Training in progress, step 549, 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 327040
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58a0752a539301f8231388b5248e6531648f9975f002794e531455585c622d59
|
| 3 |
size 327040
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 739578
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:338969914b555067bbdd012f0ebbd42b778c3238c15c283f19e85842c8428a71
|
| 3 |
size 739578
|
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:c76786f9373c21c8c5330fb28b216bb4b1ec219a1bd8e351c0901d7218a888a8
|
| 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:e07ef84832650ecf91c0ca599350004d37065dd55fdbc5ee94db4ce7b36df9f3
|
| 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": 183,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2593,6 +2593,1295 @@
|
|
| 2593 |
"eval_samples_per_second": 63.412,
|
| 2594 |
"eval_steps_per_second": 31.809,
|
| 2595 |
"step": 366
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2596 |
}
|
| 2597 |
],
|
| 2598 |
"logging_steps": 1,
|
|
@@ -2612,7 +3901,7 @@
|
|
| 2612 |
"attributes": {}
|
| 2613 |
}
|
| 2614 |
},
|
| 2615 |
-
"total_flos":
|
| 2616 |
"train_batch_size": 2,
|
| 2617 |
"trial_name": null,
|
| 2618 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7536032944406315,
|
| 5 |
"eval_steps": 183,
|
| 6 |
+
"global_step": 549,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2593 |
"eval_samples_per_second": 63.412,
|
| 2594 |
"eval_steps_per_second": 31.809,
|
| 2595 |
"step": 366
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 0.5037748798901853,
|
| 2599 |
+
"grad_norm": 0.04130493476986885,
|
| 2600 |
+
"learning_rate": 0.00010109232620893592,
|
| 2601 |
+
"loss": 11.8494,
|
| 2602 |
+
"step": 367
|
| 2603 |
+
},
|
| 2604 |
+
{
|
| 2605 |
+
"epoch": 0.5051475634866164,
|
| 2606 |
+
"grad_norm": 0.05430886521935463,
|
| 2607 |
+
"learning_rate": 0.00010065540406715483,
|
| 2608 |
+
"loss": 11.8419,
|
| 2609 |
+
"step": 368
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 0.5065202470830473,
|
| 2613 |
+
"grad_norm": 0.05343049392104149,
|
| 2614 |
+
"learning_rate": 0.00010021846941269179,
|
| 2615 |
+
"loss": 11.8495,
|
| 2616 |
+
"step": 369
|
| 2617 |
+
},
|
| 2618 |
+
{
|
| 2619 |
+
"epoch": 0.5078929306794784,
|
| 2620 |
+
"grad_norm": 0.04137863591313362,
|
| 2621 |
+
"learning_rate": 9.978153058730823e-05,
|
| 2622 |
+
"loss": 11.8486,
|
| 2623 |
+
"step": 370
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 0.5092656142759094,
|
| 2627 |
+
"grad_norm": 0.06110731139779091,
|
| 2628 |
+
"learning_rate": 9.934459593284517e-05,
|
| 2629 |
+
"loss": 11.8562,
|
| 2630 |
+
"step": 371
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 0.5106382978723404,
|
| 2634 |
+
"grad_norm": 0.04892272502183914,
|
| 2635 |
+
"learning_rate": 9.89076737910641e-05,
|
| 2636 |
+
"loss": 11.8475,
|
| 2637 |
+
"step": 372
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 0.5120109814687714,
|
| 2641 |
+
"grad_norm": 0.051346611231565475,
|
| 2642 |
+
"learning_rate": 9.847077250348749e-05,
|
| 2643 |
+
"loss": 11.8422,
|
| 2644 |
+
"step": 373
|
| 2645 |
+
},
|
| 2646 |
+
{
|
| 2647 |
+
"epoch": 0.5133836650652025,
|
| 2648 |
+
"grad_norm": 0.05730092525482178,
|
| 2649 |
+
"learning_rate": 9.803390041123979e-05,
|
| 2650 |
+
"loss": 11.8519,
|
| 2651 |
+
"step": 374
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 0.5147563486616334,
|
| 2655 |
+
"grad_norm": 0.06608743965625763,
|
| 2656 |
+
"learning_rate": 9.759706585488797e-05,
|
| 2657 |
+
"loss": 11.8449,
|
| 2658 |
+
"step": 375
|
| 2659 |
+
},
|
| 2660 |
+
{
|
| 2661 |
+
"epoch": 0.5161290322580645,
|
| 2662 |
+
"grad_norm": 0.052331529557704926,
|
| 2663 |
+
"learning_rate": 9.716027717428243e-05,
|
| 2664 |
+
"loss": 11.8476,
|
| 2665 |
+
"step": 376
|
| 2666 |
+
},
|
| 2667 |
+
{
|
| 2668 |
+
"epoch": 0.5175017158544956,
|
| 2669 |
+
"grad_norm": 0.06584908813238144,
|
| 2670 |
+
"learning_rate": 9.672354270839778e-05,
|
| 2671 |
+
"loss": 11.8452,
|
| 2672 |
+
"step": 377
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 0.5188743994509266,
|
| 2676 |
+
"grad_norm": 0.04150731861591339,
|
| 2677 |
+
"learning_rate": 9.628687079517348e-05,
|
| 2678 |
+
"loss": 11.845,
|
| 2679 |
+
"step": 378
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 0.5202470830473576,
|
| 2683 |
+
"grad_norm": 0.0465887077152729,
|
| 2684 |
+
"learning_rate": 9.585026977135487e-05,
|
| 2685 |
+
"loss": 11.8417,
|
| 2686 |
+
"step": 379
|
| 2687 |
+
},
|
| 2688 |
+
{
|
| 2689 |
+
"epoch": 0.5216197666437886,
|
| 2690 |
+
"grad_norm": 0.04746214300394058,
|
| 2691 |
+
"learning_rate": 9.541374797233381e-05,
|
| 2692 |
+
"loss": 11.8484,
|
| 2693 |
+
"step": 380
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 0.5229924502402197,
|
| 2697 |
+
"grad_norm": 0.056160494685173035,
|
| 2698 |
+
"learning_rate": 9.497731373198966e-05,
|
| 2699 |
+
"loss": 11.8527,
|
| 2700 |
+
"step": 381
|
| 2701 |
+
},
|
| 2702 |
+
{
|
| 2703 |
+
"epoch": 0.5243651338366506,
|
| 2704 |
+
"grad_norm": 0.04341844096779823,
|
| 2705 |
+
"learning_rate": 9.45409753825302e-05,
|
| 2706 |
+
"loss": 11.8463,
|
| 2707 |
+
"step": 382
|
| 2708 |
+
},
|
| 2709 |
+
{
|
| 2710 |
+
"epoch": 0.5257378174330817,
|
| 2711 |
+
"grad_norm": 0.06022747606039047,
|
| 2712 |
+
"learning_rate": 9.41047412543324e-05,
|
| 2713 |
+
"loss": 11.8479,
|
| 2714 |
+
"step": 383
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"epoch": 0.5271105010295127,
|
| 2718 |
+
"grad_norm": 0.053729359060525894,
|
| 2719 |
+
"learning_rate": 9.366861967578364e-05,
|
| 2720 |
+
"loss": 11.8415,
|
| 2721 |
+
"step": 384
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 0.5284831846259437,
|
| 2725 |
+
"grad_norm": 0.047009535133838654,
|
| 2726 |
+
"learning_rate": 9.323261897312238e-05,
|
| 2727 |
+
"loss": 11.8521,
|
| 2728 |
+
"step": 385
|
| 2729 |
+
},
|
| 2730 |
+
{
|
| 2731 |
+
"epoch": 0.5298558682223747,
|
| 2732 |
+
"grad_norm": 0.05042191594839096,
|
| 2733 |
+
"learning_rate": 9.279674747027952e-05,
|
| 2734 |
+
"loss": 11.8471,
|
| 2735 |
+
"step": 386
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 0.5312285518188058,
|
| 2739 |
+
"grad_norm": 0.05017239600419998,
|
| 2740 |
+
"learning_rate": 9.236101348871921e-05,
|
| 2741 |
+
"loss": 11.8415,
|
| 2742 |
+
"step": 387
|
| 2743 |
+
},
|
| 2744 |
+
{
|
| 2745 |
+
"epoch": 0.5326012354152367,
|
| 2746 |
+
"grad_norm": 0.05985607951879501,
|
| 2747 |
+
"learning_rate": 9.192542534728021e-05,
|
| 2748 |
+
"loss": 11.8475,
|
| 2749 |
+
"step": 388
|
| 2750 |
+
},
|
| 2751 |
+
{
|
| 2752 |
+
"epoch": 0.5339739190116678,
|
| 2753 |
+
"grad_norm": 0.0469660609960556,
|
| 2754 |
+
"learning_rate": 9.148999136201686e-05,
|
| 2755 |
+
"loss": 11.8462,
|
| 2756 |
+
"step": 389
|
| 2757 |
+
},
|
| 2758 |
+
{
|
| 2759 |
+
"epoch": 0.5353466026080989,
|
| 2760 |
+
"grad_norm": 0.0472818948328495,
|
| 2761 |
+
"learning_rate": 9.105471984604055e-05,
|
| 2762 |
+
"loss": 11.8457,
|
| 2763 |
+
"step": 390
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 0.5367192862045299,
|
| 2767 |
+
"grad_norm": 0.04820844903588295,
|
| 2768 |
+
"learning_rate": 9.061961910936078e-05,
|
| 2769 |
+
"loss": 11.8378,
|
| 2770 |
+
"step": 391
|
| 2771 |
+
},
|
| 2772 |
+
{
|
| 2773 |
+
"epoch": 0.5380919698009609,
|
| 2774 |
+
"grad_norm": 0.04612083360552788,
|
| 2775 |
+
"learning_rate": 9.018469745872661e-05,
|
| 2776 |
+
"loss": 11.8472,
|
| 2777 |
+
"step": 392
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 0.5394646533973919,
|
| 2781 |
+
"grad_norm": 0.05291563645005226,
|
| 2782 |
+
"learning_rate": 8.974996319746814e-05,
|
| 2783 |
+
"loss": 11.8543,
|
| 2784 |
+
"step": 393
|
| 2785 |
+
},
|
| 2786 |
+
{
|
| 2787 |
+
"epoch": 0.540837336993823,
|
| 2788 |
+
"grad_norm": 0.045743975788354874,
|
| 2789 |
+
"learning_rate": 8.931542462533783e-05,
|
| 2790 |
+
"loss": 11.8477,
|
| 2791 |
+
"step": 394
|
| 2792 |
+
},
|
| 2793 |
+
{
|
| 2794 |
+
"epoch": 0.5422100205902539,
|
| 2795 |
+
"grad_norm": 0.04367321729660034,
|
| 2796 |
+
"learning_rate": 8.88810900383522e-05,
|
| 2797 |
+
"loss": 11.8468,
|
| 2798 |
+
"step": 395
|
| 2799 |
+
},
|
| 2800 |
+
{
|
| 2801 |
+
"epoch": 0.543582704186685,
|
| 2802 |
+
"grad_norm": 0.04755084216594696,
|
| 2803 |
+
"learning_rate": 8.84469677286333e-05,
|
| 2804 |
+
"loss": 11.8404,
|
| 2805 |
+
"step": 396
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 0.544955387783116,
|
| 2809 |
+
"grad_norm": 0.05640660971403122,
|
| 2810 |
+
"learning_rate": 8.801306598425059e-05,
|
| 2811 |
+
"loss": 11.8361,
|
| 2812 |
+
"step": 397
|
| 2813 |
+
},
|
| 2814 |
+
{
|
| 2815 |
+
"epoch": 0.546328071379547,
|
| 2816 |
+
"grad_norm": 0.050247903913259506,
|
| 2817 |
+
"learning_rate": 8.757939308906248e-05,
|
| 2818 |
+
"loss": 11.8424,
|
| 2819 |
+
"step": 398
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 0.547700754975978,
|
| 2823 |
+
"grad_norm": 0.04780413582921028,
|
| 2824 |
+
"learning_rate": 8.714595732255834e-05,
|
| 2825 |
+
"loss": 11.8393,
|
| 2826 |
+
"step": 399
|
| 2827 |
+
},
|
| 2828 |
+
{
|
| 2829 |
+
"epoch": 0.5490734385724091,
|
| 2830 |
+
"grad_norm": 0.053548794239759445,
|
| 2831 |
+
"learning_rate": 8.671276695970043e-05,
|
| 2832 |
+
"loss": 11.8346,
|
| 2833 |
+
"step": 400
|
| 2834 |
+
},
|
| 2835 |
+
{
|
| 2836 |
+
"epoch": 0.55044612216884,
|
| 2837 |
+
"grad_norm": 0.046638406813144684,
|
| 2838 |
+
"learning_rate": 8.627983027076573e-05,
|
| 2839 |
+
"loss": 11.8557,
|
| 2840 |
+
"step": 401
|
| 2841 |
+
},
|
| 2842 |
+
{
|
| 2843 |
+
"epoch": 0.5518188057652711,
|
| 2844 |
+
"grad_norm": 0.05789132043719292,
|
| 2845 |
+
"learning_rate": 8.584715552118842e-05,
|
| 2846 |
+
"loss": 11.8492,
|
| 2847 |
+
"step": 402
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 0.5531914893617021,
|
| 2851 |
+
"grad_norm": 0.05767640471458435,
|
| 2852 |
+
"learning_rate": 8.541475097140167e-05,
|
| 2853 |
+
"loss": 11.8533,
|
| 2854 |
+
"step": 403
|
| 2855 |
+
},
|
| 2856 |
+
{
|
| 2857 |
+
"epoch": 0.5545641729581332,
|
| 2858 |
+
"grad_norm": 0.061708979308605194,
|
| 2859 |
+
"learning_rate": 8.498262487668015e-05,
|
| 2860 |
+
"loss": 11.8448,
|
| 2861 |
+
"step": 404
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 0.5559368565545642,
|
| 2865 |
+
"grad_norm": 0.05089671537280083,
|
| 2866 |
+
"learning_rate": 8.455078548698243e-05,
|
| 2867 |
+
"loss": 11.8368,
|
| 2868 |
+
"step": 405
|
| 2869 |
+
},
|
| 2870 |
+
{
|
| 2871 |
+
"epoch": 0.5573095401509952,
|
| 2872 |
+
"grad_norm": 0.049617160111665726,
|
| 2873 |
+
"learning_rate": 8.411924104679339e-05,
|
| 2874 |
+
"loss": 11.8408,
|
| 2875 |
+
"step": 406
|
| 2876 |
+
},
|
| 2877 |
+
{
|
| 2878 |
+
"epoch": 0.5586822237474263,
|
| 2879 |
+
"grad_norm": 0.05457446724176407,
|
| 2880 |
+
"learning_rate": 8.368799979496694e-05,
|
| 2881 |
+
"loss": 11.849,
|
| 2882 |
+
"step": 407
|
| 2883 |
+
},
|
| 2884 |
+
{
|
| 2885 |
+
"epoch": 0.5600549073438572,
|
| 2886 |
+
"grad_norm": 0.060327596962451935,
|
| 2887 |
+
"learning_rate": 8.325706996456854e-05,
|
| 2888 |
+
"loss": 11.8468,
|
| 2889 |
+
"step": 408
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 0.5614275909402883,
|
| 2893 |
+
"grad_norm": 0.0542745478451252,
|
| 2894 |
+
"learning_rate": 8.282645978271823e-05,
|
| 2895 |
+
"loss": 11.85,
|
| 2896 |
+
"step": 409
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"epoch": 0.5628002745367193,
|
| 2900 |
+
"grad_norm": 0.05724021792411804,
|
| 2901 |
+
"learning_rate": 8.23961774704334e-05,
|
| 2902 |
+
"loss": 11.8392,
|
| 2903 |
+
"step": 410
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 0.5641729581331503,
|
| 2907 |
+
"grad_norm": 0.06460297852754593,
|
| 2908 |
+
"learning_rate": 8.196623124247197e-05,
|
| 2909 |
+
"loss": 11.8381,
|
| 2910 |
+
"step": 411
|
| 2911 |
+
},
|
| 2912 |
+
{
|
| 2913 |
+
"epoch": 0.5655456417295813,
|
| 2914 |
+
"grad_norm": 0.0422474630177021,
|
| 2915 |
+
"learning_rate": 8.153662930717538e-05,
|
| 2916 |
+
"loss": 11.8307,
|
| 2917 |
+
"step": 412
|
| 2918 |
+
},
|
| 2919 |
+
{
|
| 2920 |
+
"epoch": 0.5669183253260124,
|
| 2921 |
+
"grad_norm": 0.05747657269239426,
|
| 2922 |
+
"learning_rate": 8.110737986631209e-05,
|
| 2923 |
+
"loss": 11.8482,
|
| 2924 |
+
"step": 413
|
| 2925 |
+
},
|
| 2926 |
+
{
|
| 2927 |
+
"epoch": 0.5682910089224433,
|
| 2928 |
+
"grad_norm": 0.06213986501097679,
|
| 2929 |
+
"learning_rate": 8.06784911149209e-05,
|
| 2930 |
+
"loss": 11.8531,
|
| 2931 |
+
"step": 414
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 0.5696636925188744,
|
| 2935 |
+
"grad_norm": 0.05039874091744423,
|
| 2936 |
+
"learning_rate": 8.024997124115437e-05,
|
| 2937 |
+
"loss": 11.8473,
|
| 2938 |
+
"step": 415
|
| 2939 |
+
},
|
| 2940 |
+
{
|
| 2941 |
+
"epoch": 0.5710363761153054,
|
| 2942 |
+
"grad_norm": 0.059983327984809875,
|
| 2943 |
+
"learning_rate": 7.982182842612274e-05,
|
| 2944 |
+
"loss": 11.8464,
|
| 2945 |
+
"step": 416
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 0.5724090597117364,
|
| 2949 |
+
"grad_norm": 0.06043354049324989,
|
| 2950 |
+
"learning_rate": 7.939407084373751e-05,
|
| 2951 |
+
"loss": 11.8493,
|
| 2952 |
+
"step": 417
|
| 2953 |
+
},
|
| 2954 |
+
{
|
| 2955 |
+
"epoch": 0.5737817433081674,
|
| 2956 |
+
"grad_norm": 0.050205111503601074,
|
| 2957 |
+
"learning_rate": 7.89667066605556e-05,
|
| 2958 |
+
"loss": 11.8526,
|
| 2959 |
+
"step": 418
|
| 2960 |
+
},
|
| 2961 |
+
{
|
| 2962 |
+
"epoch": 0.5751544269045985,
|
| 2963 |
+
"grad_norm": 0.06297186762094498,
|
| 2964 |
+
"learning_rate": 7.853974403562323e-05,
|
| 2965 |
+
"loss": 11.8473,
|
| 2966 |
+
"step": 419
|
| 2967 |
+
},
|
| 2968 |
+
{
|
| 2969 |
+
"epoch": 0.5765271105010296,
|
| 2970 |
+
"grad_norm": 0.04371930658817291,
|
| 2971 |
+
"learning_rate": 7.811319112032027e-05,
|
| 2972 |
+
"loss": 11.8421,
|
| 2973 |
+
"step": 420
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 0.5778997940974605,
|
| 2977 |
+
"grad_norm": 0.053633809089660645,
|
| 2978 |
+
"learning_rate": 7.768705605820466e-05,
|
| 2979 |
+
"loss": 11.8411,
|
| 2980 |
+
"step": 421
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"epoch": 0.5792724776938916,
|
| 2984 |
+
"grad_norm": 0.054461780935525894,
|
| 2985 |
+
"learning_rate": 7.726134698485679e-05,
|
| 2986 |
+
"loss": 11.8522,
|
| 2987 |
+
"step": 422
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 0.5806451612903226,
|
| 2991 |
+
"grad_norm": 0.046723928302526474,
|
| 2992 |
+
"learning_rate": 7.683607202772433e-05,
|
| 2993 |
+
"loss": 11.8378,
|
| 2994 |
+
"step": 423
|
| 2995 |
+
},
|
| 2996 |
+
{
|
| 2997 |
+
"epoch": 0.5820178448867536,
|
| 2998 |
+
"grad_norm": 0.047051671892404556,
|
| 2999 |
+
"learning_rate": 7.641123930596693e-05,
|
| 3000 |
+
"loss": 11.847,
|
| 3001 |
+
"step": 424
|
| 3002 |
+
},
|
| 3003 |
+
{
|
| 3004 |
+
"epoch": 0.5833905284831846,
|
| 3005 |
+
"grad_norm": 0.04326484724879265,
|
| 3006 |
+
"learning_rate": 7.598685693030136e-05,
|
| 3007 |
+
"loss": 11.8405,
|
| 3008 |
+
"step": 425
|
| 3009 |
+
},
|
| 3010 |
+
{
|
| 3011 |
+
"epoch": 0.5847632120796157,
|
| 3012 |
+
"grad_norm": 0.05661354586482048,
|
| 3013 |
+
"learning_rate": 7.556293300284647e-05,
|
| 3014 |
+
"loss": 11.8423,
|
| 3015 |
+
"step": 426
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 0.5861358956760466,
|
| 3019 |
+
"grad_norm": 0.06154714152216911,
|
| 3020 |
+
"learning_rate": 7.513947561696877e-05,
|
| 3021 |
+
"loss": 11.8456,
|
| 3022 |
+
"step": 427
|
| 3023 |
+
},
|
| 3024 |
+
{
|
| 3025 |
+
"epoch": 0.5875085792724777,
|
| 3026 |
+
"grad_norm": 0.048859577625989914,
|
| 3027 |
+
"learning_rate": 7.471649285712768e-05,
|
| 3028 |
+
"loss": 11.8409,
|
| 3029 |
+
"step": 428
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 0.5888812628689087,
|
| 3033 |
+
"grad_norm": 0.05322885513305664,
|
| 3034 |
+
"learning_rate": 7.429399279872122e-05,
|
| 3035 |
+
"loss": 11.8399,
|
| 3036 |
+
"step": 429
|
| 3037 |
+
},
|
| 3038 |
+
{
|
| 3039 |
+
"epoch": 0.5902539464653397,
|
| 3040 |
+
"grad_norm": 0.049934811890125275,
|
| 3041 |
+
"learning_rate": 7.387198350793201e-05,
|
| 3042 |
+
"loss": 11.8503,
|
| 3043 |
+
"step": 430
|
| 3044 |
+
},
|
| 3045 |
+
{
|
| 3046 |
+
"epoch": 0.5916266300617707,
|
| 3047 |
+
"grad_norm": 0.05650179833173752,
|
| 3048 |
+
"learning_rate": 7.345047304157304e-05,
|
| 3049 |
+
"loss": 11.8447,
|
| 3050 |
+
"step": 431
|
| 3051 |
+
},
|
| 3052 |
+
{
|
| 3053 |
+
"epoch": 0.5929993136582018,
|
| 3054 |
+
"grad_norm": 0.052234046161174774,
|
| 3055 |
+
"learning_rate": 7.302946944693408e-05,
|
| 3056 |
+
"loss": 11.8406,
|
| 3057 |
+
"step": 432
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 0.5943719972546329,
|
| 3061 |
+
"grad_norm": 0.044755928218364716,
|
| 3062 |
+
"learning_rate": 7.260898076162782e-05,
|
| 3063 |
+
"loss": 11.8425,
|
| 3064 |
+
"step": 433
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 0.5957446808510638,
|
| 3068 |
+
"grad_norm": 0.04504930600523949,
|
| 3069 |
+
"learning_rate": 7.218901501343658e-05,
|
| 3070 |
+
"loss": 11.8322,
|
| 3071 |
+
"step": 434
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 0.5971173644474949,
|
| 3075 |
+
"grad_norm": 0.0498984158039093,
|
| 3076 |
+
"learning_rate": 7.176958022015902e-05,
|
| 3077 |
+
"loss": 11.845,
|
| 3078 |
+
"step": 435
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"epoch": 0.5984900480439259,
|
| 3082 |
+
"grad_norm": 0.04789600521326065,
|
| 3083 |
+
"learning_rate": 7.135068438945704e-05,
|
| 3084 |
+
"loss": 11.8435,
|
| 3085 |
+
"step": 436
|
| 3086 |
+
},
|
| 3087 |
+
{
|
| 3088 |
+
"epoch": 0.5998627316403569,
|
| 3089 |
+
"grad_norm": 0.054135389626026154,
|
| 3090 |
+
"learning_rate": 7.093233551870287e-05,
|
| 3091 |
+
"loss": 11.8375,
|
| 3092 |
+
"step": 437
|
| 3093 |
+
},
|
| 3094 |
+
{
|
| 3095 |
+
"epoch": 0.6012354152367879,
|
| 3096 |
+
"grad_norm": 0.049439262598752975,
|
| 3097 |
+
"learning_rate": 7.05145415948264e-05,
|
| 3098 |
+
"loss": 11.8466,
|
| 3099 |
+
"step": 438
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 0.602608098833219,
|
| 3103 |
+
"grad_norm": 0.050558481365442276,
|
| 3104 |
+
"learning_rate": 7.009731059416287e-05,
|
| 3105 |
+
"loss": 11.8419,
|
| 3106 |
+
"step": 439
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 0.6039807824296499,
|
| 3110 |
+
"grad_norm": 0.046476006507873535,
|
| 3111 |
+
"learning_rate": 6.968065048230028e-05,
|
| 3112 |
+
"loss": 11.8412,
|
| 3113 |
+
"step": 440
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 0.605353466026081,
|
| 3117 |
+
"grad_norm": 0.04960538074374199,
|
| 3118 |
+
"learning_rate": 6.926456921392753e-05,
|
| 3119 |
+
"loss": 11.8434,
|
| 3120 |
+
"step": 441
|
| 3121 |
+
},
|
| 3122 |
+
{
|
| 3123 |
+
"epoch": 0.606726149622512,
|
| 3124 |
+
"grad_norm": 0.06765194982290268,
|
| 3125 |
+
"learning_rate": 6.884907473268254e-05,
|
| 3126 |
+
"loss": 11.8448,
|
| 3127 |
+
"step": 442
|
| 3128 |
+
},
|
| 3129 |
+
{
|
| 3130 |
+
"epoch": 0.608098833218943,
|
| 3131 |
+
"grad_norm": 0.05157052353024483,
|
| 3132 |
+
"learning_rate": 6.843417497100046e-05,
|
| 3133 |
+
"loss": 11.8465,
|
| 3134 |
+
"step": 443
|
| 3135 |
+
},
|
| 3136 |
+
{
|
| 3137 |
+
"epoch": 0.609471516815374,
|
| 3138 |
+
"grad_norm": 0.05726620927453041,
|
| 3139 |
+
"learning_rate": 6.801987784996244e-05,
|
| 3140 |
+
"loss": 11.8322,
|
| 3141 |
+
"step": 444
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 0.6108442004118051,
|
| 3145 |
+
"grad_norm": 0.04885038360953331,
|
| 3146 |
+
"learning_rate": 6.760619127914417e-05,
|
| 3147 |
+
"loss": 11.8428,
|
| 3148 |
+
"step": 445
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 0.6122168840082362,
|
| 3152 |
+
"grad_norm": 0.06902112066745758,
|
| 3153 |
+
"learning_rate": 6.719312315646507e-05,
|
| 3154 |
+
"loss": 11.8371,
|
| 3155 |
+
"step": 446
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 0.6135895676046671,
|
| 3159 |
+
"grad_norm": 0.05013713613152504,
|
| 3160 |
+
"learning_rate": 6.678068136803736e-05,
|
| 3161 |
+
"loss": 11.8309,
|
| 3162 |
+
"step": 447
|
| 3163 |
+
},
|
| 3164 |
+
{
|
| 3165 |
+
"epoch": 0.6149622512010982,
|
| 3166 |
+
"grad_norm": 0.04684825986623764,
|
| 3167 |
+
"learning_rate": 6.636887378801563e-05,
|
| 3168 |
+
"loss": 11.8422,
|
| 3169 |
+
"step": 448
|
| 3170 |
+
},
|
| 3171 |
+
{
|
| 3172 |
+
"epoch": 0.6163349347975292,
|
| 3173 |
+
"grad_norm": 0.06438250839710236,
|
| 3174 |
+
"learning_rate": 6.595770827844637e-05,
|
| 3175 |
+
"loss": 11.8434,
|
| 3176 |
+
"step": 449
|
| 3177 |
+
},
|
| 3178 |
+
{
|
| 3179 |
+
"epoch": 0.6177076183939602,
|
| 3180 |
+
"grad_norm": 0.05015719309449196,
|
| 3181 |
+
"learning_rate": 6.554719268911804e-05,
|
| 3182 |
+
"loss": 11.8507,
|
| 3183 |
+
"step": 450
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 0.6190803019903912,
|
| 3187 |
+
"grad_norm": 0.0466039776802063,
|
| 3188 |
+
"learning_rate": 6.513733485741102e-05,
|
| 3189 |
+
"loss": 11.8532,
|
| 3190 |
+
"step": 451
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 0.6204529855868223,
|
| 3194 |
+
"grad_norm": 0.05561694875359535,
|
| 3195 |
+
"learning_rate": 6.472814260814814e-05,
|
| 3196 |
+
"loss": 11.8339,
|
| 3197 |
+
"step": 452
|
| 3198 |
+
},
|
| 3199 |
+
{
|
| 3200 |
+
"epoch": 0.6218256691832532,
|
| 3201 |
+
"grad_norm": 0.04746030271053314,
|
| 3202 |
+
"learning_rate": 6.431962375344522e-05,
|
| 3203 |
+
"loss": 11.8401,
|
| 3204 |
+
"step": 453
|
| 3205 |
+
},
|
| 3206 |
+
{
|
| 3207 |
+
"epoch": 0.6231983527796843,
|
| 3208 |
+
"grad_norm": 0.051803771406412125,
|
| 3209 |
+
"learning_rate": 6.39117860925619e-05,
|
| 3210 |
+
"loss": 11.844,
|
| 3211 |
+
"step": 454
|
| 3212 |
+
},
|
| 3213 |
+
{
|
| 3214 |
+
"epoch": 0.6245710363761153,
|
| 3215 |
+
"grad_norm": 0.04621079936623573,
|
| 3216 |
+
"learning_rate": 6.350463741175281e-05,
|
| 3217 |
+
"loss": 11.8461,
|
| 3218 |
+
"step": 455
|
| 3219 |
+
},
|
| 3220 |
+
{
|
| 3221 |
+
"epoch": 0.6259437199725463,
|
| 3222 |
+
"grad_norm": 0.05719459801912308,
|
| 3223 |
+
"learning_rate": 6.309818548411881e-05,
|
| 3224 |
+
"loss": 11.8389,
|
| 3225 |
+
"step": 456
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 0.6273164035689773,
|
| 3229 |
+
"grad_norm": 0.07033354043960571,
|
| 3230 |
+
"learning_rate": 6.269243806945878e-05,
|
| 3231 |
+
"loss": 11.8494,
|
| 3232 |
+
"step": 457
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 0.6286890871654084,
|
| 3236 |
+
"grad_norm": 0.05988188832998276,
|
| 3237 |
+
"learning_rate": 6.228740291412121e-05,
|
| 3238 |
+
"loss": 11.8504,
|
| 3239 |
+
"step": 458
|
| 3240 |
+
},
|
| 3241 |
+
{
|
| 3242 |
+
"epoch": 0.6300617707618394,
|
| 3243 |
+
"grad_norm": 0.05682065710425377,
|
| 3244 |
+
"learning_rate": 6.188308775085654e-05,
|
| 3245 |
+
"loss": 11.8448,
|
| 3246 |
+
"step": 459
|
| 3247 |
+
},
|
| 3248 |
+
{
|
| 3249 |
+
"epoch": 0.6314344543582704,
|
| 3250 |
+
"grad_norm": 0.04063468426465988,
|
| 3251 |
+
"learning_rate": 6.147950029866946e-05,
|
| 3252 |
+
"loss": 11.8406,
|
| 3253 |
+
"step": 460
|
| 3254 |
+
},
|
| 3255 |
+
{
|
| 3256 |
+
"epoch": 0.6328071379547014,
|
| 3257 |
+
"grad_norm": 0.04565683752298355,
|
| 3258 |
+
"learning_rate": 6.107664826267146e-05,
|
| 3259 |
+
"loss": 11.8468,
|
| 3260 |
+
"step": 461
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"epoch": 0.6341798215511325,
|
| 3264 |
+
"grad_norm": 0.05042997747659683,
|
| 3265 |
+
"learning_rate": 6.067453933393383e-05,
|
| 3266 |
+
"loss": 11.8417,
|
| 3267 |
+
"step": 462
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 0.6355525051475635,
|
| 3271 |
+
"grad_norm": 0.052602965384721756,
|
| 3272 |
+
"learning_rate": 6.027318118934078e-05,
|
| 3273 |
+
"loss": 11.8433,
|
| 3274 |
+
"step": 463
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 0.6369251887439945,
|
| 3278 |
+
"grad_norm": 0.043272774666547775,
|
| 3279 |
+
"learning_rate": 5.987258149144288e-05,
|
| 3280 |
+
"loss": 11.8442,
|
| 3281 |
+
"step": 464
|
| 3282 |
+
},
|
| 3283 |
+
{
|
| 3284 |
+
"epoch": 0.6382978723404256,
|
| 3285 |
+
"grad_norm": 0.04724358022212982,
|
| 3286 |
+
"learning_rate": 5.9472747888310834e-05,
|
| 3287 |
+
"loss": 11.843,
|
| 3288 |
+
"step": 465
|
| 3289 |
+
},
|
| 3290 |
+
{
|
| 3291 |
+
"epoch": 0.6396705559368565,
|
| 3292 |
+
"grad_norm": 0.06730905920267105,
|
| 3293 |
+
"learning_rate": 5.907368801338927e-05,
|
| 3294 |
+
"loss": 11.8506,
|
| 3295 |
+
"step": 466
|
| 3296 |
+
},
|
| 3297 |
+
{
|
| 3298 |
+
"epoch": 0.6410432395332876,
|
| 3299 |
+
"grad_norm": 0.05358872935175896,
|
| 3300 |
+
"learning_rate": 5.86754094853513e-05,
|
| 3301 |
+
"loss": 11.8339,
|
| 3302 |
+
"step": 467
|
| 3303 |
+
},
|
| 3304 |
+
{
|
| 3305 |
+
"epoch": 0.6424159231297186,
|
| 3306 |
+
"grad_norm": 0.048429373651742935,
|
| 3307 |
+
"learning_rate": 5.827791990795278e-05,
|
| 3308 |
+
"loss": 11.8374,
|
| 3309 |
+
"step": 468
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 0.6437886067261496,
|
| 3313 |
+
"grad_norm": 0.042113129049539566,
|
| 3314 |
+
"learning_rate": 5.7881226869887314e-05,
|
| 3315 |
+
"loss": 11.8472,
|
| 3316 |
+
"step": 469
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 0.6451612903225806,
|
| 3320 |
+
"grad_norm": 0.04021785780787468,
|
| 3321 |
+
"learning_rate": 5.748533794464142e-05,
|
| 3322 |
+
"loss": 11.8391,
|
| 3323 |
+
"step": 470
|
| 3324 |
+
},
|
| 3325 |
+
{
|
| 3326 |
+
"epoch": 0.6465339739190117,
|
| 3327 |
+
"grad_norm": 0.04441021382808685,
|
| 3328 |
+
"learning_rate": 5.709026069034973e-05,
|
| 3329 |
+
"loss": 11.85,
|
| 3330 |
+
"step": 471
|
| 3331 |
+
},
|
| 3332 |
+
{
|
| 3333 |
+
"epoch": 0.6479066575154427,
|
| 3334 |
+
"grad_norm": 0.04287193715572357,
|
| 3335 |
+
"learning_rate": 5.669600264965088e-05,
|
| 3336 |
+
"loss": 11.8451,
|
| 3337 |
+
"step": 472
|
| 3338 |
+
},
|
| 3339 |
+
{
|
| 3340 |
+
"epoch": 0.6492793411118737,
|
| 3341 |
+
"grad_norm": 0.05335507541894913,
|
| 3342 |
+
"learning_rate": 5.630257134954338e-05,
|
| 3343 |
+
"loss": 11.837,
|
| 3344 |
+
"step": 473
|
| 3345 |
+
},
|
| 3346 |
+
{
|
| 3347 |
+
"epoch": 0.6506520247083047,
|
| 3348 |
+
"grad_norm": 0.052487812936306,
|
| 3349 |
+
"learning_rate": 5.59099743012421e-05,
|
| 3350 |
+
"loss": 11.8472,
|
| 3351 |
+
"step": 474
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 0.6520247083047358,
|
| 3355 |
+
"grad_norm": 0.048992034047842026,
|
| 3356 |
+
"learning_rate": 5.551821900003461e-05,
|
| 3357 |
+
"loss": 11.8343,
|
| 3358 |
+
"step": 475
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 0.6533973919011667,
|
| 3362 |
+
"grad_norm": 0.054023195058107376,
|
| 3363 |
+
"learning_rate": 5.512731292513835e-05,
|
| 3364 |
+
"loss": 11.8395,
|
| 3365 |
+
"step": 476
|
| 3366 |
+
},
|
| 3367 |
+
{
|
| 3368 |
+
"epoch": 0.6547700754975978,
|
| 3369 |
+
"grad_norm": 0.04266708716750145,
|
| 3370 |
+
"learning_rate": 5.473726353955764e-05,
|
| 3371 |
+
"loss": 11.8343,
|
| 3372 |
+
"step": 477
|
| 3373 |
+
},
|
| 3374 |
+
{
|
| 3375 |
+
"epoch": 0.6561427590940289,
|
| 3376 |
+
"grad_norm": 0.039765555411577225,
|
| 3377 |
+
"learning_rate": 5.434807828994123e-05,
|
| 3378 |
+
"loss": 11.8335,
|
| 3379 |
+
"step": 478
|
| 3380 |
+
},
|
| 3381 |
+
{
|
| 3382 |
+
"epoch": 0.6575154426904598,
|
| 3383 |
+
"grad_norm": 0.058559659868478775,
|
| 3384 |
+
"learning_rate": 5.39597646064402e-05,
|
| 3385 |
+
"loss": 11.8384,
|
| 3386 |
+
"step": 479
|
| 3387 |
+
},
|
| 3388 |
+
{
|
| 3389 |
+
"epoch": 0.6588881262868909,
|
| 3390 |
+
"grad_norm": 0.046752121299505234,
|
| 3391 |
+
"learning_rate": 5.357232990256618e-05,
|
| 3392 |
+
"loss": 11.8371,
|
| 3393 |
+
"step": 480
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 0.6602608098833219,
|
| 3397 |
+
"grad_norm": 0.056967172771692276,
|
| 3398 |
+
"learning_rate": 5.318578157504959e-05,
|
| 3399 |
+
"loss": 11.8564,
|
| 3400 |
+
"step": 481
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 0.6616334934797529,
|
| 3404 |
+
"grad_norm": 0.04283031076192856,
|
| 3405 |
+
"learning_rate": 5.280012700369856e-05,
|
| 3406 |
+
"loss": 11.8438,
|
| 3407 |
+
"step": 482
|
| 3408 |
+
},
|
| 3409 |
+
{
|
| 3410 |
+
"epoch": 0.6630061770761839,
|
| 3411 |
+
"grad_norm": 0.05448300763964653,
|
| 3412 |
+
"learning_rate": 5.2415373551258054e-05,
|
| 3413 |
+
"loss": 11.8302,
|
| 3414 |
+
"step": 483
|
| 3415 |
+
},
|
| 3416 |
+
{
|
| 3417 |
+
"epoch": 0.664378860672615,
|
| 3418 |
+
"grad_norm": 0.04729095473885536,
|
| 3419 |
+
"learning_rate": 5.203152856326933e-05,
|
| 3420 |
+
"loss": 11.8415,
|
| 3421 |
+
"step": 484
|
| 3422 |
+
},
|
| 3423 |
+
{
|
| 3424 |
+
"epoch": 0.665751544269046,
|
| 3425 |
+
"grad_norm": 0.06074047461152077,
|
| 3426 |
+
"learning_rate": 5.164859936792955e-05,
|
| 3427 |
+
"loss": 11.8346,
|
| 3428 |
+
"step": 485
|
| 3429 |
+
},
|
| 3430 |
+
{
|
| 3431 |
+
"epoch": 0.667124227865477,
|
| 3432 |
+
"grad_norm": 0.0658525601029396,
|
| 3433 |
+
"learning_rate": 5.1266593275952034e-05,
|
| 3434 |
+
"loss": 11.84,
|
| 3435 |
+
"step": 486
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 0.668496911461908,
|
| 3439 |
+
"grad_norm": 0.055728498846292496,
|
| 3440 |
+
"learning_rate": 5.088551758042655e-05,
|
| 3441 |
+
"loss": 11.8355,
|
| 3442 |
+
"step": 487
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 0.6698695950583391,
|
| 3446 |
+
"grad_norm": 0.07013114541769028,
|
| 3447 |
+
"learning_rate": 5.050537955668027e-05,
|
| 3448 |
+
"loss": 11.8376,
|
| 3449 |
+
"step": 488
|
| 3450 |
+
},
|
| 3451 |
+
{
|
| 3452 |
+
"epoch": 0.67124227865477,
|
| 3453 |
+
"grad_norm": 0.05220917984843254,
|
| 3454 |
+
"learning_rate": 5.012618646213869e-05,
|
| 3455 |
+
"loss": 11.8395,
|
| 3456 |
+
"step": 489
|
| 3457 |
+
},
|
| 3458 |
+
{
|
| 3459 |
+
"epoch": 0.6726149622512011,
|
| 3460 |
+
"grad_norm": 0.05545530095696449,
|
| 3461 |
+
"learning_rate": 4.9747945536187145e-05,
|
| 3462 |
+
"loss": 11.8468,
|
| 3463 |
+
"step": 490
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 0.6739876458476322,
|
| 3467 |
+
"grad_norm": 0.042175162583589554,
|
| 3468 |
+
"learning_rate": 4.937066400003256e-05,
|
| 3469 |
+
"loss": 11.8492,
|
| 3470 |
+
"step": 491
|
| 3471 |
+
},
|
| 3472 |
+
{
|
| 3473 |
+
"epoch": 0.6753603294440631,
|
| 3474 |
+
"grad_norm": 0.044837336987257004,
|
| 3475 |
+
"learning_rate": 4.899434905656563e-05,
|
| 3476 |
+
"loss": 11.8396,
|
| 3477 |
+
"step": 492
|
| 3478 |
+
},
|
| 3479 |
+
{
|
| 3480 |
+
"epoch": 0.6767330130404942,
|
| 3481 |
+
"grad_norm": 0.05794432386755943,
|
| 3482 |
+
"learning_rate": 4.861900789022339e-05,
|
| 3483 |
+
"loss": 11.839,
|
| 3484 |
+
"step": 493
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 0.6781056966369252,
|
| 3488 |
+
"grad_norm": 0.051717609167099,
|
| 3489 |
+
"learning_rate": 4.824464766685182e-05,
|
| 3490 |
+
"loss": 11.8364,
|
| 3491 |
+
"step": 494
|
| 3492 |
+
},
|
| 3493 |
+
{
|
| 3494 |
+
"epoch": 0.6794783802333562,
|
| 3495 |
+
"grad_norm": 0.04744836688041687,
|
| 3496 |
+
"learning_rate": 4.7871275533569223e-05,
|
| 3497 |
+
"loss": 11.8452,
|
| 3498 |
+
"step": 495
|
| 3499 |
+
},
|
| 3500 |
+
{
|
| 3501 |
+
"epoch": 0.6808510638297872,
|
| 3502 |
+
"grad_norm": 0.05374274402856827,
|
| 3503 |
+
"learning_rate": 4.749889861862977e-05,
|
| 3504 |
+
"loss": 11.8328,
|
| 3505 |
+
"step": 496
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 0.6822237474262183,
|
| 3509 |
+
"grad_norm": 0.07115243375301361,
|
| 3510 |
+
"learning_rate": 4.712752403128729e-05,
|
| 3511 |
+
"loss": 11.8418,
|
| 3512 |
+
"step": 497
|
| 3513 |
+
},
|
| 3514 |
+
{
|
| 3515 |
+
"epoch": 0.6835964310226493,
|
| 3516 |
+
"grad_norm": 0.04187103733420372,
|
| 3517 |
+
"learning_rate": 4.6757158861659714e-05,
|
| 3518 |
+
"loss": 11.8354,
|
| 3519 |
+
"step": 498
|
| 3520 |
+
},
|
| 3521 |
+
{
|
| 3522 |
+
"epoch": 0.6849691146190803,
|
| 3523 |
+
"grad_norm": 0.0765848234295845,
|
| 3524 |
+
"learning_rate": 4.638781018059356e-05,
|
| 3525 |
+
"loss": 11.8432,
|
| 3526 |
+
"step": 499
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 0.6863417982155113,
|
| 3530 |
+
"grad_norm": 0.05868866667151451,
|
| 3531 |
+
"learning_rate": 4.601948503952896e-05,
|
| 3532 |
+
"loss": 11.8356,
|
| 3533 |
+
"step": 500
|
| 3534 |
+
},
|
| 3535 |
+
{
|
| 3536 |
+
"epoch": 0.6877144818119424,
|
| 3537 |
+
"grad_norm": 0.051075566560029984,
|
| 3538 |
+
"learning_rate": 4.565219047036521e-05,
|
| 3539 |
+
"loss": 11.8426,
|
| 3540 |
+
"step": 501
|
| 3541 |
+
},
|
| 3542 |
+
{
|
| 3543 |
+
"epoch": 0.6890871654083733,
|
| 3544 |
+
"grad_norm": 0.04912392050027847,
|
| 3545 |
+
"learning_rate": 4.5285933485326226e-05,
|
| 3546 |
+
"loss": 11.8339,
|
| 3547 |
+
"step": 502
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 0.6904598490048044,
|
| 3551 |
+
"grad_norm": 0.058397840708494186,
|
| 3552 |
+
"learning_rate": 4.492072107682699e-05,
|
| 3553 |
+
"loss": 11.8503,
|
| 3554 |
+
"step": 503
|
| 3555 |
+
},
|
| 3556 |
+
{
|
| 3557 |
+
"epoch": 0.6918325326012355,
|
| 3558 |
+
"grad_norm": 0.05081508681178093,
|
| 3559 |
+
"learning_rate": 4.4556560217339774e-05,
|
| 3560 |
+
"loss": 11.8387,
|
| 3561 |
+
"step": 504
|
| 3562 |
+
},
|
| 3563 |
+
{
|
| 3564 |
+
"epoch": 0.6932052161976664,
|
| 3565 |
+
"grad_norm": 0.05394862964749336,
|
| 3566 |
+
"learning_rate": 4.419345785926119e-05,
|
| 3567 |
+
"loss": 11.8442,
|
| 3568 |
+
"step": 505
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 0.6945778997940975,
|
| 3572 |
+
"grad_norm": 0.050071366131305695,
|
| 3573 |
+
"learning_rate": 4.383142093477938e-05,
|
| 3574 |
+
"loss": 11.8376,
|
| 3575 |
+
"step": 506
|
| 3576 |
+
},
|
| 3577 |
+
{
|
| 3578 |
+
"epoch": 0.6959505833905285,
|
| 3579 |
+
"grad_norm": 0.03961324319243431,
|
| 3580 |
+
"learning_rate": 4.3470456355741786e-05,
|
| 3581 |
+
"loss": 11.8482,
|
| 3582 |
+
"step": 507
|
| 3583 |
+
},
|
| 3584 |
+
{
|
| 3585 |
+
"epoch": 0.6973232669869595,
|
| 3586 |
+
"grad_norm": 0.07291043549776077,
|
| 3587 |
+
"learning_rate": 4.311057101352304e-05,
|
| 3588 |
+
"loss": 11.8388,
|
| 3589 |
+
"step": 508
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 0.6986959505833905,
|
| 3593 |
+
"grad_norm": 0.057433102279901505,
|
| 3594 |
+
"learning_rate": 4.275177177889347e-05,
|
| 3595 |
+
"loss": 11.8364,
|
| 3596 |
+
"step": 509
|
| 3597 |
+
},
|
| 3598 |
+
{
|
| 3599 |
+
"epoch": 0.7000686341798216,
|
| 3600 |
+
"grad_norm": 0.04764627292752266,
|
| 3601 |
+
"learning_rate": 4.239406550188791e-05,
|
| 3602 |
+
"loss": 11.8416,
|
| 3603 |
+
"step": 510
|
| 3604 |
+
},
|
| 3605 |
+
{
|
| 3606 |
+
"epoch": 0.7014413177762526,
|
| 3607 |
+
"grad_norm": 0.05604006350040436,
|
| 3608 |
+
"learning_rate": 4.203745901167504e-05,
|
| 3609 |
+
"loss": 11.8492,
|
| 3610 |
+
"step": 511
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 0.7028140013726836,
|
| 3614 |
+
"grad_norm": 0.05157355219125748,
|
| 3615 |
+
"learning_rate": 4.168195911642677e-05,
|
| 3616 |
+
"loss": 11.8424,
|
| 3617 |
+
"step": 512
|
| 3618 |
+
},
|
| 3619 |
+
{
|
| 3620 |
+
"epoch": 0.7041866849691146,
|
| 3621 |
+
"grad_norm": 0.05089060217142105,
|
| 3622 |
+
"learning_rate": 4.1327572603188415e-05,
|
| 3623 |
+
"loss": 11.8404,
|
| 3624 |
+
"step": 513
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"epoch": 0.7055593685655457,
|
| 3628 |
+
"grad_norm": 0.057208992540836334,
|
| 3629 |
+
"learning_rate": 4.097430623774921e-05,
|
| 3630 |
+
"loss": 11.8366,
|
| 3631 |
+
"step": 514
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 0.7069320521619766,
|
| 3635 |
+
"grad_norm": 0.06669998168945312,
|
| 3636 |
+
"learning_rate": 4.062216676451285e-05,
|
| 3637 |
+
"loss": 11.8294,
|
| 3638 |
+
"step": 515
|
| 3639 |
+
},
|
| 3640 |
+
{
|
| 3641 |
+
"epoch": 0.7083047357584077,
|
| 3642 |
+
"grad_norm": 0.056854214519262314,
|
| 3643 |
+
"learning_rate": 4.027116090636909e-05,
|
| 3644 |
+
"loss": 11.8429,
|
| 3645 |
+
"step": 516
|
| 3646 |
+
},
|
| 3647 |
+
{
|
| 3648 |
+
"epoch": 0.7096774193548387,
|
| 3649 |
+
"grad_norm": 0.055773284286260605,
|
| 3650 |
+
"learning_rate": 3.9921295364565094e-05,
|
| 3651 |
+
"loss": 11.843,
|
| 3652 |
+
"step": 517
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 0.7110501029512697,
|
| 3656 |
+
"grad_norm": 0.0448710061609745,
|
| 3657 |
+
"learning_rate": 3.9572576818577637e-05,
|
| 3658 |
+
"loss": 11.8395,
|
| 3659 |
+
"step": 518
|
| 3660 |
+
},
|
| 3661 |
+
{
|
| 3662 |
+
"epoch": 0.7124227865477007,
|
| 3663 |
+
"grad_norm": 0.0698017105460167,
|
| 3664 |
+
"learning_rate": 3.922501192598557e-05,
|
| 3665 |
+
"loss": 11.8371,
|
| 3666 |
+
"step": 519
|
| 3667 |
+
},
|
| 3668 |
+
{
|
| 3669 |
+
"epoch": 0.7137954701441318,
|
| 3670 |
+
"grad_norm": 0.044730111956596375,
|
| 3671 |
+
"learning_rate": 3.8878607322342674e-05,
|
| 3672 |
+
"loss": 11.836,
|
| 3673 |
+
"step": 520
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 0.7151681537405628,
|
| 3677 |
+
"grad_norm": 0.060508184134960175,
|
| 3678 |
+
"learning_rate": 3.853336962105112e-05,
|
| 3679 |
+
"loss": 11.8392,
|
| 3680 |
+
"step": 521
|
| 3681 |
+
},
|
| 3682 |
+
{
|
| 3683 |
+
"epoch": 0.7165408373369938,
|
| 3684 |
+
"grad_norm": 0.07923214137554169,
|
| 3685 |
+
"learning_rate": 3.8189305413234975e-05,
|
| 3686 |
+
"loss": 11.8567,
|
| 3687 |
+
"step": 522
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 0.7179135209334249,
|
| 3691 |
+
"grad_norm": 0.04985347017645836,
|
| 3692 |
+
"learning_rate": 3.784642126761457e-05,
|
| 3693 |
+
"loss": 11.8479,
|
| 3694 |
+
"step": 523
|
| 3695 |
+
},
|
| 3696 |
+
{
|
| 3697 |
+
"epoch": 0.7192862045298559,
|
| 3698 |
+
"grad_norm": 0.06870623677968979,
|
| 3699 |
+
"learning_rate": 3.750472373038093e-05,
|
| 3700 |
+
"loss": 11.8393,
|
| 3701 |
+
"step": 524
|
| 3702 |
+
},
|
| 3703 |
+
{
|
| 3704 |
+
"epoch": 0.7206588881262869,
|
| 3705 |
+
"grad_norm": 0.04539593681693077,
|
| 3706 |
+
"learning_rate": 3.7164219325070995e-05,
|
| 3707 |
+
"loss": 11.8419,
|
| 3708 |
+
"step": 525
|
| 3709 |
+
},
|
| 3710 |
+
{
|
| 3711 |
+
"epoch": 0.7220315717227179,
|
| 3712 |
+
"grad_norm": 0.06020331010222435,
|
| 3713 |
+
"learning_rate": 3.682491455244294e-05,
|
| 3714 |
+
"loss": 11.8403,
|
| 3715 |
+
"step": 526
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 0.723404255319149,
|
| 3719 |
+
"grad_norm": 0.04615195468068123,
|
| 3720 |
+
"learning_rate": 3.6486815890352e-05,
|
| 3721 |
+
"loss": 11.8389,
|
| 3722 |
+
"step": 527
|
| 3723 |
+
},
|
| 3724 |
+
{
|
| 3725 |
+
"epoch": 0.7247769389155799,
|
| 3726 |
+
"grad_norm": 0.050100456923246384,
|
| 3727 |
+
"learning_rate": 3.6149929793626954e-05,
|
| 3728 |
+
"loss": 11.8412,
|
| 3729 |
+
"step": 528
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 0.726149622512011,
|
| 3733 |
+
"grad_norm": 0.06047607213258743,
|
| 3734 |
+
"learning_rate": 3.581426269394679e-05,
|
| 3735 |
+
"loss": 11.8482,
|
| 3736 |
+
"step": 529
|
| 3737 |
+
},
|
| 3738 |
+
{
|
| 3739 |
+
"epoch": 0.727522306108442,
|
| 3740 |
+
"grad_norm": 0.04392620921134949,
|
| 3741 |
+
"learning_rate": 3.5479820999718036e-05,
|
| 3742 |
+
"loss": 11.839,
|
| 3743 |
+
"step": 530
|
| 3744 |
+
},
|
| 3745 |
+
{
|
| 3746 |
+
"epoch": 0.728894989704873,
|
| 3747 |
+
"grad_norm": 0.0495496466755867,
|
| 3748 |
+
"learning_rate": 3.514661109595222e-05,
|
| 3749 |
+
"loss": 11.841,
|
| 3750 |
+
"step": 531
|
| 3751 |
+
},
|
| 3752 |
+
{
|
| 3753 |
+
"epoch": 0.730267673301304,
|
| 3754 |
+
"grad_norm": 0.051535751670598984,
|
| 3755 |
+
"learning_rate": 3.481463934414412e-05,
|
| 3756 |
+
"loss": 11.8436,
|
| 3757 |
+
"step": 532
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 0.7316403568977351,
|
| 3761 |
+
"grad_norm": 0.05691013112664223,
|
| 3762 |
+
"learning_rate": 3.4483912082150225e-05,
|
| 3763 |
+
"loss": 11.8418,
|
| 3764 |
+
"step": 533
|
| 3765 |
+
},
|
| 3766 |
+
{
|
| 3767 |
+
"epoch": 0.733013040494166,
|
| 3768 |
+
"grad_norm": 0.07426922768354416,
|
| 3769 |
+
"learning_rate": 3.4154435624067904e-05,
|
| 3770 |
+
"loss": 11.8454,
|
| 3771 |
+
"step": 534
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 0.7343857240905971,
|
| 3775 |
+
"grad_norm": 0.042437512427568436,
|
| 3776 |
+
"learning_rate": 3.3826216260114604e-05,
|
| 3777 |
+
"loss": 11.8413,
|
| 3778 |
+
"step": 535
|
| 3779 |
+
},
|
| 3780 |
+
{
|
| 3781 |
+
"epoch": 0.7357584076870282,
|
| 3782 |
+
"grad_norm": 0.06647706031799316,
|
| 3783 |
+
"learning_rate": 3.349926025650796e-05,
|
| 3784 |
+
"loss": 11.8375,
|
| 3785 |
+
"step": 536
|
| 3786 |
+
},
|
| 3787 |
+
{
|
| 3788 |
+
"epoch": 0.7371310912834592,
|
| 3789 |
+
"grad_norm": 0.051694292575120926,
|
| 3790 |
+
"learning_rate": 3.3173573855346043e-05,
|
| 3791 |
+
"loss": 11.8425,
|
| 3792 |
+
"step": 537
|
| 3793 |
+
},
|
| 3794 |
+
{
|
| 3795 |
+
"epoch": 0.7385037748798902,
|
| 3796 |
+
"grad_norm": 0.05568219721317291,
|
| 3797 |
+
"learning_rate": 3.2849163274488314e-05,
|
| 3798 |
+
"loss": 11.8442,
|
| 3799 |
+
"step": 538
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 0.7398764584763212,
|
| 3803 |
+
"grad_norm": 0.05254996940493584,
|
| 3804 |
+
"learning_rate": 3.252603470743676e-05,
|
| 3805 |
+
"loss": 11.8468,
|
| 3806 |
+
"step": 539
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"epoch": 0.7412491420727523,
|
| 3810 |
+
"grad_norm": 0.05059533566236496,
|
| 3811 |
+
"learning_rate": 3.220419432321783e-05,
|
| 3812 |
+
"loss": 11.8336,
|
| 3813 |
+
"step": 540
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 0.7426218256691832,
|
| 3817 |
+
"grad_norm": 0.04756971076130867,
|
| 3818 |
+
"learning_rate": 3.188364826626449e-05,
|
| 3819 |
+
"loss": 11.8369,
|
| 3820 |
+
"step": 541
|
| 3821 |
+
},
|
| 3822 |
+
{
|
| 3823 |
+
"epoch": 0.7439945092656143,
|
| 3824 |
+
"grad_norm": 0.053694888949394226,
|
| 3825 |
+
"learning_rate": 3.156440265629899e-05,
|
| 3826 |
+
"loss": 11.8314,
|
| 3827 |
+
"step": 542
|
| 3828 |
+
},
|
| 3829 |
+
{
|
| 3830 |
+
"epoch": 0.7453671928620453,
|
| 3831 |
+
"grad_norm": 0.04741639643907547,
|
| 3832 |
+
"learning_rate": 3.124646358821599e-05,
|
| 3833 |
+
"loss": 11.8371,
|
| 3834 |
+
"step": 543
|
| 3835 |
+
},
|
| 3836 |
+
{
|
| 3837 |
+
"epoch": 0.7467398764584763,
|
| 3838 |
+
"grad_norm": 0.04849841818213463,
|
| 3839 |
+
"learning_rate": 3.092983713196639e-05,
|
| 3840 |
+
"loss": 11.8434,
|
| 3841 |
+
"step": 544
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 0.7481125600549073,
|
| 3845 |
+
"grad_norm": 0.05285770818591118,
|
| 3846 |
+
"learning_rate": 3.061452933244112e-05,
|
| 3847 |
+
"loss": 11.8268,
|
| 3848 |
+
"step": 545
|
| 3849 |
+
},
|
| 3850 |
+
{
|
| 3851 |
+
"epoch": 0.7494852436513384,
|
| 3852 |
+
"grad_norm": 0.05604279786348343,
|
| 3853 |
+
"learning_rate": 3.0300546209355995e-05,
|
| 3854 |
+
"loss": 11.8241,
|
| 3855 |
+
"step": 546
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 0.7508579272477693,
|
| 3859 |
+
"grad_norm": 0.06074153259396553,
|
| 3860 |
+
"learning_rate": 2.998789375713663e-05,
|
| 3861 |
+
"loss": 11.8336,
|
| 3862 |
+
"step": 547
|
| 3863 |
+
},
|
| 3864 |
+
{
|
| 3865 |
+
"epoch": 0.7522306108442004,
|
| 3866 |
+
"grad_norm": 0.05577109381556511,
|
| 3867 |
+
"learning_rate": 2.9676577944804175e-05,
|
| 3868 |
+
"loss": 11.8356,
|
| 3869 |
+
"step": 548
|
| 3870 |
+
},
|
| 3871 |
+
{
|
| 3872 |
+
"epoch": 0.7536032944406315,
|
| 3873 |
+
"grad_norm": 0.06328094750642776,
|
| 3874 |
+
"learning_rate": 2.9366604715861156e-05,
|
| 3875 |
+
"loss": 11.8373,
|
| 3876 |
+
"step": 549
|
| 3877 |
+
},
|
| 3878 |
+
{
|
| 3879 |
+
"epoch": 0.7536032944406315,
|
| 3880 |
+
"eval_loss": 11.839405059814453,
|
| 3881 |
+
"eval_runtime": 4.8521,
|
| 3882 |
+
"eval_samples_per_second": 63.271,
|
| 3883 |
+
"eval_steps_per_second": 31.739,
|
| 3884 |
+
"step": 549
|
| 3885 |
}
|
| 3886 |
],
|
| 3887 |
"logging_steps": 1,
|
|
|
|
| 3901 |
"attributes": {}
|
| 3902 |
}
|
| 3903 |
},
|
| 3904 |
+
"total_flos": 120838054477824.0,
|
| 3905 |
"train_batch_size": 2,
|
| 3906 |
"trial_name": null,
|
| 3907 |
"trial_params": null
|