Training in progress, step 626, checkpoint
Browse files
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 80013120
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01d7218f949bedf308d12f9c2950468200781c4efa5986fb875fe56960d2154d
|
| 3 |
size 80013120
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 41120084
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:137e7ec4d974d11661475f258a688c51395eab9825636ab5c100cb79c582469d
|
| 3 |
size 41120084
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:16488c6650b1c765ea62eeec5f572c77ea27de32fc87bfb1732715dd1732f390
|
| 3 |
size 14244
|
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:eda4edf49d712f7f7fe92801c041a6b5b3549336fb75bb6697c4ba84aa04be0e
|
| 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": 313,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2206,6 +2206,2205 @@
|
|
| 2206 |
"eval_samples_per_second": 13.243,
|
| 2207 |
"eval_steps_per_second": 6.634,
|
| 2208 |
"step": 313
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2209 |
}
|
| 2210 |
],
|
| 2211 |
"logging_steps": 1,
|
|
@@ -2225,7 +4424,7 @@
|
|
| 2225 |
"attributes": {}
|
| 2226 |
}
|
| 2227 |
},
|
| 2228 |
-
"total_flos":
|
| 2229 |
"train_batch_size": 2,
|
| 2230 |
"trial_name": null,
|
| 2231 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.5003996802557954,
|
| 5 |
"eval_steps": 313,
|
| 6 |
+
"global_step": 626,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2206 |
"eval_samples_per_second": 13.243,
|
| 2207 |
"eval_steps_per_second": 6.634,
|
| 2208 |
"step": 313
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 0.2509992006394884,
|
| 2212 |
+
"grad_norm": 0.3559674322605133,
|
| 2213 |
+
"learning_rate": 0.0001718205569778497,
|
| 2214 |
+
"loss": 1.1891,
|
| 2215 |
+
"step": 314
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 0.2517985611510791,
|
| 2219 |
+
"grad_norm": 0.40814101696014404,
|
| 2220 |
+
"learning_rate": 0.0001716441773566966,
|
| 2221 |
+
"loss": 1.1652,
|
| 2222 |
+
"step": 315
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.25259792166266987,
|
| 2226 |
+
"grad_norm": 0.3883415162563324,
|
| 2227 |
+
"learning_rate": 0.00017146733860429612,
|
| 2228 |
+
"loss": 1.1556,
|
| 2229 |
+
"step": 316
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.2533972821742606,
|
| 2233 |
+
"grad_norm": 0.37391719222068787,
|
| 2234 |
+
"learning_rate": 0.0001712900418539183,
|
| 2235 |
+
"loss": 1.1343,
|
| 2236 |
+
"step": 317
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 0.2541966426858513,
|
| 2240 |
+
"grad_norm": 0.36951127648353577,
|
| 2241 |
+
"learning_rate": 0.00017111228824176825,
|
| 2242 |
+
"loss": 1.1139,
|
| 2243 |
+
"step": 318
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 0.25499600319744203,
|
| 2247 |
+
"grad_norm": 0.3808392584323883,
|
| 2248 |
+
"learning_rate": 0.0001709340789069788,
|
| 2249 |
+
"loss": 1.1057,
|
| 2250 |
+
"step": 319
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 0.2557953637090328,
|
| 2254 |
+
"grad_norm": 0.4488889276981354,
|
| 2255 |
+
"learning_rate": 0.00017075541499160334,
|
| 2256 |
+
"loss": 1.1723,
|
| 2257 |
+
"step": 320
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 0.2565947242206235,
|
| 2261 |
+
"grad_norm": 0.4316102862358093,
|
| 2262 |
+
"learning_rate": 0.0001705762976406084,
|
| 2263 |
+
"loss": 1.164,
|
| 2264 |
+
"step": 321
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.2573940847322142,
|
| 2268 |
+
"grad_norm": 0.4768711030483246,
|
| 2269 |
+
"learning_rate": 0.00017039672800186638,
|
| 2270 |
+
"loss": 1.0935,
|
| 2271 |
+
"step": 322
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.25819344524380494,
|
| 2275 |
+
"grad_norm": 0.4212109446525574,
|
| 2276 |
+
"learning_rate": 0.00017021670722614817,
|
| 2277 |
+
"loss": 1.0126,
|
| 2278 |
+
"step": 323
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 0.2589928057553957,
|
| 2282 |
+
"grad_norm": 0.4388445019721985,
|
| 2283 |
+
"learning_rate": 0.0001700362364671157,
|
| 2284 |
+
"loss": 1.2362,
|
| 2285 |
+
"step": 324
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.2597921662669864,
|
| 2289 |
+
"grad_norm": 0.4593929350376129,
|
| 2290 |
+
"learning_rate": 0.0001698553168813147,
|
| 2291 |
+
"loss": 1.323,
|
| 2292 |
+
"step": 325
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 0.26059152677857716,
|
| 2296 |
+
"grad_norm": 0.507211446762085,
|
| 2297 |
+
"learning_rate": 0.00016967394962816713,
|
| 2298 |
+
"loss": 1.2032,
|
| 2299 |
+
"step": 326
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 0.26139088729016785,
|
| 2303 |
+
"grad_norm": 0.4374110698699951,
|
| 2304 |
+
"learning_rate": 0.0001694921358699639,
|
| 2305 |
+
"loss": 0.8798,
|
| 2306 |
+
"step": 327
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.2621902478017586,
|
| 2310 |
+
"grad_norm": 0.558111310005188,
|
| 2311 |
+
"learning_rate": 0.00016930987677185727,
|
| 2312 |
+
"loss": 1.2127,
|
| 2313 |
+
"step": 328
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.26298960831334933,
|
| 2317 |
+
"grad_norm": 0.4802536070346832,
|
| 2318 |
+
"learning_rate": 0.00016912717350185355,
|
| 2319 |
+
"loss": 1.1708,
|
| 2320 |
+
"step": 329
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.2637889688249401,
|
| 2324 |
+
"grad_norm": 0.572343111038208,
|
| 2325 |
+
"learning_rate": 0.00016894402723080537,
|
| 2326 |
+
"loss": 1.2319,
|
| 2327 |
+
"step": 330
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.26458832933653076,
|
| 2331 |
+
"grad_norm": 0.5261095762252808,
|
| 2332 |
+
"learning_rate": 0.00016876043913240452,
|
| 2333 |
+
"loss": 1.1315,
|
| 2334 |
+
"step": 331
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.2653876898481215,
|
| 2338 |
+
"grad_norm": 0.5496971607208252,
|
| 2339 |
+
"learning_rate": 0.00016857641038317404,
|
| 2340 |
+
"loss": 1.1368,
|
| 2341 |
+
"step": 332
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.26618705035971224,
|
| 2345 |
+
"grad_norm": 0.5394718647003174,
|
| 2346 |
+
"learning_rate": 0.00016839194216246108,
|
| 2347 |
+
"loss": 1.0986,
|
| 2348 |
+
"step": 333
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.266986410871303,
|
| 2352 |
+
"grad_norm": 0.514948844909668,
|
| 2353 |
+
"learning_rate": 0.00016820703565242898,
|
| 2354 |
+
"loss": 1.0841,
|
| 2355 |
+
"step": 334
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.26778577138289367,
|
| 2359 |
+
"grad_norm": 0.6216516494750977,
|
| 2360 |
+
"learning_rate": 0.0001680216920380499,
|
| 2361 |
+
"loss": 0.8411,
|
| 2362 |
+
"step": 335
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.2685851318944844,
|
| 2366 |
+
"grad_norm": 0.6901413798332214,
|
| 2367 |
+
"learning_rate": 0.00016783591250709728,
|
| 2368 |
+
"loss": 1.0341,
|
| 2369 |
+
"step": 336
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.26938449240607515,
|
| 2373 |
+
"grad_norm": 0.6468019485473633,
|
| 2374 |
+
"learning_rate": 0.00016764969825013793,
|
| 2375 |
+
"loss": 0.8504,
|
| 2376 |
+
"step": 337
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.2701838529176659,
|
| 2380 |
+
"grad_norm": 0.8816102743148804,
|
| 2381 |
+
"learning_rate": 0.0001674630504605248,
|
| 2382 |
+
"loss": 1.1692,
|
| 2383 |
+
"step": 338
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.2709832134292566,
|
| 2387 |
+
"grad_norm": 0.677508533000946,
|
| 2388 |
+
"learning_rate": 0.00016727597033438902,
|
| 2389 |
+
"loss": 1.0148,
|
| 2390 |
+
"step": 339
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.2717825739408473,
|
| 2394 |
+
"grad_norm": 0.8252014517784119,
|
| 2395 |
+
"learning_rate": 0.00016708845907063236,
|
| 2396 |
+
"loss": 1.1856,
|
| 2397 |
+
"step": 340
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.27258193445243806,
|
| 2401 |
+
"grad_norm": 0.8761686682701111,
|
| 2402 |
+
"learning_rate": 0.0001669005178709195,
|
| 2403 |
+
"loss": 1.168,
|
| 2404 |
+
"step": 341
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.2733812949640288,
|
| 2408 |
+
"grad_norm": 0.9532282948493958,
|
| 2409 |
+
"learning_rate": 0.00016671214793967048,
|
| 2410 |
+
"loss": 1.1328,
|
| 2411 |
+
"step": 342
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.2741806554756195,
|
| 2415 |
+
"grad_norm": 1.1391276121139526,
|
| 2416 |
+
"learning_rate": 0.00016652335048405274,
|
| 2417 |
+
"loss": 1.1273,
|
| 2418 |
+
"step": 343
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.2749800159872102,
|
| 2422 |
+
"grad_norm": 0.8748697638511658,
|
| 2423 |
+
"learning_rate": 0.00016633412671397357,
|
| 2424 |
+
"loss": 0.9519,
|
| 2425 |
+
"step": 344
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.27577937649880097,
|
| 2429 |
+
"grad_norm": 0.9585081934928894,
|
| 2430 |
+
"learning_rate": 0.00016614447784207227,
|
| 2431 |
+
"loss": 1.2655,
|
| 2432 |
+
"step": 345
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.2765787370103917,
|
| 2436 |
+
"grad_norm": 1.1829018592834473,
|
| 2437 |
+
"learning_rate": 0.00016595440508371237,
|
| 2438 |
+
"loss": 1.4466,
|
| 2439 |
+
"step": 346
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.2773780975219824,
|
| 2443 |
+
"grad_norm": 1.3926594257354736,
|
| 2444 |
+
"learning_rate": 0.00016576390965697405,
|
| 2445 |
+
"loss": 0.9497,
|
| 2446 |
+
"step": 347
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.27817745803357313,
|
| 2450 |
+
"grad_norm": 1.3197882175445557,
|
| 2451 |
+
"learning_rate": 0.00016557299278264584,
|
| 2452 |
+
"loss": 1.625,
|
| 2453 |
+
"step": 348
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.2789768185451639,
|
| 2457 |
+
"grad_norm": 1.5563266277313232,
|
| 2458 |
+
"learning_rate": 0.0001653816556842174,
|
| 2459 |
+
"loss": 1.3074,
|
| 2460 |
+
"step": 349
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.2797761790567546,
|
| 2464 |
+
"grad_norm": 2.229512929916382,
|
| 2465 |
+
"learning_rate": 0.00016518989958787126,
|
| 2466 |
+
"loss": 0.8772,
|
| 2467 |
+
"step": 350
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.2805755395683453,
|
| 2471 |
+
"grad_norm": 0.32201531529426575,
|
| 2472 |
+
"learning_rate": 0.00016499772572247515,
|
| 2473 |
+
"loss": 0.8794,
|
| 2474 |
+
"step": 351
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.28137490007993604,
|
| 2478 |
+
"grad_norm": 0.3720848560333252,
|
| 2479 |
+
"learning_rate": 0.000164805135319574,
|
| 2480 |
+
"loss": 1.0193,
|
| 2481 |
+
"step": 352
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.2821742605915268,
|
| 2485 |
+
"grad_norm": 0.44173169136047363,
|
| 2486 |
+
"learning_rate": 0.0001646121296133822,
|
| 2487 |
+
"loss": 1.1591,
|
| 2488 |
+
"step": 353
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.2829736211031175,
|
| 2492 |
+
"grad_norm": 0.3597632646560669,
|
| 2493 |
+
"learning_rate": 0.00016441870984077554,
|
| 2494 |
+
"loss": 1.0938,
|
| 2495 |
+
"step": 354
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.2837729816147082,
|
| 2499 |
+
"grad_norm": 0.4087126851081848,
|
| 2500 |
+
"learning_rate": 0.00016422487724128342,
|
| 2501 |
+
"loss": 1.0694,
|
| 2502 |
+
"step": 355
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.28457234212629895,
|
| 2506 |
+
"grad_norm": 0.528968095779419,
|
| 2507 |
+
"learning_rate": 0.00016403063305708078,
|
| 2508 |
+
"loss": 1.024,
|
| 2509 |
+
"step": 356
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.2853717026378897,
|
| 2513 |
+
"grad_norm": 0.46297335624694824,
|
| 2514 |
+
"learning_rate": 0.0001638359785329802,
|
| 2515 |
+
"loss": 1.1045,
|
| 2516 |
+
"step": 357
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.28617106314948043,
|
| 2520 |
+
"grad_norm": 0.3513520359992981,
|
| 2521 |
+
"learning_rate": 0.00016364091491642398,
|
| 2522 |
+
"loss": 1.204,
|
| 2523 |
+
"step": 358
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.2869704236610711,
|
| 2527 |
+
"grad_norm": 0.3250669240951538,
|
| 2528 |
+
"learning_rate": 0.000163445443457476,
|
| 2529 |
+
"loss": 1.0709,
|
| 2530 |
+
"step": 359
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.28776978417266186,
|
| 2534 |
+
"grad_norm": 0.3269745707511902,
|
| 2535 |
+
"learning_rate": 0.00016324956540881384,
|
| 2536 |
+
"loss": 1.0644,
|
| 2537 |
+
"step": 360
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.2885691446842526,
|
| 2541 |
+
"grad_norm": 0.30179765820503235,
|
| 2542 |
+
"learning_rate": 0.00016305328202572076,
|
| 2543 |
+
"loss": 1.0858,
|
| 2544 |
+
"step": 361
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.28936850519584334,
|
| 2548 |
+
"grad_norm": 0.35633420944213867,
|
| 2549 |
+
"learning_rate": 0.0001628565945660775,
|
| 2550 |
+
"loss": 1.0879,
|
| 2551 |
+
"step": 362
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.290167865707434,
|
| 2555 |
+
"grad_norm": 0.3494223356246948,
|
| 2556 |
+
"learning_rate": 0.00016265950429035443,
|
| 2557 |
+
"loss": 1.0941,
|
| 2558 |
+
"step": 363
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.29096722621902477,
|
| 2562 |
+
"grad_norm": 0.3416435420513153,
|
| 2563 |
+
"learning_rate": 0.00016246201246160327,
|
| 2564 |
+
"loss": 1.185,
|
| 2565 |
+
"step": 364
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.2917665867306155,
|
| 2569 |
+
"grad_norm": 0.4749598503112793,
|
| 2570 |
+
"learning_rate": 0.00016226412034544912,
|
| 2571 |
+
"loss": 1.1142,
|
| 2572 |
+
"step": 365
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.29256594724220625,
|
| 2576 |
+
"grad_norm": 0.3291066884994507,
|
| 2577 |
+
"learning_rate": 0.00016206582921008233,
|
| 2578 |
+
"loss": 1.0666,
|
| 2579 |
+
"step": 366
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.293365307753797,
|
| 2583 |
+
"grad_norm": 0.3534018099308014,
|
| 2584 |
+
"learning_rate": 0.00016186714032625035,
|
| 2585 |
+
"loss": 1.041,
|
| 2586 |
+
"step": 367
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.2941646682653877,
|
| 2590 |
+
"grad_norm": 0.44766634702682495,
|
| 2591 |
+
"learning_rate": 0.00016166805496724965,
|
| 2592 |
+
"loss": 1.124,
|
| 2593 |
+
"step": 368
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.2949640287769784,
|
| 2597 |
+
"grad_norm": 0.43692246079444885,
|
| 2598 |
+
"learning_rate": 0.00016146857440891744,
|
| 2599 |
+
"loss": 1.1488,
|
| 2600 |
+
"step": 369
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.29576338928856916,
|
| 2604 |
+
"grad_norm": 0.546944260597229,
|
| 2605 |
+
"learning_rate": 0.00016126869992962357,
|
| 2606 |
+
"loss": 1.0826,
|
| 2607 |
+
"step": 370
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.2965627498001599,
|
| 2611 |
+
"grad_norm": 0.3757582902908325,
|
| 2612 |
+
"learning_rate": 0.00016106843281026237,
|
| 2613 |
+
"loss": 1.2105,
|
| 2614 |
+
"step": 371
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 0.2973621103117506,
|
| 2618 |
+
"grad_norm": 0.4158326983451843,
|
| 2619 |
+
"learning_rate": 0.00016086777433424435,
|
| 2620 |
+
"loss": 1.0807,
|
| 2621 |
+
"step": 372
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.2981614708233413,
|
| 2625 |
+
"grad_norm": 0.41738831996917725,
|
| 2626 |
+
"learning_rate": 0.00016066672578748814,
|
| 2627 |
+
"loss": 1.0334,
|
| 2628 |
+
"step": 373
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 0.29896083133493206,
|
| 2632 |
+
"grad_norm": 0.4042329490184784,
|
| 2633 |
+
"learning_rate": 0.00016046528845841196,
|
| 2634 |
+
"loss": 1.0007,
|
| 2635 |
+
"step": 374
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 0.2997601918465228,
|
| 2639 |
+
"grad_norm": 0.43798747658729553,
|
| 2640 |
+
"learning_rate": 0.00016026346363792567,
|
| 2641 |
+
"loss": 0.8887,
|
| 2642 |
+
"step": 375
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 0.3005595523581135,
|
| 2646 |
+
"grad_norm": 0.45503515005111694,
|
| 2647 |
+
"learning_rate": 0.00016006125261942229,
|
| 2648 |
+
"loss": 1.002,
|
| 2649 |
+
"step": 376
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 0.30135891286970423,
|
| 2653 |
+
"grad_norm": 0.44098415970802307,
|
| 2654 |
+
"learning_rate": 0.00015985865669876988,
|
| 2655 |
+
"loss": 0.9226,
|
| 2656 |
+
"step": 377
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 0.302158273381295,
|
| 2660 |
+
"grad_norm": 0.5401691198348999,
|
| 2661 |
+
"learning_rate": 0.00015965567717430304,
|
| 2662 |
+
"loss": 1.0046,
|
| 2663 |
+
"step": 378
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.3029576338928857,
|
| 2667 |
+
"grad_norm": 0.4791525602340698,
|
| 2668 |
+
"learning_rate": 0.00015945231534681483,
|
| 2669 |
+
"loss": 1.0362,
|
| 2670 |
+
"step": 379
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 0.3037569944044764,
|
| 2674 |
+
"grad_norm": 0.5511587858200073,
|
| 2675 |
+
"learning_rate": 0.00015924857251954806,
|
| 2676 |
+
"loss": 1.0618,
|
| 2677 |
+
"step": 380
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 0.30455635491606714,
|
| 2681 |
+
"grad_norm": 0.5481637716293335,
|
| 2682 |
+
"learning_rate": 0.00015904444999818745,
|
| 2683 |
+
"loss": 1.0005,
|
| 2684 |
+
"step": 381
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 0.3053557154276579,
|
| 2688 |
+
"grad_norm": 0.6237247586250305,
|
| 2689 |
+
"learning_rate": 0.0001588399490908508,
|
| 2690 |
+
"loss": 0.885,
|
| 2691 |
+
"step": 382
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 0.3061550759392486,
|
| 2695 |
+
"grad_norm": 0.7181023955345154,
|
| 2696 |
+
"learning_rate": 0.00015863507110808086,
|
| 2697 |
+
"loss": 1.4557,
|
| 2698 |
+
"step": 383
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 0.3069544364508393,
|
| 2702 |
+
"grad_norm": 0.5080341100692749,
|
| 2703 |
+
"learning_rate": 0.00015842981736283686,
|
| 2704 |
+
"loss": 0.8882,
|
| 2705 |
+
"step": 384
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 0.30775379696243005,
|
| 2709 |
+
"grad_norm": 0.6519491672515869,
|
| 2710 |
+
"learning_rate": 0.0001582241891704861,
|
| 2711 |
+
"loss": 1.1977,
|
| 2712 |
+
"step": 385
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 0.3085531574740208,
|
| 2716 |
+
"grad_norm": 0.6563478708267212,
|
| 2717 |
+
"learning_rate": 0.0001580181878487955,
|
| 2718 |
+
"loss": 1.0473,
|
| 2719 |
+
"step": 386
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 0.30935251798561153,
|
| 2723 |
+
"grad_norm": 0.6795956492424011,
|
| 2724 |
+
"learning_rate": 0.00015781181471792322,
|
| 2725 |
+
"loss": 1.1955,
|
| 2726 |
+
"step": 387
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 0.3101518784972022,
|
| 2730 |
+
"grad_norm": 0.6601610779762268,
|
| 2731 |
+
"learning_rate": 0.00015760507110041015,
|
| 2732 |
+
"loss": 1.0557,
|
| 2733 |
+
"step": 388
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 0.31095123900879296,
|
| 2737 |
+
"grad_norm": 0.9046745896339417,
|
| 2738 |
+
"learning_rate": 0.00015739795832117143,
|
| 2739 |
+
"loss": 1.1948,
|
| 2740 |
+
"step": 389
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 0.3117505995203837,
|
| 2744 |
+
"grad_norm": 0.8437576293945312,
|
| 2745 |
+
"learning_rate": 0.00015719047770748788,
|
| 2746 |
+
"loss": 0.8749,
|
| 2747 |
+
"step": 390
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 0.31254996003197444,
|
| 2751 |
+
"grad_norm": 1.1434983015060425,
|
| 2752 |
+
"learning_rate": 0.00015698263058899778,
|
| 2753 |
+
"loss": 1.158,
|
| 2754 |
+
"step": 391
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 0.3133493205435651,
|
| 2758 |
+
"grad_norm": 0.8605363965034485,
|
| 2759 |
+
"learning_rate": 0.000156774418297688,
|
| 2760 |
+
"loss": 1.2285,
|
| 2761 |
+
"step": 392
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 0.31414868105515587,
|
| 2765 |
+
"grad_norm": 0.8296773433685303,
|
| 2766 |
+
"learning_rate": 0.00015656584216788567,
|
| 2767 |
+
"loss": 1.2332,
|
| 2768 |
+
"step": 393
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 0.3149480415667466,
|
| 2772 |
+
"grad_norm": 1.2430847883224487,
|
| 2773 |
+
"learning_rate": 0.00015635690353624955,
|
| 2774 |
+
"loss": 1.1903,
|
| 2775 |
+
"step": 394
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 0.31574740207833735,
|
| 2779 |
+
"grad_norm": 1.0443687438964844,
|
| 2780 |
+
"learning_rate": 0.0001561476037417615,
|
| 2781 |
+
"loss": 1.311,
|
| 2782 |
+
"step": 395
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 0.31654676258992803,
|
| 2786 |
+
"grad_norm": 1.023246169090271,
|
| 2787 |
+
"learning_rate": 0.00015593794412571796,
|
| 2788 |
+
"loss": 1.3116,
|
| 2789 |
+
"step": 396
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 0.3173461231015188,
|
| 2793 |
+
"grad_norm": 1.2342387437820435,
|
| 2794 |
+
"learning_rate": 0.00015572792603172115,
|
| 2795 |
+
"loss": 1.1111,
|
| 2796 |
+
"step": 397
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 0.3181454836131095,
|
| 2800 |
+
"grad_norm": 1.0979640483856201,
|
| 2801 |
+
"learning_rate": 0.00015551755080567075,
|
| 2802 |
+
"loss": 1.2079,
|
| 2803 |
+
"step": 398
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 0.31894484412470026,
|
| 2807 |
+
"grad_norm": 1.2995789051055908,
|
| 2808 |
+
"learning_rate": 0.00015530681979575496,
|
| 2809 |
+
"loss": 1.2902,
|
| 2810 |
+
"step": 399
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 0.31974420463629094,
|
| 2814 |
+
"grad_norm": 1.4143871068954468,
|
| 2815 |
+
"learning_rate": 0.00015509573435244214,
|
| 2816 |
+
"loss": 1.1578,
|
| 2817 |
+
"step": 400
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 0.3205435651478817,
|
| 2821 |
+
"grad_norm": 0.37532997131347656,
|
| 2822 |
+
"learning_rate": 0.00015488429582847192,
|
| 2823 |
+
"loss": 0.8548,
|
| 2824 |
+
"step": 401
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 0.3213429256594724,
|
| 2828 |
+
"grad_norm": 0.4204126000404358,
|
| 2829 |
+
"learning_rate": 0.00015467250557884684,
|
| 2830 |
+
"loss": 1.0921,
|
| 2831 |
+
"step": 402
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 0.32214228617106316,
|
| 2835 |
+
"grad_norm": 0.38710692524909973,
|
| 2836 |
+
"learning_rate": 0.0001544603649608232,
|
| 2837 |
+
"loss": 1.1638,
|
| 2838 |
+
"step": 403
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 0.3229416466826539,
|
| 2842 |
+
"grad_norm": 0.40149056911468506,
|
| 2843 |
+
"learning_rate": 0.0001542478753339029,
|
| 2844 |
+
"loss": 1.0853,
|
| 2845 |
+
"step": 404
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 0.3237410071942446,
|
| 2849 |
+
"grad_norm": 0.4281335175037384,
|
| 2850 |
+
"learning_rate": 0.0001540350380598243,
|
| 2851 |
+
"loss": 0.9679,
|
| 2852 |
+
"step": 405
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 0.32454036770583533,
|
| 2856 |
+
"grad_norm": 0.3761771023273468,
|
| 2857 |
+
"learning_rate": 0.00015382185450255365,
|
| 2858 |
+
"loss": 1.1288,
|
| 2859 |
+
"step": 406
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 0.32533972821742607,
|
| 2863 |
+
"grad_norm": 0.4560992121696472,
|
| 2864 |
+
"learning_rate": 0.00015360832602827644,
|
| 2865 |
+
"loss": 1.12,
|
| 2866 |
+
"step": 407
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 0.3261390887290168,
|
| 2870 |
+
"grad_norm": 0.35900700092315674,
|
| 2871 |
+
"learning_rate": 0.00015339445400538852,
|
| 2872 |
+
"loss": 1.0102,
|
| 2873 |
+
"step": 408
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.3269384492406075,
|
| 2877 |
+
"grad_norm": 0.3479507267475128,
|
| 2878 |
+
"learning_rate": 0.00015318023980448735,
|
| 2879 |
+
"loss": 1.2253,
|
| 2880 |
+
"step": 409
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 0.32773780975219824,
|
| 2884 |
+
"grad_norm": 0.3966832458972931,
|
| 2885 |
+
"learning_rate": 0.00015296568479836325,
|
| 2886 |
+
"loss": 1.2279,
|
| 2887 |
+
"step": 410
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 0.328537170263789,
|
| 2891 |
+
"grad_norm": 0.3351828157901764,
|
| 2892 |
+
"learning_rate": 0.00015275079036199062,
|
| 2893 |
+
"loss": 1.161,
|
| 2894 |
+
"step": 411
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 0.3293365307753797,
|
| 2898 |
+
"grad_norm": 0.3129248023033142,
|
| 2899 |
+
"learning_rate": 0.000152535557872519,
|
| 2900 |
+
"loss": 1.0208,
|
| 2901 |
+
"step": 412
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 0.3301358912869704,
|
| 2905 |
+
"grad_norm": 0.3181811273097992,
|
| 2906 |
+
"learning_rate": 0.00015231998870926446,
|
| 2907 |
+
"loss": 1.1992,
|
| 2908 |
+
"step": 413
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 0.33093525179856115,
|
| 2912 |
+
"grad_norm": 0.32170534133911133,
|
| 2913 |
+
"learning_rate": 0.0001521040842537006,
|
| 2914 |
+
"loss": 1.1188,
|
| 2915 |
+
"step": 414
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 0.3317346123101519,
|
| 2919 |
+
"grad_norm": 0.348155677318573,
|
| 2920 |
+
"learning_rate": 0.0001518878458894497,
|
| 2921 |
+
"loss": 1.1693,
|
| 2922 |
+
"step": 415
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 0.33253397282174263,
|
| 2926 |
+
"grad_norm": 0.468597948551178,
|
| 2927 |
+
"learning_rate": 0.00015167127500227397,
|
| 2928 |
+
"loss": 1.041,
|
| 2929 |
+
"step": 416
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 0.3333333333333333,
|
| 2933 |
+
"grad_norm": 0.35079798102378845,
|
| 2934 |
+
"learning_rate": 0.0001514543729800664,
|
| 2935 |
+
"loss": 0.8822,
|
| 2936 |
+
"step": 417
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 0.33413269384492406,
|
| 2940 |
+
"grad_norm": 0.36019131541252136,
|
| 2941 |
+
"learning_rate": 0.0001512371412128424,
|
| 2942 |
+
"loss": 0.9667,
|
| 2943 |
+
"step": 418
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 0.3349320543565148,
|
| 2947 |
+
"grad_norm": 0.37374162673950195,
|
| 2948 |
+
"learning_rate": 0.00015101958109273015,
|
| 2949 |
+
"loss": 1.0106,
|
| 2950 |
+
"step": 419
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 0.33573141486810554,
|
| 2954 |
+
"grad_norm": 0.39020803570747375,
|
| 2955 |
+
"learning_rate": 0.0001508016940139624,
|
| 2956 |
+
"loss": 1.1184,
|
| 2957 |
+
"step": 420
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 0.3365307753796962,
|
| 2961 |
+
"grad_norm": 0.4587428569793701,
|
| 2962 |
+
"learning_rate": 0.00015058348137286703,
|
| 2963 |
+
"loss": 1.3182,
|
| 2964 |
+
"step": 421
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 0.33733013589128696,
|
| 2968 |
+
"grad_norm": 0.40048739314079285,
|
| 2969 |
+
"learning_rate": 0.00015036494456785836,
|
| 2970 |
+
"loss": 1.1058,
|
| 2971 |
+
"step": 422
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 0.3381294964028777,
|
| 2975 |
+
"grad_norm": 0.4465924799442291,
|
| 2976 |
+
"learning_rate": 0.00015014608499942808,
|
| 2977 |
+
"loss": 1.0426,
|
| 2978 |
+
"step": 423
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 0.33892885691446845,
|
| 2982 |
+
"grad_norm": 0.39607423543930054,
|
| 2983 |
+
"learning_rate": 0.00014992690407013634,
|
| 2984 |
+
"loss": 0.9871,
|
| 2985 |
+
"step": 424
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 0.33972821742605913,
|
| 2989 |
+
"grad_norm": 0.3925151824951172,
|
| 2990 |
+
"learning_rate": 0.0001497074031846027,
|
| 2991 |
+
"loss": 1.0183,
|
| 2992 |
+
"step": 425
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 0.3405275779376499,
|
| 2996 |
+
"grad_norm": 0.40714192390441895,
|
| 2997 |
+
"learning_rate": 0.00014948758374949713,
|
| 2998 |
+
"loss": 1.1662,
|
| 2999 |
+
"step": 426
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 0.3413269384492406,
|
| 3003 |
+
"grad_norm": 0.5418148636817932,
|
| 3004 |
+
"learning_rate": 0.0001492674471735311,
|
| 3005 |
+
"loss": 0.9804,
|
| 3006 |
+
"step": 427
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 0.34212629896083135,
|
| 3010 |
+
"grad_norm": 0.5795961022377014,
|
| 3011 |
+
"learning_rate": 0.00014904699486744847,
|
| 3012 |
+
"loss": 1.2834,
|
| 3013 |
+
"step": 428
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 0.34292565947242204,
|
| 3017 |
+
"grad_norm": 0.4905664920806885,
|
| 3018 |
+
"learning_rate": 0.0001488262282440163,
|
| 3019 |
+
"loss": 1.0091,
|
| 3020 |
+
"step": 429
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 0.3437250199840128,
|
| 3024 |
+
"grad_norm": 0.5001060366630554,
|
| 3025 |
+
"learning_rate": 0.00014860514871801618,
|
| 3026 |
+
"loss": 1.1577,
|
| 3027 |
+
"step": 430
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 0.3445243804956035,
|
| 3031 |
+
"grad_norm": 0.7573508620262146,
|
| 3032 |
+
"learning_rate": 0.0001483837577062348,
|
| 3033 |
+
"loss": 0.9795,
|
| 3034 |
+
"step": 431
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 0.34532374100719426,
|
| 3038 |
+
"grad_norm": 0.4972291886806488,
|
| 3039 |
+
"learning_rate": 0.00014816205662745497,
|
| 3040 |
+
"loss": 1.1244,
|
| 3041 |
+
"step": 432
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 0.34612310151878495,
|
| 3045 |
+
"grad_norm": 0.5998781323432922,
|
| 3046 |
+
"learning_rate": 0.00014794004690244669,
|
| 3047 |
+
"loss": 1.1667,
|
| 3048 |
+
"step": 433
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 0.3469224620303757,
|
| 3052 |
+
"grad_norm": 0.6669907569885254,
|
| 3053 |
+
"learning_rate": 0.00014771772995395777,
|
| 3054 |
+
"loss": 1.1025,
|
| 3055 |
+
"step": 434
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 0.34772182254196643,
|
| 3059 |
+
"grad_norm": 0.6359307169914246,
|
| 3060 |
+
"learning_rate": 0.00014749510720670506,
|
| 3061 |
+
"loss": 0.9878,
|
| 3062 |
+
"step": 435
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 0.34852118305355717,
|
| 3066 |
+
"grad_norm": 0.7467820048332214,
|
| 3067 |
+
"learning_rate": 0.00014727218008736486,
|
| 3068 |
+
"loss": 1.1126,
|
| 3069 |
+
"step": 436
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 0.34932054356514786,
|
| 3073 |
+
"grad_norm": 0.6224162578582764,
|
| 3074 |
+
"learning_rate": 0.00014704895002456424,
|
| 3075 |
+
"loss": 1.0024,
|
| 3076 |
+
"step": 437
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 0.3501199040767386,
|
| 3080 |
+
"grad_norm": 0.70039963722229,
|
| 3081 |
+
"learning_rate": 0.00014682541844887164,
|
| 3082 |
+
"loss": 1.219,
|
| 3083 |
+
"step": 438
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 0.35091926458832934,
|
| 3087 |
+
"grad_norm": 0.7462148070335388,
|
| 3088 |
+
"learning_rate": 0.00014660158679278765,
|
| 3089 |
+
"loss": 1.106,
|
| 3090 |
+
"step": 439
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 0.3517186250999201,
|
| 3094 |
+
"grad_norm": 0.7545698881149292,
|
| 3095 |
+
"learning_rate": 0.00014637745649073602,
|
| 3096 |
+
"loss": 1.2432,
|
| 3097 |
+
"step": 440
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 0.35251798561151076,
|
| 3101 |
+
"grad_norm": 0.7020461559295654,
|
| 3102 |
+
"learning_rate": 0.00014615302897905434,
|
| 3103 |
+
"loss": 1.1659,
|
| 3104 |
+
"step": 441
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 0.3533173461231015,
|
| 3108 |
+
"grad_norm": 0.7818939685821533,
|
| 3109 |
+
"learning_rate": 0.0001459283056959849,
|
| 3110 |
+
"loss": 0.9831,
|
| 3111 |
+
"step": 442
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 0.35411670663469225,
|
| 3115 |
+
"grad_norm": 0.8227357268333435,
|
| 3116 |
+
"learning_rate": 0.00014570328808166539,
|
| 3117 |
+
"loss": 1.049,
|
| 3118 |
+
"step": 443
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 0.354916067146283,
|
| 3122 |
+
"grad_norm": 0.8649093508720398,
|
| 3123 |
+
"learning_rate": 0.0001454779775781197,
|
| 3124 |
+
"loss": 1.3937,
|
| 3125 |
+
"step": 444
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 0.35571542765787373,
|
| 3129 |
+
"grad_norm": 0.8531369566917419,
|
| 3130 |
+
"learning_rate": 0.00014525237562924877,
|
| 3131 |
+
"loss": 1.2183,
|
| 3132 |
+
"step": 445
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 0.3565147881694644,
|
| 3136 |
+
"grad_norm": 1.2276400327682495,
|
| 3137 |
+
"learning_rate": 0.00014502648368082123,
|
| 3138 |
+
"loss": 1.4021,
|
| 3139 |
+
"step": 446
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 0.35731414868105515,
|
| 3143 |
+
"grad_norm": 1.1620177030563354,
|
| 3144 |
+
"learning_rate": 0.00014480030318046416,
|
| 3145 |
+
"loss": 1.3332,
|
| 3146 |
+
"step": 447
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 0.3581135091926459,
|
| 3150 |
+
"grad_norm": 1.2599079608917236,
|
| 3151 |
+
"learning_rate": 0.00014457383557765386,
|
| 3152 |
+
"loss": 1.4229,
|
| 3153 |
+
"step": 448
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 0.35891286970423664,
|
| 3157 |
+
"grad_norm": 1.5748292207717896,
|
| 3158 |
+
"learning_rate": 0.0001443470823237064,
|
| 3159 |
+
"loss": 0.9769,
|
| 3160 |
+
"step": 449
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 0.3597122302158273,
|
| 3164 |
+
"grad_norm": 1.4127042293548584,
|
| 3165 |
+
"learning_rate": 0.0001441200448717687,
|
| 3166 |
+
"loss": 0.8914,
|
| 3167 |
+
"step": 450
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 0.36051159072741806,
|
| 3171 |
+
"grad_norm": 0.3052649199962616,
|
| 3172 |
+
"learning_rate": 0.0001438927246768087,
|
| 3173 |
+
"loss": 0.8243,
|
| 3174 |
+
"step": 451
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 0.3613109512390088,
|
| 3178 |
+
"grad_norm": 0.38650140166282654,
|
| 3179 |
+
"learning_rate": 0.0001436651231956064,
|
| 3180 |
+
"loss": 1.1737,
|
| 3181 |
+
"step": 452
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 0.36211031175059955,
|
| 3185 |
+
"grad_norm": 0.3861415982246399,
|
| 3186 |
+
"learning_rate": 0.00014343724188674447,
|
| 3187 |
+
"loss": 1.1948,
|
| 3188 |
+
"step": 453
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 0.36290967226219023,
|
| 3192 |
+
"grad_norm": 0.42397868633270264,
|
| 3193 |
+
"learning_rate": 0.00014320908221059875,
|
| 3194 |
+
"loss": 1.0755,
|
| 3195 |
+
"step": 454
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 0.36370903277378097,
|
| 3199 |
+
"grad_norm": 0.33729004859924316,
|
| 3200 |
+
"learning_rate": 0.00014298064562932904,
|
| 3201 |
+
"loss": 1.0324,
|
| 3202 |
+
"step": 455
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 0.3645083932853717,
|
| 3206 |
+
"grad_norm": 0.46672719717025757,
|
| 3207 |
+
"learning_rate": 0.0001427519336068698,
|
| 3208 |
+
"loss": 1.1945,
|
| 3209 |
+
"step": 456
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 0.36530775379696245,
|
| 3213 |
+
"grad_norm": 0.36421141028404236,
|
| 3214 |
+
"learning_rate": 0.0001425229476089204,
|
| 3215 |
+
"loss": 1.1335,
|
| 3216 |
+
"step": 457
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 0.36610711430855314,
|
| 3220 |
+
"grad_norm": 0.4340507686138153,
|
| 3221 |
+
"learning_rate": 0.00014229368910293627,
|
| 3222 |
+
"loss": 1.0877,
|
| 3223 |
+
"step": 458
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 0.3669064748201439,
|
| 3227 |
+
"grad_norm": 0.34189310669898987,
|
| 3228 |
+
"learning_rate": 0.00014206415955811898,
|
| 3229 |
+
"loss": 1.0999,
|
| 3230 |
+
"step": 459
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 0.3677058353317346,
|
| 3234 |
+
"grad_norm": 0.3289022147655487,
|
| 3235 |
+
"learning_rate": 0.00014183436044540713,
|
| 3236 |
+
"loss": 1.055,
|
| 3237 |
+
"step": 460
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 0.36850519584332536,
|
| 3241 |
+
"grad_norm": 0.3434530198574066,
|
| 3242 |
+
"learning_rate": 0.00014160429323746687,
|
| 3243 |
+
"loss": 1.0693,
|
| 3244 |
+
"step": 461
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 0.36930455635491605,
|
| 3248 |
+
"grad_norm": 0.3331538438796997,
|
| 3249 |
+
"learning_rate": 0.00014137395940868244,
|
| 3250 |
+
"loss": 1.0318,
|
| 3251 |
+
"step": 462
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 0.3701039168665068,
|
| 3255 |
+
"grad_norm": 0.3151646852493286,
|
| 3256 |
+
"learning_rate": 0.00014114336043514661,
|
| 3257 |
+
"loss": 1.1513,
|
| 3258 |
+
"step": 463
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 0.37090327737809753,
|
| 3262 |
+
"grad_norm": 0.3027523159980774,
|
| 3263 |
+
"learning_rate": 0.0001409124977946515,
|
| 3264 |
+
"loss": 1.2295,
|
| 3265 |
+
"step": 464
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 0.37170263788968827,
|
| 3269 |
+
"grad_norm": 0.3899821639060974,
|
| 3270 |
+
"learning_rate": 0.00014068137296667884,
|
| 3271 |
+
"loss": 1.3589,
|
| 3272 |
+
"step": 465
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 0.37250199840127896,
|
| 3276 |
+
"grad_norm": 0.4125242233276367,
|
| 3277 |
+
"learning_rate": 0.0001404499874323905,
|
| 3278 |
+
"loss": 1.0456,
|
| 3279 |
+
"step": 466
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 0.3733013589128697,
|
| 3283 |
+
"grad_norm": 0.3297238051891327,
|
| 3284 |
+
"learning_rate": 0.00014021834267461932,
|
| 3285 |
+
"loss": 1.2101,
|
| 3286 |
+
"step": 467
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 0.37410071942446044,
|
| 3290 |
+
"grad_norm": 0.40019673109054565,
|
| 3291 |
+
"learning_rate": 0.00013998644017785913,
|
| 3292 |
+
"loss": 1.0543,
|
| 3293 |
+
"step": 468
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 0.3749000799360512,
|
| 3297 |
+
"grad_norm": 0.36964860558509827,
|
| 3298 |
+
"learning_rate": 0.0001397542814282556,
|
| 3299 |
+
"loss": 1.2206,
|
| 3300 |
+
"step": 469
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 0.37569944044764186,
|
| 3304 |
+
"grad_norm": 0.3775829076766968,
|
| 3305 |
+
"learning_rate": 0.0001395218679135966,
|
| 3306 |
+
"loss": 0.9938,
|
| 3307 |
+
"step": 470
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 0.3764988009592326,
|
| 3311 |
+
"grad_norm": 0.3969877064228058,
|
| 3312 |
+
"learning_rate": 0.00013928920112330262,
|
| 3313 |
+
"loss": 1.126,
|
| 3314 |
+
"step": 471
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 0.37729816147082335,
|
| 3318 |
+
"grad_norm": 0.43040886521339417,
|
| 3319 |
+
"learning_rate": 0.00013905628254841729,
|
| 3320 |
+
"loss": 1.1977,
|
| 3321 |
+
"step": 472
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 0.3780975219824141,
|
| 3325 |
+
"grad_norm": 0.35382160544395447,
|
| 3326 |
+
"learning_rate": 0.00013882311368159782,
|
| 3327 |
+
"loss": 0.9983,
|
| 3328 |
+
"step": 473
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 0.37889688249400477,
|
| 3332 |
+
"grad_norm": 0.362121045589447,
|
| 3333 |
+
"learning_rate": 0.00013858969601710533,
|
| 3334 |
+
"loss": 0.9183,
|
| 3335 |
+
"step": 474
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"epoch": 0.3796962430055955,
|
| 3339 |
+
"grad_norm": 0.4062749445438385,
|
| 3340 |
+
"learning_rate": 0.00013835603105079539,
|
| 3341 |
+
"loss": 0.9459,
|
| 3342 |
+
"step": 475
|
| 3343 |
+
},
|
| 3344 |
+
{
|
| 3345 |
+
"epoch": 0.38049560351718625,
|
| 3346 |
+
"grad_norm": 0.4215144217014313,
|
| 3347 |
+
"learning_rate": 0.00013812212028010854,
|
| 3348 |
+
"loss": 1.0205,
|
| 3349 |
+
"step": 476
|
| 3350 |
+
},
|
| 3351 |
+
{
|
| 3352 |
+
"epoch": 0.381294964028777,
|
| 3353 |
+
"grad_norm": 0.49302709102630615,
|
| 3354 |
+
"learning_rate": 0.00013788796520406034,
|
| 3355 |
+
"loss": 1.0908,
|
| 3356 |
+
"step": 477
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 0.3820943245403677,
|
| 3360 |
+
"grad_norm": 0.49464038014411926,
|
| 3361 |
+
"learning_rate": 0.0001376535673232321,
|
| 3362 |
+
"loss": 1.0684,
|
| 3363 |
+
"step": 478
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 0.3828936850519584,
|
| 3367 |
+
"grad_norm": 0.4653661549091339,
|
| 3368 |
+
"learning_rate": 0.00013741892813976116,
|
| 3369 |
+
"loss": 1.1386,
|
| 3370 |
+
"step": 479
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 0.38369304556354916,
|
| 3374 |
+
"grad_norm": 0.5560212731361389,
|
| 3375 |
+
"learning_rate": 0.0001371840491573312,
|
| 3376 |
+
"loss": 1.1828,
|
| 3377 |
+
"step": 480
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 0.3844924060751399,
|
| 3381 |
+
"grad_norm": 0.46823370456695557,
|
| 3382 |
+
"learning_rate": 0.0001369489318811626,
|
| 3383 |
+
"loss": 1.1321,
|
| 3384 |
+
"step": 481
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 0.3852917665867306,
|
| 3388 |
+
"grad_norm": 0.5444937348365784,
|
| 3389 |
+
"learning_rate": 0.000136713577818003,
|
| 3390 |
+
"loss": 1.0788,
|
| 3391 |
+
"step": 482
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 0.38609112709832133,
|
| 3395 |
+
"grad_norm": 0.5422117710113525,
|
| 3396 |
+
"learning_rate": 0.0001364779884761173,
|
| 3397 |
+
"loss": 1.1162,
|
| 3398 |
+
"step": 483
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 0.38689048760991207,
|
| 3402 |
+
"grad_norm": 0.5373064875602722,
|
| 3403 |
+
"learning_rate": 0.0001362421653652784,
|
| 3404 |
+
"loss": 1.0085,
|
| 3405 |
+
"step": 484
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 0.3876898481215028,
|
| 3409 |
+
"grad_norm": 0.6702423095703125,
|
| 3410 |
+
"learning_rate": 0.00013600610999675702,
|
| 3411 |
+
"loss": 1.0995,
|
| 3412 |
+
"step": 485
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 0.38848920863309355,
|
| 3416 |
+
"grad_norm": 0.5719643235206604,
|
| 3417 |
+
"learning_rate": 0.0001357698238833126,
|
| 3418 |
+
"loss": 0.9434,
|
| 3419 |
+
"step": 486
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 0.38928856914468424,
|
| 3423 |
+
"grad_norm": 0.5578810572624207,
|
| 3424 |
+
"learning_rate": 0.00013553330853918304,
|
| 3425 |
+
"loss": 1.1579,
|
| 3426 |
+
"step": 487
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 0.390087929656275,
|
| 3430 |
+
"grad_norm": 0.7008306384086609,
|
| 3431 |
+
"learning_rate": 0.0001352965654800755,
|
| 3432 |
+
"loss": 1.0893,
|
| 3433 |
+
"step": 488
|
| 3434 |
+
},
|
| 3435 |
+
{
|
| 3436 |
+
"epoch": 0.3908872901678657,
|
| 3437 |
+
"grad_norm": 0.6100425124168396,
|
| 3438 |
+
"learning_rate": 0.0001350595962231563,
|
| 3439 |
+
"loss": 1.1744,
|
| 3440 |
+
"step": 489
|
| 3441 |
+
},
|
| 3442 |
+
{
|
| 3443 |
+
"epoch": 0.39168665067945646,
|
| 3444 |
+
"grad_norm": 0.7699117064476013,
|
| 3445 |
+
"learning_rate": 0.00013482240228704142,
|
| 3446 |
+
"loss": 1.1863,
|
| 3447 |
+
"step": 490
|
| 3448 |
+
},
|
| 3449 |
+
{
|
| 3450 |
+
"epoch": 0.39248601119104715,
|
| 3451 |
+
"grad_norm": 0.7704407572746277,
|
| 3452 |
+
"learning_rate": 0.00013458498519178664,
|
| 3453 |
+
"loss": 1.1027,
|
| 3454 |
+
"step": 491
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 0.3932853717026379,
|
| 3458 |
+
"grad_norm": 0.8181988000869751,
|
| 3459 |
+
"learning_rate": 0.00013434734645887788,
|
| 3460 |
+
"loss": 0.897,
|
| 3461 |
+
"step": 492
|
| 3462 |
+
},
|
| 3463 |
+
{
|
| 3464 |
+
"epoch": 0.39408473221422863,
|
| 3465 |
+
"grad_norm": 0.8602463603019714,
|
| 3466 |
+
"learning_rate": 0.00013410948761122143,
|
| 3467 |
+
"loss": 1.0187,
|
| 3468 |
+
"step": 493
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"epoch": 0.39488409272581937,
|
| 3472 |
+
"grad_norm": 0.8925572037696838,
|
| 3473 |
+
"learning_rate": 0.00013387141017313418,
|
| 3474 |
+
"loss": 1.1308,
|
| 3475 |
+
"step": 494
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 0.39568345323741005,
|
| 3479 |
+
"grad_norm": 0.8595069050788879,
|
| 3480 |
+
"learning_rate": 0.00013363311567033382,
|
| 3481 |
+
"loss": 1.1001,
|
| 3482 |
+
"step": 495
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 0.3964828137490008,
|
| 3486 |
+
"grad_norm": 0.7815159559249878,
|
| 3487 |
+
"learning_rate": 0.0001333946056299292,
|
| 3488 |
+
"loss": 1.0728,
|
| 3489 |
+
"step": 496
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 0.39728217426059154,
|
| 3493 |
+
"grad_norm": 0.9263238310813904,
|
| 3494 |
+
"learning_rate": 0.00013315588158041032,
|
| 3495 |
+
"loss": 1.1485,
|
| 3496 |
+
"step": 497
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 0.3980815347721823,
|
| 3500 |
+
"grad_norm": 1.121984839439392,
|
| 3501 |
+
"learning_rate": 0.0001329169450516387,
|
| 3502 |
+
"loss": 0.9351,
|
| 3503 |
+
"step": 498
|
| 3504 |
+
},
|
| 3505 |
+
{
|
| 3506 |
+
"epoch": 0.39888089528377296,
|
| 3507 |
+
"grad_norm": 1.190191388130188,
|
| 3508 |
+
"learning_rate": 0.0001326777975748376,
|
| 3509 |
+
"loss": 1.1409,
|
| 3510 |
+
"step": 499
|
| 3511 |
+
},
|
| 3512 |
+
{
|
| 3513 |
+
"epoch": 0.3996802557953637,
|
| 3514 |
+
"grad_norm": 1.6871143579483032,
|
| 3515 |
+
"learning_rate": 0.000132438440682582,
|
| 3516 |
+
"loss": 1.2688,
|
| 3517 |
+
"step": 500
|
| 3518 |
+
},
|
| 3519 |
+
{
|
| 3520 |
+
"epoch": 0.40047961630695444,
|
| 3521 |
+
"grad_norm": 0.3561265468597412,
|
| 3522 |
+
"learning_rate": 0.00013219887590878903,
|
| 3523 |
+
"loss": 1.0588,
|
| 3524 |
+
"step": 501
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 0.4012789768185452,
|
| 3528 |
+
"grad_norm": 0.37955114245414734,
|
| 3529 |
+
"learning_rate": 0.00013195910478870803,
|
| 3530 |
+
"loss": 1.0717,
|
| 3531 |
+
"step": 502
|
| 3532 |
+
},
|
| 3533 |
+
{
|
| 3534 |
+
"epoch": 0.40207833733013587,
|
| 3535 |
+
"grad_norm": 0.3907930850982666,
|
| 3536 |
+
"learning_rate": 0.00013171912885891063,
|
| 3537 |
+
"loss": 1.1347,
|
| 3538 |
+
"step": 503
|
| 3539 |
+
},
|
| 3540 |
+
{
|
| 3541 |
+
"epoch": 0.4028776978417266,
|
| 3542 |
+
"grad_norm": 0.3910665214061737,
|
| 3543 |
+
"learning_rate": 0.000131478949657281,
|
| 3544 |
+
"loss": 1.2596,
|
| 3545 |
+
"step": 504
|
| 3546 |
+
},
|
| 3547 |
+
{
|
| 3548 |
+
"epoch": 0.40367705835331735,
|
| 3549 |
+
"grad_norm": 0.3303968012332916,
|
| 3550 |
+
"learning_rate": 0.00013123856872300607,
|
| 3551 |
+
"loss": 0.8503,
|
| 3552 |
+
"step": 505
|
| 3553 |
+
},
|
| 3554 |
+
{
|
| 3555 |
+
"epoch": 0.4044764188649081,
|
| 3556 |
+
"grad_norm": 0.383739173412323,
|
| 3557 |
+
"learning_rate": 0.00013099798759656545,
|
| 3558 |
+
"loss": 1.1542,
|
| 3559 |
+
"step": 506
|
| 3560 |
+
},
|
| 3561 |
+
{
|
| 3562 |
+
"epoch": 0.4052757793764988,
|
| 3563 |
+
"grad_norm": 0.3624424934387207,
|
| 3564 |
+
"learning_rate": 0.00013075720781972175,
|
| 3565 |
+
"loss": 1.2591,
|
| 3566 |
+
"step": 507
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 0.4060751398880895,
|
| 3570 |
+
"grad_norm": 0.33615046739578247,
|
| 3571 |
+
"learning_rate": 0.00013051623093551063,
|
| 3572 |
+
"loss": 1.1455,
|
| 3573 |
+
"step": 508
|
| 3574 |
+
},
|
| 3575 |
+
{
|
| 3576 |
+
"epoch": 0.40687450039968026,
|
| 3577 |
+
"grad_norm": 0.3861737847328186,
|
| 3578 |
+
"learning_rate": 0.00013027505848823084,
|
| 3579 |
+
"loss": 1.0889,
|
| 3580 |
+
"step": 509
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"epoch": 0.407673860911271,
|
| 3584 |
+
"grad_norm": 0.43577584624290466,
|
| 3585 |
+
"learning_rate": 0.00013003369202343453,
|
| 3586 |
+
"loss": 1.1141,
|
| 3587 |
+
"step": 510
|
| 3588 |
+
},
|
| 3589 |
+
{
|
| 3590 |
+
"epoch": 0.4084732214228617,
|
| 3591 |
+
"grad_norm": 0.3891628086566925,
|
| 3592 |
+
"learning_rate": 0.0001297921330879171,
|
| 3593 |
+
"loss": 1.2212,
|
| 3594 |
+
"step": 511
|
| 3595 |
+
},
|
| 3596 |
+
{
|
| 3597 |
+
"epoch": 0.40927258193445243,
|
| 3598 |
+
"grad_norm": 0.35330793261528015,
|
| 3599 |
+
"learning_rate": 0.00012955038322970736,
|
| 3600 |
+
"loss": 1.1245,
|
| 3601 |
+
"step": 512
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"epoch": 0.41007194244604317,
|
| 3605 |
+
"grad_norm": 0.3083156645298004,
|
| 3606 |
+
"learning_rate": 0.0001293084439980578,
|
| 3607 |
+
"loss": 1.1398,
|
| 3608 |
+
"step": 513
|
| 3609 |
+
},
|
| 3610 |
+
{
|
| 3611 |
+
"epoch": 0.4108713029576339,
|
| 3612 |
+
"grad_norm": 0.3527776300907135,
|
| 3613 |
+
"learning_rate": 0.00012906631694343436,
|
| 3614 |
+
"loss": 1.0458,
|
| 3615 |
+
"step": 514
|
| 3616 |
+
},
|
| 3617 |
+
{
|
| 3618 |
+
"epoch": 0.4116706634692246,
|
| 3619 |
+
"grad_norm": 0.34124818444252014,
|
| 3620 |
+
"learning_rate": 0.00012882400361750675,
|
| 3621 |
+
"loss": 1.2349,
|
| 3622 |
+
"step": 515
|
| 3623 |
+
},
|
| 3624 |
+
{
|
| 3625 |
+
"epoch": 0.41247002398081534,
|
| 3626 |
+
"grad_norm": 0.3764817416667938,
|
| 3627 |
+
"learning_rate": 0.0001285815055731383,
|
| 3628 |
+
"loss": 1.1106,
|
| 3629 |
+
"step": 516
|
| 3630 |
+
},
|
| 3631 |
+
{
|
| 3632 |
+
"epoch": 0.4132693844924061,
|
| 3633 |
+
"grad_norm": 0.35788124799728394,
|
| 3634 |
+
"learning_rate": 0.00012833882436437624,
|
| 3635 |
+
"loss": 1.0105,
|
| 3636 |
+
"step": 517
|
| 3637 |
+
},
|
| 3638 |
+
{
|
| 3639 |
+
"epoch": 0.4140687450039968,
|
| 3640 |
+
"grad_norm": 0.35195162892341614,
|
| 3641 |
+
"learning_rate": 0.00012809596154644148,
|
| 3642 |
+
"loss": 1.034,
|
| 3643 |
+
"step": 518
|
| 3644 |
+
},
|
| 3645 |
+
{
|
| 3646 |
+
"epoch": 0.4148681055155875,
|
| 3647 |
+
"grad_norm": 0.3743147552013397,
|
| 3648 |
+
"learning_rate": 0.00012785291867571883,
|
| 3649 |
+
"loss": 1.1551,
|
| 3650 |
+
"step": 519
|
| 3651 |
+
},
|
| 3652 |
+
{
|
| 3653 |
+
"epoch": 0.41566746602717825,
|
| 3654 |
+
"grad_norm": 0.37264588475227356,
|
| 3655 |
+
"learning_rate": 0.00012760969730974694,
|
| 3656 |
+
"loss": 1.084,
|
| 3657 |
+
"step": 520
|
| 3658 |
+
},
|
| 3659 |
+
{
|
| 3660 |
+
"epoch": 0.416466826538769,
|
| 3661 |
+
"grad_norm": 0.4359087646007538,
|
| 3662 |
+
"learning_rate": 0.0001273662990072083,
|
| 3663 |
+
"loss": 1.1565,
|
| 3664 |
+
"step": 521
|
| 3665 |
+
},
|
| 3666 |
+
{
|
| 3667 |
+
"epoch": 0.4172661870503597,
|
| 3668 |
+
"grad_norm": 0.4252093732357025,
|
| 3669 |
+
"learning_rate": 0.00012712272532791946,
|
| 3670 |
+
"loss": 1.1465,
|
| 3671 |
+
"step": 522
|
| 3672 |
+
},
|
| 3673 |
+
{
|
| 3674 |
+
"epoch": 0.4180655475619504,
|
| 3675 |
+
"grad_norm": 0.4450690746307373,
|
| 3676 |
+
"learning_rate": 0.0001268789778328207,
|
| 3677 |
+
"loss": 1.051,
|
| 3678 |
+
"step": 523
|
| 3679 |
+
},
|
| 3680 |
+
{
|
| 3681 |
+
"epoch": 0.41886490807354115,
|
| 3682 |
+
"grad_norm": 0.4181789755821228,
|
| 3683 |
+
"learning_rate": 0.0001266350580839664,
|
| 3684 |
+
"loss": 0.94,
|
| 3685 |
+
"step": 524
|
| 3686 |
+
},
|
| 3687 |
+
{
|
| 3688 |
+
"epoch": 0.4196642685851319,
|
| 3689 |
+
"grad_norm": 0.43827134370803833,
|
| 3690 |
+
"learning_rate": 0.0001263909676445145,
|
| 3691 |
+
"loss": 1.1064,
|
| 3692 |
+
"step": 525
|
| 3693 |
+
},
|
| 3694 |
+
{
|
| 3695 |
+
"epoch": 0.42046362909672264,
|
| 3696 |
+
"grad_norm": 0.43333086371421814,
|
| 3697 |
+
"learning_rate": 0.0001261467080787171,
|
| 3698 |
+
"loss": 1.0093,
|
| 3699 |
+
"step": 526
|
| 3700 |
+
},
|
| 3701 |
+
{
|
| 3702 |
+
"epoch": 0.4212629896083134,
|
| 3703 |
+
"grad_norm": 0.4483776092529297,
|
| 3704 |
+
"learning_rate": 0.0001259022809519101,
|
| 3705 |
+
"loss": 0.9522,
|
| 3706 |
+
"step": 527
|
| 3707 |
+
},
|
| 3708 |
+
{
|
| 3709 |
+
"epoch": 0.42206235011990406,
|
| 3710 |
+
"grad_norm": 0.42380601167678833,
|
| 3711 |
+
"learning_rate": 0.00012565768783050305,
|
| 3712 |
+
"loss": 1.1873,
|
| 3713 |
+
"step": 528
|
| 3714 |
+
},
|
| 3715 |
+
{
|
| 3716 |
+
"epoch": 0.4228617106314948,
|
| 3717 |
+
"grad_norm": 0.46552905440330505,
|
| 3718 |
+
"learning_rate": 0.00012541293028196944,
|
| 3719 |
+
"loss": 1.1422,
|
| 3720 |
+
"step": 529
|
| 3721 |
+
},
|
| 3722 |
+
{
|
| 3723 |
+
"epoch": 0.42366107114308554,
|
| 3724 |
+
"grad_norm": 0.4734083414077759,
|
| 3725 |
+
"learning_rate": 0.0001251680098748365,
|
| 3726 |
+
"loss": 1.1511,
|
| 3727 |
+
"step": 530
|
| 3728 |
+
},
|
| 3729 |
+
{
|
| 3730 |
+
"epoch": 0.4244604316546763,
|
| 3731 |
+
"grad_norm": 0.5062898993492126,
|
| 3732 |
+
"learning_rate": 0.00012492292817867497,
|
| 3733 |
+
"loss": 0.8868,
|
| 3734 |
+
"step": 531
|
| 3735 |
+
},
|
| 3736 |
+
{
|
| 3737 |
+
"epoch": 0.42525979216626697,
|
| 3738 |
+
"grad_norm": 0.57962566614151,
|
| 3739 |
+
"learning_rate": 0.00012467768676408935,
|
| 3740 |
+
"loss": 1.072,
|
| 3741 |
+
"step": 532
|
| 3742 |
+
},
|
| 3743 |
+
{
|
| 3744 |
+
"epoch": 0.4260591526778577,
|
| 3745 |
+
"grad_norm": 0.5128533244132996,
|
| 3746 |
+
"learning_rate": 0.00012443228720270767,
|
| 3747 |
+
"loss": 0.9567,
|
| 3748 |
+
"step": 533
|
| 3749 |
+
},
|
| 3750 |
+
{
|
| 3751 |
+
"epoch": 0.42685851318944845,
|
| 3752 |
+
"grad_norm": 0.5683857202529907,
|
| 3753 |
+
"learning_rate": 0.00012418673106717143,
|
| 3754 |
+
"loss": 1.2017,
|
| 3755 |
+
"step": 534
|
| 3756 |
+
},
|
| 3757 |
+
{
|
| 3758 |
+
"epoch": 0.4276578737010392,
|
| 3759 |
+
"grad_norm": 0.5700757503509521,
|
| 3760 |
+
"learning_rate": 0.00012394101993112557,
|
| 3761 |
+
"loss": 1.2576,
|
| 3762 |
+
"step": 535
|
| 3763 |
+
},
|
| 3764 |
+
{
|
| 3765 |
+
"epoch": 0.4284572342126299,
|
| 3766 |
+
"grad_norm": 0.7162390351295471,
|
| 3767 |
+
"learning_rate": 0.00012369515536920825,
|
| 3768 |
+
"loss": 1.0045,
|
| 3769 |
+
"step": 536
|
| 3770 |
+
},
|
| 3771 |
+
{
|
| 3772 |
+
"epoch": 0.4292565947242206,
|
| 3773 |
+
"grad_norm": 0.594262421131134,
|
| 3774 |
+
"learning_rate": 0.00012344913895704097,
|
| 3775 |
+
"loss": 0.8915,
|
| 3776 |
+
"step": 537
|
| 3777 |
+
},
|
| 3778 |
+
{
|
| 3779 |
+
"epoch": 0.43005595523581136,
|
| 3780 |
+
"grad_norm": 0.6075459718704224,
|
| 3781 |
+
"learning_rate": 0.00012320297227121833,
|
| 3782 |
+
"loss": 1.1498,
|
| 3783 |
+
"step": 538
|
| 3784 |
+
},
|
| 3785 |
+
{
|
| 3786 |
+
"epoch": 0.4308553157474021,
|
| 3787 |
+
"grad_norm": 0.7022448182106018,
|
| 3788 |
+
"learning_rate": 0.0001229566568892979,
|
| 3789 |
+
"loss": 1.1091,
|
| 3790 |
+
"step": 539
|
| 3791 |
+
},
|
| 3792 |
+
{
|
| 3793 |
+
"epoch": 0.4316546762589928,
|
| 3794 |
+
"grad_norm": 0.8270899057388306,
|
| 3795 |
+
"learning_rate": 0.00012271019438979023,
|
| 3796 |
+
"loss": 1.1842,
|
| 3797 |
+
"step": 540
|
| 3798 |
+
},
|
| 3799 |
+
{
|
| 3800 |
+
"epoch": 0.4324540367705835,
|
| 3801 |
+
"grad_norm": 0.7911314368247986,
|
| 3802 |
+
"learning_rate": 0.00012246358635214868,
|
| 3803 |
+
"loss": 1.0775,
|
| 3804 |
+
"step": 541
|
| 3805 |
+
},
|
| 3806 |
+
{
|
| 3807 |
+
"epoch": 0.43325339728217427,
|
| 3808 |
+
"grad_norm": 0.6665616035461426,
|
| 3809 |
+
"learning_rate": 0.0001222168343567592,
|
| 3810 |
+
"loss": 1.0582,
|
| 3811 |
+
"step": 542
|
| 3812 |
+
},
|
| 3813 |
+
{
|
| 3814 |
+
"epoch": 0.434052757793765,
|
| 3815 |
+
"grad_norm": 0.7318521738052368,
|
| 3816 |
+
"learning_rate": 0.00012196993998493043,
|
| 3817 |
+
"loss": 1.1444,
|
| 3818 |
+
"step": 543
|
| 3819 |
+
},
|
| 3820 |
+
{
|
| 3821 |
+
"epoch": 0.4348521183053557,
|
| 3822 |
+
"grad_norm": 0.7249645590782166,
|
| 3823 |
+
"learning_rate": 0.00012172290481888331,
|
| 3824 |
+
"loss": 1.086,
|
| 3825 |
+
"step": 544
|
| 3826 |
+
},
|
| 3827 |
+
{
|
| 3828 |
+
"epoch": 0.43565147881694644,
|
| 3829 |
+
"grad_norm": 0.8864445090293884,
|
| 3830 |
+
"learning_rate": 0.00012147573044174112,
|
| 3831 |
+
"loss": 1.3008,
|
| 3832 |
+
"step": 545
|
| 3833 |
+
},
|
| 3834 |
+
{
|
| 3835 |
+
"epoch": 0.4364508393285372,
|
| 3836 |
+
"grad_norm": 0.7858752012252808,
|
| 3837 |
+
"learning_rate": 0.00012122841843751924,
|
| 3838 |
+
"loss": 1.1622,
|
| 3839 |
+
"step": 546
|
| 3840 |
+
},
|
| 3841 |
+
{
|
| 3842 |
+
"epoch": 0.4372501998401279,
|
| 3843 |
+
"grad_norm": 0.8661579489707947,
|
| 3844 |
+
"learning_rate": 0.00012098097039111512,
|
| 3845 |
+
"loss": 1.009,
|
| 3846 |
+
"step": 547
|
| 3847 |
+
},
|
| 3848 |
+
{
|
| 3849 |
+
"epoch": 0.4380495603517186,
|
| 3850 |
+
"grad_norm": 1.0622788667678833,
|
| 3851 |
+
"learning_rate": 0.00012073338788829787,
|
| 3852 |
+
"loss": 1.044,
|
| 3853 |
+
"step": 548
|
| 3854 |
+
},
|
| 3855 |
+
{
|
| 3856 |
+
"epoch": 0.43884892086330934,
|
| 3857 |
+
"grad_norm": 1.1994202136993408,
|
| 3858 |
+
"learning_rate": 0.0001204856725156984,
|
| 3859 |
+
"loss": 1.1495,
|
| 3860 |
+
"step": 549
|
| 3861 |
+
},
|
| 3862 |
+
{
|
| 3863 |
+
"epoch": 0.4396482813749001,
|
| 3864 |
+
"grad_norm": 1.5216366052627563,
|
| 3865 |
+
"learning_rate": 0.00012023782586079908,
|
| 3866 |
+
"loss": 1.1216,
|
| 3867 |
+
"step": 550
|
| 3868 |
+
},
|
| 3869 |
+
{
|
| 3870 |
+
"epoch": 0.4404476418864908,
|
| 3871 |
+
"grad_norm": 0.4045037627220154,
|
| 3872 |
+
"learning_rate": 0.00011998984951192361,
|
| 3873 |
+
"loss": 1.0062,
|
| 3874 |
+
"step": 551
|
| 3875 |
+
},
|
| 3876 |
+
{
|
| 3877 |
+
"epoch": 0.4412470023980815,
|
| 3878 |
+
"grad_norm": 0.39483803510665894,
|
| 3879 |
+
"learning_rate": 0.00011974174505822677,
|
| 3880 |
+
"loss": 0.9856,
|
| 3881 |
+
"step": 552
|
| 3882 |
+
},
|
| 3883 |
+
{
|
| 3884 |
+
"epoch": 0.44204636290967225,
|
| 3885 |
+
"grad_norm": 0.409810870885849,
|
| 3886 |
+
"learning_rate": 0.00011949351408968439,
|
| 3887 |
+
"loss": 1.0131,
|
| 3888 |
+
"step": 553
|
| 3889 |
+
},
|
| 3890 |
+
{
|
| 3891 |
+
"epoch": 0.442845723421263,
|
| 3892 |
+
"grad_norm": 0.3824607729911804,
|
| 3893 |
+
"learning_rate": 0.000119245158197083,
|
| 3894 |
+
"loss": 1.0047,
|
| 3895 |
+
"step": 554
|
| 3896 |
+
},
|
| 3897 |
+
{
|
| 3898 |
+
"epoch": 0.44364508393285373,
|
| 3899 |
+
"grad_norm": 0.37841466069221497,
|
| 3900 |
+
"learning_rate": 0.00011899667897200977,
|
| 3901 |
+
"loss": 1.0711,
|
| 3902 |
+
"step": 555
|
| 3903 |
+
},
|
| 3904 |
+
{
|
| 3905 |
+
"epoch": 0.4444444444444444,
|
| 3906 |
+
"grad_norm": 0.3607119023799896,
|
| 3907 |
+
"learning_rate": 0.00011874807800684215,
|
| 3908 |
+
"loss": 1.0795,
|
| 3909 |
+
"step": 556
|
| 3910 |
+
},
|
| 3911 |
+
{
|
| 3912 |
+
"epoch": 0.44524380495603516,
|
| 3913 |
+
"grad_norm": 0.3559788465499878,
|
| 3914 |
+
"learning_rate": 0.00011849935689473786,
|
| 3915 |
+
"loss": 1.2118,
|
| 3916 |
+
"step": 557
|
| 3917 |
+
},
|
| 3918 |
+
{
|
| 3919 |
+
"epoch": 0.4460431654676259,
|
| 3920 |
+
"grad_norm": 0.678111732006073,
|
| 3921 |
+
"learning_rate": 0.00011825051722962453,
|
| 3922 |
+
"loss": 1.2167,
|
| 3923 |
+
"step": 558
|
| 3924 |
+
},
|
| 3925 |
+
{
|
| 3926 |
+
"epoch": 0.44684252597921664,
|
| 3927 |
+
"grad_norm": 0.3986605107784271,
|
| 3928 |
+
"learning_rate": 0.00011800156060618955,
|
| 3929 |
+
"loss": 1.1445,
|
| 3930 |
+
"step": 559
|
| 3931 |
+
},
|
| 3932 |
+
{
|
| 3933 |
+
"epoch": 0.44764188649080733,
|
| 3934 |
+
"grad_norm": 0.36188986897468567,
|
| 3935 |
+
"learning_rate": 0.00011775248861986978,
|
| 3936 |
+
"loss": 1.0736,
|
| 3937 |
+
"step": 560
|
| 3938 |
+
},
|
| 3939 |
+
{
|
| 3940 |
+
"epoch": 0.44844124700239807,
|
| 3941 |
+
"grad_norm": 0.33992865681648254,
|
| 3942 |
+
"learning_rate": 0.00011750330286684153,
|
| 3943 |
+
"loss": 1.1707,
|
| 3944 |
+
"step": 561
|
| 3945 |
+
},
|
| 3946 |
+
{
|
| 3947 |
+
"epoch": 0.4492406075139888,
|
| 3948 |
+
"grad_norm": 0.35370391607284546,
|
| 3949 |
+
"learning_rate": 0.00011725400494401,
|
| 3950 |
+
"loss": 1.1023,
|
| 3951 |
+
"step": 562
|
| 3952 |
+
},
|
| 3953 |
+
{
|
| 3954 |
+
"epoch": 0.45003996802557955,
|
| 3955 |
+
"grad_norm": 0.3561865985393524,
|
| 3956 |
+
"learning_rate": 0.00011700459644899937,
|
| 3957 |
+
"loss": 1.2305,
|
| 3958 |
+
"step": 563
|
| 3959 |
+
},
|
| 3960 |
+
{
|
| 3961 |
+
"epoch": 0.45083932853717024,
|
| 3962 |
+
"grad_norm": 0.37104496359825134,
|
| 3963 |
+
"learning_rate": 0.00011675507898014235,
|
| 3964 |
+
"loss": 1.1114,
|
| 3965 |
+
"step": 564
|
| 3966 |
+
},
|
| 3967 |
+
{
|
| 3968 |
+
"epoch": 0.451638689048761,
|
| 3969 |
+
"grad_norm": 0.350696861743927,
|
| 3970 |
+
"learning_rate": 0.00011650545413647,
|
| 3971 |
+
"loss": 1.1613,
|
| 3972 |
+
"step": 565
|
| 3973 |
+
},
|
| 3974 |
+
{
|
| 3975 |
+
"epoch": 0.4524380495603517,
|
| 3976 |
+
"grad_norm": 0.38277414441108704,
|
| 3977 |
+
"learning_rate": 0.00011625572351770155,
|
| 3978 |
+
"loss": 1.1631,
|
| 3979 |
+
"step": 566
|
| 3980 |
+
},
|
| 3981 |
+
{
|
| 3982 |
+
"epoch": 0.45323741007194246,
|
| 3983 |
+
"grad_norm": 0.35308247804641724,
|
| 3984 |
+
"learning_rate": 0.00011600588872423405,
|
| 3985 |
+
"loss": 1.1086,
|
| 3986 |
+
"step": 567
|
| 3987 |
+
},
|
| 3988 |
+
{
|
| 3989 |
+
"epoch": 0.4540367705835332,
|
| 3990 |
+
"grad_norm": 0.3877826929092407,
|
| 3991 |
+
"learning_rate": 0.0001157559513571322,
|
| 3992 |
+
"loss": 1.1722,
|
| 3993 |
+
"step": 568
|
| 3994 |
+
},
|
| 3995 |
+
{
|
| 3996 |
+
"epoch": 0.4548361310951239,
|
| 3997 |
+
"grad_norm": 0.38781118392944336,
|
| 3998 |
+
"learning_rate": 0.00011550591301811788,
|
| 3999 |
+
"loss": 1.013,
|
| 4000 |
+
"step": 569
|
| 4001 |
+
},
|
| 4002 |
+
{
|
| 4003 |
+
"epoch": 0.4556354916067146,
|
| 4004 |
+
"grad_norm": 0.40038326382637024,
|
| 4005 |
+
"learning_rate": 0.00011525577530956029,
|
| 4006 |
+
"loss": 1.0764,
|
| 4007 |
+
"step": 570
|
| 4008 |
+
},
|
| 4009 |
+
{
|
| 4010 |
+
"epoch": 0.45643485211830537,
|
| 4011 |
+
"grad_norm": 0.3875632584095001,
|
| 4012 |
+
"learning_rate": 0.00011500553983446527,
|
| 4013 |
+
"loss": 1.1239,
|
| 4014 |
+
"step": 571
|
| 4015 |
+
},
|
| 4016 |
+
{
|
| 4017 |
+
"epoch": 0.4572342126298961,
|
| 4018 |
+
"grad_norm": 0.44367122650146484,
|
| 4019 |
+
"learning_rate": 0.00011475520819646526,
|
| 4020 |
+
"loss": 1.0437,
|
| 4021 |
+
"step": 572
|
| 4022 |
+
},
|
| 4023 |
+
{
|
| 4024 |
+
"epoch": 0.4580335731414868,
|
| 4025 |
+
"grad_norm": 0.5386861562728882,
|
| 4026 |
+
"learning_rate": 0.0001145047819998089,
|
| 4027 |
+
"loss": 0.9693,
|
| 4028 |
+
"step": 573
|
| 4029 |
+
},
|
| 4030 |
+
{
|
| 4031 |
+
"epoch": 0.45883293365307753,
|
| 4032 |
+
"grad_norm": 0.5589727163314819,
|
| 4033 |
+
"learning_rate": 0.00011425426284935091,
|
| 4034 |
+
"loss": 1.0766,
|
| 4035 |
+
"step": 574
|
| 4036 |
+
},
|
| 4037 |
+
{
|
| 4038 |
+
"epoch": 0.4596322941646683,
|
| 4039 |
+
"grad_norm": 0.48679694533348083,
|
| 4040 |
+
"learning_rate": 0.0001140036523505416,
|
| 4041 |
+
"loss": 1.1827,
|
| 4042 |
+
"step": 575
|
| 4043 |
+
},
|
| 4044 |
+
{
|
| 4045 |
+
"epoch": 0.460431654676259,
|
| 4046 |
+
"grad_norm": 0.3698740005493164,
|
| 4047 |
+
"learning_rate": 0.00011375295210941674,
|
| 4048 |
+
"loss": 1.0565,
|
| 4049 |
+
"step": 576
|
| 4050 |
+
},
|
| 4051 |
+
{
|
| 4052 |
+
"epoch": 0.4612310151878497,
|
| 4053 |
+
"grad_norm": 0.4788266718387604,
|
| 4054 |
+
"learning_rate": 0.00011350216373258722,
|
| 4055 |
+
"loss": 1.006,
|
| 4056 |
+
"step": 577
|
| 4057 |
+
},
|
| 4058 |
+
{
|
| 4059 |
+
"epoch": 0.46203037569944044,
|
| 4060 |
+
"grad_norm": 0.6378310322761536,
|
| 4061 |
+
"learning_rate": 0.00011325128882722875,
|
| 4062 |
+
"loss": 1.2113,
|
| 4063 |
+
"step": 578
|
| 4064 |
+
},
|
| 4065 |
+
{
|
| 4066 |
+
"epoch": 0.4628297362110312,
|
| 4067 |
+
"grad_norm": 0.5069262981414795,
|
| 4068 |
+
"learning_rate": 0.00011300032900107147,
|
| 4069 |
+
"loss": 0.9833,
|
| 4070 |
+
"step": 579
|
| 4071 |
+
},
|
| 4072 |
+
{
|
| 4073 |
+
"epoch": 0.4636290967226219,
|
| 4074 |
+
"grad_norm": 0.44473326206207275,
|
| 4075 |
+
"learning_rate": 0.00011274928586238986,
|
| 4076 |
+
"loss": 1.0615,
|
| 4077 |
+
"step": 580
|
| 4078 |
+
},
|
| 4079 |
+
{
|
| 4080 |
+
"epoch": 0.4644284572342126,
|
| 4081 |
+
"grad_norm": 0.5024760961532593,
|
| 4082 |
+
"learning_rate": 0.00011249816101999227,
|
| 4083 |
+
"loss": 0.9796,
|
| 4084 |
+
"step": 581
|
| 4085 |
+
},
|
| 4086 |
+
{
|
| 4087 |
+
"epoch": 0.46522781774580335,
|
| 4088 |
+
"grad_norm": 0.537169337272644,
|
| 4089 |
+
"learning_rate": 0.00011224695608321056,
|
| 4090 |
+
"loss": 1.0713,
|
| 4091 |
+
"step": 582
|
| 4092 |
+
},
|
| 4093 |
+
{
|
| 4094 |
+
"epoch": 0.4660271782573941,
|
| 4095 |
+
"grad_norm": 0.6242894530296326,
|
| 4096 |
+
"learning_rate": 0.00011199567266188999,
|
| 4097 |
+
"loss": 1.3315,
|
| 4098 |
+
"step": 583
|
| 4099 |
+
},
|
| 4100 |
+
{
|
| 4101 |
+
"epoch": 0.46682653876898483,
|
| 4102 |
+
"grad_norm": 0.6122602820396423,
|
| 4103 |
+
"learning_rate": 0.00011174431236637874,
|
| 4104 |
+
"loss": 1.0745,
|
| 4105 |
+
"step": 584
|
| 4106 |
+
},
|
| 4107 |
+
{
|
| 4108 |
+
"epoch": 0.4676258992805755,
|
| 4109 |
+
"grad_norm": 0.5972760319709778,
|
| 4110 |
+
"learning_rate": 0.00011149287680751763,
|
| 4111 |
+
"loss": 1.1155,
|
| 4112 |
+
"step": 585
|
| 4113 |
+
},
|
| 4114 |
+
{
|
| 4115 |
+
"epoch": 0.46842525979216626,
|
| 4116 |
+
"grad_norm": 0.6527268290519714,
|
| 4117 |
+
"learning_rate": 0.0001112413675966298,
|
| 4118 |
+
"loss": 1.0451,
|
| 4119 |
+
"step": 586
|
| 4120 |
+
},
|
| 4121 |
+
{
|
| 4122 |
+
"epoch": 0.469224620303757,
|
| 4123 |
+
"grad_norm": 0.662009060382843,
|
| 4124 |
+
"learning_rate": 0.0001109897863455104,
|
| 4125 |
+
"loss": 1.1162,
|
| 4126 |
+
"step": 587
|
| 4127 |
+
},
|
| 4128 |
+
{
|
| 4129 |
+
"epoch": 0.47002398081534774,
|
| 4130 |
+
"grad_norm": 0.7552705407142639,
|
| 4131 |
+
"learning_rate": 0.00011073813466641632,
|
| 4132 |
+
"loss": 0.9946,
|
| 4133 |
+
"step": 588
|
| 4134 |
+
},
|
| 4135 |
+
{
|
| 4136 |
+
"epoch": 0.4708233413269384,
|
| 4137 |
+
"grad_norm": 0.7745950818061829,
|
| 4138 |
+
"learning_rate": 0.00011048641417205563,
|
| 4139 |
+
"loss": 1.1047,
|
| 4140 |
+
"step": 589
|
| 4141 |
+
},
|
| 4142 |
+
{
|
| 4143 |
+
"epoch": 0.47162270183852917,
|
| 4144 |
+
"grad_norm": 0.6825327277183533,
|
| 4145 |
+
"learning_rate": 0.00011023462647557751,
|
| 4146 |
+
"loss": 1.0597,
|
| 4147 |
+
"step": 590
|
| 4148 |
+
},
|
| 4149 |
+
{
|
| 4150 |
+
"epoch": 0.4724220623501199,
|
| 4151 |
+
"grad_norm": 0.9221934080123901,
|
| 4152 |
+
"learning_rate": 0.0001099827731905618,
|
| 4153 |
+
"loss": 1.0257,
|
| 4154 |
+
"step": 591
|
| 4155 |
+
},
|
| 4156 |
+
{
|
| 4157 |
+
"epoch": 0.47322142286171065,
|
| 4158 |
+
"grad_norm": 0.8638830184936523,
|
| 4159 |
+
"learning_rate": 0.0001097308559310087,
|
| 4160 |
+
"loss": 1.0615,
|
| 4161 |
+
"step": 592
|
| 4162 |
+
},
|
| 4163 |
+
{
|
| 4164 |
+
"epoch": 0.47402078337330134,
|
| 4165 |
+
"grad_norm": 0.921091616153717,
|
| 4166 |
+
"learning_rate": 0.00010947887631132823,
|
| 4167 |
+
"loss": 1.0958,
|
| 4168 |
+
"step": 593
|
| 4169 |
+
},
|
| 4170 |
+
{
|
| 4171 |
+
"epoch": 0.4748201438848921,
|
| 4172 |
+
"grad_norm": 0.9311054348945618,
|
| 4173 |
+
"learning_rate": 0.00010922683594633021,
|
| 4174 |
+
"loss": 1.1697,
|
| 4175 |
+
"step": 594
|
| 4176 |
+
},
|
| 4177 |
+
{
|
| 4178 |
+
"epoch": 0.4756195043964828,
|
| 4179 |
+
"grad_norm": 1.0105490684509277,
|
| 4180 |
+
"learning_rate": 0.00010897473645121372,
|
| 4181 |
+
"loss": 1.1954,
|
| 4182 |
+
"step": 595
|
| 4183 |
+
},
|
| 4184 |
+
{
|
| 4185 |
+
"epoch": 0.47641886490807356,
|
| 4186 |
+
"grad_norm": 0.9977965354919434,
|
| 4187 |
+
"learning_rate": 0.00010872257944155669,
|
| 4188 |
+
"loss": 1.1052,
|
| 4189 |
+
"step": 596
|
| 4190 |
+
},
|
| 4191 |
+
{
|
| 4192 |
+
"epoch": 0.47721822541966424,
|
| 4193 |
+
"grad_norm": 0.8218235969543457,
|
| 4194 |
+
"learning_rate": 0.00010847036653330568,
|
| 4195 |
+
"loss": 1.0016,
|
| 4196 |
+
"step": 597
|
| 4197 |
+
},
|
| 4198 |
+
{
|
| 4199 |
+
"epoch": 0.478017585931255,
|
| 4200 |
+
"grad_norm": 1.238542914390564,
|
| 4201 |
+
"learning_rate": 0.00010821809934276554,
|
| 4202 |
+
"loss": 1.1577,
|
| 4203 |
+
"step": 598
|
| 4204 |
+
},
|
| 4205 |
+
{
|
| 4206 |
+
"epoch": 0.4788169464428457,
|
| 4207 |
+
"grad_norm": 1.1989014148712158,
|
| 4208 |
+
"learning_rate": 0.00010796577948658892,
|
| 4209 |
+
"loss": 1.0534,
|
| 4210 |
+
"step": 599
|
| 4211 |
+
},
|
| 4212 |
+
{
|
| 4213 |
+
"epoch": 0.47961630695443647,
|
| 4214 |
+
"grad_norm": 1.591269612312317,
|
| 4215 |
+
"learning_rate": 0.00010771340858176594,
|
| 4216 |
+
"loss": 1.3851,
|
| 4217 |
+
"step": 600
|
| 4218 |
+
},
|
| 4219 |
+
{
|
| 4220 |
+
"epoch": 0.48041566746602715,
|
| 4221 |
+
"grad_norm": 0.3721429109573364,
|
| 4222 |
+
"learning_rate": 0.00010746098824561397,
|
| 4223 |
+
"loss": 0.9508,
|
| 4224 |
+
"step": 601
|
| 4225 |
+
},
|
| 4226 |
+
{
|
| 4227 |
+
"epoch": 0.4812150279776179,
|
| 4228 |
+
"grad_norm": 0.40401971340179443,
|
| 4229 |
+
"learning_rate": 0.00010720852009576705,
|
| 4230 |
+
"loss": 0.8829,
|
| 4231 |
+
"step": 602
|
| 4232 |
+
},
|
| 4233 |
+
{
|
| 4234 |
+
"epoch": 0.48201438848920863,
|
| 4235 |
+
"grad_norm": 0.4219840466976166,
|
| 4236 |
+
"learning_rate": 0.00010695600575016571,
|
| 4237 |
+
"loss": 0.9494,
|
| 4238 |
+
"step": 603
|
| 4239 |
+
},
|
| 4240 |
+
{
|
| 4241 |
+
"epoch": 0.4828137490007994,
|
| 4242 |
+
"grad_norm": 0.3684583604335785,
|
| 4243 |
+
"learning_rate": 0.0001067034468270465,
|
| 4244 |
+
"loss": 0.9306,
|
| 4245 |
+
"step": 604
|
| 4246 |
+
},
|
| 4247 |
+
{
|
| 4248 |
+
"epoch": 0.48361310951239006,
|
| 4249 |
+
"grad_norm": 0.3637100160121918,
|
| 4250 |
+
"learning_rate": 0.00010645084494493165,
|
| 4251 |
+
"loss": 1.1087,
|
| 4252 |
+
"step": 605
|
| 4253 |
+
},
|
| 4254 |
+
{
|
| 4255 |
+
"epoch": 0.4844124700239808,
|
| 4256 |
+
"grad_norm": 0.4035799503326416,
|
| 4257 |
+
"learning_rate": 0.00010619820172261867,
|
| 4258 |
+
"loss": 1.3074,
|
| 4259 |
+
"step": 606
|
| 4260 |
+
},
|
| 4261 |
+
{
|
| 4262 |
+
"epoch": 0.48521183053557154,
|
| 4263 |
+
"grad_norm": 0.39611637592315674,
|
| 4264 |
+
"learning_rate": 0.00010594551877917003,
|
| 4265 |
+
"loss": 0.9484,
|
| 4266 |
+
"step": 607
|
| 4267 |
+
},
|
| 4268 |
+
{
|
| 4269 |
+
"epoch": 0.4860111910471623,
|
| 4270 |
+
"grad_norm": 0.45287951827049255,
|
| 4271 |
+
"learning_rate": 0.00010569279773390273,
|
| 4272 |
+
"loss": 1.0403,
|
| 4273 |
+
"step": 608
|
| 4274 |
+
},
|
| 4275 |
+
{
|
| 4276 |
+
"epoch": 0.486810551558753,
|
| 4277 |
+
"grad_norm": 0.3595340847969055,
|
| 4278 |
+
"learning_rate": 0.00010544004020637799,
|
| 4279 |
+
"loss": 1.0235,
|
| 4280 |
+
"step": 609
|
| 4281 |
+
},
|
| 4282 |
+
{
|
| 4283 |
+
"epoch": 0.4876099120703437,
|
| 4284 |
+
"grad_norm": 0.3542434275150299,
|
| 4285 |
+
"learning_rate": 0.00010518724781639073,
|
| 4286 |
+
"loss": 1.2554,
|
| 4287 |
+
"step": 610
|
| 4288 |
+
},
|
| 4289 |
+
{
|
| 4290 |
+
"epoch": 0.48840927258193445,
|
| 4291 |
+
"grad_norm": 0.35258346796035767,
|
| 4292 |
+
"learning_rate": 0.00010493442218395936,
|
| 4293 |
+
"loss": 1.1231,
|
| 4294 |
+
"step": 611
|
| 4295 |
+
},
|
| 4296 |
+
{
|
| 4297 |
+
"epoch": 0.4892086330935252,
|
| 4298 |
+
"grad_norm": 0.3523050844669342,
|
| 4299 |
+
"learning_rate": 0.0001046815649293154,
|
| 4300 |
+
"loss": 1.0851,
|
| 4301 |
+
"step": 612
|
| 4302 |
+
},
|
| 4303 |
+
{
|
| 4304 |
+
"epoch": 0.49000799360511593,
|
| 4305 |
+
"grad_norm": 0.336111456155777,
|
| 4306 |
+
"learning_rate": 0.00010442867767289281,
|
| 4307 |
+
"loss": 1.0403,
|
| 4308 |
+
"step": 613
|
| 4309 |
+
},
|
| 4310 |
+
{
|
| 4311 |
+
"epoch": 0.4908073541167066,
|
| 4312 |
+
"grad_norm": 0.3554600775241852,
|
| 4313 |
+
"learning_rate": 0.000104175762035318,
|
| 4314 |
+
"loss": 1.1551,
|
| 4315 |
+
"step": 614
|
| 4316 |
+
},
|
| 4317 |
+
{
|
| 4318 |
+
"epoch": 0.49160671462829736,
|
| 4319 |
+
"grad_norm": 0.35645225644111633,
|
| 4320 |
+
"learning_rate": 0.00010392281963739918,
|
| 4321 |
+
"loss": 1.1704,
|
| 4322 |
+
"step": 615
|
| 4323 |
+
},
|
| 4324 |
+
{
|
| 4325 |
+
"epoch": 0.4924060751398881,
|
| 4326 |
+
"grad_norm": 0.4191342890262604,
|
| 4327 |
+
"learning_rate": 0.0001036698521001161,
|
| 4328 |
+
"loss": 1.0728,
|
| 4329 |
+
"step": 616
|
| 4330 |
+
},
|
| 4331 |
+
{
|
| 4332 |
+
"epoch": 0.49320543565147884,
|
| 4333 |
+
"grad_norm": 0.3663305342197418,
|
| 4334 |
+
"learning_rate": 0.00010341686104460954,
|
| 4335 |
+
"loss": 1.0764,
|
| 4336 |
+
"step": 617
|
| 4337 |
+
},
|
| 4338 |
+
{
|
| 4339 |
+
"epoch": 0.4940047961630695,
|
| 4340 |
+
"grad_norm": 0.3945816159248352,
|
| 4341 |
+
"learning_rate": 0.00010316384809217108,
|
| 4342 |
+
"loss": 1.0224,
|
| 4343 |
+
"step": 618
|
| 4344 |
+
},
|
| 4345 |
+
{
|
| 4346 |
+
"epoch": 0.49480415667466027,
|
| 4347 |
+
"grad_norm": 0.39029353857040405,
|
| 4348 |
+
"learning_rate": 0.0001029108148642326,
|
| 4349 |
+
"loss": 1.0606,
|
| 4350 |
+
"step": 619
|
| 4351 |
+
},
|
| 4352 |
+
{
|
| 4353 |
+
"epoch": 0.495603517186251,
|
| 4354 |
+
"grad_norm": 0.3495793342590332,
|
| 4355 |
+
"learning_rate": 0.00010265776298235586,
|
| 4356 |
+
"loss": 1.1024,
|
| 4357 |
+
"step": 620
|
| 4358 |
+
},
|
| 4359 |
+
{
|
| 4360 |
+
"epoch": 0.49640287769784175,
|
| 4361 |
+
"grad_norm": 0.3891185224056244,
|
| 4362 |
+
"learning_rate": 0.00010240469406822221,
|
| 4363 |
+
"loss": 0.9662,
|
| 4364 |
+
"step": 621
|
| 4365 |
+
},
|
| 4366 |
+
{
|
| 4367 |
+
"epoch": 0.49720223820943243,
|
| 4368 |
+
"grad_norm": 0.41916343569755554,
|
| 4369 |
+
"learning_rate": 0.00010215160974362223,
|
| 4370 |
+
"loss": 1.117,
|
| 4371 |
+
"step": 622
|
| 4372 |
+
},
|
| 4373 |
+
{
|
| 4374 |
+
"epoch": 0.4980015987210232,
|
| 4375 |
+
"grad_norm": 0.44954174757003784,
|
| 4376 |
+
"learning_rate": 0.0001018985116304451,
|
| 4377 |
+
"loss": 0.9685,
|
| 4378 |
+
"step": 623
|
| 4379 |
+
},
|
| 4380 |
+
{
|
| 4381 |
+
"epoch": 0.4988009592326139,
|
| 4382 |
+
"grad_norm": 0.40312501788139343,
|
| 4383 |
+
"learning_rate": 0.00010164540135066845,
|
| 4384 |
+
"loss": 1.022,
|
| 4385 |
+
"step": 624
|
| 4386 |
+
},
|
| 4387 |
+
{
|
| 4388 |
+
"epoch": 0.49960031974420466,
|
| 4389 |
+
"grad_norm": 0.4184950590133667,
|
| 4390 |
+
"learning_rate": 0.00010139228052634789,
|
| 4391 |
+
"loss": 1.0082,
|
| 4392 |
+
"step": 625
|
| 4393 |
+
},
|
| 4394 |
+
{
|
| 4395 |
+
"epoch": 0.5003996802557954,
|
| 4396 |
+
"grad_norm": 0.3949190676212311,
|
| 4397 |
+
"learning_rate": 0.00010113915077960658,
|
| 4398 |
+
"loss": 1.1159,
|
| 4399 |
+
"step": 626
|
| 4400 |
+
},
|
| 4401 |
+
{
|
| 4402 |
+
"epoch": 0.5003996802557954,
|
| 4403 |
+
"eval_loss": 1.0933643579483032,
|
| 4404 |
+
"eval_runtime": 39.8289,
|
| 4405 |
+
"eval_samples_per_second": 13.232,
|
| 4406 |
+
"eval_steps_per_second": 6.628,
|
| 4407 |
+
"step": 626
|
| 4408 |
}
|
| 4409 |
],
|
| 4410 |
"logging_steps": 1,
|
|
|
|
| 4424 |
"attributes": {}
|
| 4425 |
}
|
| 4426 |
},
|
| 4427 |
+
"total_flos": 2.738333182674862e+17,
|
| 4428 |
"train_batch_size": 2,
|
| 4429 |
"trial_name": null,
|
| 4430 |
"trial_params": null
|