8.231485669561371e-05,
|
| 3308 |
+
"loss": 0.0,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.584970377299657,
|
| 3313 |
+
"grad_norm": NaN,
|
| 3314 |
+
"learning_rate": 8.190328514050365e-05,
|
| 3315 |
+
"loss": 0.0,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.5862176488930465,
|
| 3320 |
+
"grad_norm": NaN,
|
| 3321 |
+
"learning_rate": 8.1492030264187e-05,
|
| 3322 |
+
"loss": 0.0,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.587464920486436,
|
| 3327 |
+
"grad_norm": NaN,
|
| 3328 |
+
"learning_rate": 8.108109926331238e-05,
|
| 3329 |
+
"loss": 0.0,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.5887121920798254,
|
| 3334 |
+
"grad_norm": NaN,
|
| 3335 |
+
"learning_rate": 8.067049932886084e-05,
|
| 3336 |
+
"loss": 0.0,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.5899594636732148,
|
| 3341 |
+
"grad_norm": NaN,
|
| 3342 |
+
"learning_rate": 8.026023764601999e-05,
|
| 3343 |
+
"loss": 0.0,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.5912067352666043,
|
| 3348 |
+
"grad_norm": NaN,
|
| 3349 |
+
"learning_rate": 7.985032139405836e-05,
|
| 3350 |
+
"loss": 0.0,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.5924540068599937,
|
| 3355 |
+
"grad_norm": NaN,
|
| 3356 |
+
"learning_rate": 7.944075774619963e-05,
|
| 3357 |
+
"loss": 0.0,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.5937012784533833,
|
| 3362 |
+
"grad_norm": NaN,
|
| 3363 |
+
"learning_rate": 7.903155386949723e-05,
|
| 3364 |
+
"loss": 0.0,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.5949485500467727,
|
| 3369 |
+
"grad_norm": NaN,
|
| 3370 |
+
"learning_rate": 7.862271692470884e-05,
|
| 3371 |
+
"loss": 0.0,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.5961958216401622,
|
| 3376 |
+
"grad_norm": NaN,
|
| 3377 |
+
"learning_rate": 7.821425406617106e-05,
|
| 3378 |
+
"loss": 0.0,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.5974430932335516,
|
| 3383 |
+
"grad_norm": NaN,
|
| 3384 |
+
"learning_rate": 7.780617244167432e-05,
|
| 3385 |
+
"loss": 0.0,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.598690364826941,
|
| 3390 |
+
"grad_norm": NaN,
|
| 3391 |
+
"learning_rate": 7.739847919233781e-05,
|
| 3392 |
+
"loss": 0.0,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.5999376364203305,
|
| 3397 |
+
"grad_norm": NaN,
|
| 3398 |
+
"learning_rate": 7.699118145248434e-05,
|
| 3399 |
+
"loss": 0.0,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.60118490801372,
|
| 3404 |
+
"grad_norm": NaN,
|
| 3405 |
+
"learning_rate": 7.658428634951562e-05,
|
| 3406 |
+
"loss": 0.0,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.6024321796071095,
|
| 3411 |
+
"grad_norm": NaN,
|
| 3412 |
+
"learning_rate": 7.617780100378756e-05,
|
| 3413 |
+
"loss": 0.0,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.6036794512004989,
|
| 3418 |
+
"grad_norm": NaN,
|
| 3419 |
+
"learning_rate": 7.57717325284856e-05,
|
| 3420 |
+
"loss": 0.0,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.6049267227938884,
|
| 3425 |
+
"grad_norm": NaN,
|
| 3426 |
+
"learning_rate": 7.536608802950027e-05,
|
| 3427 |
+
"loss": 0.0,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.6061739943872778,
|
| 3432 |
+
"grad_norm": NaN,
|
| 3433 |
+
"learning_rate": 7.496087460530285e-05,
|
| 3434 |
+
"loss": 0.0,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.6074212659806673,
|
| 3439 |
+
"grad_norm": NaN,
|
| 3440 |
+
"learning_rate": 7.455609934682116e-05,
|
| 3441 |
+
"loss": 0.0,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.6086685375740567,
|
| 3446 |
+
"grad_norm": NaN,
|
| 3447 |
+
"learning_rate": 7.415176933731536e-05,
|
| 3448 |
+
"loss": 0.0,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.6099158091674463,
|
| 3453 |
+
"grad_norm": NaN,
|
| 3454 |
+
"learning_rate": 7.374789165225416e-05,
|
| 3455 |
+
"loss": 0.0,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.6111630807608357,
|
| 3460 |
+
"grad_norm": NaN,
|
| 3461 |
+
"learning_rate": 7.334447335919096e-05,
|
| 3462 |
+
"loss": 0.0,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.6124103523542251,
|
| 3467 |
+
"grad_norm": NaN,
|
| 3468 |
+
"learning_rate": 7.294152151764006e-05,
|
| 3469 |
+
"loss": 0.0,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.6136576239476146,
|
| 3474 |
+
"grad_norm": NaN,
|
| 3475 |
+
"learning_rate": 7.253904317895332e-05,
|
| 3476 |
+
"loss": 0.0,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.614904895541004,
|
| 3481 |
+
"grad_norm": NaN,
|
| 3482 |
+
"learning_rate": 7.21370453861966e-05,
|
| 3483 |
+
"loss": 0.0,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.6161521671343935,
|
| 3488 |
+
"grad_norm": NaN,
|
| 3489 |
+
"learning_rate": 7.173553517402652e-05,
|
| 3490 |
+
"loss": 0.0,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.617399438727783,
|
| 3495 |
+
"grad_norm": NaN,
|
| 3496 |
+
"learning_rate": 7.133451956856751e-05,
|
| 3497 |
+
"loss": 0.0,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.6186467103211725,
|
| 3502 |
+
"grad_norm": NaN,
|
| 3503 |
+
"learning_rate": 7.093400558728871e-05,
|
| 3504 |
+
"loss": 0.0,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.6198939819145619,
|
| 3509 |
+
"grad_norm": NaN,
|
| 3510 |
+
"learning_rate": 7.053400023888115e-05,
|
| 3511 |
+
"loss": 0.0,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.6211412535079514,
|
| 3516 |
+
"grad_norm": NaN,
|
| 3517 |
+
"learning_rate": 7.013451052313534e-05,
|
| 3518 |
+
"loss": 0.0,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.6223885251013408,
|
| 3523 |
+
"grad_norm": NaN,
|
| 3524 |
+
"learning_rate": 6.973554343081846e-05,
|
| 3525 |
+
"loss": 0.0,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.6236357966947302,
|
| 3530 |
+
"grad_norm": NaN,
|
| 3531 |
+
"learning_rate": 6.933710594355225e-05,
|
| 3532 |
+
"loss": 0.0,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.6248830682881198,
|
| 3537 |
+
"grad_norm": NaN,
|
| 3538 |
+
"learning_rate": 6.893920503369068e-05,
|
| 3539 |
+
"loss": 0.0,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.6261303398815092,
|
| 3544 |
+
"grad_norm": NaN,
|
| 3545 |
+
"learning_rate": 6.854184766419812e-05,
|
| 3546 |
+
"loss": 0.0,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.6273776114748987,
|
| 3551 |
+
"grad_norm": NaN,
|
| 3552 |
+
"learning_rate": 6.814504078852729e-05,
|
| 3553 |
+
"loss": 0.0,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.6286248830682881,
|
| 3558 |
+
"grad_norm": NaN,
|
| 3559 |
+
"learning_rate": 6.774879135049787e-05,
|
| 3560 |
+
"loss": 0.0,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.6298721546616776,
|
| 3565 |
+
"grad_norm": NaN,
|
| 3566 |
+
"learning_rate": 6.735310628417461e-05,
|
| 3567 |
+
"loss": 0.0,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.631119426255067,
|
| 3572 |
+
"grad_norm": NaN,
|
| 3573 |
+
"learning_rate": 6.695799251374625e-05,
|
| 3574 |
+
"loss": 0.0,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.6323666978484564,
|
| 3579 |
+
"grad_norm": NaN,
|
| 3580 |
+
"learning_rate": 6.656345695340431e-05,
|
| 3581 |
+
"loss": 0.0,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.633613969441846,
|
| 3586 |
+
"grad_norm": NaN,
|
| 3587 |
+
"learning_rate": 6.616950650722205e-05,
|
| 3588 |
+
"loss": 0.0,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.6348612410352354,
|
| 3593 |
+
"grad_norm": NaN,
|
| 3594 |
+
"learning_rate": 6.577614806903365e-05,
|
| 3595 |
+
"loss": 0.0,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.6361085126286249,
|
| 3600 |
+
"grad_norm": NaN,
|
| 3601 |
+
"learning_rate": 6.538338852231367e-05,
|
| 3602 |
+
"loss": 0.0,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.6373557842220143,
|
| 3607 |
+
"grad_norm": NaN,
|
| 3608 |
+
"learning_rate": 6.499123474005647e-05,
|
| 3609 |
+
"loss": 0.0,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.6386030558154038,
|
| 3614 |
+
"grad_norm": NaN,
|
| 3615 |
+
"learning_rate": 6.4599693584656e-05,
|
| 3616 |
+
"loss": 0.0,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.6398503274087932,
|
| 3621 |
+
"grad_norm": NaN,
|
| 3622 |
+
"learning_rate": 6.420877190778569e-05,
|
| 3623 |
+
"loss": 0.0,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.6410975990021828,
|
| 3628 |
+
"grad_norm": NaN,
|
| 3629 |
+
"learning_rate": 6.381847655027864e-05,
|
| 3630 |
+
"loss": 0.0,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.6423448705955722,
|
| 3635 |
+
"grad_norm": NaN,
|
| 3636 |
+
"learning_rate": 6.342881434200765e-05,
|
| 3637 |
+
"loss": 0.0,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.6435921421889617,
|
| 3642 |
+
"grad_norm": NaN,
|
| 3643 |
+
"learning_rate": 6.303979210176614e-05,
|
| 3644 |
+
"loss": 0.0,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.6448394137823511,
|
| 3649 |
+
"grad_norm": NaN,
|
| 3650 |
+
"learning_rate": 6.26514166371484e-05,
|
| 3651 |
+
"loss": 0.0,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.6460866853757405,
|
| 3656 |
+
"grad_norm": NaN,
|
| 3657 |
+
"learning_rate": 6.226369474443072e-05,
|
| 3658 |
+
"loss": 0.0,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.64733395696913,
|
| 3663 |
+
"grad_norm": NaN,
|
| 3664 |
+
"learning_rate": 6.18766332084523e-05,
|
| 3665 |
+
"loss": 0.0,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.6485812285625195,
|
| 3670 |
+
"grad_norm": NaN,
|
| 3671 |
+
"learning_rate": 6.149023880249665e-05,
|
| 3672 |
+
"loss": 0.0,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.649828500155909,
|
| 3677 |
+
"grad_norm": NaN,
|
| 3678 |
+
"learning_rate": 6.110451828817298e-05,
|
| 3679 |
+
"loss": 0.0,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.6510757717492984,
|
| 3684 |
+
"grad_norm": NaN,
|
| 3685 |
+
"learning_rate": 6.071947841529801e-05,
|
| 3686 |
+
"loss": 0.0,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.6523230433426879,
|
| 3691 |
+
"grad_norm": NaN,
|
| 3692 |
+
"learning_rate": 6.03351259217776e-05,
|
| 3693 |
+
"loss": 0.0,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.6535703149360773,
|
| 3698 |
+
"grad_norm": NaN,
|
| 3699 |
+
"learning_rate": 5.995146753348909e-05,
|
| 3700 |
+
"loss": 0.0,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.6548175865294668,
|
| 3705 |
+
"grad_norm": NaN,
|
| 3706 |
+
"learning_rate": 5.9568509964163464e-05,
|
| 3707 |
+
"loss": 0.0,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.6560648581228562,
|
| 3712 |
+
"grad_norm": NaN,
|
| 3713 |
+
"learning_rate": 5.9186259915267916e-05,
|
| 3714 |
+
"loss": 0.0,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.6573121297162458,
|
| 3719 |
+
"grad_norm": NaN,
|
| 3720 |
+
"learning_rate": 5.880472407588857e-05,
|
| 3721 |
+
"loss": 0.0,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.6585594013096352,
|
| 3726 |
+
"grad_norm": NaN,
|
| 3727 |
+
"learning_rate": 5.842390912261344e-05,
|
| 3728 |
+
"loss": 0.0,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.6598066729030246,
|
| 3733 |
+
"grad_norm": NaN,
|
| 3734 |
+
"learning_rate": 5.8043821719415534e-05,
|
| 3735 |
+
"loss": 0.0,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.6610539444964141,
|
| 3740 |
+
"grad_norm": NaN,
|
| 3741 |
+
"learning_rate": 5.7664468517536395e-05,
|
| 3742 |
+
"loss": 0.0,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.6623012160898035,
|
| 3747 |
+
"grad_norm": NaN,
|
| 3748 |
+
"learning_rate": 5.728585615536946e-05,
|
| 3749 |
+
"loss": 0.0,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.663548487683193,
|
| 3754 |
+
"grad_norm": NaN,
|
| 3755 |
+
"learning_rate": 5.6907991258344e-05,
|
| 3756 |
+
"loss": 0.0,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.6647957592765825,
|
| 3761 |
+
"grad_norm": NaN,
|
| 3762 |
+
"learning_rate": 5.6530880438809494e-05,
|
| 3763 |
+
"loss": 0.0,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.666043030869972,
|
| 3768 |
+
"grad_norm": NaN,
|
| 3769 |
+
"learning_rate": 5.615453029591935e-05,
|
| 3770 |
+
"loss": 0.0,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.6672903024633614,
|
| 3775 |
+
"grad_norm": NaN,
|
| 3776 |
+
"learning_rate": 5.5778947415515784e-05,
|
| 3777 |
+
"loss": 0.0,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.6685375740567508,
|
| 3782 |
+
"grad_norm": NaN,
|
| 3783 |
+
"learning_rate": 5.540413837001459e-05,
|
| 3784 |
+
"loss": 0.0,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.6697848456501403,
|
| 3789 |
+
"grad_norm": NaN,
|
| 3790 |
+
"learning_rate": 5.50301097182899e-05,
|
| 3791 |
+
"loss": 0.0,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.6710321172435297,
|
| 3796 |
+
"grad_norm": NaN,
|
| 3797 |
+
"learning_rate": 5.465686800555967e-05,
|
| 3798 |
+
"loss": 0.0,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.6722793888369193,
|
| 3803 |
+
"grad_norm": NaN,
|
| 3804 |
+
"learning_rate": 5.4284419763271e-05,
|
| 3805 |
+
"loss": 0.0,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.6735266604303087,
|
| 3810 |
+
"grad_norm": NaN,
|
| 3811 |
+
"learning_rate": 5.391277150898575e-05,
|
| 3812 |
+
"loss": 0.0,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.6747739320236982,
|
| 3817 |
+
"grad_norm": NaN,
|
| 3818 |
+
"learning_rate": 5.354192974626674e-05,
|
| 3819 |
+
"loss": 0.0,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.6760212036170876,
|
| 3824 |
+
"grad_norm": NaN,
|
| 3825 |
+
"learning_rate": 5.317190096456368e-05,
|
| 3826 |
+
"loss": 0.0,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.6772684752104771,
|
| 3831 |
+
"grad_norm": NaN,
|
| 3832 |
+
"learning_rate": 5.2802691639099834e-05,
|
| 3833 |
+
"loss": 0.0,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.6785157468038665,
|
| 3838 |
+
"grad_norm": NaN,
|
| 3839 |
+
"learning_rate": 5.24343082307585e-05,
|
| 3840 |
+
"loss": 0.0,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.679763018397256,
|
| 3845 |
+
"grad_norm": NaN,
|
| 3846 |
+
"learning_rate": 5.206675718597012e-05,
|
| 3847 |
+
"loss": 0.0,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.6810102899906455,
|
| 3852 |
+
"grad_norm": NaN,
|
| 3853 |
+
"learning_rate": 5.1700044936599434e-05,
|
| 3854 |
+
"loss": 0.0,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.6822575615840349,
|
| 3859 |
+
"grad_norm": NaN,
|
| 3860 |
+
"learning_rate": 5.133417789983277e-05,
|
| 3861 |
+
"loss": 0.0,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.6835048331774244,
|
| 3866 |
+
"grad_norm": NaN,
|
| 3867 |
+
"learning_rate": 5.0969162478066055e-05,
|
| 3868 |
+
"loss": 0.0,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.6847521047708138,
|
| 3873 |
+
"grad_norm": NaN,
|
| 3874 |
+
"learning_rate": 5.060500505879244e-05,
|
| 3875 |
+
"loss": 0.0,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.6859993763642033,
|
| 3880 |
+
"grad_norm": NaN,
|
| 3881 |
+
"learning_rate": 5.0241712014490684e-05,
|
| 3882 |
+
"loss": 0.0,
|
| 3883 |
+
"step": 550
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 0.6872466479575927,
|
| 3887 |
+
"grad_norm": NaN,
|
| 3888 |
+
"learning_rate": 4.9879289702513845e-05,
|
| 3889 |
+
"loss": 0.0,
|
| 3890 |
+
"step": 551
|
| 3891 |
+
},
|
| 3892 |
+
{
|
| 3893 |
+
"epoch": 0.6884939195509823,
|
| 3894 |
+
"grad_norm": NaN,
|
| 3895 |
+
"learning_rate": 4.95177444649776e-05,
|
| 3896 |
+
"loss": 0.0,
|
| 3897 |
+
"step": 552
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 0.6897411911443717,
|
| 3901 |
+
"grad_norm": NaN,
|
| 3902 |
+
"learning_rate": 4.9157082628649545e-05,
|
| 3903 |
+
"loss": 0.0,
|
| 3904 |
+
"step": 553
|
| 3905 |
+
},
|
| 3906 |
+
{
|
| 3907 |
+
"epoch": 0.6909884627377612,
|
| 3908 |
+
"grad_norm": NaN,
|
| 3909 |
+
"learning_rate": 4.87973105048385e-05,
|
| 3910 |
+
"loss": 0.0,
|
| 3911 |
+
"step": 554
|
| 3912 |
+
},
|
| 3913 |
+
{
|
| 3914 |
+
"epoch": 0.6922357343311506,
|
| 3915 |
+
"grad_norm": NaN,
|
| 3916 |
+
"learning_rate": 4.8438434389283895e-05,
|
| 3917 |
+
"loss": 0.0,
|
| 3918 |
+
"step": 555
|
| 3919 |
+
},
|
| 3920 |
+
{
|
| 3921 |
+
"epoch": 0.69348300592454,
|
| 3922 |
+
"grad_norm": NaN,
|
| 3923 |
+
"learning_rate": 4.8080460562045736e-05,
|
| 3924 |
+
"loss": 0.0,
|
| 3925 |
+
"step": 556
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 0.6947302775179295,
|
| 3929 |
+
"grad_norm": NaN,
|
| 3930 |
+
"learning_rate": 4.7723395287394746e-05,
|
| 3931 |
+
"loss": 0.0,
|
| 3932 |
+
"step": 557
|
| 3933 |
+
},
|
| 3934 |
+
{
|
| 3935 |
+
"epoch": 0.695977549111319,
|
| 3936 |
+
"grad_norm": NaN,
|
| 3937 |
+
"learning_rate": 4.736724481370248e-05,
|
| 3938 |
+
"loss": 0.0,
|
| 3939 |
+
"step": 558
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 0.6972248207047085,
|
| 3943 |
+
"grad_norm": NaN,
|
| 3944 |
+
"learning_rate": 4.701201537333237e-05,
|
| 3945 |
+
"loss": 0.0,
|
| 3946 |
+
"step": 559
|
| 3947 |
+
},
|
| 3948 |
+
{
|
| 3949 |
+
"epoch": 0.6984720922980979,
|
| 3950 |
+
"grad_norm": NaN,
|
| 3951 |
+
"learning_rate": 4.6657713182530316e-05,
|
| 3952 |
+
"loss": 0.0,
|
| 3953 |
+
"step": 560
|
| 3954 |
+
},
|
| 3955 |
+
{
|
| 3956 |
+
"epoch": 0.6997193638914874,
|
| 3957 |
+
"grad_norm": NaN,
|
| 3958 |
+
"learning_rate": 4.630434444131615e-05,
|
| 3959 |
+
"loss": 0.0,
|
| 3960 |
+
"step": 561
|
| 3961 |
+
},
|
| 3962 |
+
{
|
| 3963 |
+
"epoch": 0.7009666354848768,
|
| 3964 |
+
"grad_norm": NaN,
|
| 3965 |
+
"learning_rate": 4.595191533337494e-05,
|
| 3966 |
+
"loss": 0.0,
|
| 3967 |
+
"step": 562
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 0.7022139070782663,
|
| 3971 |
+
"grad_norm": NaN,
|
| 3972 |
+
"learning_rate": 4.560043202594899e-05,
|
| 3973 |
+
"loss": 0.0,
|
| 3974 |
+
"step": 563
|
| 3975 |
+
},
|
| 3976 |
+
{
|
| 3977 |
+
"epoch": 0.7034611786716558,
|
| 3978 |
+
"grad_norm": NaN,
|
| 3979 |
+
"learning_rate": 4.524990066972982e-05,
|
| 3980 |
+
"loss": 0.0,
|
| 3981 |
+
"step": 564
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 0.7047084502650452,
|
| 3985 |
+
"grad_norm": NaN,
|
| 3986 |
+
"learning_rate": 4.4900327398750363e-05,
|
| 3987 |
+
"loss": 0.0,
|
| 3988 |
+
"step": 565
|
| 3989 |
+
},
|
| 3990 |
+
{
|
| 3991 |
+
"epoch": 0.7059557218584347,
|
| 3992 |
+
"grad_norm": NaN,
|
| 3993 |
+
"learning_rate": 4.4551718330278006e-05,
|
| 3994 |
+
"loss": 0.0,
|
| 3995 |
+
"step": 566
|
| 3996 |
+
},
|
| 3997 |
+
{
|
| 3998 |
+
"epoch": 0.7072029934518241,
|
| 3999 |
+
"grad_norm": NaN,
|
| 4000 |
+
"learning_rate": 4.4204079564707144e-05,
|
| 4001 |
+
"loss": 0.0,
|
| 4002 |
+
"step": 567
|
| 4003 |
+
},
|
| 4004 |
+
{
|
| 4005 |
+
"epoch": 0.7084502650452136,
|
| 4006 |
+
"grad_norm": NaN,
|
| 4007 |
+
"learning_rate": 4.3857417185452644e-05,
|
| 4008 |
+
"loss": 0.0,
|
| 4009 |
+
"step": 568
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 0.709697536638603,
|
| 4013 |
+
"grad_norm": NaN,
|
| 4014 |
+
"learning_rate": 4.351173725884351e-05,
|
| 4015 |
+
"loss": 0.0,
|
| 4016 |
+
"step": 569
|
| 4017 |
+
},
|
| 4018 |
+
{
|
| 4019 |
+
"epoch": 0.7109448082319925,
|
| 4020 |
+
"grad_norm": NaN,
|
| 4021 |
+
"learning_rate": 4.3167045834016326e-05,
|
| 4022 |
+
"loss": 0.0,
|
| 4023 |
+
"step": 570
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 0.712192079825382,
|
| 4027 |
+
"grad_norm": NaN,
|
| 4028 |
+
"learning_rate": 4.282334894280986e-05,
|
| 4029 |
+
"loss": 0.0,
|
| 4030 |
+
"step": 571
|
| 4031 |
+
},
|
| 4032 |
+
{
|
| 4033 |
+
"epoch": 0.7134393514187715,
|
| 4034 |
+
"grad_norm": NaN,
|
| 4035 |
+
"learning_rate": 4.2480652599659154e-05,
|
| 4036 |
+
"loss": 0.0,
|
| 4037 |
+
"step": 572
|
| 4038 |
+
},
|
| 4039 |
+
{
|
| 4040 |
+
"epoch": 0.7146866230121609,
|
| 4041 |
+
"grad_norm": NaN,
|
| 4042 |
+
"learning_rate": 4.213896280149041e-05,
|
| 4043 |
+
"loss": 0.0,
|
| 4044 |
+
"step": 573
|
| 4045 |
+
},
|
| 4046 |
+
{
|
| 4047 |
+
"epoch": 0.7159338946055503,
|
| 4048 |
+
"grad_norm": NaN,
|
| 4049 |
+
"learning_rate": 4.179828552761617e-05,
|
| 4050 |
+
"loss": 0.0,
|
| 4051 |
+
"step": 574
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 0.7171811661989398,
|
| 4055 |
+
"grad_norm": NaN,
|
| 4056 |
+
"learning_rate": 4.1458626739630526e-05,
|
| 4057 |
+
"loss": 0.0,
|
| 4058 |
+
"step": 575
|
| 4059 |
+
},
|
| 4060 |
+
{
|
| 4061 |
+
"epoch": 0.7184284377923292,
|
| 4062 |
+
"grad_norm": NaN,
|
| 4063 |
+
"learning_rate": 4.1119992381304754e-05,
|
| 4064 |
+
"loss": 0.0,
|
| 4065 |
+
"step": 576
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 0.7196757093857188,
|
| 4069 |
+
"grad_norm": NaN,
|
| 4070 |
+
"learning_rate": 4.078238837848352e-05,
|
| 4071 |
+
"loss": 0.0,
|
| 4072 |
+
"step": 577
|
| 4073 |
+
},
|
| 4074 |
+
{
|
| 4075 |
+
"epoch": 0.7209229809791082,
|
| 4076 |
+
"grad_norm": NaN,
|
| 4077 |
+
"learning_rate": 4.04458206389809e-05,
|
| 4078 |
+
"loss": 0.0,
|
| 4079 |
+
"step": 578
|
| 4080 |
+
},
|
| 4081 |
+
{
|
| 4082 |
+
"epoch": 0.7221702525724977,
|
| 4083 |
+
"grad_norm": NaN,
|
| 4084 |
+
"learning_rate": 4.011029505247732e-05,
|
| 4085 |
+
"loss": 0.0,
|
| 4086 |
+
"step": 579
|
| 4087 |
+
},
|
| 4088 |
+
{
|
| 4089 |
+
"epoch": 0.7234175241658871,
|
| 4090 |
+
"grad_norm": NaN,
|
| 4091 |
+
"learning_rate": 3.977581749041616e-05,
|
| 4092 |
+
"loss": 0.0,
|
| 4093 |
+
"step": 580
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 0.7246647957592766,
|
| 4097 |
+
"grad_norm": NaN,
|
| 4098 |
+
"learning_rate": 3.9442393805901245e-05,
|
| 4099 |
+
"loss": 0.0,
|
| 4100 |
+
"step": 581
|
| 4101 |
+
},
|
| 4102 |
+
{
|
| 4103 |
+
"epoch": 0.725912067352666,
|
| 4104 |
+
"grad_norm": NaN,
|
| 4105 |
+
"learning_rate": 3.91100298335944e-05,
|
| 4106 |
+
"loss": 0.0,
|
| 4107 |
+
"step": 582
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 0.7271593389460556,
|
| 4111 |
+
"grad_norm": NaN,
|
| 4112 |
+
"learning_rate": 3.877873138961311e-05,
|
| 4113 |
+
"loss": 0.0,
|
| 4114 |
+
"step": 583
|
| 4115 |
+
},
|
| 4116 |
+
{
|
| 4117 |
+
"epoch": 0.728406610539445,
|
| 4118 |
+
"grad_norm": NaN,
|
| 4119 |
+
"learning_rate": 3.844850427142914e-05,
|
| 4120 |
+
"loss": 0.0,
|
| 4121 |
+
"step": 584
|
| 4122 |
+
},
|
| 4123 |
+
{
|
| 4124 |
+
"epoch": 0.7296538821328344,
|
| 4125 |
+
"grad_norm": NaN,
|
| 4126 |
+
"learning_rate": 3.811935425776667e-05,
|
| 4127 |
+
"loss": 0.0,
|
| 4128 |
+
"step": 585
|
| 4129 |
+
},
|
| 4130 |
+
{
|
| 4131 |
+
"epoch": 0.7309011537262239,
|
| 4132 |
+
"grad_norm": NaN,
|
| 4133 |
+
"learning_rate": 3.779128710850151e-05,
|
| 4134 |
+
"loss": 0.0,
|
| 4135 |
+
"step": 586
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 0.7321484253196133,
|
| 4139 |
+
"grad_norm": NaN,
|
| 4140 |
+
"learning_rate": 3.7464308564560106e-05,
|
| 4141 |
+
"loss": 0.0,
|
| 4142 |
+
"step": 587
|
| 4143 |
+
},
|
| 4144 |
+
{
|
| 4145 |
+
"epoch": 0.7333956969130028,
|
| 4146 |
+
"grad_norm": NaN,
|
| 4147 |
+
"learning_rate": 3.71384243478191e-05,
|
| 4148 |
+
"loss": 0.0,
|
| 4149 |
+
"step": 588
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 0.7346429685063922,
|
| 4153 |
+
"grad_norm": NaN,
|
| 4154 |
+
"learning_rate": 3.681364016100535e-05,
|
| 4155 |
+
"loss": 0.0,
|
| 4156 |
+
"step": 589
|
| 4157 |
+
},
|
| 4158 |
+
{
|
| 4159 |
+
"epoch": 0.7358902400997818,
|
| 4160 |
+
"grad_norm": NaN,
|
| 4161 |
+
"learning_rate": 3.64899616875959e-05,
|
| 4162 |
+
"loss": 0.0,
|
| 4163 |
+
"step": 590
|
| 4164 |
+
},
|
| 4165 |
+
{
|
| 4166 |
+
"epoch": 0.7371375116931712,
|
| 4167 |
+
"grad_norm": NaN,
|
| 4168 |
+
"learning_rate": 3.616739459171866e-05,
|
| 4169 |
+
"loss": 0.0,
|
| 4170 |
+
"step": 591
|
| 4171 |
+
},
|
| 4172 |
+
{
|
| 4173 |
+
"epoch": 0.7383847832865607,
|
| 4174 |
+
"grad_norm": NaN,
|
| 4175 |
+
"learning_rate": 3.5845944518053376e-05,
|
| 4176 |
+
"loss": 0.0,
|
| 4177 |
+
"step": 592
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 0.7396320548799501,
|
| 4181 |
+
"grad_norm": NaN,
|
| 4182 |
+
"learning_rate": 3.552561709173266e-05,
|
| 4183 |
+
"loss": 0.0,
|
| 4184 |
+
"step": 593
|
| 4185 |
+
},
|
| 4186 |
+
{
|
| 4187 |
+
"epoch": 0.7408793264733395,
|
| 4188 |
+
"grad_norm": NaN,
|
| 4189 |
+
"learning_rate": 3.520641791824374e-05,
|
| 4190 |
+
"loss": 0.0,
|
| 4191 |
+
"step": 594
|
| 4192 |
+
},
|
| 4193 |
+
{
|
| 4194 |
+
"epoch": 0.742126598066729,
|
| 4195 |
+
"grad_norm": NaN,
|
| 4196 |
+
"learning_rate": 3.488835258333014e-05,
|
| 4197 |
+
"loss": 0.0,
|
| 4198 |
+
"step": 595
|
| 4199 |
+
},
|
| 4200 |
+
{
|
| 4201 |
+
"epoch": 0.7433738696601185,
|
| 4202 |
+
"grad_norm": NaN,
|
| 4203 |
+
"learning_rate": 3.4571426652894144e-05,
|
| 4204 |
+
"loss": 0.0,
|
| 4205 |
+
"step": 596
|
| 4206 |
+
},
|
| 4207 |
+
{
|
| 4208 |
+
"epoch": 0.744621141253508,
|
| 4209 |
+
"grad_norm": NaN,
|
| 4210 |
+
"learning_rate": 3.4255645672899325e-05,
|
| 4211 |
+
"loss": 0.0,
|
| 4212 |
+
"step": 597
|
| 4213 |
+
},
|
| 4214 |
+
{
|
| 4215 |
+
"epoch": 0.7458684128468974,
|
| 4216 |
+
"grad_norm": NaN,
|
| 4217 |
+
"learning_rate": 3.3941015169273524e-05,
|
| 4218 |
+
"loss": 0.0,
|
| 4219 |
+
"step": 598
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 0.7471156844402869,
|
| 4223 |
+
"grad_norm": NaN,
|
| 4224 |
+
"learning_rate": 3.362754064781202e-05,
|
| 4225 |
+
"loss": 0.0,
|
| 4226 |
+
"step": 599
|
| 4227 |
+
},
|
| 4228 |
+
{
|
| 4229 |
+
"epoch": 0.7483629560336763,
|
| 4230 |
+
"grad_norm": NaN,
|
| 4231 |
+
"learning_rate": 3.331522759408138e-05,
|
| 4232 |
+
"loss": 0.0,
|
| 4233 |
+
"step": 600
|
| 4234 |
+
},
|
| 4235 |
+
{
|
| 4236 |
+
"epoch": 0.7496102276270658,
|
| 4237 |
+
"grad_norm": NaN,
|
| 4238 |
+
"learning_rate": 3.300408147332327e-05,
|
| 4239 |
+
"loss": 0.0,
|
| 4240 |
+
"step": 601
|
| 4241 |
+
},
|
| 4242 |
+
{
|
| 4243 |
+
"epoch": 0.7508574992204553,
|
| 4244 |
+
"grad_norm": NaN,
|
| 4245 |
+
"learning_rate": 3.269410773035903e-05,
|
| 4246 |
+
"loss": 0.0,
|
| 4247 |
+
"step": 602
|
| 4248 |
+
},
|
| 4249 |
+
{
|
| 4250 |
+
"epoch": 0.7521047708138447,
|
| 4251 |
+
"grad_norm": NaN,
|
| 4252 |
+
"learning_rate": 3.238531178949417e-05,
|
| 4253 |
+
"loss": 0.0,
|
| 4254 |
+
"step": 603
|
| 4255 |
+
},
|
| 4256 |
+
{
|
| 4257 |
+
"epoch": 0.7521047708138447,
|
| 4258 |
+
"eval_loss": NaN,
|
| 4259 |
+
"eval_runtime": 60.7212,
|
| 4260 |
+
"eval_samples_per_second": 22.249,
|
| 4261 |
+
"eval_steps_per_second": 5.566,
|
| 4262 |
+
"step": 603
|
| 4263 |
}
|
| 4264 |
],
|
| 4265 |
"logging_steps": 1,
|
|
|
|
| 4279 |
"attributes": {}
|
| 4280 |
}
|
| 4281 |
},
|
| 4282 |
+
"total_flos": 2.483866847590482e+17,
|
| 4283 |
"train_batch_size": 2,
|
| 4284 |
"trial_name": null,
|
| 4285 |
"trial_params": null
|