Training in progress, step 10000, checkpoint
Browse files
last-checkpoint/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 223144592
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f1172332b4f4cf2ce6af9c26f710924116b9939c47b97977e6e6ac30b8704561
|
| 3 |
size 223144592
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 440235130
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:924f3d098447b1c44810fa13bd846863671b54c81e144bf0b30111f0ccdbc2f9
|
| 3 |
size 440235130
|
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 sha256:6b83c8e2618548ad0a6cb55d5f9e419ee82c921eaf0cb178e6e4f4700b370bc6
|
| 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:fa6158b7f4e5c150fe6ade940825c51c282bf26d832e5e702ff0db258a477b79
|
| 3 |
size 1064
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": 95.4926921263555,
|
| 3 |
"best_model_checkpoint": "./iteboshi_temp/checkpoint-4000",
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -1167,6 +1167,1746 @@
|
|
| 1167 |
"eval_steps_per_second": 1.188,
|
| 1168 |
"eval_wer": 95.4926921263555,
|
| 1169 |
"step": 4000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1170 |
}
|
| 1171 |
],
|
| 1172 |
"logging_steps": 25,
|
|
@@ -1181,12 +2921,12 @@
|
|
| 1181 |
"should_evaluate": false,
|
| 1182 |
"should_log": false,
|
| 1183 |
"should_save": true,
|
| 1184 |
-
"should_training_stop":
|
| 1185 |
},
|
| 1186 |
"attributes": {}
|
| 1187 |
}
|
| 1188 |
},
|
| 1189 |
-
"total_flos":
|
| 1190 |
"train_batch_size": 12,
|
| 1191 |
"trial_name": null,
|
| 1192 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": 95.4926921263555,
|
| 3 |
"best_model_checkpoint": "./iteboshi_temp/checkpoint-4000",
|
| 4 |
+
"epoch": 16.50206440957886,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 10000,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1167 |
"eval_steps_per_second": 1.188,
|
| 1168 |
"eval_wer": 95.4926921263555,
|
| 1169 |
"step": 4000
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 6.642444260941371,
|
| 1173 |
+
"grad_norm": 7.094240665435791,
|
| 1174 |
+
"learning_rate": 1.2581052631578949e-05,
|
| 1175 |
+
"loss": 0.5365,
|
| 1176 |
+
"step": 4025
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 6.68373245251858,
|
| 1180 |
+
"grad_norm": 8.255016326904297,
|
| 1181 |
+
"learning_rate": 1.252842105263158e-05,
|
| 1182 |
+
"loss": 0.5768,
|
| 1183 |
+
"step": 4050
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 6.725020644095789,
|
| 1187 |
+
"grad_norm": 7.434171676635742,
|
| 1188 |
+
"learning_rate": 1.2475789473684212e-05,
|
| 1189 |
+
"loss": 0.5892,
|
| 1190 |
+
"step": 4075
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 6.766308835672998,
|
| 1194 |
+
"grad_norm": 6.989074230194092,
|
| 1195 |
+
"learning_rate": 1.2423157894736844e-05,
|
| 1196 |
+
"loss": 0.5494,
|
| 1197 |
+
"step": 4100
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 6.807597027250207,
|
| 1201 |
+
"grad_norm": 6.808770656585693,
|
| 1202 |
+
"learning_rate": 1.2370526315789476e-05,
|
| 1203 |
+
"loss": 0.5281,
|
| 1204 |
+
"step": 4125
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 6.848885218827416,
|
| 1208 |
+
"grad_norm": 7.470989227294922,
|
| 1209 |
+
"learning_rate": 1.2317894736842107e-05,
|
| 1210 |
+
"loss": 0.5499,
|
| 1211 |
+
"step": 4150
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 6.890173410404624,
|
| 1215 |
+
"grad_norm": 7.557090759277344,
|
| 1216 |
+
"learning_rate": 1.2265263157894737e-05,
|
| 1217 |
+
"loss": 0.5413,
|
| 1218 |
+
"step": 4175
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 6.931461601981833,
|
| 1222 |
+
"grad_norm": 7.504119396209717,
|
| 1223 |
+
"learning_rate": 1.2212631578947369e-05,
|
| 1224 |
+
"loss": 0.513,
|
| 1225 |
+
"step": 4200
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 6.972749793559042,
|
| 1229 |
+
"grad_norm": 6.75087308883667,
|
| 1230 |
+
"learning_rate": 1.216e-05,
|
| 1231 |
+
"loss": 0.5343,
|
| 1232 |
+
"step": 4225
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 7.013212221304707,
|
| 1236 |
+
"grad_norm": 5.9934515953063965,
|
| 1237 |
+
"learning_rate": 1.2107368421052632e-05,
|
| 1238 |
+
"loss": 0.4879,
|
| 1239 |
+
"step": 4250
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 7.054500412881916,
|
| 1243 |
+
"grad_norm": 6.634411334991455,
|
| 1244 |
+
"learning_rate": 1.2054736842105264e-05,
|
| 1245 |
+
"loss": 0.4454,
|
| 1246 |
+
"step": 4275
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 7.095788604459124,
|
| 1250 |
+
"grad_norm": 7.614193916320801,
|
| 1251 |
+
"learning_rate": 1.2002105263157895e-05,
|
| 1252 |
+
"loss": 0.4392,
|
| 1253 |
+
"step": 4300
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 7.137076796036333,
|
| 1257 |
+
"grad_norm": 5.324835777282715,
|
| 1258 |
+
"learning_rate": 1.1949473684210527e-05,
|
| 1259 |
+
"loss": 0.456,
|
| 1260 |
+
"step": 4325
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 7.178364987613542,
|
| 1264 |
+
"grad_norm": 6.1424241065979,
|
| 1265 |
+
"learning_rate": 1.189684210526316e-05,
|
| 1266 |
+
"loss": 0.46,
|
| 1267 |
+
"step": 4350
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 7.219653179190751,
|
| 1271 |
+
"grad_norm": 7.1868767738342285,
|
| 1272 |
+
"learning_rate": 1.1844210526315792e-05,
|
| 1273 |
+
"loss": 0.4591,
|
| 1274 |
+
"step": 4375
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 7.26094137076796,
|
| 1278 |
+
"grad_norm": 6.41041374206543,
|
| 1279 |
+
"learning_rate": 1.1791578947368424e-05,
|
| 1280 |
+
"loss": 0.4618,
|
| 1281 |
+
"step": 4400
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 7.302229562345169,
|
| 1285 |
+
"grad_norm": 6.095521450042725,
|
| 1286 |
+
"learning_rate": 1.1738947368421054e-05,
|
| 1287 |
+
"loss": 0.4692,
|
| 1288 |
+
"step": 4425
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 7.343517753922378,
|
| 1292 |
+
"grad_norm": 6.523583889007568,
|
| 1293 |
+
"learning_rate": 1.1686315789473685e-05,
|
| 1294 |
+
"loss": 0.4631,
|
| 1295 |
+
"step": 4450
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 7.384805945499587,
|
| 1299 |
+
"grad_norm": 6.192616939544678,
|
| 1300 |
+
"learning_rate": 1.1633684210526317e-05,
|
| 1301 |
+
"loss": 0.4689,
|
| 1302 |
+
"step": 4475
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 7.426094137076796,
|
| 1306 |
+
"grad_norm": 5.833901882171631,
|
| 1307 |
+
"learning_rate": 1.1581052631578948e-05,
|
| 1308 |
+
"loss": 0.4472,
|
| 1309 |
+
"step": 4500
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 7.467382328654005,
|
| 1313 |
+
"grad_norm": 7.350675582885742,
|
| 1314 |
+
"learning_rate": 1.152842105263158e-05,
|
| 1315 |
+
"loss": 0.472,
|
| 1316 |
+
"step": 4525
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 7.508670520231214,
|
| 1320 |
+
"grad_norm": 6.825387477874756,
|
| 1321 |
+
"learning_rate": 1.1475789473684212e-05,
|
| 1322 |
+
"loss": 0.4773,
|
| 1323 |
+
"step": 4550
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 7.549958711808423,
|
| 1327 |
+
"grad_norm": 7.2338762283325195,
|
| 1328 |
+
"learning_rate": 1.1423157894736843e-05,
|
| 1329 |
+
"loss": 0.4515,
|
| 1330 |
+
"step": 4575
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 7.591246903385632,
|
| 1334 |
+
"grad_norm": 7.510733604431152,
|
| 1335 |
+
"learning_rate": 1.1370526315789473e-05,
|
| 1336 |
+
"loss": 0.4676,
|
| 1337 |
+
"step": 4600
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 7.632535094962841,
|
| 1341 |
+
"grad_norm": 6.926485538482666,
|
| 1342 |
+
"learning_rate": 1.1317894736842105e-05,
|
| 1343 |
+
"loss": 0.4692,
|
| 1344 |
+
"step": 4625
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 7.673823286540049,
|
| 1348 |
+
"grad_norm": 7.200405120849609,
|
| 1349 |
+
"learning_rate": 1.1265263157894737e-05,
|
| 1350 |
+
"loss": 0.4874,
|
| 1351 |
+
"step": 4650
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 7.715111478117258,
|
| 1355 |
+
"grad_norm": 7.665420055389404,
|
| 1356 |
+
"learning_rate": 1.1212631578947368e-05,
|
| 1357 |
+
"loss": 0.4515,
|
| 1358 |
+
"step": 4675
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 7.756399669694467,
|
| 1362 |
+
"grad_norm": 6.377760410308838,
|
| 1363 |
+
"learning_rate": 1.1160000000000002e-05,
|
| 1364 |
+
"loss": 0.4639,
|
| 1365 |
+
"step": 4700
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 7.797687861271676,
|
| 1369 |
+
"grad_norm": 7.134575366973877,
|
| 1370 |
+
"learning_rate": 1.1107368421052633e-05,
|
| 1371 |
+
"loss": 0.4543,
|
| 1372 |
+
"step": 4725
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 7.838976052848885,
|
| 1376 |
+
"grad_norm": 6.138679027557373,
|
| 1377 |
+
"learning_rate": 1.1054736842105265e-05,
|
| 1378 |
+
"loss": 0.4276,
|
| 1379 |
+
"step": 4750
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 7.880264244426094,
|
| 1383 |
+
"grad_norm": 6.741207122802734,
|
| 1384 |
+
"learning_rate": 1.1002105263157896e-05,
|
| 1385 |
+
"loss": 0.4975,
|
| 1386 |
+
"step": 4775
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 7.921552436003303,
|
| 1390 |
+
"grad_norm": 6.514609336853027,
|
| 1391 |
+
"learning_rate": 1.0949473684210528e-05,
|
| 1392 |
+
"loss": 0.4841,
|
| 1393 |
+
"step": 4800
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 7.962840627580512,
|
| 1397 |
+
"grad_norm": 6.2714009284973145,
|
| 1398 |
+
"learning_rate": 1.089684210526316e-05,
|
| 1399 |
+
"loss": 0.4658,
|
| 1400 |
+
"step": 4825
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 8.003303055326176,
|
| 1404 |
+
"grad_norm": 7.009129524230957,
|
| 1405 |
+
"learning_rate": 1.084421052631579e-05,
|
| 1406 |
+
"loss": 0.4919,
|
| 1407 |
+
"step": 4850
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 8.044591246903385,
|
| 1411 |
+
"grad_norm": 6.334277153015137,
|
| 1412 |
+
"learning_rate": 1.0791578947368421e-05,
|
| 1413 |
+
"loss": 0.3873,
|
| 1414 |
+
"step": 4875
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 8.085879438480594,
|
| 1418 |
+
"grad_norm": 6.403853416442871,
|
| 1419 |
+
"learning_rate": 1.0738947368421053e-05,
|
| 1420 |
+
"loss": 0.3961,
|
| 1421 |
+
"step": 4900
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 8.127167630057803,
|
| 1425 |
+
"grad_norm": 6.143258571624756,
|
| 1426 |
+
"learning_rate": 1.0686315789473685e-05,
|
| 1427 |
+
"loss": 0.3699,
|
| 1428 |
+
"step": 4925
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 8.168455821635012,
|
| 1432 |
+
"grad_norm": 6.0696539878845215,
|
| 1433 |
+
"learning_rate": 1.0633684210526316e-05,
|
| 1434 |
+
"loss": 0.3858,
|
| 1435 |
+
"step": 4950
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 8.209744013212221,
|
| 1439 |
+
"grad_norm": 6.363874435424805,
|
| 1440 |
+
"learning_rate": 1.0581052631578948e-05,
|
| 1441 |
+
"loss": 0.392,
|
| 1442 |
+
"step": 4975
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 8.25103220478943,
|
| 1446 |
+
"grad_norm": 5.990923881530762,
|
| 1447 |
+
"learning_rate": 1.052842105263158e-05,
|
| 1448 |
+
"loss": 0.4182,
|
| 1449 |
+
"step": 5000
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 8.25103220478943,
|
| 1453 |
+
"eval_cer": 47.292882280788234,
|
| 1454 |
+
"eval_loss": 1.0573757886886597,
|
| 1455 |
+
"eval_runtime": 732.0504,
|
| 1456 |
+
"eval_samples_per_second": 14.454,
|
| 1457 |
+
"eval_steps_per_second": 1.205,
|
| 1458 |
+
"eval_wer": 96.23762376237623,
|
| 1459 |
+
"step": 5000
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 8.29232039636664,
|
| 1463 |
+
"grad_norm": 5.979728698730469,
|
| 1464 |
+
"learning_rate": 1.047578947368421e-05,
|
| 1465 |
+
"loss": 0.38,
|
| 1466 |
+
"step": 5025
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 8.333608587943848,
|
| 1470 |
+
"grad_norm": 6.235442161560059,
|
| 1471 |
+
"learning_rate": 1.0423157894736844e-05,
|
| 1472 |
+
"loss": 0.3823,
|
| 1473 |
+
"step": 5050
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 8.374896779521057,
|
| 1477 |
+
"grad_norm": 6.0435471534729,
|
| 1478 |
+
"learning_rate": 1.0370526315789476e-05,
|
| 1479 |
+
"loss": 0.3889,
|
| 1480 |
+
"step": 5075
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 8.416184971098266,
|
| 1484 |
+
"grad_norm": 5.2331624031066895,
|
| 1485 |
+
"learning_rate": 1.0317894736842106e-05,
|
| 1486 |
+
"loss": 0.4002,
|
| 1487 |
+
"step": 5100
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 8.457473162675475,
|
| 1491 |
+
"grad_norm": 5.645880699157715,
|
| 1492 |
+
"learning_rate": 1.0265263157894738e-05,
|
| 1493 |
+
"loss": 0.395,
|
| 1494 |
+
"step": 5125
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 8.498761354252684,
|
| 1498 |
+
"grad_norm": 6.611478805541992,
|
| 1499 |
+
"learning_rate": 1.021263157894737e-05,
|
| 1500 |
+
"loss": 0.3964,
|
| 1501 |
+
"step": 5150
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 8.540049545829893,
|
| 1505 |
+
"grad_norm": 5.404575347900391,
|
| 1506 |
+
"learning_rate": 1.0160000000000001e-05,
|
| 1507 |
+
"loss": 0.3886,
|
| 1508 |
+
"step": 5175
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.581337737407102,
|
| 1512 |
+
"grad_norm": 7.283960819244385,
|
| 1513 |
+
"learning_rate": 1.0107368421052633e-05,
|
| 1514 |
+
"loss": 0.4193,
|
| 1515 |
+
"step": 5200
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 8.622625928984311,
|
| 1519 |
+
"grad_norm": 6.334190368652344,
|
| 1520 |
+
"learning_rate": 1.0054736842105264e-05,
|
| 1521 |
+
"loss": 0.4129,
|
| 1522 |
+
"step": 5225
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 8.66391412056152,
|
| 1526 |
+
"grad_norm": 6.392534255981445,
|
| 1527 |
+
"learning_rate": 1.0002105263157896e-05,
|
| 1528 |
+
"loss": 0.3832,
|
| 1529 |
+
"step": 5250
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 8.705202312138729,
|
| 1533 |
+
"grad_norm": 6.049492359161377,
|
| 1534 |
+
"learning_rate": 9.949473684210526e-06,
|
| 1535 |
+
"loss": 0.4008,
|
| 1536 |
+
"step": 5275
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 8.746490503715938,
|
| 1540 |
+
"grad_norm": 5.881133556365967,
|
| 1541 |
+
"learning_rate": 9.89684210526316e-06,
|
| 1542 |
+
"loss": 0.3927,
|
| 1543 |
+
"step": 5300
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 8.787778695293147,
|
| 1547 |
+
"grad_norm": 6.470844745635986,
|
| 1548 |
+
"learning_rate": 9.84421052631579e-06,
|
| 1549 |
+
"loss": 0.4036,
|
| 1550 |
+
"step": 5325
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 8.829066886870356,
|
| 1554 |
+
"grad_norm": 6.188573837280273,
|
| 1555 |
+
"learning_rate": 9.791578947368422e-06,
|
| 1556 |
+
"loss": 0.394,
|
| 1557 |
+
"step": 5350
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 8.870355078447563,
|
| 1561 |
+
"grad_norm": 6.691900730133057,
|
| 1562 |
+
"learning_rate": 9.738947368421054e-06,
|
| 1563 |
+
"loss": 0.399,
|
| 1564 |
+
"step": 5375
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 8.911643270024772,
|
| 1568 |
+
"grad_norm": 6.903393745422363,
|
| 1569 |
+
"learning_rate": 9.686315789473684e-06,
|
| 1570 |
+
"loss": 0.3999,
|
| 1571 |
+
"step": 5400
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 8.952931461601981,
|
| 1575 |
+
"grad_norm": 6.473091125488281,
|
| 1576 |
+
"learning_rate": 9.633684210526316e-06,
|
| 1577 |
+
"loss": 0.3833,
|
| 1578 |
+
"step": 5425
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 8.99421965317919,
|
| 1582 |
+
"grad_norm": 6.615601062774658,
|
| 1583 |
+
"learning_rate": 9.581052631578947e-06,
|
| 1584 |
+
"loss": 0.3961,
|
| 1585 |
+
"step": 5450
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 9.034682080924856,
|
| 1589 |
+
"grad_norm": 5.409454345703125,
|
| 1590 |
+
"learning_rate": 9.52842105263158e-06,
|
| 1591 |
+
"loss": 0.3184,
|
| 1592 |
+
"step": 5475
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 9.075970272502065,
|
| 1596 |
+
"grad_norm": 4.625083923339844,
|
| 1597 |
+
"learning_rate": 9.475789473684212e-06,
|
| 1598 |
+
"loss": 0.308,
|
| 1599 |
+
"step": 5500
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 9.117258464079274,
|
| 1603 |
+
"grad_norm": 6.347021579742432,
|
| 1604 |
+
"learning_rate": 9.423157894736842e-06,
|
| 1605 |
+
"loss": 0.3087,
|
| 1606 |
+
"step": 5525
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 9.158546655656481,
|
| 1610 |
+
"grad_norm": 6.036218166351318,
|
| 1611 |
+
"learning_rate": 9.370526315789474e-06,
|
| 1612 |
+
"loss": 0.3361,
|
| 1613 |
+
"step": 5550
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 9.19983484723369,
|
| 1617 |
+
"grad_norm": 6.974196910858154,
|
| 1618 |
+
"learning_rate": 9.317894736842105e-06,
|
| 1619 |
+
"loss": 0.3277,
|
| 1620 |
+
"step": 5575
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 9.2411230388109,
|
| 1624 |
+
"grad_norm": 6.37134313583374,
|
| 1625 |
+
"learning_rate": 9.265263157894737e-06,
|
| 1626 |
+
"loss": 0.3227,
|
| 1627 |
+
"step": 5600
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 9.282411230388108,
|
| 1631 |
+
"grad_norm": 5.647594928741455,
|
| 1632 |
+
"learning_rate": 9.21263157894737e-06,
|
| 1633 |
+
"loss": 0.3408,
|
| 1634 |
+
"step": 5625
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 9.323699421965317,
|
| 1638 |
+
"grad_norm": 5.690819263458252,
|
| 1639 |
+
"learning_rate": 9.16e-06,
|
| 1640 |
+
"loss": 0.3431,
|
| 1641 |
+
"step": 5650
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 9.364987613542526,
|
| 1645 |
+
"grad_norm": 6.394257545471191,
|
| 1646 |
+
"learning_rate": 9.107368421052632e-06,
|
| 1647 |
+
"loss": 0.3684,
|
| 1648 |
+
"step": 5675
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 9.406275805119735,
|
| 1652 |
+
"grad_norm": 5.489740371704102,
|
| 1653 |
+
"learning_rate": 9.054736842105264e-06,
|
| 1654 |
+
"loss": 0.3401,
|
| 1655 |
+
"step": 5700
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 9.447563996696944,
|
| 1659 |
+
"grad_norm": 5.952576637268066,
|
| 1660 |
+
"learning_rate": 9.002105263157895e-06,
|
| 1661 |
+
"loss": 0.3485,
|
| 1662 |
+
"step": 5725
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 9.488852188274153,
|
| 1666 |
+
"grad_norm": 5.653362274169922,
|
| 1667 |
+
"learning_rate": 8.949473684210527e-06,
|
| 1668 |
+
"loss": 0.349,
|
| 1669 |
+
"step": 5750
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 9.530140379851362,
|
| 1673 |
+
"grad_norm": 6.067948341369629,
|
| 1674 |
+
"learning_rate": 8.896842105263159e-06,
|
| 1675 |
+
"loss": 0.3629,
|
| 1676 |
+
"step": 5775
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 9.571428571428571,
|
| 1680 |
+
"grad_norm": 5.475063800811768,
|
| 1681 |
+
"learning_rate": 8.84421052631579e-06,
|
| 1682 |
+
"loss": 0.3207,
|
| 1683 |
+
"step": 5800
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 9.61271676300578,
|
| 1687 |
+
"grad_norm": 6.7783308029174805,
|
| 1688 |
+
"learning_rate": 8.791578947368422e-06,
|
| 1689 |
+
"loss": 0.3454,
|
| 1690 |
+
"step": 5825
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 9.65400495458299,
|
| 1694 |
+
"grad_norm": 5.636470317840576,
|
| 1695 |
+
"learning_rate": 8.738947368421053e-06,
|
| 1696 |
+
"loss": 0.3408,
|
| 1697 |
+
"step": 5850
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 9.695293146160198,
|
| 1701 |
+
"grad_norm": 5.290091037750244,
|
| 1702 |
+
"learning_rate": 8.686315789473685e-06,
|
| 1703 |
+
"loss": 0.3229,
|
| 1704 |
+
"step": 5875
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 9.736581337737407,
|
| 1708 |
+
"grad_norm": 5.795228481292725,
|
| 1709 |
+
"learning_rate": 8.633684210526317e-06,
|
| 1710 |
+
"loss": 0.36,
|
| 1711 |
+
"step": 5900
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 9.777869529314616,
|
| 1715 |
+
"grad_norm": 5.348466396331787,
|
| 1716 |
+
"learning_rate": 8.581052631578948e-06,
|
| 1717 |
+
"loss": 0.3358,
|
| 1718 |
+
"step": 5925
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 9.819157720891825,
|
| 1722 |
+
"grad_norm": 5.813995361328125,
|
| 1723 |
+
"learning_rate": 8.528421052631578e-06,
|
| 1724 |
+
"loss": 0.349,
|
| 1725 |
+
"step": 5950
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 9.860445912469034,
|
| 1729 |
+
"grad_norm": 5.806293964385986,
|
| 1730 |
+
"learning_rate": 8.475789473684212e-06,
|
| 1731 |
+
"loss": 0.3493,
|
| 1732 |
+
"step": 5975
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 9.901734104046243,
|
| 1736 |
+
"grad_norm": 6.348217487335205,
|
| 1737 |
+
"learning_rate": 8.423157894736843e-06,
|
| 1738 |
+
"loss": 0.3164,
|
| 1739 |
+
"step": 6000
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 9.901734104046243,
|
| 1743 |
+
"eval_cer": 49.44172503525042,
|
| 1744 |
+
"eval_loss": 1.0615819692611694,
|
| 1745 |
+
"eval_runtime": 712.2161,
|
| 1746 |
+
"eval_samples_per_second": 14.856,
|
| 1747 |
+
"eval_steps_per_second": 1.238,
|
| 1748 |
+
"eval_wer": 96.38849599245638,
|
| 1749 |
+
"step": 6000
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 9.943022295623452,
|
| 1753 |
+
"grad_norm": 5.729813575744629,
|
| 1754 |
+
"learning_rate": 8.370526315789475e-06,
|
| 1755 |
+
"loss": 0.3253,
|
| 1756 |
+
"step": 6025
|
| 1757 |
+
},
|
| 1758 |
+
{
|
| 1759 |
+
"epoch": 9.984310487200661,
|
| 1760 |
+
"grad_norm": 7.191168785095215,
|
| 1761 |
+
"learning_rate": 8.317894736842107e-06,
|
| 1762 |
+
"loss": 0.3433,
|
| 1763 |
+
"step": 6050
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 10.024772914946325,
|
| 1767 |
+
"grad_norm": 6.767040729522705,
|
| 1768 |
+
"learning_rate": 8.265263157894737e-06,
|
| 1769 |
+
"loss": 0.2875,
|
| 1770 |
+
"step": 6075
|
| 1771 |
+
},
|
| 1772 |
+
{
|
| 1773 |
+
"epoch": 10.066061106523534,
|
| 1774 |
+
"grad_norm": 5.048575401306152,
|
| 1775 |
+
"learning_rate": 8.212631578947368e-06,
|
| 1776 |
+
"loss": 0.2773,
|
| 1777 |
+
"step": 6100
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 10.107349298100743,
|
| 1781 |
+
"grad_norm": 4.836114883422852,
|
| 1782 |
+
"learning_rate": 8.16e-06,
|
| 1783 |
+
"loss": 0.2889,
|
| 1784 |
+
"step": 6125
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 10.148637489677952,
|
| 1788 |
+
"grad_norm": 4.830049514770508,
|
| 1789 |
+
"learning_rate": 8.107368421052633e-06,
|
| 1790 |
+
"loss": 0.3004,
|
| 1791 |
+
"step": 6150
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 10.189925681255161,
|
| 1795 |
+
"grad_norm": 5.718193054199219,
|
| 1796 |
+
"learning_rate": 8.054736842105265e-06,
|
| 1797 |
+
"loss": 0.2989,
|
| 1798 |
+
"step": 6175
|
| 1799 |
+
},
|
| 1800 |
+
{
|
| 1801 |
+
"epoch": 10.23121387283237,
|
| 1802 |
+
"grad_norm": 7.24569034576416,
|
| 1803 |
+
"learning_rate": 8.002105263157895e-06,
|
| 1804 |
+
"loss": 0.2712,
|
| 1805 |
+
"step": 6200
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 10.27250206440958,
|
| 1809 |
+
"grad_norm": 6.074246883392334,
|
| 1810 |
+
"learning_rate": 7.949473684210526e-06,
|
| 1811 |
+
"loss": 0.2899,
|
| 1812 |
+
"step": 6225
|
| 1813 |
+
},
|
| 1814 |
+
{
|
| 1815 |
+
"epoch": 10.313790255986788,
|
| 1816 |
+
"grad_norm": 5.600971698760986,
|
| 1817 |
+
"learning_rate": 7.896842105263158e-06,
|
| 1818 |
+
"loss": 0.3,
|
| 1819 |
+
"step": 6250
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 10.355078447563997,
|
| 1823 |
+
"grad_norm": 6.002727508544922,
|
| 1824 |
+
"learning_rate": 7.84421052631579e-06,
|
| 1825 |
+
"loss": 0.2895,
|
| 1826 |
+
"step": 6275
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 10.396366639141206,
|
| 1830 |
+
"grad_norm": 5.616031169891357,
|
| 1831 |
+
"learning_rate": 7.793684210526316e-06,
|
| 1832 |
+
"loss": 0.3123,
|
| 1833 |
+
"step": 6300
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 10.437654830718415,
|
| 1837 |
+
"grad_norm": 5.227489948272705,
|
| 1838 |
+
"learning_rate": 7.741052631578948e-06,
|
| 1839 |
+
"loss": 0.2796,
|
| 1840 |
+
"step": 6325
|
| 1841 |
+
},
|
| 1842 |
+
{
|
| 1843 |
+
"epoch": 10.478943022295624,
|
| 1844 |
+
"grad_norm": 5.6762261390686035,
|
| 1845 |
+
"learning_rate": 7.68842105263158e-06,
|
| 1846 |
+
"loss": 0.3061,
|
| 1847 |
+
"step": 6350
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 10.520231213872833,
|
| 1851 |
+
"grad_norm": 4.9320197105407715,
|
| 1852 |
+
"learning_rate": 7.635789473684211e-06,
|
| 1853 |
+
"loss": 0.2779,
|
| 1854 |
+
"step": 6375
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 10.561519405450042,
|
| 1858 |
+
"grad_norm": 5.518511772155762,
|
| 1859 |
+
"learning_rate": 7.583157894736842e-06,
|
| 1860 |
+
"loss": 0.2861,
|
| 1861 |
+
"step": 6400
|
| 1862 |
+
},
|
| 1863 |
+
{
|
| 1864 |
+
"epoch": 10.602807597027251,
|
| 1865 |
+
"grad_norm": 5.470175266265869,
|
| 1866 |
+
"learning_rate": 7.5305263157894745e-06,
|
| 1867 |
+
"loss": 0.279,
|
| 1868 |
+
"step": 6425
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 10.64409578860446,
|
| 1872 |
+
"grad_norm": 4.248904228210449,
|
| 1873 |
+
"learning_rate": 7.477894736842106e-06,
|
| 1874 |
+
"loss": 0.3054,
|
| 1875 |
+
"step": 6450
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 10.685383980181667,
|
| 1879 |
+
"grad_norm": 6.3203935623168945,
|
| 1880 |
+
"learning_rate": 7.425263157894738e-06,
|
| 1881 |
+
"loss": 0.3168,
|
| 1882 |
+
"step": 6475
|
| 1883 |
+
},
|
| 1884 |
+
{
|
| 1885 |
+
"epoch": 10.726672171758876,
|
| 1886 |
+
"grad_norm": 5.901068687438965,
|
| 1887 |
+
"learning_rate": 7.3726315789473694e-06,
|
| 1888 |
+
"loss": 0.3103,
|
| 1889 |
+
"step": 6500
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 10.767960363336085,
|
| 1893 |
+
"grad_norm": 6.379432678222656,
|
| 1894 |
+
"learning_rate": 7.32e-06,
|
| 1895 |
+
"loss": 0.2715,
|
| 1896 |
+
"step": 6525
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 10.809248554913294,
|
| 1900 |
+
"grad_norm": 5.802628040313721,
|
| 1901 |
+
"learning_rate": 7.267368421052632e-06,
|
| 1902 |
+
"loss": 0.2981,
|
| 1903 |
+
"step": 6550
|
| 1904 |
+
},
|
| 1905 |
+
{
|
| 1906 |
+
"epoch": 10.850536746490503,
|
| 1907 |
+
"grad_norm": 6.295346260070801,
|
| 1908 |
+
"learning_rate": 7.2147368421052635e-06,
|
| 1909 |
+
"loss": 0.2676,
|
| 1910 |
+
"step": 6575
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 10.891824938067712,
|
| 1914 |
+
"grad_norm": 5.411930561065674,
|
| 1915 |
+
"learning_rate": 7.162105263157896e-06,
|
| 1916 |
+
"loss": 0.2809,
|
| 1917 |
+
"step": 6600
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 10.933113129644921,
|
| 1921 |
+
"grad_norm": 5.577861309051514,
|
| 1922 |
+
"learning_rate": 7.109473684210528e-06,
|
| 1923 |
+
"loss": 0.2473,
|
| 1924 |
+
"step": 6625
|
| 1925 |
+
},
|
| 1926 |
+
{
|
| 1927 |
+
"epoch": 10.97440132122213,
|
| 1928 |
+
"grad_norm": 5.272998332977295,
|
| 1929 |
+
"learning_rate": 7.056842105263158e-06,
|
| 1930 |
+
"loss": 0.2931,
|
| 1931 |
+
"step": 6650
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 11.014863748967795,
|
| 1935 |
+
"grad_norm": 5.220921993255615,
|
| 1936 |
+
"learning_rate": 7.00421052631579e-06,
|
| 1937 |
+
"loss": 0.2772,
|
| 1938 |
+
"step": 6675
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 11.056151940545003,
|
| 1942 |
+
"grad_norm": 4.028062343597412,
|
| 1943 |
+
"learning_rate": 6.951578947368422e-06,
|
| 1944 |
+
"loss": 0.2511,
|
| 1945 |
+
"step": 6700
|
| 1946 |
+
},
|
| 1947 |
+
{
|
| 1948 |
+
"epoch": 11.097440132122212,
|
| 1949 |
+
"grad_norm": 4.835925579071045,
|
| 1950 |
+
"learning_rate": 6.8989473684210525e-06,
|
| 1951 |
+
"loss": 0.2502,
|
| 1952 |
+
"step": 6725
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 11.138728323699421,
|
| 1956 |
+
"grad_norm": 5.920368194580078,
|
| 1957 |
+
"learning_rate": 6.846315789473684e-06,
|
| 1958 |
+
"loss": 0.2438,
|
| 1959 |
+
"step": 6750
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 11.18001651527663,
|
| 1963 |
+
"grad_norm": 5.114356517791748,
|
| 1964 |
+
"learning_rate": 6.793684210526317e-06,
|
| 1965 |
+
"loss": 0.2478,
|
| 1966 |
+
"step": 6775
|
| 1967 |
+
},
|
| 1968 |
+
{
|
| 1969 |
+
"epoch": 11.22130470685384,
|
| 1970 |
+
"grad_norm": 4.871094703674316,
|
| 1971 |
+
"learning_rate": 6.741052631578948e-06,
|
| 1972 |
+
"loss": 0.2504,
|
| 1973 |
+
"step": 6800
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 11.262592898431048,
|
| 1977 |
+
"grad_norm": 4.20628547668457,
|
| 1978 |
+
"learning_rate": 6.68842105263158e-06,
|
| 1979 |
+
"loss": 0.2316,
|
| 1980 |
+
"step": 6825
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 11.303881090008257,
|
| 1984 |
+
"grad_norm": 5.95331335067749,
|
| 1985 |
+
"learning_rate": 6.635789473684211e-06,
|
| 1986 |
+
"loss": 0.2524,
|
| 1987 |
+
"step": 6850
|
| 1988 |
+
},
|
| 1989 |
+
{
|
| 1990 |
+
"epoch": 11.345169281585466,
|
| 1991 |
+
"grad_norm": 5.462084770202637,
|
| 1992 |
+
"learning_rate": 6.583157894736842e-06,
|
| 1993 |
+
"loss": 0.2488,
|
| 1994 |
+
"step": 6875
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 11.386457473162675,
|
| 1998 |
+
"grad_norm": 4.980792999267578,
|
| 1999 |
+
"learning_rate": 6.530526315789474e-06,
|
| 2000 |
+
"loss": 0.2344,
|
| 2001 |
+
"step": 6900
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 11.427745664739884,
|
| 2005 |
+
"grad_norm": 5.869212627410889,
|
| 2006 |
+
"learning_rate": 6.477894736842106e-06,
|
| 2007 |
+
"loss": 0.2484,
|
| 2008 |
+
"step": 6925
|
| 2009 |
+
},
|
| 2010 |
+
{
|
| 2011 |
+
"epoch": 11.469033856317093,
|
| 2012 |
+
"grad_norm": 5.4397873878479,
|
| 2013 |
+
"learning_rate": 6.425263157894738e-06,
|
| 2014 |
+
"loss": 0.2523,
|
| 2015 |
+
"step": 6950
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 11.510322047894302,
|
| 2019 |
+
"grad_norm": 5.890199661254883,
|
| 2020 |
+
"learning_rate": 6.372631578947369e-06,
|
| 2021 |
+
"loss": 0.2565,
|
| 2022 |
+
"step": 6975
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 11.551610239471511,
|
| 2026 |
+
"grad_norm": 4.83267879486084,
|
| 2027 |
+
"learning_rate": 6.3200000000000005e-06,
|
| 2028 |
+
"loss": 0.2319,
|
| 2029 |
+
"step": 7000
|
| 2030 |
+
},
|
| 2031 |
+
{
|
| 2032 |
+
"epoch": 11.551610239471511,
|
| 2033 |
+
"eval_cer": 49.55349466371671,
|
| 2034 |
+
"eval_loss": 1.0929207801818848,
|
| 2035 |
+
"eval_runtime": 718.5035,
|
| 2036 |
+
"eval_samples_per_second": 14.726,
|
| 2037 |
+
"eval_steps_per_second": 1.228,
|
| 2038 |
+
"eval_wer": 96.25648279113625,
|
| 2039 |
+
"step": 7000
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 11.59289843104872,
|
| 2043 |
+
"grad_norm": 6.01038932800293,
|
| 2044 |
+
"learning_rate": 6.267368421052632e-06,
|
| 2045 |
+
"loss": 0.2703,
|
| 2046 |
+
"step": 7025
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 11.63418662262593,
|
| 2050 |
+
"grad_norm": 4.515256404876709,
|
| 2051 |
+
"learning_rate": 6.214736842105264e-06,
|
| 2052 |
+
"loss": 0.2497,
|
| 2053 |
+
"step": 7050
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 11.675474814203138,
|
| 2057 |
+
"grad_norm": 5.914691925048828,
|
| 2058 |
+
"learning_rate": 6.1621052631578945e-06,
|
| 2059 |
+
"loss": 0.2374,
|
| 2060 |
+
"step": 7075
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 11.716763005780347,
|
| 2064 |
+
"grad_norm": 5.570655822753906,
|
| 2065 |
+
"learning_rate": 6.109473684210527e-06,
|
| 2066 |
+
"loss": 0.253,
|
| 2067 |
+
"step": 7100
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 11.758051197357556,
|
| 2071 |
+
"grad_norm": 5.9783244132995605,
|
| 2072 |
+
"learning_rate": 6.056842105263159e-06,
|
| 2073 |
+
"loss": 0.2364,
|
| 2074 |
+
"step": 7125
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 11.799339388934765,
|
| 2078 |
+
"grad_norm": 5.595921516418457,
|
| 2079 |
+
"learning_rate": 6.00421052631579e-06,
|
| 2080 |
+
"loss": 0.2375,
|
| 2081 |
+
"step": 7150
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 11.840627580511974,
|
| 2085 |
+
"grad_norm": 4.966527938842773,
|
| 2086 |
+
"learning_rate": 5.951578947368422e-06,
|
| 2087 |
+
"loss": 0.2652,
|
| 2088 |
+
"step": 7175
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 11.881915772089183,
|
| 2092 |
+
"grad_norm": 4.243671894073486,
|
| 2093 |
+
"learning_rate": 5.898947368421053e-06,
|
| 2094 |
+
"loss": 0.2348,
|
| 2095 |
+
"step": 7200
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 11.923203963666392,
|
| 2099 |
+
"grad_norm": 4.7744669914245605,
|
| 2100 |
+
"learning_rate": 5.846315789473684e-06,
|
| 2101 |
+
"loss": 0.2461,
|
| 2102 |
+
"step": 7225
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.964492155243601,
|
| 2106 |
+
"grad_norm": 4.887286186218262,
|
| 2107 |
+
"learning_rate": 5.793684210526316e-06,
|
| 2108 |
+
"loss": 0.2466,
|
| 2109 |
+
"step": 7250
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 12.004954582989265,
|
| 2113 |
+
"grad_norm": 4.767704963684082,
|
| 2114 |
+
"learning_rate": 5.7410526315789485e-06,
|
| 2115 |
+
"loss": 0.2628,
|
| 2116 |
+
"step": 7275
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 12.046242774566474,
|
| 2120 |
+
"grad_norm": 3.731494426727295,
|
| 2121 |
+
"learning_rate": 5.68842105263158e-06,
|
| 2122 |
+
"loss": 0.2121,
|
| 2123 |
+
"step": 7300
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 12.087530966143683,
|
| 2127 |
+
"grad_norm": 4.329957962036133,
|
| 2128 |
+
"learning_rate": 5.635789473684211e-06,
|
| 2129 |
+
"loss": 0.208,
|
| 2130 |
+
"step": 7325
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 12.128819157720892,
|
| 2134 |
+
"grad_norm": 4.992996692657471,
|
| 2135 |
+
"learning_rate": 5.5831578947368425e-06,
|
| 2136 |
+
"loss": 0.2007,
|
| 2137 |
+
"step": 7350
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 12.170107349298101,
|
| 2141 |
+
"grad_norm": 4.756143093109131,
|
| 2142 |
+
"learning_rate": 5.530526315789474e-06,
|
| 2143 |
+
"loss": 0.2084,
|
| 2144 |
+
"step": 7375
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 12.21139554087531,
|
| 2148 |
+
"grad_norm": 5.0002827644348145,
|
| 2149 |
+
"learning_rate": 5.477894736842105e-06,
|
| 2150 |
+
"loss": 0.2285,
|
| 2151 |
+
"step": 7400
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 12.25268373245252,
|
| 2155 |
+
"grad_norm": 3.5156126022338867,
|
| 2156 |
+
"learning_rate": 5.425263157894737e-06,
|
| 2157 |
+
"loss": 0.214,
|
| 2158 |
+
"step": 7425
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 12.293971924029728,
|
| 2162 |
+
"grad_norm": 4.908341884613037,
|
| 2163 |
+
"learning_rate": 5.372631578947369e-06,
|
| 2164 |
+
"loss": 0.2023,
|
| 2165 |
+
"step": 7450
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 12.335260115606937,
|
| 2169 |
+
"grad_norm": 5.120205879211426,
|
| 2170 |
+
"learning_rate": 5.320000000000001e-06,
|
| 2171 |
+
"loss": 0.2199,
|
| 2172 |
+
"step": 7475
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 12.376548307184144,
|
| 2176 |
+
"grad_norm": 5.24783992767334,
|
| 2177 |
+
"learning_rate": 5.267368421052632e-06,
|
| 2178 |
+
"loss": 0.1918,
|
| 2179 |
+
"step": 7500
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 12.417836498761353,
|
| 2183 |
+
"grad_norm": 4.663553714752197,
|
| 2184 |
+
"learning_rate": 5.214736842105263e-06,
|
| 2185 |
+
"loss": 0.2018,
|
| 2186 |
+
"step": 7525
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 12.459124690338562,
|
| 2190 |
+
"grad_norm": 4.562893867492676,
|
| 2191 |
+
"learning_rate": 5.162105263157895e-06,
|
| 2192 |
+
"loss": 0.2141,
|
| 2193 |
+
"step": 7550
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 12.500412881915771,
|
| 2197 |
+
"grad_norm": 5.8733062744140625,
|
| 2198 |
+
"learning_rate": 5.1094736842105264e-06,
|
| 2199 |
+
"loss": 0.2359,
|
| 2200 |
+
"step": 7575
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 12.54170107349298,
|
| 2204 |
+
"grad_norm": 7.215207576751709,
|
| 2205 |
+
"learning_rate": 5.056842105263158e-06,
|
| 2206 |
+
"loss": 0.2422,
|
| 2207 |
+
"step": 7600
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 12.58298926507019,
|
| 2211 |
+
"grad_norm": 4.633525371551514,
|
| 2212 |
+
"learning_rate": 5.0042105263157906e-06,
|
| 2213 |
+
"loss": 0.2188,
|
| 2214 |
+
"step": 7625
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 12.624277456647398,
|
| 2218 |
+
"grad_norm": 5.8343305587768555,
|
| 2219 |
+
"learning_rate": 4.951578947368421e-06,
|
| 2220 |
+
"loss": 0.2169,
|
| 2221 |
+
"step": 7650
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 12.665565648224607,
|
| 2225 |
+
"grad_norm": 3.9599239826202393,
|
| 2226 |
+
"learning_rate": 4.898947368421053e-06,
|
| 2227 |
+
"loss": 0.2275,
|
| 2228 |
+
"step": 7675
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 12.706853839801816,
|
| 2232 |
+
"grad_norm": 5.128805637359619,
|
| 2233 |
+
"learning_rate": 4.846315789473685e-06,
|
| 2234 |
+
"loss": 0.2095,
|
| 2235 |
+
"step": 7700
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 12.748142031379025,
|
| 2239 |
+
"grad_norm": 4.6408162117004395,
|
| 2240 |
+
"learning_rate": 4.793684210526316e-06,
|
| 2241 |
+
"loss": 0.1983,
|
| 2242 |
+
"step": 7725
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 12.789430222956234,
|
| 2246 |
+
"grad_norm": 4.684074878692627,
|
| 2247 |
+
"learning_rate": 4.741052631578948e-06,
|
| 2248 |
+
"loss": 0.2247,
|
| 2249 |
+
"step": 7750
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 12.830718414533443,
|
| 2253 |
+
"grad_norm": 4.256105422973633,
|
| 2254 |
+
"learning_rate": 4.6884210526315795e-06,
|
| 2255 |
+
"loss": 0.2164,
|
| 2256 |
+
"step": 7775
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 12.872006606110652,
|
| 2260 |
+
"grad_norm": 4.794992446899414,
|
| 2261 |
+
"learning_rate": 4.63578947368421e-06,
|
| 2262 |
+
"loss": 0.2133,
|
| 2263 |
+
"step": 7800
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 12.913294797687861,
|
| 2267 |
+
"grad_norm": 6.137476921081543,
|
| 2268 |
+
"learning_rate": 4.583157894736843e-06,
|
| 2269 |
+
"loss": 0.2434,
|
| 2270 |
+
"step": 7825
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 12.95458298926507,
|
| 2274 |
+
"grad_norm": 4.048770427703857,
|
| 2275 |
+
"learning_rate": 4.5305263157894744e-06,
|
| 2276 |
+
"loss": 0.2031,
|
| 2277 |
+
"step": 7850
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 12.995871180842279,
|
| 2281 |
+
"grad_norm": 4.465664863586426,
|
| 2282 |
+
"learning_rate": 4.477894736842105e-06,
|
| 2283 |
+
"loss": 0.2296,
|
| 2284 |
+
"step": 7875
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 13.036333608587944,
|
| 2288 |
+
"grad_norm": 4.5229668617248535,
|
| 2289 |
+
"learning_rate": 4.425263157894737e-06,
|
| 2290 |
+
"loss": 0.189,
|
| 2291 |
+
"step": 7900
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 13.077621800165153,
|
| 2295 |
+
"grad_norm": 3.7040534019470215,
|
| 2296 |
+
"learning_rate": 4.3726315789473685e-06,
|
| 2297 |
+
"loss": 0.1951,
|
| 2298 |
+
"step": 7925
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 13.118909991742361,
|
| 2302 |
+
"grad_norm": 4.726925373077393,
|
| 2303 |
+
"learning_rate": 4.32e-06,
|
| 2304 |
+
"loss": 0.1743,
|
| 2305 |
+
"step": 7950
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 13.16019818331957,
|
| 2309 |
+
"grad_norm": 4.432371616363525,
|
| 2310 |
+
"learning_rate": 4.267368421052632e-06,
|
| 2311 |
+
"loss": 0.1948,
|
| 2312 |
+
"step": 7975
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 13.20148637489678,
|
| 2316 |
+
"grad_norm": 6.360472679138184,
|
| 2317 |
+
"learning_rate": 4.214736842105263e-06,
|
| 2318 |
+
"loss": 0.1899,
|
| 2319 |
+
"step": 8000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 13.20148637489678,
|
| 2323 |
+
"eval_cer": 48.673667075533345,
|
| 2324 |
+
"eval_loss": 1.1223334074020386,
|
| 2325 |
+
"eval_runtime": 698.1183,
|
| 2326 |
+
"eval_samples_per_second": 15.156,
|
| 2327 |
+
"eval_steps_per_second": 1.263,
|
| 2328 |
+
"eval_wer": 97.27487034417727,
|
| 2329 |
+
"step": 8000
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 13.242774566473988,
|
| 2333 |
+
"grad_norm": 4.355146408081055,
|
| 2334 |
+
"learning_rate": 4.162105263157895e-06,
|
| 2335 |
+
"loss": 0.207,
|
| 2336 |
+
"step": 8025
|
| 2337 |
+
},
|
| 2338 |
+
{
|
| 2339 |
+
"epoch": 13.284062758051197,
|
| 2340 |
+
"grad_norm": 4.026797771453857,
|
| 2341 |
+
"learning_rate": 4.109473684210527e-06,
|
| 2342 |
+
"loss": 0.1656,
|
| 2343 |
+
"step": 8050
|
| 2344 |
+
},
|
| 2345 |
+
{
|
| 2346 |
+
"epoch": 13.325350949628406,
|
| 2347 |
+
"grad_norm": 4.001533031463623,
|
| 2348 |
+
"learning_rate": 4.056842105263158e-06,
|
| 2349 |
+
"loss": 0.1899,
|
| 2350 |
+
"step": 8075
|
| 2351 |
+
},
|
| 2352 |
+
{
|
| 2353 |
+
"epoch": 13.366639141205615,
|
| 2354 |
+
"grad_norm": 3.8727056980133057,
|
| 2355 |
+
"learning_rate": 4.00421052631579e-06,
|
| 2356 |
+
"loss": 0.1919,
|
| 2357 |
+
"step": 8100
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 13.407927332782824,
|
| 2361 |
+
"grad_norm": 4.644744873046875,
|
| 2362 |
+
"learning_rate": 3.951578947368422e-06,
|
| 2363 |
+
"loss": 0.1987,
|
| 2364 |
+
"step": 8125
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 13.449215524360033,
|
| 2368 |
+
"grad_norm": 4.7693071365356445,
|
| 2369 |
+
"learning_rate": 3.898947368421052e-06,
|
| 2370 |
+
"loss": 0.185,
|
| 2371 |
+
"step": 8150
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 13.490503715937242,
|
| 2375 |
+
"grad_norm": 5.065967082977295,
|
| 2376 |
+
"learning_rate": 3.846315789473685e-06,
|
| 2377 |
+
"loss": 0.1811,
|
| 2378 |
+
"step": 8175
|
| 2379 |
+
},
|
| 2380 |
+
{
|
| 2381 |
+
"epoch": 13.531791907514451,
|
| 2382 |
+
"grad_norm": 3.682410955429077,
|
| 2383 |
+
"learning_rate": 3.793684210526316e-06,
|
| 2384 |
+
"loss": 0.2118,
|
| 2385 |
+
"step": 8200
|
| 2386 |
+
},
|
| 2387 |
+
{
|
| 2388 |
+
"epoch": 13.57308009909166,
|
| 2389 |
+
"grad_norm": 4.680816650390625,
|
| 2390 |
+
"learning_rate": 3.7410526315789473e-06,
|
| 2391 |
+
"loss": 0.1995,
|
| 2392 |
+
"step": 8225
|
| 2393 |
+
},
|
| 2394 |
+
{
|
| 2395 |
+
"epoch": 13.61436829066887,
|
| 2396 |
+
"grad_norm": 5.313553333282471,
|
| 2397 |
+
"learning_rate": 3.6884210526315794e-06,
|
| 2398 |
+
"loss": 0.1871,
|
| 2399 |
+
"step": 8250
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 13.655656482246078,
|
| 2403 |
+
"grad_norm": 4.104048728942871,
|
| 2404 |
+
"learning_rate": 3.635789473684211e-06,
|
| 2405 |
+
"loss": 0.1861,
|
| 2406 |
+
"step": 8275
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 13.696944673823287,
|
| 2410 |
+
"grad_norm": 4.6280293464660645,
|
| 2411 |
+
"learning_rate": 3.5831578947368422e-06,
|
| 2412 |
+
"loss": 0.1936,
|
| 2413 |
+
"step": 8300
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 13.738232865400496,
|
| 2417 |
+
"grad_norm": 5.779093265533447,
|
| 2418 |
+
"learning_rate": 3.5305263157894743e-06,
|
| 2419 |
+
"loss": 0.1899,
|
| 2420 |
+
"step": 8325
|
| 2421 |
+
},
|
| 2422 |
+
{
|
| 2423 |
+
"epoch": 13.779521056977705,
|
| 2424 |
+
"grad_norm": 3.84360933303833,
|
| 2425 |
+
"learning_rate": 3.4778947368421055e-06,
|
| 2426 |
+
"loss": 0.1937,
|
| 2427 |
+
"step": 8350
|
| 2428 |
+
},
|
| 2429 |
+
{
|
| 2430 |
+
"epoch": 13.820809248554914,
|
| 2431 |
+
"grad_norm": 5.995066165924072,
|
| 2432 |
+
"learning_rate": 3.425263157894737e-06,
|
| 2433 |
+
"loss": 0.1976,
|
| 2434 |
+
"step": 8375
|
| 2435 |
+
},
|
| 2436 |
+
{
|
| 2437 |
+
"epoch": 13.862097440132121,
|
| 2438 |
+
"grad_norm": 5.045938491821289,
|
| 2439 |
+
"learning_rate": 3.3726315789473683e-06,
|
| 2440 |
+
"loss": 0.177,
|
| 2441 |
+
"step": 8400
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 13.90338563170933,
|
| 2445 |
+
"grad_norm": 4.3831658363342285,
|
| 2446 |
+
"learning_rate": 3.3200000000000004e-06,
|
| 2447 |
+
"loss": 0.1831,
|
| 2448 |
+
"step": 8425
|
| 2449 |
+
},
|
| 2450 |
+
{
|
| 2451 |
+
"epoch": 13.94467382328654,
|
| 2452 |
+
"grad_norm": 5.6045002937316895,
|
| 2453 |
+
"learning_rate": 3.267368421052632e-06,
|
| 2454 |
+
"loss": 0.177,
|
| 2455 |
+
"step": 8450
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 13.985962014863748,
|
| 2459 |
+
"grad_norm": 5.313602447509766,
|
| 2460 |
+
"learning_rate": 3.2147368421052633e-06,
|
| 2461 |
+
"loss": 0.2106,
|
| 2462 |
+
"step": 8475
|
| 2463 |
+
},
|
| 2464 |
+
{
|
| 2465 |
+
"epoch": 14.026424442609414,
|
| 2466 |
+
"grad_norm": 4.277268409729004,
|
| 2467 |
+
"learning_rate": 3.1621052631578953e-06,
|
| 2468 |
+
"loss": 0.1697,
|
| 2469 |
+
"step": 8500
|
| 2470 |
+
},
|
| 2471 |
+
{
|
| 2472 |
+
"epoch": 14.067712634186623,
|
| 2473 |
+
"grad_norm": 4.43281364440918,
|
| 2474 |
+
"learning_rate": 3.1094736842105265e-06,
|
| 2475 |
+
"loss": 0.1655,
|
| 2476 |
+
"step": 8525
|
| 2477 |
+
},
|
| 2478 |
+
{
|
| 2479 |
+
"epoch": 14.109000825763832,
|
| 2480 |
+
"grad_norm": 5.5066304206848145,
|
| 2481 |
+
"learning_rate": 3.056842105263158e-06,
|
| 2482 |
+
"loss": 0.1611,
|
| 2483 |
+
"step": 8550
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 14.15028901734104,
|
| 2487 |
+
"grad_norm": 5.53204345703125,
|
| 2488 |
+
"learning_rate": 3.0042105263157894e-06,
|
| 2489 |
+
"loss": 0.1808,
|
| 2490 |
+
"step": 8575
|
| 2491 |
+
},
|
| 2492 |
+
{
|
| 2493 |
+
"epoch": 14.191577208918249,
|
| 2494 |
+
"grad_norm": 4.399189472198486,
|
| 2495 |
+
"learning_rate": 2.9515789473684214e-06,
|
| 2496 |
+
"loss": 0.1487,
|
| 2497 |
+
"step": 8600
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 14.232865400495458,
|
| 2501 |
+
"grad_norm": 4.710522174835205,
|
| 2502 |
+
"learning_rate": 2.8989473684210526e-06,
|
| 2503 |
+
"loss": 0.1684,
|
| 2504 |
+
"step": 8625
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 14.274153592072667,
|
| 2508 |
+
"grad_norm": 4.319856643676758,
|
| 2509 |
+
"learning_rate": 2.8463157894736843e-06,
|
| 2510 |
+
"loss": 0.1626,
|
| 2511 |
+
"step": 8650
|
| 2512 |
+
},
|
| 2513 |
+
{
|
| 2514 |
+
"epoch": 14.315441783649876,
|
| 2515 |
+
"grad_norm": 5.691439628601074,
|
| 2516 |
+
"learning_rate": 2.7936842105263163e-06,
|
| 2517 |
+
"loss": 0.1734,
|
| 2518 |
+
"step": 8675
|
| 2519 |
+
},
|
| 2520 |
+
{
|
| 2521 |
+
"epoch": 14.356729975227084,
|
| 2522 |
+
"grad_norm": 5.005030632019043,
|
| 2523 |
+
"learning_rate": 2.7410526315789476e-06,
|
| 2524 |
+
"loss": 0.1627,
|
| 2525 |
+
"step": 8700
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 14.398018166804293,
|
| 2529 |
+
"grad_norm": 4.130620002746582,
|
| 2530 |
+
"learning_rate": 2.688421052631579e-06,
|
| 2531 |
+
"loss": 0.1718,
|
| 2532 |
+
"step": 8725
|
| 2533 |
+
},
|
| 2534 |
+
{
|
| 2535 |
+
"epoch": 14.439306358381502,
|
| 2536 |
+
"grad_norm": 3.437147617340088,
|
| 2537 |
+
"learning_rate": 2.635789473684211e-06,
|
| 2538 |
+
"loss": 0.1756,
|
| 2539 |
+
"step": 8750
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 14.480594549958711,
|
| 2543 |
+
"grad_norm": 4.4854044914245605,
|
| 2544 |
+
"learning_rate": 2.5831578947368425e-06,
|
| 2545 |
+
"loss": 0.1671,
|
| 2546 |
+
"step": 8775
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 14.52188274153592,
|
| 2550 |
+
"grad_norm": 5.259097099304199,
|
| 2551 |
+
"learning_rate": 2.5305263157894737e-06,
|
| 2552 |
+
"loss": 0.1782,
|
| 2553 |
+
"step": 8800
|
| 2554 |
+
},
|
| 2555 |
+
{
|
| 2556 |
+
"epoch": 14.56317093311313,
|
| 2557 |
+
"grad_norm": 3.7995481491088867,
|
| 2558 |
+
"learning_rate": 2.4778947368421053e-06,
|
| 2559 |
+
"loss": 0.1684,
|
| 2560 |
+
"step": 8825
|
| 2561 |
+
},
|
| 2562 |
+
{
|
| 2563 |
+
"epoch": 14.604459124690338,
|
| 2564 |
+
"grad_norm": 4.331813335418701,
|
| 2565 |
+
"learning_rate": 2.425263157894737e-06,
|
| 2566 |
+
"loss": 0.1757,
|
| 2567 |
+
"step": 8850
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 14.645747316267547,
|
| 2571 |
+
"grad_norm": 5.367282867431641,
|
| 2572 |
+
"learning_rate": 2.3726315789473686e-06,
|
| 2573 |
+
"loss": 0.1796,
|
| 2574 |
+
"step": 8875
|
| 2575 |
+
},
|
| 2576 |
+
{
|
| 2577 |
+
"epoch": 14.687035507844756,
|
| 2578 |
+
"grad_norm": 5.606546401977539,
|
| 2579 |
+
"learning_rate": 2.3200000000000002e-06,
|
| 2580 |
+
"loss": 0.1911,
|
| 2581 |
+
"step": 8900
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 14.728323699421965,
|
| 2585 |
+
"grad_norm": 3.9387214183807373,
|
| 2586 |
+
"learning_rate": 2.267368421052632e-06,
|
| 2587 |
+
"loss": 0.1666,
|
| 2588 |
+
"step": 8925
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 14.769611890999174,
|
| 2592 |
+
"grad_norm": 4.426852703094482,
|
| 2593 |
+
"learning_rate": 2.2147368421052635e-06,
|
| 2594 |
+
"loss": 0.1717,
|
| 2595 |
+
"step": 8950
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 14.810900082576383,
|
| 2599 |
+
"grad_norm": 3.4610788822174072,
|
| 2600 |
+
"learning_rate": 2.1621052631578947e-06,
|
| 2601 |
+
"loss": 0.168,
|
| 2602 |
+
"step": 8975
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 14.852188274153592,
|
| 2606 |
+
"grad_norm": 4.5235676765441895,
|
| 2607 |
+
"learning_rate": 2.1094736842105264e-06,
|
| 2608 |
+
"loss": 0.1425,
|
| 2609 |
+
"step": 9000
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 14.852188274153592,
|
| 2613 |
+
"eval_cer": 48.44840828585512,
|
| 2614 |
+
"eval_loss": 1.142202377319336,
|
| 2615 |
+
"eval_runtime": 701.7243,
|
| 2616 |
+
"eval_samples_per_second": 15.079,
|
| 2617 |
+
"eval_steps_per_second": 1.257,
|
| 2618 |
+
"eval_wer": 96.61480433757662,
|
| 2619 |
+
"step": 9000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 14.893476465730801,
|
| 2623 |
+
"grad_norm": 4.106043815612793,
|
| 2624 |
+
"learning_rate": 2.058947368421053e-06,
|
| 2625 |
+
"loss": 0.1853,
|
| 2626 |
+
"step": 9025
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 14.93476465730801,
|
| 2630 |
+
"grad_norm": 4.604458808898926,
|
| 2631 |
+
"learning_rate": 2.0063157894736843e-06,
|
| 2632 |
+
"loss": 0.1725,
|
| 2633 |
+
"step": 9050
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 14.97605284888522,
|
| 2637 |
+
"grad_norm": 5.01749324798584,
|
| 2638 |
+
"learning_rate": 1.953684210526316e-06,
|
| 2639 |
+
"loss": 0.1566,
|
| 2640 |
+
"step": 9075
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 15.016515276630884,
|
| 2644 |
+
"grad_norm": 3.69610857963562,
|
| 2645 |
+
"learning_rate": 1.9010526315789476e-06,
|
| 2646 |
+
"loss": 0.1791,
|
| 2647 |
+
"step": 9100
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 15.057803468208093,
|
| 2651 |
+
"grad_norm": 3.566075563430786,
|
| 2652 |
+
"learning_rate": 1.848421052631579e-06,
|
| 2653 |
+
"loss": 0.1626,
|
| 2654 |
+
"step": 9125
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 15.099091659785302,
|
| 2658 |
+
"grad_norm": 3.8838043212890625,
|
| 2659 |
+
"learning_rate": 1.7957894736842108e-06,
|
| 2660 |
+
"loss": 0.1507,
|
| 2661 |
+
"step": 9150
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 15.14037985136251,
|
| 2665 |
+
"grad_norm": 4.462072372436523,
|
| 2666 |
+
"learning_rate": 1.7431578947368423e-06,
|
| 2667 |
+
"loss": 0.1682,
|
| 2668 |
+
"step": 9175
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 15.18166804293972,
|
| 2672 |
+
"grad_norm": 5.543285846710205,
|
| 2673 |
+
"learning_rate": 1.6905263157894739e-06,
|
| 2674 |
+
"loss": 0.141,
|
| 2675 |
+
"step": 9200
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 15.222956234516928,
|
| 2679 |
+
"grad_norm": 4.328714847564697,
|
| 2680 |
+
"learning_rate": 1.6378947368421053e-06,
|
| 2681 |
+
"loss": 0.1747,
|
| 2682 |
+
"step": 9225
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 15.264244426094137,
|
| 2686 |
+
"grad_norm": 3.924700975418091,
|
| 2687 |
+
"learning_rate": 1.585263157894737e-06,
|
| 2688 |
+
"loss": 0.1727,
|
| 2689 |
+
"step": 9250
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 15.305532617671346,
|
| 2693 |
+
"grad_norm": 4.647414207458496,
|
| 2694 |
+
"learning_rate": 1.5326315789473686e-06,
|
| 2695 |
+
"loss": 0.1512,
|
| 2696 |
+
"step": 9275
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 15.346820809248555,
|
| 2700 |
+
"grad_norm": 3.2526192665100098,
|
| 2701 |
+
"learning_rate": 1.48e-06,
|
| 2702 |
+
"loss": 0.1521,
|
| 2703 |
+
"step": 9300
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 15.388109000825764,
|
| 2707 |
+
"grad_norm": 3.697683334350586,
|
| 2708 |
+
"learning_rate": 1.4273684210526317e-06,
|
| 2709 |
+
"loss": 0.1498,
|
| 2710 |
+
"step": 9325
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 15.429397192402973,
|
| 2714 |
+
"grad_norm": 3.709627866744995,
|
| 2715 |
+
"learning_rate": 1.374736842105263e-06,
|
| 2716 |
+
"loss": 0.1654,
|
| 2717 |
+
"step": 9350
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 15.470685383980182,
|
| 2721 |
+
"grad_norm": 3.7767739295959473,
|
| 2722 |
+
"learning_rate": 1.322105263157895e-06,
|
| 2723 |
+
"loss": 0.1466,
|
| 2724 |
+
"step": 9375
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 15.511973575557391,
|
| 2728 |
+
"grad_norm": 3.6882431507110596,
|
| 2729 |
+
"learning_rate": 1.2694736842105266e-06,
|
| 2730 |
+
"loss": 0.1496,
|
| 2731 |
+
"step": 9400
|
| 2732 |
+
},
|
| 2733 |
+
{
|
| 2734 |
+
"epoch": 15.5532617671346,
|
| 2735 |
+
"grad_norm": 3.9242100715637207,
|
| 2736 |
+
"learning_rate": 1.216842105263158e-06,
|
| 2737 |
+
"loss": 0.1491,
|
| 2738 |
+
"step": 9425
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 15.59454995871181,
|
| 2742 |
+
"grad_norm": 5.024509429931641,
|
| 2743 |
+
"learning_rate": 1.1642105263157896e-06,
|
| 2744 |
+
"loss": 0.1549,
|
| 2745 |
+
"step": 9450
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 15.635838150289018,
|
| 2749 |
+
"grad_norm": 4.048088550567627,
|
| 2750 |
+
"learning_rate": 1.1115789473684213e-06,
|
| 2751 |
+
"loss": 0.1564,
|
| 2752 |
+
"step": 9475
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 15.677126341866225,
|
| 2756 |
+
"grad_norm": 4.535757541656494,
|
| 2757 |
+
"learning_rate": 1.0589473684210527e-06,
|
| 2758 |
+
"loss": 0.1548,
|
| 2759 |
+
"step": 9500
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 15.718414533443434,
|
| 2763 |
+
"grad_norm": 4.47163200378418,
|
| 2764 |
+
"learning_rate": 1.0063157894736843e-06,
|
| 2765 |
+
"loss": 0.1521,
|
| 2766 |
+
"step": 9525
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 15.759702725020643,
|
| 2770 |
+
"grad_norm": 3.4826717376708984,
|
| 2771 |
+
"learning_rate": 9.536842105263158e-07,
|
| 2772 |
+
"loss": 0.151,
|
| 2773 |
+
"step": 9550
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 15.800990916597852,
|
| 2777 |
+
"grad_norm": 3.6962897777557373,
|
| 2778 |
+
"learning_rate": 9.010526315789474e-07,
|
| 2779 |
+
"loss": 0.1611,
|
| 2780 |
+
"step": 9575
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 15.842279108175061,
|
| 2784 |
+
"grad_norm": 3.6767818927764893,
|
| 2785 |
+
"learning_rate": 8.48421052631579e-07,
|
| 2786 |
+
"loss": 0.1594,
|
| 2787 |
+
"step": 9600
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 15.88356729975227,
|
| 2791 |
+
"grad_norm": 2.950598955154419,
|
| 2792 |
+
"learning_rate": 7.957894736842107e-07,
|
| 2793 |
+
"loss": 0.1583,
|
| 2794 |
+
"step": 9625
|
| 2795 |
+
},
|
| 2796 |
+
{
|
| 2797 |
+
"epoch": 15.92485549132948,
|
| 2798 |
+
"grad_norm": 4.535677909851074,
|
| 2799 |
+
"learning_rate": 7.431578947368422e-07,
|
| 2800 |
+
"loss": 0.1743,
|
| 2801 |
+
"step": 9650
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 15.966143682906688,
|
| 2805 |
+
"grad_norm": 3.578261613845825,
|
| 2806 |
+
"learning_rate": 6.905263157894737e-07,
|
| 2807 |
+
"loss": 0.1279,
|
| 2808 |
+
"step": 9675
|
| 2809 |
+
},
|
| 2810 |
+
{
|
| 2811 |
+
"epoch": 16.006606110652353,
|
| 2812 |
+
"grad_norm": 3.0115976333618164,
|
| 2813 |
+
"learning_rate": 6.378947368421053e-07,
|
| 2814 |
+
"loss": 0.1442,
|
| 2815 |
+
"step": 9700
|
| 2816 |
+
},
|
| 2817 |
+
{
|
| 2818 |
+
"epoch": 16.04789430222956,
|
| 2819 |
+
"grad_norm": 4.302647590637207,
|
| 2820 |
+
"learning_rate": 5.852631578947369e-07,
|
| 2821 |
+
"loss": 0.1524,
|
| 2822 |
+
"step": 9725
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 16.08918249380677,
|
| 2826 |
+
"grad_norm": 3.516667366027832,
|
| 2827 |
+
"learning_rate": 5.326315789473684e-07,
|
| 2828 |
+
"loss": 0.1431,
|
| 2829 |
+
"step": 9750
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 16.13047068538398,
|
| 2833 |
+
"grad_norm": 3.802489757537842,
|
| 2834 |
+
"learning_rate": 4.800000000000001e-07,
|
| 2835 |
+
"loss": 0.1607,
|
| 2836 |
+
"step": 9775
|
| 2837 |
+
},
|
| 2838 |
+
{
|
| 2839 |
+
"epoch": 16.17175887696119,
|
| 2840 |
+
"grad_norm": 2.9078941345214844,
|
| 2841 |
+
"learning_rate": 4.273684210526316e-07,
|
| 2842 |
+
"loss": 0.134,
|
| 2843 |
+
"step": 9800
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 16.213047068538398,
|
| 2847 |
+
"grad_norm": 3.3042197227478027,
|
| 2848 |
+
"learning_rate": 3.7473684210526323e-07,
|
| 2849 |
+
"loss": 0.1614,
|
| 2850 |
+
"step": 9825
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 16.254335260115607,
|
| 2854 |
+
"grad_norm": 4.696882724761963,
|
| 2855 |
+
"learning_rate": 3.2210526315789476e-07,
|
| 2856 |
+
"loss": 0.1379,
|
| 2857 |
+
"step": 9850
|
| 2858 |
+
},
|
| 2859 |
+
{
|
| 2860 |
+
"epoch": 16.295623451692816,
|
| 2861 |
+
"grad_norm": 4.522494316101074,
|
| 2862 |
+
"learning_rate": 2.6947368421052635e-07,
|
| 2863 |
+
"loss": 0.1586,
|
| 2864 |
+
"step": 9875
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 16.336911643270025,
|
| 2868 |
+
"grad_norm": 4.257038593292236,
|
| 2869 |
+
"learning_rate": 2.168421052631579e-07,
|
| 2870 |
+
"loss": 0.1423,
|
| 2871 |
+
"step": 9900
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 16.378199834847234,
|
| 2875 |
+
"grad_norm": 4.159156799316406,
|
| 2876 |
+
"learning_rate": 1.642105263157895e-07,
|
| 2877 |
+
"loss": 0.1617,
|
| 2878 |
+
"step": 9925
|
| 2879 |
+
},
|
| 2880 |
+
{
|
| 2881 |
+
"epoch": 16.419488026424442,
|
| 2882 |
+
"grad_norm": 4.338253974914551,
|
| 2883 |
+
"learning_rate": 1.1157894736842106e-07,
|
| 2884 |
+
"loss": 0.1597,
|
| 2885 |
+
"step": 9950
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 16.46077621800165,
|
| 2889 |
+
"grad_norm": 5.437859058380127,
|
| 2890 |
+
"learning_rate": 5.8947368421052637e-08,
|
| 2891 |
+
"loss": 0.1629,
|
| 2892 |
+
"step": 9975
|
| 2893 |
+
},
|
| 2894 |
+
{
|
| 2895 |
+
"epoch": 16.50206440957886,
|
| 2896 |
+
"grad_norm": 4.178699970245361,
|
| 2897 |
+
"learning_rate": 6.315789473684211e-09,
|
| 2898 |
+
"loss": 0.161,
|
| 2899 |
+
"step": 10000
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 16.50206440957886,
|
| 2903 |
+
"eval_cer": 49.64348354407163,
|
| 2904 |
+
"eval_loss": 1.1519368886947632,
|
| 2905 |
+
"eval_runtime": 708.8725,
|
| 2906 |
+
"eval_samples_per_second": 14.927,
|
| 2907 |
+
"eval_steps_per_second": 1.244,
|
| 2908 |
+
"eval_wer": 96.77510608203677,
|
| 2909 |
+
"step": 10000
|
| 2910 |
}
|
| 2911 |
],
|
| 2912 |
"logging_steps": 25,
|
|
|
|
| 2921 |
"should_evaluate": false,
|
| 2922 |
"should_log": false,
|
| 2923 |
"should_save": true,
|
| 2924 |
+
"should_training_stop": true
|
| 2925 |
},
|
| 2926 |
"attributes": {}
|
| 2927 |
}
|
| 2928 |
},
|
| 2929 |
+
"total_flos": 1.949150849531904e+19,
|
| 2930 |
"train_batch_size": 12,
|
| 2931 |
"trial_name": null,
|
| 2932 |
"trial_params": null
|