Training in progress, step 534, 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 125918320
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b9f81d0efa5188128ee4bac973f68e17b7b28562e9fad5ac815d55ab076a9676
|
| 3 |
size 125918320
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 64684244
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:956b80a47a080bf4e4914f80e77ef9e40fae68aa309e96b2a4a5fe2d27a0c5ef
|
| 3 |
size 64684244
|
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 sha256:62313ddb8dc64325159ef14adb7791952640fd6b66568d544210bad8634941a8
|
| 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:acd7f4bc9eb4b305e7ac32f54a84c099be33ac10feb45323ce141b57fcbba77d
|
| 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": 500,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
@@ -2499,6 +2499,1252 @@
|
|
| 2499 |
"learning_rate": 5.044497892100981e-05,
|
| 2500 |
"loss": 0.5968,
|
| 2501 |
"step": 356
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2502 |
}
|
| 2503 |
],
|
| 2504 |
"logging_steps": 1,
|
|
@@ -2518,7 +3764,7 @@
|
|
| 2518 |
"attributes": {}
|
| 2519 |
}
|
| 2520 |
},
|
| 2521 |
-
"total_flos":
|
| 2522 |
"train_batch_size": 4,
|
| 2523 |
"trial_name": null,
|
| 2524 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.7518479408658922,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 534,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 2499 |
"learning_rate": 5.044497892100981e-05,
|
| 2500 |
"loss": 0.5968,
|
| 2501 |
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.5026399155227033,
|
| 2505 |
+
"grad_norm": 23.293493270874023,
|
| 2506 |
+
"learning_rate": 5.022249166329632e-05,
|
| 2507 |
+
"loss": 1.3873,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.504047870468145,
|
| 2512 |
+
"grad_norm": 15.999886512756348,
|
| 2513 |
+
"learning_rate": 5e-05,
|
| 2514 |
+
"loss": 0.9723,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.5054558254135868,
|
| 2519 |
+
"grad_norm": 29.49577522277832,
|
| 2520 |
+
"learning_rate": 4.9777508336703705e-05,
|
| 2521 |
+
"loss": 1.6385,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.5068637803590285,
|
| 2526 |
+
"grad_norm": 17.99614143371582,
|
| 2527 |
+
"learning_rate": 4.9555021078990206e-05,
|
| 2528 |
+
"loss": 0.9292,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.5082717353044702,
|
| 2533 |
+
"grad_norm": 15.118061065673828,
|
| 2534 |
+
"learning_rate": 4.9332542632355114e-05,
|
| 2535 |
+
"loss": 0.7235,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.509679690249912,
|
| 2540 |
+
"grad_norm": 16.40995979309082,
|
| 2541 |
+
"learning_rate": 4.911007740211953e-05,
|
| 2542 |
+
"loss": 0.976,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.5110876451953538,
|
| 2547 |
+
"grad_norm": 21.50313949584961,
|
| 2548 |
+
"learning_rate": 4.888762979334285e-05,
|
| 2549 |
+
"loss": 0.9848,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.5124956001407955,
|
| 2554 |
+
"grad_norm": 17.15250015258789,
|
| 2555 |
+
"learning_rate": 4.866520421073556e-05,
|
| 2556 |
+
"loss": 0.9352,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.5139035550862372,
|
| 2561 |
+
"grad_norm": 10.609292030334473,
|
| 2562 |
+
"learning_rate": 4.844280505857202e-05,
|
| 2563 |
+
"loss": 0.6787,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.515311510031679,
|
| 2568 |
+
"grad_norm": 24.926061630249023,
|
| 2569 |
+
"learning_rate": 4.822043674060323e-05,
|
| 2570 |
+
"loss": 0.9806,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.5167194649771207,
|
| 2575 |
+
"grad_norm": 10.759723663330078,
|
| 2576 |
+
"learning_rate": 4.799810365996961e-05,
|
| 2577 |
+
"loss": 0.7693,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.5181274199225625,
|
| 2582 |
+
"grad_norm": 12.087748527526855,
|
| 2583 |
+
"learning_rate": 4.7775810219113856e-05,
|
| 2584 |
+
"loss": 1.1873,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.5195353748680043,
|
| 2589 |
+
"grad_norm": 15.061980247497559,
|
| 2590 |
+
"learning_rate": 4.7553560819693775e-05,
|
| 2591 |
+
"loss": 1.1541,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.5209433298134459,
|
| 2596 |
+
"grad_norm": 12.949942588806152,
|
| 2597 |
+
"learning_rate": 4.7331359862495087e-05,
|
| 2598 |
+
"loss": 0.9467,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.5223512847588877,
|
| 2603 |
+
"grad_norm": 7.894239902496338,
|
| 2604 |
+
"learning_rate": 4.7109211747344274e-05,
|
| 2605 |
+
"loss": 0.5866,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.5237592397043295,
|
| 2610 |
+
"grad_norm": 5.802740097045898,
|
| 2611 |
+
"learning_rate": 4.688712087302155e-05,
|
| 2612 |
+
"loss": 0.5254,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.5251671946497712,
|
| 2617 |
+
"grad_norm": 3.9965198040008545,
|
| 2618 |
+
"learning_rate": 4.6665091637173653e-05,
|
| 2619 |
+
"loss": 0.3989,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.5265751495952129,
|
| 2624 |
+
"grad_norm": 11.480055809020996,
|
| 2625 |
+
"learning_rate": 4.644312843622681e-05,
|
| 2626 |
+
"loss": 0.8184,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.5279831045406547,
|
| 2631 |
+
"grad_norm": 16.241785049438477,
|
| 2632 |
+
"learning_rate": 4.622123566529969e-05,
|
| 2633 |
+
"loss": 1.0828,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.5293910594860964,
|
| 2638 |
+
"grad_norm": 9.330286026000977,
|
| 2639 |
+
"learning_rate": 4.599941771811638e-05,
|
| 2640 |
+
"loss": 0.7499,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.5307990144315382,
|
| 2645 |
+
"grad_norm": 12.609516143798828,
|
| 2646 |
+
"learning_rate": 4.577767898691937e-05,
|
| 2647 |
+
"loss": 0.7155,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.53220696937698,
|
| 2652 |
+
"grad_norm": 10.434843063354492,
|
| 2653 |
+
"learning_rate": 4.555602386238257e-05,
|
| 2654 |
+
"loss": 0.7718,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.5336149243224216,
|
| 2659 |
+
"grad_norm": 4.869112968444824,
|
| 2660 |
+
"learning_rate": 4.533445673352438e-05,
|
| 2661 |
+
"loss": 0.3643,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.5350228792678634,
|
| 2666 |
+
"grad_norm": 14.049832344055176,
|
| 2667 |
+
"learning_rate": 4.511298198762085e-05,
|
| 2668 |
+
"loss": 0.8233,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.5364308342133052,
|
| 2673 |
+
"grad_norm": 7.066025257110596,
|
| 2674 |
+
"learning_rate": 4.4891604010118655e-05,
|
| 2675 |
+
"loss": 0.3357,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.5378387891587469,
|
| 2680 |
+
"grad_norm": 20.782114028930664,
|
| 2681 |
+
"learning_rate": 4.467032718454838e-05,
|
| 2682 |
+
"loss": 0.8549,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.5392467441041887,
|
| 2687 |
+
"grad_norm": 7.494271278381348,
|
| 2688 |
+
"learning_rate": 4.444915589243775e-05,
|
| 2689 |
+
"loss": 0.5325,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.5406546990496304,
|
| 2694 |
+
"grad_norm": 25.774547576904297,
|
| 2695 |
+
"learning_rate": 4.4228094513224725e-05,
|
| 2696 |
+
"loss": 0.9002,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.5420626539950721,
|
| 2701 |
+
"grad_norm": 10.338793754577637,
|
| 2702 |
+
"learning_rate": 4.4007147424170914e-05,
|
| 2703 |
+
"loss": 0.6071,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.5434706089405139,
|
| 2708 |
+
"grad_norm": 68.29295349121094,
|
| 2709 |
+
"learning_rate": 4.378631900027482e-05,
|
| 2710 |
+
"loss": 1.8258,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.5448785638859557,
|
| 2715 |
+
"grad_norm": 69.48816680908203,
|
| 2716 |
+
"learning_rate": 4.356561361418529e-05,
|
| 2717 |
+
"loss": 1.15,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.5462865188313974,
|
| 2722 |
+
"grad_norm": 39.41474533081055,
|
| 2723 |
+
"learning_rate": 4.3345035636114854e-05,
|
| 2724 |
+
"loss": 0.5735,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.5476944737768391,
|
| 2729 |
+
"grad_norm": 39.44690704345703,
|
| 2730 |
+
"learning_rate": 4.312458943375319e-05,
|
| 2731 |
+
"loss": 0.9276,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.5491024287222809,
|
| 2736 |
+
"grad_norm": 85.98686981201172,
|
| 2737 |
+
"learning_rate": 4.290427937218078e-05,
|
| 2738 |
+
"loss": 1.4463,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.5505103836677226,
|
| 2743 |
+
"grad_norm": 38.16764831542969,
|
| 2744 |
+
"learning_rate": 4.268410981378223e-05,
|
| 2745 |
+
"loss": 0.5563,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.5519183386131644,
|
| 2750 |
+
"grad_norm": 20.098003387451172,
|
| 2751 |
+
"learning_rate": 4.246408511816014e-05,
|
| 2752 |
+
"loss": 0.9429,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.5533262935586061,
|
| 2757 |
+
"grad_norm": 8.55196475982666,
|
| 2758 |
+
"learning_rate": 4.224420964204856e-05,
|
| 2759 |
+
"loss": 0.6951,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.5547342485040478,
|
| 2764 |
+
"grad_norm": 8.005936622619629,
|
| 2765 |
+
"learning_rate": 4.202448773922693e-05,
|
| 2766 |
+
"loss": 0.4835,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.5561422034494896,
|
| 2771 |
+
"grad_norm": 17.085020065307617,
|
| 2772 |
+
"learning_rate": 4.1804923760433715e-05,
|
| 2773 |
+
"loss": 1.3263,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.5575501583949314,
|
| 2778 |
+
"grad_norm": 47.16075134277344,
|
| 2779 |
+
"learning_rate": 4.158552205328029e-05,
|
| 2780 |
+
"loss": 1.0402,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.5589581133403732,
|
| 2785 |
+
"grad_norm": 9.869107246398926,
|
| 2786 |
+
"learning_rate": 4.1366286962164904e-05,
|
| 2787 |
+
"loss": 0.7081,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.5603660682858148,
|
| 2792 |
+
"grad_norm": 11.739812850952148,
|
| 2793 |
+
"learning_rate": 4.114722282818663e-05,
|
| 2794 |
+
"loss": 0.9259,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.5617740232312566,
|
| 2799 |
+
"grad_norm": 18.6031494140625,
|
| 2800 |
+
"learning_rate": 4.092833398905939e-05,
|
| 2801 |
+
"loss": 0.8313,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.5631819781766984,
|
| 2806 |
+
"grad_norm": 29.32172393798828,
|
| 2807 |
+
"learning_rate": 4.070962477902602e-05,
|
| 2808 |
+
"loss": 1.7005,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.5645899331221401,
|
| 2813 |
+
"grad_norm": 37.216609954833984,
|
| 2814 |
+
"learning_rate": 4.04910995287726e-05,
|
| 2815 |
+
"loss": 0.3046,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.5659978880675819,
|
| 2820 |
+
"grad_norm": 3.302595853805542,
|
| 2821 |
+
"learning_rate": 4.0272762565342526e-05,
|
| 2822 |
+
"loss": 0.227,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.5674058430130235,
|
| 2827 |
+
"grad_norm": 11.610830307006836,
|
| 2828 |
+
"learning_rate": 4.005461821205092e-05,
|
| 2829 |
+
"loss": 0.622,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.5688137979584653,
|
| 2834 |
+
"grad_norm": 11.117403984069824,
|
| 2835 |
+
"learning_rate": 3.983667078839901e-05,
|
| 2836 |
+
"loss": 0.7481,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.5702217529039071,
|
| 2841 |
+
"grad_norm": 12.287819862365723,
|
| 2842 |
+
"learning_rate": 3.961892460998863e-05,
|
| 2843 |
+
"loss": 0.9908,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.5716297078493489,
|
| 2848 |
+
"grad_norm": 12.712357521057129,
|
| 2849 |
+
"learning_rate": 3.940138398843669e-05,
|
| 2850 |
+
"loss": 0.7259,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.5730376627947905,
|
| 2855 |
+
"grad_norm": 7.371547222137451,
|
| 2856 |
+
"learning_rate": 3.9184053231289856e-05,
|
| 2857 |
+
"loss": 0.6341,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.5744456177402323,
|
| 2862 |
+
"grad_norm": 4.734263896942139,
|
| 2863 |
+
"learning_rate": 3.896693664193923e-05,
|
| 2864 |
+
"loss": 0.2962,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.575853572685674,
|
| 2869 |
+
"grad_norm": 7.0978264808654785,
|
| 2870 |
+
"learning_rate": 3.875003851953523e-05,
|
| 2871 |
+
"loss": 0.5214,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.5772615276311158,
|
| 2876 |
+
"grad_norm": 5.476172924041748,
|
| 2877 |
+
"learning_rate": 3.85333631589023e-05,
|
| 2878 |
+
"loss": 0.4687,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.5786694825765576,
|
| 2883 |
+
"grad_norm": 13.698134422302246,
|
| 2884 |
+
"learning_rate": 3.8316914850453975e-05,
|
| 2885 |
+
"loss": 1.2396,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.5800774375219993,
|
| 2890 |
+
"grad_norm": 18.33237648010254,
|
| 2891 |
+
"learning_rate": 3.810069788010793e-05,
|
| 2892 |
+
"loss": 1.7078,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.581485392467441,
|
| 2897 |
+
"grad_norm": 16.64714241027832,
|
| 2898 |
+
"learning_rate": 3.788471652920105e-05,
|
| 2899 |
+
"loss": 1.4505,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.5828933474128828,
|
| 2904 |
+
"grad_norm": 12.553242683410645,
|
| 2905 |
+
"learning_rate": 3.7668975074404715e-05,
|
| 2906 |
+
"loss": 1.0523,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.5843013023583246,
|
| 2911 |
+
"grad_norm": 10.081384658813477,
|
| 2912 |
+
"learning_rate": 3.745347778764008e-05,
|
| 2913 |
+
"loss": 0.7268,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.5857092573037663,
|
| 2918 |
+
"grad_norm": 7.674866676330566,
|
| 2919 |
+
"learning_rate": 3.723822893599355e-05,
|
| 2920 |
+
"loss": 0.4389,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.587117212249208,
|
| 2925 |
+
"grad_norm": 13.804383277893066,
|
| 2926 |
+
"learning_rate": 3.702323278163217e-05,
|
| 2927 |
+
"loss": 0.5402,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.5885251671946498,
|
| 2932 |
+
"grad_norm": 10.615878105163574,
|
| 2933 |
+
"learning_rate": 3.680849358171933e-05,
|
| 2934 |
+
"loss": 0.9286,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.5899331221400915,
|
| 2939 |
+
"grad_norm": 9.648208618164062,
|
| 2940 |
+
"learning_rate": 3.659401558833042e-05,
|
| 2941 |
+
"loss": 0.5558,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.5913410770855333,
|
| 2946 |
+
"grad_norm": 78.71580505371094,
|
| 2947 |
+
"learning_rate": 3.63798030483687e-05,
|
| 2948 |
+
"loss": 2.0743,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.592749032030975,
|
| 2953 |
+
"grad_norm": 20.05280113220215,
|
| 2954 |
+
"learning_rate": 3.616586020348108e-05,
|
| 2955 |
+
"loss": 1.4442,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.5941569869764167,
|
| 2960 |
+
"grad_norm": 20.129953384399414,
|
| 2961 |
+
"learning_rate": 3.595219128997427e-05,
|
| 2962 |
+
"loss": 1.4081,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.5955649419218585,
|
| 2967 |
+
"grad_norm": 16.63743019104004,
|
| 2968 |
+
"learning_rate": 3.57388005387308e-05,
|
| 2969 |
+
"loss": 1.0921,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.5969728968673003,
|
| 2974 |
+
"grad_norm": 13.33332633972168,
|
| 2975 |
+
"learning_rate": 3.5525692175125284e-05,
|
| 2976 |
+
"loss": 0.8281,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.598380851812742,
|
| 2981 |
+
"grad_norm": 9.779206275939941,
|
| 2982 |
+
"learning_rate": 3.531287041894076e-05,
|
| 2983 |
+
"loss": 0.5832,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.5997888067581837,
|
| 2988 |
+
"grad_norm": 7.802907466888428,
|
| 2989 |
+
"learning_rate": 3.510033948428508e-05,
|
| 2990 |
+
"loss": 0.5807,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.6011967617036255,
|
| 2995 |
+
"grad_norm": 5.158997535705566,
|
| 2996 |
+
"learning_rate": 3.488810357950757e-05,
|
| 2997 |
+
"loss": 0.3516,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.6026047166490672,
|
| 3002 |
+
"grad_norm": 7.417584419250488,
|
| 3003 |
+
"learning_rate": 3.4676166907115595e-05,
|
| 3004 |
+
"loss": 0.5142,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.604012671594509,
|
| 3009 |
+
"grad_norm": 5.034250259399414,
|
| 3010 |
+
"learning_rate": 3.446453366369136e-05,
|
| 3011 |
+
"loss": 0.3203,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.6054206265399508,
|
| 3016 |
+
"grad_norm": 19.075397491455078,
|
| 3017 |
+
"learning_rate": 3.42532080398089e-05,
|
| 3018 |
+
"loss": 0.5543,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.6068285814853924,
|
| 3023 |
+
"grad_norm": 5.768478870391846,
|
| 3024 |
+
"learning_rate": 3.4042194219951e-05,
|
| 3025 |
+
"loss": 0.4293,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.6082365364308342,
|
| 3030 |
+
"grad_norm": 19.794889450073242,
|
| 3031 |
+
"learning_rate": 3.383149638242639e-05,
|
| 3032 |
+
"loss": 0.6996,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.609644491376276,
|
| 3037 |
+
"grad_norm": 13.78519344329834,
|
| 3038 |
+
"learning_rate": 3.362111869928696e-05,
|
| 3039 |
+
"loss": 0.8824,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.6110524463217177,
|
| 3044 |
+
"grad_norm": 19.145423889160156,
|
| 3045 |
+
"learning_rate": 3.3411065336245273e-05,
|
| 3046 |
+
"loss": 1.2345,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.6124604012671595,
|
| 3051 |
+
"grad_norm": 9.899946212768555,
|
| 3052 |
+
"learning_rate": 3.320134045259192e-05,
|
| 3053 |
+
"loss": 0.8943,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.6138683562126012,
|
| 3058 |
+
"grad_norm": 35.387332916259766,
|
| 3059 |
+
"learning_rate": 3.299194820111326e-05,
|
| 3060 |
+
"loss": 1.3449,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.6152763111580429,
|
| 3065 |
+
"grad_norm": 12.40296459197998,
|
| 3066 |
+
"learning_rate": 3.2782892728009144e-05,
|
| 3067 |
+
"loss": 0.7122,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.6166842661034847,
|
| 3072 |
+
"grad_norm": 82.74261474609375,
|
| 3073 |
+
"learning_rate": 3.257417817281086e-05,
|
| 3074 |
+
"loss": 1.9809,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.6180922210489265,
|
| 3079 |
+
"grad_norm": 16.26070213317871,
|
| 3080 |
+
"learning_rate": 3.236580866829914e-05,
|
| 3081 |
+
"loss": 1.0837,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.6195001759943681,
|
| 3086 |
+
"grad_norm": 28.62177848815918,
|
| 3087 |
+
"learning_rate": 3.215778834042229e-05,
|
| 3088 |
+
"loss": 1.2744,
|
| 3089 |
+
"step": 440
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.6209081309398099,
|
| 3093 |
+
"grad_norm": 13.872343063354492,
|
| 3094 |
+
"learning_rate": 3.195012130821461e-05,
|
| 3095 |
+
"loss": 0.6844,
|
| 3096 |
+
"step": 441
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.6223160858852517,
|
| 3100 |
+
"grad_norm": 39.674625396728516,
|
| 3101 |
+
"learning_rate": 3.174281168371469e-05,
|
| 3102 |
+
"loss": 1.3668,
|
| 3103 |
+
"step": 442
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.6237240408306934,
|
| 3107 |
+
"grad_norm": 18.342069625854492,
|
| 3108 |
+
"learning_rate": 3.153586357188404e-05,
|
| 3109 |
+
"loss": 1.1236,
|
| 3110 |
+
"step": 443
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.6251319957761352,
|
| 3114 |
+
"grad_norm": 20.320011138916016,
|
| 3115 |
+
"learning_rate": 3.1329281070525826e-05,
|
| 3116 |
+
"loss": 0.8428,
|
| 3117 |
+
"step": 444
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.6265399507215769,
|
| 3121 |
+
"grad_norm": 9.062244415283203,
|
| 3122 |
+
"learning_rate": 3.112306827020377e-05,
|
| 3123 |
+
"loss": 0.6577,
|
| 3124 |
+
"step": 445
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.6279479056670186,
|
| 3128 |
+
"grad_norm": 5.732834339141846,
|
| 3129 |
+
"learning_rate": 3.0917229254161054e-05,
|
| 3130 |
+
"loss": 0.4779,
|
| 3131 |
+
"step": 446
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.6293558606124604,
|
| 3135 |
+
"grad_norm": 5.922752380371094,
|
| 3136 |
+
"learning_rate": 3.071176809823952e-05,
|
| 3137 |
+
"loss": 0.4158,
|
| 3138 |
+
"step": 447
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.6307638155579022,
|
| 3142 |
+
"grad_norm": 22.569480895996094,
|
| 3143 |
+
"learning_rate": 3.0506688870798944e-05,
|
| 3144 |
+
"loss": 1.9873,
|
| 3145 |
+
"step": 448
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.6321717705033439,
|
| 3149 |
+
"grad_norm": 7.844541549682617,
|
| 3150 |
+
"learning_rate": 3.0301995632636537e-05,
|
| 3151 |
+
"loss": 0.7175,
|
| 3152 |
+
"step": 449
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.6335797254487856,
|
| 3156 |
+
"grad_norm": 21.511882781982422,
|
| 3157 |
+
"learning_rate": 3.0097692436906465e-05,
|
| 3158 |
+
"loss": 1.7123,
|
| 3159 |
+
"step": 450
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.6349876803942274,
|
| 3163 |
+
"grad_norm": 6.862496852874756,
|
| 3164 |
+
"learning_rate": 2.9893783329039586e-05,
|
| 3165 |
+
"loss": 0.1849,
|
| 3166 |
+
"step": 451
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.6363956353396691,
|
| 3170 |
+
"grad_norm": 4.311605930328369,
|
| 3171 |
+
"learning_rate": 2.9690272346663463e-05,
|
| 3172 |
+
"loss": 0.3234,
|
| 3173 |
+
"step": 452
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.6378035902851109,
|
| 3177 |
+
"grad_norm": 8.582545280456543,
|
| 3178 |
+
"learning_rate": 2.9487163519522264e-05,
|
| 3179 |
+
"loss": 0.7373,
|
| 3180 |
+
"step": 453
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.6392115452305526,
|
| 3184 |
+
"grad_norm": 6.1634602546691895,
|
| 3185 |
+
"learning_rate": 2.9284460869397044e-05,
|
| 3186 |
+
"loss": 0.6531,
|
| 3187 |
+
"step": 454
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.6406195001759943,
|
| 3191 |
+
"grad_norm": 5.800738334655762,
|
| 3192 |
+
"learning_rate": 2.9082168410026102e-05,
|
| 3193 |
+
"loss": 0.602,
|
| 3194 |
+
"step": 455
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.6420274551214361,
|
| 3198 |
+
"grad_norm": 11.040019035339355,
|
| 3199 |
+
"learning_rate": 2.8880290147025548e-05,
|
| 3200 |
+
"loss": 0.9513,
|
| 3201 |
+
"step": 456
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.6434354100668779,
|
| 3205 |
+
"grad_norm": 3.3194429874420166,
|
| 3206 |
+
"learning_rate": 2.8678830077809855e-05,
|
| 3207 |
+
"loss": 0.3049,
|
| 3208 |
+
"step": 457
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.6448433650123196,
|
| 3212 |
+
"grad_norm": 5.3615336418151855,
|
| 3213 |
+
"learning_rate": 2.8477792191512898e-05,
|
| 3214 |
+
"loss": 0.523,
|
| 3215 |
+
"step": 458
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.6462513199577613,
|
| 3219 |
+
"grad_norm": 9.6665678024292,
|
| 3220 |
+
"learning_rate": 2.827718046890875e-05,
|
| 3221 |
+
"loss": 0.7228,
|
| 3222 |
+
"step": 459
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.6476592749032031,
|
| 3226 |
+
"grad_norm": 5.203541278839111,
|
| 3227 |
+
"learning_rate": 2.807699888233305e-05,
|
| 3228 |
+
"loss": 0.4447,
|
| 3229 |
+
"step": 460
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.6490672298486448,
|
| 3233 |
+
"grad_norm": 6.642237186431885,
|
| 3234 |
+
"learning_rate": 2.7877251395604252e-05,
|
| 3235 |
+
"loss": 0.7796,
|
| 3236 |
+
"step": 461
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.6504751847940866,
|
| 3240 |
+
"grad_norm": 14.666544914245605,
|
| 3241 |
+
"learning_rate": 2.7677941963945087e-05,
|
| 3242 |
+
"loss": 0.9353,
|
| 3243 |
+
"step": 462
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.6518831397395284,
|
| 3247 |
+
"grad_norm": 13.554447174072266,
|
| 3248 |
+
"learning_rate": 2.747907453390443e-05,
|
| 3249 |
+
"loss": 0.7865,
|
| 3250 |
+
"step": 463
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.65329109468497,
|
| 3254 |
+
"grad_norm": 8.27875804901123,
|
| 3255 |
+
"learning_rate": 2.7280653043278914e-05,
|
| 3256 |
+
"loss": 0.7993,
|
| 3257 |
+
"step": 464
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.6546990496304118,
|
| 3261 |
+
"grad_norm": 4.644435882568359,
|
| 3262 |
+
"learning_rate": 2.7082681421035093e-05,
|
| 3263 |
+
"loss": 0.4763,
|
| 3264 |
+
"step": 465
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.6561070045758536,
|
| 3268 |
+
"grad_norm": 18.395509719848633,
|
| 3269 |
+
"learning_rate": 2.6885163587231686e-05,
|
| 3270 |
+
"loss": 0.626,
|
| 3271 |
+
"step": 466
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.6575149595212954,
|
| 3275 |
+
"grad_norm": 8.523805618286133,
|
| 3276 |
+
"learning_rate": 2.6688103452941847e-05,
|
| 3277 |
+
"loss": 0.5022,
|
| 3278 |
+
"step": 467
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.6589229144667371,
|
| 3282 |
+
"grad_norm": 5.822423934936523,
|
| 3283 |
+
"learning_rate": 2.6491504920175804e-05,
|
| 3284 |
+
"loss": 0.8194,
|
| 3285 |
+
"step": 468
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.6603308694121788,
|
| 3289 |
+
"grad_norm": 5.749691963195801,
|
| 3290 |
+
"learning_rate": 2.6295371881803504e-05,
|
| 3291 |
+
"loss": 0.7872,
|
| 3292 |
+
"step": 469
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.6617388243576205,
|
| 3296 |
+
"grad_norm": 7.810802459716797,
|
| 3297 |
+
"learning_rate": 2.6099708221477644e-05,
|
| 3298 |
+
"loss": 0.903,
|
| 3299 |
+
"step": 470
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.6631467793030623,
|
| 3303 |
+
"grad_norm": 7.922653675079346,
|
| 3304 |
+
"learning_rate": 2.59045178135567e-05,
|
| 3305 |
+
"loss": 0.7677,
|
| 3306 |
+
"step": 471
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.6645547342485041,
|
| 3310 |
+
"grad_norm": 5.2952470779418945,
|
| 3311 |
+
"learning_rate": 2.570980452302818e-05,
|
| 3312 |
+
"loss": 0.6949,
|
| 3313 |
+
"step": 472
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.6659626891939457,
|
| 3317 |
+
"grad_norm": 6.853337287902832,
|
| 3318 |
+
"learning_rate": 2.5515572205432192e-05,
|
| 3319 |
+
"loss": 0.8158,
|
| 3320 |
+
"step": 473
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.6673706441393875,
|
| 3324 |
+
"grad_norm": 5.707924842834473,
|
| 3325 |
+
"learning_rate": 2.5321824706785017e-05,
|
| 3326 |
+
"loss": 0.5228,
|
| 3327 |
+
"step": 474
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.6687785990848293,
|
| 3331 |
+
"grad_norm": 5.7441325187683105,
|
| 3332 |
+
"learning_rate": 2.5128565863503008e-05,
|
| 3333 |
+
"loss": 0.6525,
|
| 3334 |
+
"step": 475
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.670186554030271,
|
| 3338 |
+
"grad_norm": 12.442033767700195,
|
| 3339 |
+
"learning_rate": 2.4935799502326533e-05,
|
| 3340 |
+
"loss": 0.6434,
|
| 3341 |
+
"step": 476
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.6715945089757128,
|
| 3345 |
+
"grad_norm": 11.724411010742188,
|
| 3346 |
+
"learning_rate": 2.4743529440244328e-05,
|
| 3347 |
+
"loss": 0.5995,
|
| 3348 |
+
"step": 477
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.6730024639211545,
|
| 3352 |
+
"grad_norm": 2.8634560108184814,
|
| 3353 |
+
"learning_rate": 2.4551759484417853e-05,
|
| 3354 |
+
"loss": 0.4455,
|
| 3355 |
+
"step": 478
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.6744104188665963,
|
| 3359 |
+
"grad_norm": 3.936350107192993,
|
| 3360 |
+
"learning_rate": 2.436049343210585e-05,
|
| 3361 |
+
"loss": 0.4504,
|
| 3362 |
+
"step": 479
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.675818373812038,
|
| 3366 |
+
"grad_norm": 7.890109539031982,
|
| 3367 |
+
"learning_rate": 2.4169735070589262e-05,
|
| 3368 |
+
"loss": 0.4005,
|
| 3369 |
+
"step": 480
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 0.6772263287574798,
|
| 3373 |
+
"grad_norm": 5.764025688171387,
|
| 3374 |
+
"learning_rate": 2.3979488177096175e-05,
|
| 3375 |
+
"loss": 0.4319,
|
| 3376 |
+
"step": 481
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 0.6786342837029216,
|
| 3380 |
+
"grad_norm": 3.937145709991455,
|
| 3381 |
+
"learning_rate": 2.3789756518727063e-05,
|
| 3382 |
+
"loss": 0.3359,
|
| 3383 |
+
"step": 482
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 0.6800422386483632,
|
| 3387 |
+
"grad_norm": 4.763424396514893,
|
| 3388 |
+
"learning_rate": 2.360054385238012e-05,
|
| 3389 |
+
"loss": 0.4529,
|
| 3390 |
+
"step": 483
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 0.681450193593805,
|
| 3394 |
+
"grad_norm": 10.706249237060547,
|
| 3395 |
+
"learning_rate": 2.3411853924676907e-05,
|
| 3396 |
+
"loss": 0.8968,
|
| 3397 |
+
"step": 484
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 0.6828581485392468,
|
| 3401 |
+
"grad_norm": 22.618295669555664,
|
| 3402 |
+
"learning_rate": 2.3223690471888288e-05,
|
| 3403 |
+
"loss": 1.333,
|
| 3404 |
+
"step": 485
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 0.6842661034846885,
|
| 3408 |
+
"grad_norm": 25.886438369750977,
|
| 3409 |
+
"learning_rate": 2.3036057219860212e-05,
|
| 3410 |
+
"loss": 1.2769,
|
| 3411 |
+
"step": 486
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 0.6856740584301302,
|
| 3415 |
+
"grad_norm": 9.857765197753906,
|
| 3416 |
+
"learning_rate": 2.284895788394015e-05,
|
| 3417 |
+
"loss": 0.7986,
|
| 3418 |
+
"step": 487
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 0.687082013375572,
|
| 3422 |
+
"grad_norm": 8.730622291564941,
|
| 3423 |
+
"learning_rate": 2.2662396168903365e-05,
|
| 3424 |
+
"loss": 0.6074,
|
| 3425 |
+
"step": 488
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 0.6884899683210137,
|
| 3429 |
+
"grad_norm": 10.999805450439453,
|
| 3430 |
+
"learning_rate": 2.2476375768879686e-05,
|
| 3431 |
+
"loss": 0.8271,
|
| 3432 |
+
"step": 489
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 0.6898979232664555,
|
| 3436 |
+
"grad_norm": 6.877896308898926,
|
| 3437 |
+
"learning_rate": 2.2290900367280303e-05,
|
| 3438 |
+
"loss": 0.7952,
|
| 3439 |
+
"step": 490
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 0.6913058782118973,
|
| 3443 |
+
"grad_norm": 11.425212860107422,
|
| 3444 |
+
"learning_rate": 2.2105973636724747e-05,
|
| 3445 |
+
"loss": 0.9463,
|
| 3446 |
+
"step": 491
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 0.6927138331573389,
|
| 3450 |
+
"grad_norm": 56.40801239013672,
|
| 3451 |
+
"learning_rate": 2.1921599238968392e-05,
|
| 3452 |
+
"loss": 2.2503,
|
| 3453 |
+
"step": 492
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 0.6941217881027807,
|
| 3457 |
+
"grad_norm": 11.69538402557373,
|
| 3458 |
+
"learning_rate": 2.173778082482967e-05,
|
| 3459 |
+
"loss": 0.3918,
|
| 3460 |
+
"step": 493
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 0.6955297430482225,
|
| 3464 |
+
"grad_norm": 12.837867736816406,
|
| 3465 |
+
"learning_rate": 2.1554522034118015e-05,
|
| 3466 |
+
"loss": 0.4069,
|
| 3467 |
+
"step": 494
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 0.6969376979936642,
|
| 3471 |
+
"grad_norm": 17.93108367919922,
|
| 3472 |
+
"learning_rate": 2.1371826495561615e-05,
|
| 3473 |
+
"loss": 1.216,
|
| 3474 |
+
"step": 495
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 0.698345652939106,
|
| 3478 |
+
"grad_norm": 12.086804389953613,
|
| 3479 |
+
"learning_rate": 2.1189697826735693e-05,
|
| 3480 |
+
"loss": 0.7372,
|
| 3481 |
+
"step": 496
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 0.6997536078845477,
|
| 3485 |
+
"grad_norm": 10.504433631896973,
|
| 3486 |
+
"learning_rate": 2.1008139633990832e-05,
|
| 3487 |
+
"loss": 0.7175,
|
| 3488 |
+
"step": 497
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 0.7011615628299894,
|
| 3492 |
+
"grad_norm": 9.431361198425293,
|
| 3493 |
+
"learning_rate": 2.082715551238148e-05,
|
| 3494 |
+
"loss": 0.9107,
|
| 3495 |
+
"step": 498
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 0.7025695177754312,
|
| 3499 |
+
"grad_norm": 12.150747299194336,
|
| 3500 |
+
"learning_rate": 2.0646749045594903e-05,
|
| 3501 |
+
"loss": 0.6491,
|
| 3502 |
+
"step": 499
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 0.703977472720873,
|
| 3506 |
+
"grad_norm": 24.827930450439453,
|
| 3507 |
+
"learning_rate": 2.0466923805880135e-05,
|
| 3508 |
+
"loss": 1.423,
|
| 3509 |
+
"step": 500
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 0.7053854276663146,
|
| 3513 |
+
"grad_norm": 4.016744613647461,
|
| 3514 |
+
"learning_rate": 2.0287683353977294e-05,
|
| 3515 |
+
"loss": 0.1746,
|
| 3516 |
+
"step": 501
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 0.7067933826117564,
|
| 3520 |
+
"grad_norm": 9.725351333618164,
|
| 3521 |
+
"learning_rate": 2.0109031239046977e-05,
|
| 3522 |
+
"loss": 0.5166,
|
| 3523 |
+
"step": 502
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 0.7082013375571982,
|
| 3527 |
+
"grad_norm": 6.610229969024658,
|
| 3528 |
+
"learning_rate": 1.9930970998600122e-05,
|
| 3529 |
+
"loss": 0.4892,
|
| 3530 |
+
"step": 503
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 0.7096092925026399,
|
| 3534 |
+
"grad_norm": 7.100261211395264,
|
| 3535 |
+
"learning_rate": 1.9753506158427887e-05,
|
| 3536 |
+
"loss": 0.7671,
|
| 3537 |
+
"step": 504
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 0.7110172474480817,
|
| 3541 |
+
"grad_norm": 5.738232135772705,
|
| 3542 |
+
"learning_rate": 1.9576640232531785e-05,
|
| 3543 |
+
"loss": 0.6904,
|
| 3544 |
+
"step": 505
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 0.7124252023935234,
|
| 3548 |
+
"grad_norm": 7.7215962409973145,
|
| 3549 |
+
"learning_rate": 1.940037672305422e-05,
|
| 3550 |
+
"loss": 0.5436,
|
| 3551 |
+
"step": 506
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 0.7138331573389651,
|
| 3555 |
+
"grad_norm": 4.128014087677002,
|
| 3556 |
+
"learning_rate": 1.9224719120209067e-05,
|
| 3557 |
+
"loss": 0.5873,
|
| 3558 |
+
"step": 507
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 0.7152411122844069,
|
| 3562 |
+
"grad_norm": 9.871435165405273,
|
| 3563 |
+
"learning_rate": 1.9049670902212542e-05,
|
| 3564 |
+
"loss": 0.8,
|
| 3565 |
+
"step": 508
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 0.7166490672298487,
|
| 3569 |
+
"grad_norm": 2.5685479640960693,
|
| 3570 |
+
"learning_rate": 1.8875235535214415e-05,
|
| 3571 |
+
"loss": 0.2577,
|
| 3572 |
+
"step": 509
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 0.7180570221752904,
|
| 3576 |
+
"grad_norm": 6.314105033874512,
|
| 3577 |
+
"learning_rate": 1.8701416473229238e-05,
|
| 3578 |
+
"loss": 0.4349,
|
| 3579 |
+
"step": 510
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 0.7194649771207321,
|
| 3583 |
+
"grad_norm": 9.4436674118042,
|
| 3584 |
+
"learning_rate": 1.852821715806817e-05,
|
| 3585 |
+
"loss": 0.807,
|
| 3586 |
+
"step": 511
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 0.7208729320661739,
|
| 3590 |
+
"grad_norm": 6.681199550628662,
|
| 3591 |
+
"learning_rate": 1.8355641019270548e-05,
|
| 3592 |
+
"loss": 0.6379,
|
| 3593 |
+
"step": 512
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 0.7222808870116156,
|
| 3597 |
+
"grad_norm": 18.25213623046875,
|
| 3598 |
+
"learning_rate": 1.8183691474036224e-05,
|
| 3599 |
+
"loss": 1.1948,
|
| 3600 |
+
"step": 513
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 0.7236888419570574,
|
| 3604 |
+
"grad_norm": 12.817623138427734,
|
| 3605 |
+
"learning_rate": 1.801237192715778e-05,
|
| 3606 |
+
"loss": 1.1606,
|
| 3607 |
+
"step": 514
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 0.7250967969024992,
|
| 3611 |
+
"grad_norm": 9.447029113769531,
|
| 3612 |
+
"learning_rate": 1.784168577095307e-05,
|
| 3613 |
+
"loss": 0.9631,
|
| 3614 |
+
"step": 515
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 0.7265047518479408,
|
| 3618 |
+
"grad_norm": 9.531036376953125,
|
| 3619 |
+
"learning_rate": 1.767163638519819e-05,
|
| 3620 |
+
"loss": 0.9417,
|
| 3621 |
+
"step": 516
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 0.7279127067933826,
|
| 3625 |
+
"grad_norm": 4.567352294921875,
|
| 3626 |
+
"learning_rate": 1.7502227137060384e-05,
|
| 3627 |
+
"loss": 0.5713,
|
| 3628 |
+
"step": 517
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 0.7293206617388244,
|
| 3632 |
+
"grad_norm": 7.561388969421387,
|
| 3633 |
+
"learning_rate": 1.7333461381031574e-05,
|
| 3634 |
+
"loss": 0.6314,
|
| 3635 |
+
"step": 518
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 0.7307286166842661,
|
| 3639 |
+
"grad_norm": 3.648564577102661,
|
| 3640 |
+
"learning_rate": 1.7165342458861706e-05,
|
| 3641 |
+
"loss": 0.5302,
|
| 3642 |
+
"step": 519
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 0.7321365716297078,
|
| 3646 |
+
"grad_norm": 6.6868815422058105,
|
| 3647 |
+
"learning_rate": 1.699787369949275e-05,
|
| 3648 |
+
"loss": 0.685,
|
| 3649 |
+
"step": 520
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 0.7335445265751496,
|
| 3653 |
+
"grad_norm": 7.9714035987854,
|
| 3654 |
+
"learning_rate": 1.6831058418992746e-05,
|
| 3655 |
+
"loss": 0.7526,
|
| 3656 |
+
"step": 521
|
| 3657 |
+
},
|
| 3658 |
+
{
|
| 3659 |
+
"epoch": 0.7349524815205913,
|
| 3660 |
+
"grad_norm": 6.323826313018799,
|
| 3661 |
+
"learning_rate": 1.666489992049006e-05,
|
| 3662 |
+
"loss": 0.685,
|
| 3663 |
+
"step": 522
|
| 3664 |
+
},
|
| 3665 |
+
{
|
| 3666 |
+
"epoch": 0.7363604364660331,
|
| 3667 |
+
"grad_norm": 5.036644458770752,
|
| 3668 |
+
"learning_rate": 1.6499401494108124e-05,
|
| 3669 |
+
"loss": 0.6348,
|
| 3670 |
+
"step": 523
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"epoch": 0.7377683914114749,
|
| 3674 |
+
"grad_norm": 4.294615268707275,
|
| 3675 |
+
"learning_rate": 1.6334566416900104e-05,
|
| 3676 |
+
"loss": 0.5513,
|
| 3677 |
+
"step": 524
|
| 3678 |
+
},
|
| 3679 |
+
{
|
| 3680 |
+
"epoch": 0.7391763463569165,
|
| 3681 |
+
"grad_norm": 4.160864353179932,
|
| 3682 |
+
"learning_rate": 1.617039795278425e-05,
|
| 3683 |
+
"loss": 0.5917,
|
| 3684 |
+
"step": 525
|
| 3685 |
+
},
|
| 3686 |
+
{
|
| 3687 |
+
"epoch": 0.7405843013023583,
|
| 3688 |
+
"grad_norm": 4.928703308105469,
|
| 3689 |
+
"learning_rate": 1.600689935247902e-05,
|
| 3690 |
+
"loss": 0.6139,
|
| 3691 |
+
"step": 526
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 0.7419922562478001,
|
| 3695 |
+
"grad_norm": 3.4751169681549072,
|
| 3696 |
+
"learning_rate": 1.5844073853438833e-05,
|
| 3697 |
+
"loss": 0.4871,
|
| 3698 |
+
"step": 527
|
| 3699 |
+
},
|
| 3700 |
+
{
|
| 3701 |
+
"epoch": 0.7434002111932418,
|
| 3702 |
+
"grad_norm": 2.0456016063690186,
|
| 3703 |
+
"learning_rate": 1.568192467979e-05,
|
| 3704 |
+
"loss": 0.3064,
|
| 3705 |
+
"step": 528
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 0.7448081661386836,
|
| 3709 |
+
"grad_norm": 1.985282301902771,
|
| 3710 |
+
"learning_rate": 1.5520455042266807e-05,
|
| 3711 |
+
"loss": 0.2762,
|
| 3712 |
+
"step": 529
|
| 3713 |
+
},
|
| 3714 |
+
{
|
| 3715 |
+
"epoch": 0.7462161210841253,
|
| 3716 |
+
"grad_norm": 4.5425944328308105,
|
| 3717 |
+
"learning_rate": 1.535966813814799e-05,
|
| 3718 |
+
"loss": 0.4767,
|
| 3719 |
+
"step": 530
|
| 3720 |
+
},
|
| 3721 |
+
{
|
| 3722 |
+
"epoch": 0.747624076029567,
|
| 3723 |
+
"grad_norm": 3.6157071590423584,
|
| 3724 |
+
"learning_rate": 1.519956715119336e-05,
|
| 3725 |
+
"loss": 0.2803,
|
| 3726 |
+
"step": 531
|
| 3727 |
+
},
|
| 3728 |
+
{
|
| 3729 |
+
"epoch": 0.7490320309750088,
|
| 3730 |
+
"grad_norm": 4.535380840301514,
|
| 3731 |
+
"learning_rate": 1.504015525158084e-05,
|
| 3732 |
+
"loss": 0.3895,
|
| 3733 |
+
"step": 532
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 0.7504399859204506,
|
| 3737 |
+
"grad_norm": 8.593908309936523,
|
| 3738 |
+
"learning_rate": 1.4881435595843657e-05,
|
| 3739 |
+
"loss": 0.6963,
|
| 3740 |
+
"step": 533
|
| 3741 |
+
},
|
| 3742 |
+
{
|
| 3743 |
+
"epoch": 0.7518479408658922,
|
| 3744 |
+
"grad_norm": 5.495645523071289,
|
| 3745 |
+
"learning_rate": 1.4723411326807807e-05,
|
| 3746 |
+
"loss": 0.5797,
|
| 3747 |
+
"step": 534
|
| 3748 |
}
|
| 3749 |
],
|
| 3750 |
"logging_steps": 1,
|
|
|
|
| 3764 |
"attributes": {}
|
| 3765 |
}
|
| 3766 |
},
|
| 3767 |
+
"total_flos": 5.5811502310529434e+17,
|
| 3768 |
"train_batch_size": 4,
|
| 3769 |
"trial_name": null,
|
| 3770 |
"trial_params": null
|