Training in progress, step 725, 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 83945296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a9728ad149b17d6e388d14fc0092f7866b47970310ba1ee971994a0c412caa0
|
| 3 |
size 83945296
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 43123028
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8a6be952359022b7f3bd188ecda8b1550bad03314ad34d49b5081a5290540fb3
|
| 3 |
size 43123028
|
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 sha256:9cf6ba00f59cf81a848257ca605964ca94bd8077ad494096d04d397120668e04
|
| 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:cb0dd6f595260943e4a46a32458e4a547c9bb110c47e094cd85eeb5c578a29ae
|
| 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":
|
| 5 |
"eval_steps": 182,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -3861,6 +3861,1259 @@
|
|
| 3861 |
"eval_samples_per_second": 31.469,
|
| 3862 |
"eval_steps_per_second": 15.786,
|
| 3863 |
"step": 546
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3864 |
}
|
| 3865 |
],
|
| 3866 |
"logging_steps": 1,
|
|
@@ -3875,12 +5128,12 @@
|
|
| 3875 |
"should_evaluate": false,
|
| 3876 |
"should_log": false,
|
| 3877 |
"should_save": true,
|
| 3878 |
-
"should_training_stop":
|
| 3879 |
},
|
| 3880 |
"attributes": {}
|
| 3881 |
}
|
| 3882 |
},
|
| 3883 |
-
"total_flos":
|
| 3884 |
"train_batch_size": 2,
|
| 3885 |
"trial_name": null,
|
| 3886 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0006901311249137,
|
| 5 |
"eval_steps": 182,
|
| 6 |
+
"global_step": 725,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 3861 |
"eval_samples_per_second": 31.469,
|
| 3862 |
"eval_steps_per_second": 15.786,
|
| 3863 |
"step": 546
|
| 3864 |
+
},
|
| 3865 |
+
{
|
| 3866 |
+
"epoch": 0.7550034506556246,
|
| 3867 |
+
"grad_norm": 19.898162841796875,
|
| 3868 |
+
"learning_rate": 2.9056687908446102e-05,
|
| 3869 |
+
"loss": 4.1283,
|
| 3870 |
+
"step": 547
|
| 3871 |
+
},
|
| 3872 |
+
{
|
| 3873 |
+
"epoch": 0.756383712905452,
|
| 3874 |
+
"grad_norm": 20.785324096679688,
|
| 3875 |
+
"learning_rate": 2.8747708127098593e-05,
|
| 3876 |
+
"loss": 5.1127,
|
| 3877 |
+
"step": 548
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 0.7577639751552795,
|
| 3881 |
+
"grad_norm": 18.006441116333008,
|
| 3882 |
+
"learning_rate": 2.84401039255879e-05,
|
| 3883 |
+
"loss": 4.9055,
|
| 3884 |
+
"step": 549
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"epoch": 0.759144237405107,
|
| 3888 |
+
"grad_norm": 24.335613250732422,
|
| 3889 |
+
"learning_rate": 2.813388124244778e-05,
|
| 3890 |
+
"loss": 6.7394,
|
| 3891 |
+
"step": 550
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 0.7605244996549344,
|
| 3895 |
+
"grad_norm": 22.286670684814453,
|
| 3896 |
+
"learning_rate": 2.7829045989540593e-05,
|
| 3897 |
+
"loss": 5.5805,
|
| 3898 |
+
"step": 551
|
| 3899 |
+
},
|
| 3900 |
+
{
|
| 3901 |
+
"epoch": 0.7619047619047619,
|
| 3902 |
+
"grad_norm": 15.699090003967285,
|
| 3903 |
+
"learning_rate": 2.752560405194351e-05,
|
| 3904 |
+
"loss": 2.9697,
|
| 3905 |
+
"step": 552
|
| 3906 |
+
},
|
| 3907 |
+
{
|
| 3908 |
+
"epoch": 0.7632850241545893,
|
| 3909 |
+
"grad_norm": 19.606931686401367,
|
| 3910 |
+
"learning_rate": 2.7223561287834465e-05,
|
| 3911 |
+
"loss": 3.3231,
|
| 3912 |
+
"step": 553
|
| 3913 |
+
},
|
| 3914 |
+
{
|
| 3915 |
+
"epoch": 0.7646652864044169,
|
| 3916 |
+
"grad_norm": 17.65702247619629,
|
| 3917 |
+
"learning_rate": 2.692292352837942e-05,
|
| 3918 |
+
"loss": 6.1613,
|
| 3919 |
+
"step": 554
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 0.7660455486542443,
|
| 3923 |
+
"grad_norm": 21.681224822998047,
|
| 3924 |
+
"learning_rate": 2.6623696577619627e-05,
|
| 3925 |
+
"loss": 3.443,
|
| 3926 |
+
"step": 555
|
| 3927 |
+
},
|
| 3928 |
+
{
|
| 3929 |
+
"epoch": 0.7674258109040718,
|
| 3930 |
+
"grad_norm": 22.102741241455078,
|
| 3931 |
+
"learning_rate": 2.6325886212359498e-05,
|
| 3932 |
+
"loss": 4.5611,
|
| 3933 |
+
"step": 556
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 0.7688060731538993,
|
| 3937 |
+
"grad_norm": 21.978052139282227,
|
| 3938 |
+
"learning_rate": 2.602949818205539e-05,
|
| 3939 |
+
"loss": 5.8679,
|
| 3940 |
+
"step": 557
|
| 3941 |
+
},
|
| 3942 |
+
{
|
| 3943 |
+
"epoch": 0.7701863354037267,
|
| 3944 |
+
"grad_norm": 19.260032653808594,
|
| 3945 |
+
"learning_rate": 2.5734538208704195e-05,
|
| 3946 |
+
"loss": 3.9815,
|
| 3947 |
+
"step": 558
|
| 3948 |
+
},
|
| 3949 |
+
{
|
| 3950 |
+
"epoch": 0.7715665976535542,
|
| 3951 |
+
"grad_norm": 17.845951080322266,
|
| 3952 |
+
"learning_rate": 2.5441011986733166e-05,
|
| 3953 |
+
"loss": 3.3853,
|
| 3954 |
+
"step": 559
|
| 3955 |
+
},
|
| 3956 |
+
{
|
| 3957 |
+
"epoch": 0.7729468599033816,
|
| 3958 |
+
"grad_norm": 23.304794311523438,
|
| 3959 |
+
"learning_rate": 2.514892518288988e-05,
|
| 3960 |
+
"loss": 5.8599,
|
| 3961 |
+
"step": 560
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 0.7743271221532091,
|
| 3965 |
+
"grad_norm": 19.359304428100586,
|
| 3966 |
+
"learning_rate": 2.485828343613288e-05,
|
| 3967 |
+
"loss": 4.678,
|
| 3968 |
+
"step": 561
|
| 3969 |
+
},
|
| 3970 |
+
{
|
| 3971 |
+
"epoch": 0.7757073844030365,
|
| 3972 |
+
"grad_norm": 22.408767700195312,
|
| 3973 |
+
"learning_rate": 2.456909235752276e-05,
|
| 3974 |
+
"loss": 5.0614,
|
| 3975 |
+
"step": 562
|
| 3976 |
+
},
|
| 3977 |
+
{
|
| 3978 |
+
"epoch": 0.777087646652864,
|
| 3979 |
+
"grad_norm": 21.053815841674805,
|
| 3980 |
+
"learning_rate": 2.42813575301138e-05,
|
| 3981 |
+
"loss": 4.851,
|
| 3982 |
+
"step": 563
|
| 3983 |
+
},
|
| 3984 |
+
{
|
| 3985 |
+
"epoch": 0.7784679089026915,
|
| 3986 |
+
"grad_norm": 19.501859664916992,
|
| 3987 |
+
"learning_rate": 2.399508450884631e-05,
|
| 3988 |
+
"loss": 3.8913,
|
| 3989 |
+
"step": 564
|
| 3990 |
+
},
|
| 3991 |
+
{
|
| 3992 |
+
"epoch": 0.779848171152519,
|
| 3993 |
+
"grad_norm": 26.093942642211914,
|
| 3994 |
+
"learning_rate": 2.3710278820439313e-05,
|
| 3995 |
+
"loss": 4.5444,
|
| 3996 |
+
"step": 565
|
| 3997 |
+
},
|
| 3998 |
+
{
|
| 3999 |
+
"epoch": 0.7812284334023465,
|
| 4000 |
+
"grad_norm": 16.253053665161133,
|
| 4001 |
+
"learning_rate": 2.3426945963283853e-05,
|
| 4002 |
+
"loss": 3.8306,
|
| 4003 |
+
"step": 566
|
| 4004 |
+
},
|
| 4005 |
+
{
|
| 4006 |
+
"epoch": 0.782608695652174,
|
| 4007 |
+
"grad_norm": 22.449371337890625,
|
| 4008 |
+
"learning_rate": 2.3145091407336784e-05,
|
| 4009 |
+
"loss": 3.9088,
|
| 4010 |
+
"step": 567
|
| 4011 |
+
},
|
| 4012 |
+
{
|
| 4013 |
+
"epoch": 0.7839889579020014,
|
| 4014 |
+
"grad_norm": 18.785884857177734,
|
| 4015 |
+
"learning_rate": 2.2864720594015288e-05,
|
| 4016 |
+
"loss": 3.6168,
|
| 4017 |
+
"step": 568
|
| 4018 |
+
},
|
| 4019 |
+
{
|
| 4020 |
+
"epoch": 0.7853692201518289,
|
| 4021 |
+
"grad_norm": 25.96500015258789,
|
| 4022 |
+
"learning_rate": 2.2585838936091754e-05,
|
| 4023 |
+
"loss": 4.7778,
|
| 4024 |
+
"step": 569
|
| 4025 |
+
},
|
| 4026 |
+
{
|
| 4027 |
+
"epoch": 0.7867494824016563,
|
| 4028 |
+
"grad_norm": 32.132694244384766,
|
| 4029 |
+
"learning_rate": 2.2308451817589283e-05,
|
| 4030 |
+
"loss": 7.8387,
|
| 4031 |
+
"step": 570
|
| 4032 |
+
},
|
| 4033 |
+
{
|
| 4034 |
+
"epoch": 0.7881297446514838,
|
| 4035 |
+
"grad_norm": 19.92507553100586,
|
| 4036 |
+
"learning_rate": 2.2032564593677774e-05,
|
| 4037 |
+
"loss": 3.7773,
|
| 4038 |
+
"step": 571
|
| 4039 |
+
},
|
| 4040 |
+
{
|
| 4041 |
+
"epoch": 0.7895100069013112,
|
| 4042 |
+
"grad_norm": 15.472801208496094,
|
| 4043 |
+
"learning_rate": 2.175818259057045e-05,
|
| 4044 |
+
"loss": 3.1225,
|
| 4045 |
+
"step": 572
|
| 4046 |
+
},
|
| 4047 |
+
{
|
| 4048 |
+
"epoch": 0.7908902691511387,
|
| 4049 |
+
"grad_norm": 18.66532325744629,
|
| 4050 |
+
"learning_rate": 2.148531110542118e-05,
|
| 4051 |
+
"loss": 3.4943,
|
| 4052 |
+
"step": 573
|
| 4053 |
+
},
|
| 4054 |
+
{
|
| 4055 |
+
"epoch": 0.7922705314009661,
|
| 4056 |
+
"grad_norm": 19.573230743408203,
|
| 4057 |
+
"learning_rate": 2.1213955406222074e-05,
|
| 4058 |
+
"loss": 3.7143,
|
| 4059 |
+
"step": 574
|
| 4060 |
+
},
|
| 4061 |
+
{
|
| 4062 |
+
"epoch": 0.7936507936507936,
|
| 4063 |
+
"grad_norm": 24.81309700012207,
|
| 4064 |
+
"learning_rate": 2.09441207317019e-05,
|
| 4065 |
+
"loss": 4.6328,
|
| 4066 |
+
"step": 575
|
| 4067 |
+
},
|
| 4068 |
+
{
|
| 4069 |
+
"epoch": 0.7950310559006211,
|
| 4070 |
+
"grad_norm": 21.88273811340332,
|
| 4071 |
+
"learning_rate": 2.0675812291224793e-05,
|
| 4072 |
+
"loss": 4.0822,
|
| 4073 |
+
"step": 576
|
| 4074 |
+
},
|
| 4075 |
+
{
|
| 4076 |
+
"epoch": 0.7964113181504486,
|
| 4077 |
+
"grad_norm": 29.533687591552734,
|
| 4078 |
+
"learning_rate": 2.0409035264689857e-05,
|
| 4079 |
+
"loss": 5.4004,
|
| 4080 |
+
"step": 577
|
| 4081 |
+
},
|
| 4082 |
+
{
|
| 4083 |
+
"epoch": 0.7977915804002761,
|
| 4084 |
+
"grad_norm": 21.00408172607422,
|
| 4085 |
+
"learning_rate": 2.0143794802431047e-05,
|
| 4086 |
+
"loss": 3.9792,
|
| 4087 |
+
"step": 578
|
| 4088 |
+
},
|
| 4089 |
+
{
|
| 4090 |
+
"epoch": 0.7991718426501035,
|
| 4091 |
+
"grad_norm": 16.558748245239258,
|
| 4092 |
+
"learning_rate": 1.988009602511779e-05,
|
| 4093 |
+
"loss": 3.084,
|
| 4094 |
+
"step": 579
|
| 4095 |
+
},
|
| 4096 |
+
{
|
| 4097 |
+
"epoch": 0.800552104899931,
|
| 4098 |
+
"grad_norm": 20.905330657958984,
|
| 4099 |
+
"learning_rate": 1.961794402365611e-05,
|
| 4100 |
+
"loss": 4.823,
|
| 4101 |
+
"step": 580
|
| 4102 |
+
},
|
| 4103 |
+
{
|
| 4104 |
+
"epoch": 0.8019323671497585,
|
| 4105 |
+
"grad_norm": 16.98253631591797,
|
| 4106 |
+
"learning_rate": 1.935734385909028e-05,
|
| 4107 |
+
"loss": 3.5759,
|
| 4108 |
+
"step": 581
|
| 4109 |
+
},
|
| 4110 |
+
{
|
| 4111 |
+
"epoch": 0.8033126293995859,
|
| 4112 |
+
"grad_norm": 24.210498809814453,
|
| 4113 |
+
"learning_rate": 1.9098300562505266e-05,
|
| 4114 |
+
"loss": 6.3733,
|
| 4115 |
+
"step": 582
|
| 4116 |
+
},
|
| 4117 |
+
{
|
| 4118 |
+
"epoch": 0.8046928916494134,
|
| 4119 |
+
"grad_norm": 18.199722290039062,
|
| 4120 |
+
"learning_rate": 1.8840819134929465e-05,
|
| 4121 |
+
"loss": 3.6873,
|
| 4122 |
+
"step": 583
|
| 4123 |
+
},
|
| 4124 |
+
{
|
| 4125 |
+
"epoch": 0.8060731538992408,
|
| 4126 |
+
"grad_norm": 19.356443405151367,
|
| 4127 |
+
"learning_rate": 1.8584904547238212e-05,
|
| 4128 |
+
"loss": 4.3978,
|
| 4129 |
+
"step": 584
|
| 4130 |
+
},
|
| 4131 |
+
{
|
| 4132 |
+
"epoch": 0.8074534161490683,
|
| 4133 |
+
"grad_norm": 21.787994384765625,
|
| 4134 |
+
"learning_rate": 1.833056174005784e-05,
|
| 4135 |
+
"loss": 4.2661,
|
| 4136 |
+
"step": 585
|
| 4137 |
+
},
|
| 4138 |
+
{
|
| 4139 |
+
"epoch": 0.8088336783988958,
|
| 4140 |
+
"grad_norm": 24.865665435791016,
|
| 4141 |
+
"learning_rate": 1.8077795623670137e-05,
|
| 4142 |
+
"loss": 5.5552,
|
| 4143 |
+
"step": 586
|
| 4144 |
+
},
|
| 4145 |
+
{
|
| 4146 |
+
"epoch": 0.8102139406487232,
|
| 4147 |
+
"grad_norm": 21.28284454345703,
|
| 4148 |
+
"learning_rate": 1.7826611077917842e-05,
|
| 4149 |
+
"loss": 5.044,
|
| 4150 |
+
"step": 587
|
| 4151 |
+
},
|
| 4152 |
+
{
|
| 4153 |
+
"epoch": 0.8115942028985508,
|
| 4154 |
+
"grad_norm": 17.053913116455078,
|
| 4155 |
+
"learning_rate": 1.757701295211014e-05,
|
| 4156 |
+
"loss": 3.1183,
|
| 4157 |
+
"step": 588
|
| 4158 |
+
},
|
| 4159 |
+
{
|
| 4160 |
+
"epoch": 0.8129744651483782,
|
| 4161 |
+
"grad_norm": 20.05781364440918,
|
| 4162 |
+
"learning_rate": 1.7329006064929232e-05,
|
| 4163 |
+
"loss": 3.9421,
|
| 4164 |
+
"step": 589
|
| 4165 |
+
},
|
| 4166 |
+
{
|
| 4167 |
+
"epoch": 0.8143547273982057,
|
| 4168 |
+
"grad_norm": 26.919628143310547,
|
| 4169 |
+
"learning_rate": 1.7082595204337182e-05,
|
| 4170 |
+
"loss": 7.2336,
|
| 4171 |
+
"step": 590
|
| 4172 |
+
},
|
| 4173 |
+
{
|
| 4174 |
+
"epoch": 0.8157349896480331,
|
| 4175 |
+
"grad_norm": 24.17670249938965,
|
| 4176 |
+
"learning_rate": 1.6837785127483618e-05,
|
| 4177 |
+
"loss": 5.4203,
|
| 4178 |
+
"step": 591
|
| 4179 |
+
},
|
| 4180 |
+
{
|
| 4181 |
+
"epoch": 0.8171152518978606,
|
| 4182 |
+
"grad_norm": 18.08180809020996,
|
| 4183 |
+
"learning_rate": 1.659458056061378e-05,
|
| 4184 |
+
"loss": 3.136,
|
| 4185 |
+
"step": 592
|
| 4186 |
+
},
|
| 4187 |
+
{
|
| 4188 |
+
"epoch": 0.8184955141476881,
|
| 4189 |
+
"grad_norm": 25.226892471313477,
|
| 4190 |
+
"learning_rate": 1.6352986198977325e-05,
|
| 4191 |
+
"loss": 5.6002,
|
| 4192 |
+
"step": 593
|
| 4193 |
+
},
|
| 4194 |
+
{
|
| 4195 |
+
"epoch": 0.8198757763975155,
|
| 4196 |
+
"grad_norm": 21.082447052001953,
|
| 4197 |
+
"learning_rate": 1.6113006706737667e-05,
|
| 4198 |
+
"loss": 3.8211,
|
| 4199 |
+
"step": 594
|
| 4200 |
+
},
|
| 4201 |
+
{
|
| 4202 |
+
"epoch": 0.821256038647343,
|
| 4203 |
+
"grad_norm": 23.459060668945312,
|
| 4204 |
+
"learning_rate": 1.587464671688187e-05,
|
| 4205 |
+
"loss": 5.1864,
|
| 4206 |
+
"step": 595
|
| 4207 |
+
},
|
| 4208 |
+
{
|
| 4209 |
+
"epoch": 0.8226363008971704,
|
| 4210 |
+
"grad_norm": 19.926836013793945,
|
| 4211 |
+
"learning_rate": 1.563791083113142e-05,
|
| 4212 |
+
"loss": 2.9439,
|
| 4213 |
+
"step": 596
|
| 4214 |
+
},
|
| 4215 |
+
{
|
| 4216 |
+
"epoch": 0.8240165631469979,
|
| 4217 |
+
"grad_norm": 21.065765380859375,
|
| 4218 |
+
"learning_rate": 1.540280361985308e-05,
|
| 4219 |
+
"loss": 5.3703,
|
| 4220 |
+
"step": 597
|
| 4221 |
+
},
|
| 4222 |
+
{
|
| 4223 |
+
"epoch": 0.8253968253968254,
|
| 4224 |
+
"grad_norm": 24.38242530822754,
|
| 4225 |
+
"learning_rate": 1.5169329621970918e-05,
|
| 4226 |
+
"loss": 4.0207,
|
| 4227 |
+
"step": 598
|
| 4228 |
+
},
|
| 4229 |
+
{
|
| 4230 |
+
"epoch": 0.8267770876466529,
|
| 4231 |
+
"grad_norm": 24.726024627685547,
|
| 4232 |
+
"learning_rate": 1.4937493344878473e-05,
|
| 4233 |
+
"loss": 6.1451,
|
| 4234 |
+
"step": 599
|
| 4235 |
+
},
|
| 4236 |
+
{
|
| 4237 |
+
"epoch": 0.8281573498964804,
|
| 4238 |
+
"grad_norm": 20.777873992919922,
|
| 4239 |
+
"learning_rate": 1.4707299264351915e-05,
|
| 4240 |
+
"loss": 4.3696,
|
| 4241 |
+
"step": 600
|
| 4242 |
+
},
|
| 4243 |
+
{
|
| 4244 |
+
"epoch": 0.8295376121463078,
|
| 4245 |
+
"grad_norm": 22.057493209838867,
|
| 4246 |
+
"learning_rate": 1.4478751824463543e-05,
|
| 4247 |
+
"loss": 5.1269,
|
| 4248 |
+
"step": 601
|
| 4249 |
+
},
|
| 4250 |
+
{
|
| 4251 |
+
"epoch": 0.8309178743961353,
|
| 4252 |
+
"grad_norm": 21.792823791503906,
|
| 4253 |
+
"learning_rate": 1.4251855437495975e-05,
|
| 4254 |
+
"loss": 5.1566,
|
| 4255 |
+
"step": 602
|
| 4256 |
+
},
|
| 4257 |
+
{
|
| 4258 |
+
"epoch": 0.8322981366459627,
|
| 4259 |
+
"grad_norm": 18.181209564208984,
|
| 4260 |
+
"learning_rate": 1.4026614483857036e-05,
|
| 4261 |
+
"loss": 4.1663,
|
| 4262 |
+
"step": 603
|
| 4263 |
+
},
|
| 4264 |
+
{
|
| 4265 |
+
"epoch": 0.8336783988957902,
|
| 4266 |
+
"grad_norm": 23.090852737426758,
|
| 4267 |
+
"learning_rate": 1.3803033311995072e-05,
|
| 4268 |
+
"loss": 4.3103,
|
| 4269 |
+
"step": 604
|
| 4270 |
+
},
|
| 4271 |
+
{
|
| 4272 |
+
"epoch": 0.8350586611456177,
|
| 4273 |
+
"grad_norm": 22.44392967224121,
|
| 4274 |
+
"learning_rate": 1.3581116238315195e-05,
|
| 4275 |
+
"loss": 4.3511,
|
| 4276 |
+
"step": 605
|
| 4277 |
+
},
|
| 4278 |
+
{
|
| 4279 |
+
"epoch": 0.8364389233954451,
|
| 4280 |
+
"grad_norm": 19.312456130981445,
|
| 4281 |
+
"learning_rate": 1.336086754709569e-05,
|
| 4282 |
+
"loss": 3.5947,
|
| 4283 |
+
"step": 606
|
| 4284 |
+
},
|
| 4285 |
+
{
|
| 4286 |
+
"epoch": 0.8378191856452726,
|
| 4287 |
+
"grad_norm": 25.088590621948242,
|
| 4288 |
+
"learning_rate": 1.3142291490405568e-05,
|
| 4289 |
+
"loss": 4.6283,
|
| 4290 |
+
"step": 607
|
| 4291 |
+
},
|
| 4292 |
+
{
|
| 4293 |
+
"epoch": 0.8391994478951,
|
| 4294 |
+
"grad_norm": 20.34014129638672,
|
| 4295 |
+
"learning_rate": 1.2925392288022298e-05,
|
| 4296 |
+
"loss": 4.1621,
|
| 4297 |
+
"step": 608
|
| 4298 |
+
},
|
| 4299 |
+
{
|
| 4300 |
+
"epoch": 0.8405797101449275,
|
| 4301 |
+
"grad_norm": 18.029075622558594,
|
| 4302 |
+
"learning_rate": 1.2710174127350361e-05,
|
| 4303 |
+
"loss": 4.1052,
|
| 4304 |
+
"step": 609
|
| 4305 |
+
},
|
| 4306 |
+
{
|
| 4307 |
+
"epoch": 0.841959972394755,
|
| 4308 |
+
"grad_norm": 26.527965545654297,
|
| 4309 |
+
"learning_rate": 1.2496641163340562e-05,
|
| 4310 |
+
"loss": 5.4877,
|
| 4311 |
+
"step": 610
|
| 4312 |
+
},
|
| 4313 |
+
{
|
| 4314 |
+
"epoch": 0.8433402346445825,
|
| 4315 |
+
"grad_norm": 17.80951690673828,
|
| 4316 |
+
"learning_rate": 1.2284797518409575e-05,
|
| 4317 |
+
"loss": 3.6808,
|
| 4318 |
+
"step": 611
|
| 4319 |
+
},
|
| 4320 |
+
{
|
| 4321 |
+
"epoch": 0.84472049689441,
|
| 4322 |
+
"grad_norm": 23.62249183654785,
|
| 4323 |
+
"learning_rate": 1.2074647282360574e-05,
|
| 4324 |
+
"loss": 3.8818,
|
| 4325 |
+
"step": 612
|
| 4326 |
+
},
|
| 4327 |
+
{
|
| 4328 |
+
"epoch": 0.8461007591442374,
|
| 4329 |
+
"grad_norm": 22.320627212524414,
|
| 4330 |
+
"learning_rate": 1.1866194512304073e-05,
|
| 4331 |
+
"loss": 4.7351,
|
| 4332 |
+
"step": 613
|
| 4333 |
+
},
|
| 4334 |
+
{
|
| 4335 |
+
"epoch": 0.8474810213940649,
|
| 4336 |
+
"grad_norm": 25.21207046508789,
|
| 4337 |
+
"learning_rate": 1.1659443232579858e-05,
|
| 4338 |
+
"loss": 6.1288,
|
| 4339 |
+
"step": 614
|
| 4340 |
+
},
|
| 4341 |
+
{
|
| 4342 |
+
"epoch": 0.8488612836438924,
|
| 4343 |
+
"grad_norm": 21.264904022216797,
|
| 4344 |
+
"learning_rate": 1.1454397434679021e-05,
|
| 4345 |
+
"loss": 4.2007,
|
| 4346 |
+
"step": 615
|
| 4347 |
+
},
|
| 4348 |
+
{
|
| 4349 |
+
"epoch": 0.8502415458937198,
|
| 4350 |
+
"grad_norm": 20.78594398498535,
|
| 4351 |
+
"learning_rate": 1.125106107716708e-05,
|
| 4352 |
+
"loss": 4.3806,
|
| 4353 |
+
"step": 616
|
| 4354 |
+
},
|
| 4355 |
+
{
|
| 4356 |
+
"epoch": 0.8516218081435473,
|
| 4357 |
+
"grad_norm": 20.06401824951172,
|
| 4358 |
+
"learning_rate": 1.10494380856075e-05,
|
| 4359 |
+
"loss": 3.2476,
|
| 4360 |
+
"step": 617
|
| 4361 |
+
},
|
| 4362 |
+
{
|
| 4363 |
+
"epoch": 0.8530020703933747,
|
| 4364 |
+
"grad_norm": 26.73411750793457,
|
| 4365 |
+
"learning_rate": 1.0849532352485903e-05,
|
| 4366 |
+
"loss": 5.4257,
|
| 4367 |
+
"step": 618
|
| 4368 |
+
},
|
| 4369 |
+
{
|
| 4370 |
+
"epoch": 0.8543823326432022,
|
| 4371 |
+
"grad_norm": 21.02422332763672,
|
| 4372 |
+
"learning_rate": 1.0651347737134965e-05,
|
| 4373 |
+
"loss": 3.4578,
|
| 4374 |
+
"step": 619
|
| 4375 |
+
},
|
| 4376 |
+
{
|
| 4377 |
+
"epoch": 0.8557625948930296,
|
| 4378 |
+
"grad_norm": 25.98722267150879,
|
| 4379 |
+
"learning_rate": 1.0454888065659774e-05,
|
| 4380 |
+
"loss": 4.4609,
|
| 4381 |
+
"step": 620
|
| 4382 |
+
},
|
| 4383 |
+
{
|
| 4384 |
+
"epoch": 0.8571428571428571,
|
| 4385 |
+
"grad_norm": 20.993576049804688,
|
| 4386 |
+
"learning_rate": 1.026015713086418e-05,
|
| 4387 |
+
"loss": 3.4471,
|
| 4388 |
+
"step": 621
|
| 4389 |
+
},
|
| 4390 |
+
{
|
| 4391 |
+
"epoch": 0.8585231193926847,
|
| 4392 |
+
"grad_norm": 19.924787521362305,
|
| 4393 |
+
"learning_rate": 1.0067158692177326e-05,
|
| 4394 |
+
"loss": 3.8117,
|
| 4395 |
+
"step": 622
|
| 4396 |
+
},
|
| 4397 |
+
{
|
| 4398 |
+
"epoch": 0.8599033816425121,
|
| 4399 |
+
"grad_norm": 21.03854751586914,
|
| 4400 |
+
"learning_rate": 9.875896475581348e-06,
|
| 4401 |
+
"loss": 4.3201,
|
| 4402 |
+
"step": 623
|
| 4403 |
+
},
|
| 4404 |
+
{
|
| 4405 |
+
"epoch": 0.8612836438923396,
|
| 4406 |
+
"grad_norm": 21.847026824951172,
|
| 4407 |
+
"learning_rate": 9.686374173539148e-06,
|
| 4408 |
+
"loss": 4.7991,
|
| 4409 |
+
"step": 624
|
| 4410 |
+
},
|
| 4411 |
+
{
|
| 4412 |
+
"epoch": 0.862663906142167,
|
| 4413 |
+
"grad_norm": 24.594127655029297,
|
| 4414 |
+
"learning_rate": 9.49859544492332e-06,
|
| 4415 |
+
"loss": 5.826,
|
| 4416 |
+
"step": 625
|
| 4417 |
+
},
|
| 4418 |
+
{
|
| 4419 |
+
"epoch": 0.8640441683919945,
|
| 4420 |
+
"grad_norm": 21.561784744262695,
|
| 4421 |
+
"learning_rate": 9.31256391494546e-06,
|
| 4422 |
+
"loss": 4.593,
|
| 4423 |
+
"step": 626
|
| 4424 |
+
},
|
| 4425 |
+
{
|
| 4426 |
+
"epoch": 0.865424430641822,
|
| 4427 |
+
"grad_norm": 22.014583587646484,
|
| 4428 |
+
"learning_rate": 9.128283175086105e-06,
|
| 4429 |
+
"loss": 4.8819,
|
| 4430 |
+
"step": 627
|
| 4431 |
+
},
|
| 4432 |
+
{
|
| 4433 |
+
"epoch": 0.8668046928916494,
|
| 4434 |
+
"grad_norm": 26.051342010498047,
|
| 4435 |
+
"learning_rate": 8.945756783025527e-06,
|
| 4436 |
+
"loss": 5.2834,
|
| 4437 |
+
"step": 628
|
| 4438 |
+
},
|
| 4439 |
+
{
|
| 4440 |
+
"epoch": 0.8681849551414769,
|
| 4441 |
+
"grad_norm": 23.40290069580078,
|
| 4442 |
+
"learning_rate": 8.76498826257488e-06,
|
| 4443 |
+
"loss": 4.1225,
|
| 4444 |
+
"step": 629
|
| 4445 |
+
},
|
| 4446 |
+
{
|
| 4447 |
+
"epoch": 0.8695652173913043,
|
| 4448 |
+
"grad_norm": 23.62264060974121,
|
| 4449 |
+
"learning_rate": 8.585981103608342e-06,
|
| 4450 |
+
"loss": 4.5137,
|
| 4451 |
+
"step": 630
|
| 4452 |
+
},
|
| 4453 |
+
{
|
| 4454 |
+
"epoch": 0.8709454796411318,
|
| 4455 |
+
"grad_norm": 17.327234268188477,
|
| 4456 |
+
"learning_rate": 8.408738761995649e-06,
|
| 4457 |
+
"loss": 3.1094,
|
| 4458 |
+
"step": 631
|
| 4459 |
+
},
|
| 4460 |
+
{
|
| 4461 |
+
"epoch": 0.8723257418909592,
|
| 4462 |
+
"grad_norm": 17.089147567749023,
|
| 4463 |
+
"learning_rate": 8.233264659535367e-06,
|
| 4464 |
+
"loss": 3.8834,
|
| 4465 |
+
"step": 632
|
| 4466 |
+
},
|
| 4467 |
+
{
|
| 4468 |
+
"epoch": 0.8737060041407867,
|
| 4469 |
+
"grad_norm": 17.930011749267578,
|
| 4470 |
+
"learning_rate": 8.059562183888902e-06,
|
| 4471 |
+
"loss": 4.0732,
|
| 4472 |
+
"step": 633
|
| 4473 |
+
},
|
| 4474 |
+
{
|
| 4475 |
+
"epoch": 0.8750862663906143,
|
| 4476 |
+
"grad_norm": 17.872472763061523,
|
| 4477 |
+
"learning_rate": 7.887634688515e-06,
|
| 4478 |
+
"loss": 3.4639,
|
| 4479 |
+
"step": 634
|
| 4480 |
+
},
|
| 4481 |
+
{
|
| 4482 |
+
"epoch": 0.8764665286404417,
|
| 4483 |
+
"grad_norm": 29.12419319152832,
|
| 4484 |
+
"learning_rate": 7.71748549260507e-06,
|
| 4485 |
+
"loss": 5.0055,
|
| 4486 |
+
"step": 635
|
| 4487 |
+
},
|
| 4488 |
+
{
|
| 4489 |
+
"epoch": 0.8778467908902692,
|
| 4490 |
+
"grad_norm": 15.905129432678223,
|
| 4491 |
+
"learning_rate": 7.549117881019141e-06,
|
| 4492 |
+
"loss": 4.0129,
|
| 4493 |
+
"step": 636
|
| 4494 |
+
},
|
| 4495 |
+
{
|
| 4496 |
+
"epoch": 0.8792270531400966,
|
| 4497 |
+
"grad_norm": 18.55198097229004,
|
| 4498 |
+
"learning_rate": 7.382535104222366e-06,
|
| 4499 |
+
"loss": 4.0759,
|
| 4500 |
+
"step": 637
|
| 4501 |
+
},
|
| 4502 |
+
{
|
| 4503 |
+
"epoch": 0.8806073153899241,
|
| 4504 |
+
"grad_norm": 15.301910400390625,
|
| 4505 |
+
"learning_rate": 7.21774037822226e-06,
|
| 4506 |
+
"loss": 2.5912,
|
| 4507 |
+
"step": 638
|
| 4508 |
+
},
|
| 4509 |
+
{
|
| 4510 |
+
"epoch": 0.8819875776397516,
|
| 4511 |
+
"grad_norm": 23.361509323120117,
|
| 4512 |
+
"learning_rate": 7.054736884506719e-06,
|
| 4513 |
+
"loss": 5.7187,
|
| 4514 |
+
"step": 639
|
| 4515 |
+
},
|
| 4516 |
+
{
|
| 4517 |
+
"epoch": 0.883367839889579,
|
| 4518 |
+
"grad_norm": 18.59641456604004,
|
| 4519 |
+
"learning_rate": 6.893527769982499e-06,
|
| 4520 |
+
"loss": 3.4449,
|
| 4521 |
+
"step": 640
|
| 4522 |
+
},
|
| 4523 |
+
{
|
| 4524 |
+
"epoch": 0.8847481021394065,
|
| 4525 |
+
"grad_norm": 25.18804931640625,
|
| 4526 |
+
"learning_rate": 6.734116146914516e-06,
|
| 4527 |
+
"loss": 4.9088,
|
| 4528 |
+
"step": 641
|
| 4529 |
+
},
|
| 4530 |
+
{
|
| 4531 |
+
"epoch": 0.8861283643892339,
|
| 4532 |
+
"grad_norm": 22.016386032104492,
|
| 4533 |
+
"learning_rate": 6.576505092865748e-06,
|
| 4534 |
+
"loss": 4.6404,
|
| 4535 |
+
"step": 642
|
| 4536 |
+
},
|
| 4537 |
+
{
|
| 4538 |
+
"epoch": 0.8875086266390614,
|
| 4539 |
+
"grad_norm": 22.610445022583008,
|
| 4540 |
+
"learning_rate": 6.420697650637752e-06,
|
| 4541 |
+
"loss": 5.5401,
|
| 4542 |
+
"step": 643
|
| 4543 |
+
},
|
| 4544 |
+
{
|
| 4545 |
+
"epoch": 0.8888888888888888,
|
| 4546 |
+
"grad_norm": 22.13422966003418,
|
| 4547 |
+
"learning_rate": 6.266696828212071e-06,
|
| 4548 |
+
"loss": 3.569,
|
| 4549 |
+
"step": 644
|
| 4550 |
+
},
|
| 4551 |
+
{
|
| 4552 |
+
"epoch": 0.8902691511387164,
|
| 4553 |
+
"grad_norm": 20.980058670043945,
|
| 4554 |
+
"learning_rate": 6.1145055986920105e-06,
|
| 4555 |
+
"loss": 5.7634,
|
| 4556 |
+
"step": 645
|
| 4557 |
+
},
|
| 4558 |
+
{
|
| 4559 |
+
"epoch": 0.8916494133885439,
|
| 4560 |
+
"grad_norm": 21.698143005371094,
|
| 4561 |
+
"learning_rate": 5.964126900245359e-06,
|
| 4562 |
+
"loss": 3.6239,
|
| 4563 |
+
"step": 646
|
| 4564 |
+
},
|
| 4565 |
+
{
|
| 4566 |
+
"epoch": 0.8930296756383713,
|
| 4567 |
+
"grad_norm": 19.81194305419922,
|
| 4568 |
+
"learning_rate": 5.8155636360475385e-06,
|
| 4569 |
+
"loss": 4.6039,
|
| 4570 |
+
"step": 647
|
| 4571 |
+
},
|
| 4572 |
+
{
|
| 4573 |
+
"epoch": 0.8944099378881988,
|
| 4574 |
+
"grad_norm": 15.131420135498047,
|
| 4575 |
+
"learning_rate": 5.668818674225685e-06,
|
| 4576 |
+
"loss": 2.5671,
|
| 4577 |
+
"step": 648
|
| 4578 |
+
},
|
| 4579 |
+
{
|
| 4580 |
+
"epoch": 0.8957902001380262,
|
| 4581 |
+
"grad_norm": 17.206186294555664,
|
| 4582 |
+
"learning_rate": 5.523894847803235e-06,
|
| 4583 |
+
"loss": 3.0977,
|
| 4584 |
+
"step": 649
|
| 4585 |
+
},
|
| 4586 |
+
{
|
| 4587 |
+
"epoch": 0.8971704623878537,
|
| 4588 |
+
"grad_norm": 18.428306579589844,
|
| 4589 |
+
"learning_rate": 5.3807949546451404e-06,
|
| 4590 |
+
"loss": 3.2133,
|
| 4591 |
+
"step": 650
|
| 4592 |
+
},
|
| 4593 |
+
{
|
| 4594 |
+
"epoch": 0.8985507246376812,
|
| 4595 |
+
"grad_norm": 15.980635643005371,
|
| 4596 |
+
"learning_rate": 5.23952175740402e-06,
|
| 4597 |
+
"loss": 3.5965,
|
| 4598 |
+
"step": 651
|
| 4599 |
+
},
|
| 4600 |
+
{
|
| 4601 |
+
"epoch": 0.8999309868875086,
|
| 4602 |
+
"grad_norm": 22.935367584228516,
|
| 4603 |
+
"learning_rate": 5.100077983466667e-06,
|
| 4604 |
+
"loss": 7.5044,
|
| 4605 |
+
"step": 652
|
| 4606 |
+
},
|
| 4607 |
+
{
|
| 4608 |
+
"epoch": 0.9013112491373361,
|
| 4609 |
+
"grad_norm": 18.72053337097168,
|
| 4610 |
+
"learning_rate": 4.9624663249014825e-06,
|
| 4611 |
+
"loss": 3.108,
|
| 4612 |
+
"step": 653
|
| 4613 |
+
},
|
| 4614 |
+
{
|
| 4615 |
+
"epoch": 0.9026915113871635,
|
| 4616 |
+
"grad_norm": 20.732444763183594,
|
| 4617 |
+
"learning_rate": 4.826689438406495e-06,
|
| 4618 |
+
"loss": 4.3188,
|
| 4619 |
+
"step": 654
|
| 4620 |
+
},
|
| 4621 |
+
{
|
| 4622 |
+
"epoch": 0.904071773636991,
|
| 4623 |
+
"grad_norm": 18.4609375,
|
| 4624 |
+
"learning_rate": 4.692749945258057e-06,
|
| 4625 |
+
"loss": 2.5593,
|
| 4626 |
+
"step": 655
|
| 4627 |
+
},
|
| 4628 |
+
{
|
| 4629 |
+
"epoch": 0.9054520358868184,
|
| 4630 |
+
"grad_norm": 21.71295738220215,
|
| 4631 |
+
"learning_rate": 4.560650431260239e-06,
|
| 4632 |
+
"loss": 4.629,
|
| 4633 |
+
"step": 656
|
| 4634 |
+
},
|
| 4635 |
+
{
|
| 4636 |
+
"epoch": 0.906832298136646,
|
| 4637 |
+
"grad_norm": 22.27686309814453,
|
| 4638 |
+
"learning_rate": 4.4303934466948805e-06,
|
| 4639 |
+
"loss": 5.1413,
|
| 4640 |
+
"step": 657
|
| 4641 |
+
},
|
| 4642 |
+
{
|
| 4643 |
+
"epoch": 0.9082125603864735,
|
| 4644 |
+
"grad_norm": 19.66610336303711,
|
| 4645 |
+
"learning_rate": 4.301981506272457e-06,
|
| 4646 |
+
"loss": 3.9924,
|
| 4647 |
+
"step": 658
|
| 4648 |
+
},
|
| 4649 |
+
{
|
| 4650 |
+
"epoch": 0.9095928226363009,
|
| 4651 |
+
"grad_norm": 17.331315994262695,
|
| 4652 |
+
"learning_rate": 4.175417089083378e-06,
|
| 4653 |
+
"loss": 3.9328,
|
| 4654 |
+
"step": 659
|
| 4655 |
+
},
|
| 4656 |
+
{
|
| 4657 |
+
"epoch": 0.9109730848861284,
|
| 4658 |
+
"grad_norm": 16.983257293701172,
|
| 4659 |
+
"learning_rate": 4.050702638550275e-06,
|
| 4660 |
+
"loss": 4.0252,
|
| 4661 |
+
"step": 660
|
| 4662 |
+
},
|
| 4663 |
+
{
|
| 4664 |
+
"epoch": 0.9123533471359558,
|
| 4665 |
+
"grad_norm": 19.764429092407227,
|
| 4666 |
+
"learning_rate": 3.927840562380692e-06,
|
| 4667 |
+
"loss": 2.7445,
|
| 4668 |
+
"step": 661
|
| 4669 |
+
},
|
| 4670 |
+
{
|
| 4671 |
+
"epoch": 0.9137336093857833,
|
| 4672 |
+
"grad_norm": 18.17449378967285,
|
| 4673 |
+
"learning_rate": 3.8068332325207457e-06,
|
| 4674 |
+
"loss": 3.8537,
|
| 4675 |
+
"step": 662
|
| 4676 |
+
},
|
| 4677 |
+
{
|
| 4678 |
+
"epoch": 0.9151138716356108,
|
| 4679 |
+
"grad_norm": 20.223976135253906,
|
| 4680 |
+
"learning_rate": 3.6876829851092087e-06,
|
| 4681 |
+
"loss": 3.5495,
|
| 4682 |
+
"step": 663
|
| 4683 |
+
},
|
| 4684 |
+
{
|
| 4685 |
+
"epoch": 0.9164941338854382,
|
| 4686 |
+
"grad_norm": 25.060184478759766,
|
| 4687 |
+
"learning_rate": 3.5703921204324865e-06,
|
| 4688 |
+
"loss": 5.1044,
|
| 4689 |
+
"step": 664
|
| 4690 |
+
},
|
| 4691 |
+
{
|
| 4692 |
+
"epoch": 0.9178743961352657,
|
| 4693 |
+
"grad_norm": 21.926902770996094,
|
| 4694 |
+
"learning_rate": 3.454962902880199e-06,
|
| 4695 |
+
"loss": 5.1936,
|
| 4696 |
+
"step": 665
|
| 4697 |
+
},
|
| 4698 |
+
{
|
| 4699 |
+
"epoch": 0.9192546583850931,
|
| 4700 |
+
"grad_norm": 23.566516876220703,
|
| 4701 |
+
"learning_rate": 3.3413975609013713e-06,
|
| 4702 |
+
"loss": 6.3522,
|
| 4703 |
+
"step": 666
|
| 4704 |
+
},
|
| 4705 |
+
{
|
| 4706 |
+
"epoch": 0.9206349206349206,
|
| 4707 |
+
"grad_norm": 30.516544342041016,
|
| 4708 |
+
"learning_rate": 3.2296982869616133e-06,
|
| 4709 |
+
"loss": 6.1352,
|
| 4710 |
+
"step": 667
|
| 4711 |
+
},
|
| 4712 |
+
{
|
| 4713 |
+
"epoch": 0.9220151828847482,
|
| 4714 |
+
"grad_norm": 27.719097137451172,
|
| 4715 |
+
"learning_rate": 3.11986723750054e-06,
|
| 4716 |
+
"loss": 6.3657,
|
| 4717 |
+
"step": 668
|
| 4718 |
+
},
|
| 4719 |
+
{
|
| 4720 |
+
"epoch": 0.9233954451345756,
|
| 4721 |
+
"grad_norm": 21.81649398803711,
|
| 4722 |
+
"learning_rate": 3.011906532890352e-06,
|
| 4723 |
+
"loss": 4.6634,
|
| 4724 |
+
"step": 669
|
| 4725 |
+
},
|
| 4726 |
+
{
|
| 4727 |
+
"epoch": 0.9247757073844031,
|
| 4728 |
+
"grad_norm": 19.389427185058594,
|
| 4729 |
+
"learning_rate": 2.905818257394799e-06,
|
| 4730 |
+
"loss": 3.3689,
|
| 4731 |
+
"step": 670
|
| 4732 |
+
},
|
| 4733 |
+
{
|
| 4734 |
+
"epoch": 0.9261559696342305,
|
| 4735 |
+
"grad_norm": 22.784374237060547,
|
| 4736 |
+
"learning_rate": 2.801604459128926e-06,
|
| 4737 |
+
"loss": 3.927,
|
| 4738 |
+
"step": 671
|
| 4739 |
+
},
|
| 4740 |
+
{
|
| 4741 |
+
"epoch": 0.927536231884058,
|
| 4742 |
+
"grad_norm": 23.757984161376953,
|
| 4743 |
+
"learning_rate": 2.6992671500196132e-06,
|
| 4744 |
+
"loss": 4.9386,
|
| 4745 |
+
"step": 672
|
| 4746 |
+
},
|
| 4747 |
+
{
|
| 4748 |
+
"epoch": 0.9289164941338854,
|
| 4749 |
+
"grad_norm": 24.24261474609375,
|
| 4750 |
+
"learning_rate": 2.5988083057666533e-06,
|
| 4751 |
+
"loss": 4.3493,
|
| 4752 |
+
"step": 673
|
| 4753 |
+
},
|
| 4754 |
+
{
|
| 4755 |
+
"epoch": 0.9302967563837129,
|
| 4756 |
+
"grad_norm": 21.458877563476562,
|
| 4757 |
+
"learning_rate": 2.5002298658046487e-06,
|
| 4758 |
+
"loss": 4.5866,
|
| 4759 |
+
"step": 674
|
| 4760 |
+
},
|
| 4761 |
+
{
|
| 4762 |
+
"epoch": 0.9316770186335404,
|
| 4763 |
+
"grad_norm": 19.24845314025879,
|
| 4764 |
+
"learning_rate": 2.4035337332655506e-06,
|
| 4765 |
+
"loss": 5.569,
|
| 4766 |
+
"step": 675
|
| 4767 |
+
},
|
| 4768 |
+
{
|
| 4769 |
+
"epoch": 0.9330572808833678,
|
| 4770 |
+
"grad_norm": 27.00035285949707,
|
| 4771 |
+
"learning_rate": 2.308721774941991e-06,
|
| 4772 |
+
"loss": 4.5707,
|
| 4773 |
+
"step": 676
|
| 4774 |
+
},
|
| 4775 |
+
{
|
| 4776 |
+
"epoch": 0.9344375431331953,
|
| 4777 |
+
"grad_norm": 23.597713470458984,
|
| 4778 |
+
"learning_rate": 2.2157958212510877e-06,
|
| 4779 |
+
"loss": 3.8271,
|
| 4780 |
+
"step": 677
|
| 4781 |
+
},
|
| 4782 |
+
{
|
| 4783 |
+
"epoch": 0.9358178053830227,
|
| 4784 |
+
"grad_norm": 18.218555450439453,
|
| 4785 |
+
"learning_rate": 2.124757666199273e-06,
|
| 4786 |
+
"loss": 3.6037,
|
| 4787 |
+
"step": 678
|
| 4788 |
+
},
|
| 4789 |
+
{
|
| 4790 |
+
"epoch": 0.9371980676328503,
|
| 4791 |
+
"grad_norm": 13.22622299194336,
|
| 4792 |
+
"learning_rate": 2.035609067347566e-06,
|
| 4793 |
+
"loss": 2.1873,
|
| 4794 |
+
"step": 679
|
| 4795 |
+
},
|
| 4796 |
+
{
|
| 4797 |
+
"epoch": 0.9385783298826778,
|
| 4798 |
+
"grad_norm": 18.116043090820312,
|
| 4799 |
+
"learning_rate": 1.9483517457776436e-06,
|
| 4800 |
+
"loss": 4.0654,
|
| 4801 |
+
"step": 680
|
| 4802 |
+
},
|
| 4803 |
+
{
|
| 4804 |
+
"epoch": 0.9399585921325052,
|
| 4805 |
+
"grad_norm": 17.52006721496582,
|
| 4806 |
+
"learning_rate": 1.8629873860586566e-06,
|
| 4807 |
+
"loss": 3.4956,
|
| 4808 |
+
"step": 681
|
| 4809 |
+
},
|
| 4810 |
+
{
|
| 4811 |
+
"epoch": 0.9413388543823327,
|
| 4812 |
+
"grad_norm": 22.48879051208496,
|
| 4813 |
+
"learning_rate": 1.779517636214678e-06,
|
| 4814 |
+
"loss": 3.2213,
|
| 4815 |
+
"step": 682
|
| 4816 |
+
},
|
| 4817 |
+
{
|
| 4818 |
+
"epoch": 0.9427191166321601,
|
| 4819 |
+
"grad_norm": 16.940576553344727,
|
| 4820 |
+
"learning_rate": 1.6979441076928836e-06,
|
| 4821 |
+
"loss": 3.7022,
|
| 4822 |
+
"step": 683
|
| 4823 |
+
},
|
| 4824 |
+
{
|
| 4825 |
+
"epoch": 0.9440993788819876,
|
| 4826 |
+
"grad_norm": 19.83792495727539,
|
| 4827 |
+
"learning_rate": 1.6182683753324435e-06,
|
| 4828 |
+
"loss": 6.5674,
|
| 4829 |
+
"step": 684
|
| 4830 |
+
},
|
| 4831 |
+
{
|
| 4832 |
+
"epoch": 0.945479641131815,
|
| 4833 |
+
"grad_norm": 24.09896469116211,
|
| 4834 |
+
"learning_rate": 1.5404919773341576e-06,
|
| 4835 |
+
"loss": 3.5441,
|
| 4836 |
+
"step": 685
|
| 4837 |
+
},
|
| 4838 |
+
{
|
| 4839 |
+
"epoch": 0.9468599033816425,
|
| 4840 |
+
"grad_norm": 22.972850799560547,
|
| 4841 |
+
"learning_rate": 1.4646164152307018e-06,
|
| 4842 |
+
"loss": 6.0495,
|
| 4843 |
+
"step": 686
|
| 4844 |
+
},
|
| 4845 |
+
{
|
| 4846 |
+
"epoch": 0.94824016563147,
|
| 4847 |
+
"grad_norm": 28.845718383789062,
|
| 4848 |
+
"learning_rate": 1.3906431538576626e-06,
|
| 4849 |
+
"loss": 6.6508,
|
| 4850 |
+
"step": 687
|
| 4851 |
+
},
|
| 4852 |
+
{
|
| 4853 |
+
"epoch": 0.9496204278812974,
|
| 4854 |
+
"grad_norm": 19.374427795410156,
|
| 4855 |
+
"learning_rate": 1.3185736213252809e-06,
|
| 4856 |
+
"loss": 3.758,
|
| 4857 |
+
"step": 688
|
| 4858 |
+
},
|
| 4859 |
+
{
|
| 4860 |
+
"epoch": 0.9510006901311249,
|
| 4861 |
+
"grad_norm": 24.994707107543945,
|
| 4862 |
+
"learning_rate": 1.2484092089908305e-06,
|
| 4863 |
+
"loss": 5.0294,
|
| 4864 |
+
"step": 689
|
| 4865 |
+
},
|
| 4866 |
+
{
|
| 4867 |
+
"epoch": 0.9523809523809523,
|
| 4868 |
+
"grad_norm": 23.40306282043457,
|
| 4869 |
+
"learning_rate": 1.1801512714318285e-06,
|
| 4870 |
+
"loss": 4.6954,
|
| 4871 |
+
"step": 690
|
| 4872 |
+
},
|
| 4873 |
+
{
|
| 4874 |
+
"epoch": 0.9537612146307799,
|
| 4875 |
+
"grad_norm": 28.18086051940918,
|
| 4876 |
+
"learning_rate": 1.113801126419789e-06,
|
| 4877 |
+
"loss": 5.1102,
|
| 4878 |
+
"step": 691
|
| 4879 |
+
},
|
| 4880 |
+
{
|
| 4881 |
+
"epoch": 0.9551414768806074,
|
| 4882 |
+
"grad_norm": 23.489946365356445,
|
| 4883 |
+
"learning_rate": 1.0493600548948878e-06,
|
| 4884 |
+
"loss": 4.7169,
|
| 4885 |
+
"step": 692
|
| 4886 |
+
},
|
| 4887 |
+
{
|
| 4888 |
+
"epoch": 0.9565217391304348,
|
| 4889 |
+
"grad_norm": 19.926145553588867,
|
| 4890 |
+
"learning_rate": 9.8682930094115e-07,
|
| 4891 |
+
"loss": 4.3657,
|
| 4892 |
+
"step": 693
|
| 4893 |
+
},
|
| 4894 |
+
{
|
| 4895 |
+
"epoch": 0.9579020013802623,
|
| 4896 |
+
"grad_norm": 21.520130157470703,
|
| 4897 |
+
"learning_rate": 9.262100717624677e-07,
|
| 4898 |
+
"loss": 3.7835,
|
| 4899 |
+
"step": 694
|
| 4900 |
+
},
|
| 4901 |
+
{
|
| 4902 |
+
"epoch": 0.9592822636300897,
|
| 4903 |
+
"grad_norm": 23.506120681762695,
|
| 4904 |
+
"learning_rate": 8.675035376593088e-07,
|
| 4905 |
+
"loss": 5.4438,
|
| 4906 |
+
"step": 695
|
| 4907 |
+
},
|
| 4908 |
+
{
|
| 4909 |
+
"epoch": 0.9606625258799172,
|
| 4910 |
+
"grad_norm": 39.26706314086914,
|
| 4911 |
+
"learning_rate": 8.107108320060674e-07,
|
| 4912 |
+
"loss": 5.6457,
|
| 4913 |
+
"step": 696
|
| 4914 |
+
},
|
| 4915 |
+
{
|
| 4916 |
+
"epoch": 0.9620427881297446,
|
| 4917 |
+
"grad_norm": 20.84963607788086,
|
| 4918 |
+
"learning_rate": 7.558330512292378e-07,
|
| 4919 |
+
"loss": 5.1845,
|
| 4920 |
+
"step": 697
|
| 4921 |
+
},
|
| 4922 |
+
{
|
| 4923 |
+
"epoch": 0.9634230503795721,
|
| 4924 |
+
"grad_norm": 20.25655174255371,
|
| 4925 |
+
"learning_rate": 7.028712547862526e-07,
|
| 4926 |
+
"loss": 3.9212,
|
| 4927 |
+
"step": 698
|
| 4928 |
+
},
|
| 4929 |
+
{
|
| 4930 |
+
"epoch": 0.9648033126293996,
|
| 4931 |
+
"grad_norm": 21.61455535888672,
|
| 4932 |
+
"learning_rate": 6.518264651449779e-07,
|
| 4933 |
+
"loss": 5.3809,
|
| 4934 |
+
"step": 699
|
| 4935 |
+
},
|
| 4936 |
+
{
|
| 4937 |
+
"epoch": 0.966183574879227,
|
| 4938 |
+
"grad_norm": 19.310720443725586,
|
| 4939 |
+
"learning_rate": 6.026996677640062e-07,
|
| 4940 |
+
"loss": 3.6578,
|
| 4941 |
+
"step": 700
|
| 4942 |
+
},
|
| 4943 |
+
{
|
| 4944 |
+
"epoch": 0.9675638371290545,
|
| 4945 |
+
"grad_norm": 20.59223175048828,
|
| 4946 |
+
"learning_rate": 5.554918110736273e-07,
|
| 4947 |
+
"loss": 4.8361,
|
| 4948 |
+
"step": 701
|
| 4949 |
+
},
|
| 4950 |
+
{
|
| 4951 |
+
"epoch": 0.968944099378882,
|
| 4952 |
+
"grad_norm": 18.125131607055664,
|
| 4953 |
+
"learning_rate": 5.102038064575099e-07,
|
| 4954 |
+
"loss": 2.8242,
|
| 4955 |
+
"step": 702
|
| 4956 |
+
},
|
| 4957 |
+
{
|
| 4958 |
+
"epoch": 0.9703243616287095,
|
| 4959 |
+
"grad_norm": 19.294221878051758,
|
| 4960 |
+
"learning_rate": 4.668365282351372e-07,
|
| 4961 |
+
"loss": 3.2637,
|
| 4962 |
+
"step": 703
|
| 4963 |
+
},
|
| 4964 |
+
{
|
| 4965 |
+
"epoch": 0.971704623878537,
|
| 4966 |
+
"grad_norm": 15.094534873962402,
|
| 4967 |
+
"learning_rate": 4.2539081364488807e-07,
|
| 4968 |
+
"loss": 3.3754,
|
| 4969 |
+
"step": 704
|
| 4970 |
+
},
|
| 4971 |
+
{
|
| 4972 |
+
"epoch": 0.9730848861283644,
|
| 4973 |
+
"grad_norm": 19.29479217529297,
|
| 4974 |
+
"learning_rate": 3.8586746282788244e-07,
|
| 4975 |
+
"loss": 4.4401,
|
| 4976 |
+
"step": 705
|
| 4977 |
+
},
|
| 4978 |
+
{
|
| 4979 |
+
"epoch": 0.9744651483781919,
|
| 4980 |
+
"grad_norm": 27.921985626220703,
|
| 4981 |
+
"learning_rate": 3.4826723881257186e-07,
|
| 4982 |
+
"loss": 5.3932,
|
| 4983 |
+
"step": 706
|
| 4984 |
+
},
|
| 4985 |
+
{
|
| 4986 |
+
"epoch": 0.9758454106280193,
|
| 4987 |
+
"grad_norm": 26.087324142456055,
|
| 4988 |
+
"learning_rate": 3.1259086749992895e-07,
|
| 4989 |
+
"loss": 5.4936,
|
| 4990 |
+
"step": 707
|
| 4991 |
+
},
|
| 4992 |
+
{
|
| 4993 |
+
"epoch": 0.9772256728778468,
|
| 4994 |
+
"grad_norm": 19.32054901123047,
|
| 4995 |
+
"learning_rate": 2.7883903764953644e-07,
|
| 4996 |
+
"loss": 3.5865,
|
| 4997 |
+
"step": 708
|
| 4998 |
+
},
|
| 4999 |
+
{
|
| 5000 |
+
"epoch": 0.9786059351276742,
|
| 5001 |
+
"grad_norm": 19.498537063598633,
|
| 5002 |
+
"learning_rate": 2.470124008661978e-07,
|
| 5003 |
+
"loss": 3.1686,
|
| 5004 |
+
"step": 709
|
| 5005 |
+
},
|
| 5006 |
+
{
|
| 5007 |
+
"epoch": 0.9799861973775017,
|
| 5008 |
+
"grad_norm": 22.79661750793457,
|
| 5009 |
+
"learning_rate": 2.171115715874139e-07,
|
| 5010 |
+
"loss": 4.4749,
|
| 5011 |
+
"step": 710
|
| 5012 |
+
},
|
| 5013 |
+
{
|
| 5014 |
+
"epoch": 0.9813664596273292,
|
| 5015 |
+
"grad_norm": 16.39584732055664,
|
| 5016 |
+
"learning_rate": 1.8913712707149256e-07,
|
| 5017 |
+
"loss": 3.414,
|
| 5018 |
+
"step": 711
|
| 5019 |
+
},
|
| 5020 |
+
{
|
| 5021 |
+
"epoch": 0.9827467218771566,
|
| 5022 |
+
"grad_norm": 19.80657196044922,
|
| 5023 |
+
"learning_rate": 1.630896073864352e-07,
|
| 5024 |
+
"loss": 4.0456,
|
| 5025 |
+
"step": 712
|
| 5026 |
+
},
|
| 5027 |
+
{
|
| 5028 |
+
"epoch": 0.9841269841269841,
|
| 5029 |
+
"grad_norm": 24.56593894958496,
|
| 5030 |
+
"learning_rate": 1.3896951539945636e-07,
|
| 5031 |
+
"loss": 5.6586,
|
| 5032 |
+
"step": 713
|
| 5033 |
+
},
|
| 5034 |
+
{
|
| 5035 |
+
"epoch": 0.9855072463768116,
|
| 5036 |
+
"grad_norm": 23.019926071166992,
|
| 5037 |
+
"learning_rate": 1.1677731676733584e-07,
|
| 5038 |
+
"loss": 4.3073,
|
| 5039 |
+
"step": 714
|
| 5040 |
+
},
|
| 5041 |
+
{
|
| 5042 |
+
"epoch": 0.9868875086266391,
|
| 5043 |
+
"grad_norm": 26.53896713256836,
|
| 5044 |
+
"learning_rate": 9.65134399274037e-08,
|
| 5045 |
+
"loss": 5.958,
|
| 5046 |
+
"step": 715
|
| 5047 |
+
},
|
| 5048 |
+
{
|
| 5049 |
+
"epoch": 0.9882677708764666,
|
| 5050 |
+
"grad_norm": 15.73496150970459,
|
| 5051 |
+
"learning_rate": 7.817827608924688e-08,
|
| 5052 |
+
"loss": 2.9358,
|
| 5053 |
+
"step": 716
|
| 5054 |
+
},
|
| 5055 |
+
{
|
| 5056 |
+
"epoch": 0.989648033126294,
|
| 5057 |
+
"grad_norm": 23.03923797607422,
|
| 5058 |
+
"learning_rate": 6.177217922718192e-08,
|
| 5059 |
+
"loss": 4.2186,
|
| 5060 |
+
"step": 717
|
| 5061 |
+
},
|
| 5062 |
+
{
|
| 5063 |
+
"epoch": 0.9910282953761215,
|
| 5064 |
+
"grad_norm": 20.209564208984375,
|
| 5065 |
+
"learning_rate": 4.7295466073427053e-08,
|
| 5066 |
+
"loss": 4.3526,
|
| 5067 |
+
"step": 718
|
| 5068 |
+
},
|
| 5069 |
+
{
|
| 5070 |
+
"epoch": 0.9924085576259489,
|
| 5071 |
+
"grad_norm": 17.545196533203125,
|
| 5072 |
+
"learning_rate": 3.474841611197377e-08,
|
| 5073 |
+
"loss": 3.5361,
|
| 5074 |
+
"step": 719
|
| 5075 |
+
},
|
| 5076 |
+
{
|
| 5077 |
+
"epoch": 0.9937888198757764,
|
| 5078 |
+
"grad_norm": 16.188581466674805,
|
| 5079 |
+
"learning_rate": 2.4131271573191172e-08,
|
| 5080 |
+
"loss": 2.5456,
|
| 5081 |
+
"step": 720
|
| 5082 |
+
},
|
| 5083 |
+
{
|
| 5084 |
+
"epoch": 0.9951690821256038,
|
| 5085 |
+
"grad_norm": 17.67585563659668,
|
| 5086 |
+
"learning_rate": 1.5444237429140807e-08,
|
| 5087 |
+
"loss": 3.162,
|
| 5088 |
+
"step": 721
|
| 5089 |
+
},
|
| 5090 |
+
{
|
| 5091 |
+
"epoch": 0.9965493443754313,
|
| 5092 |
+
"grad_norm": 21.626819610595703,
|
| 5093 |
+
"learning_rate": 8.687481389657582e-09,
|
| 5094 |
+
"loss": 4.8709,
|
| 5095 |
+
"step": 722
|
| 5096 |
+
},
|
| 5097 |
+
{
|
| 5098 |
+
"epoch": 0.9979296066252588,
|
| 5099 |
+
"grad_norm": 18.768949508666992,
|
| 5100 |
+
"learning_rate": 3.861133899063507e-09,
|
| 5101 |
+
"loss": 3.8211,
|
| 5102 |
+
"step": 723
|
| 5103 |
+
},
|
| 5104 |
+
{
|
| 5105 |
+
"epoch": 0.9993098688750862,
|
| 5106 |
+
"grad_norm": 20.41644287109375,
|
| 5107 |
+
"learning_rate": 9.65288133669695e-10,
|
| 5108 |
+
"loss": 3.3884,
|
| 5109 |
+
"step": 724
|
| 5110 |
+
},
|
| 5111 |
+
{
|
| 5112 |
+
"epoch": 1.0006901311249137,
|
| 5113 |
+
"grad_norm": 21.81919288635254,
|
| 5114 |
+
"learning_rate": 0.0,
|
| 5115 |
+
"loss": 3.69,
|
| 5116 |
+
"step": 725
|
| 5117 |
}
|
| 5118 |
],
|
| 5119 |
"logging_steps": 1,
|
|
|
|
| 5128 |
"should_evaluate": false,
|
| 5129 |
"should_log": false,
|
| 5130 |
"should_save": true,
|
| 5131 |
+
"should_training_stop": true
|
| 5132 |
},
|
| 5133 |
"attributes": {}
|
| 5134 |
}
|
| 5135 |
},
|
| 5136 |
+
"total_flos": 1.2704672407722394e+17,
|
| 5137 |
"train_batch_size": 2,
|
| 5138 |
"trial_name": null,
|
| 5139 |
"trial_params": null
|