Training in progress, step 440, checkpoint
Browse files- last-checkpoint/adapter_model.safetensors +1 -1
- last-checkpoint/optimizer.pt +1 -1
- last-checkpoint/rng_state_0.pth +1 -1
- last-checkpoint/rng_state_1.pth +1 -1
- last-checkpoint/rng_state_2.pth +1 -1
- last-checkpoint/rng_state_3.pth +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +1411 -4
last-checkpoint/adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4286680
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3eb5c3228d5f169d3e566f341e1de01ec7e78408c8fecbc1639b3c2b3cbec950
|
| 3 |
size 4286680
|
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 8583659
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58c37d2fccdce3632e266132ad2b89189f9a9e1ca4b9fd0df854408688d6ab55
|
| 3 |
size 8583659
|
last-checkpoint/rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd028d3cffde00dd4a7c26cac86475631732cdb6a5cc82b9a5501c0ce419bf57
|
| 3 |
size 15024
|
last-checkpoint/rng_state_1.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5cd4496eeda825da88221f896955530b92634bc738e1e71ab630e2caa9dc3167
|
| 3 |
size 15024
|
last-checkpoint/rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:84224df08fab8840835a04924367865c53711750c088c8d49882edab4dbeea47
|
| 3 |
size 15024
|
last-checkpoint/rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 15024
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a5a19b7546bf0a06091a07d7a5fd8a6937ce53b062869e943ea4f610c95f3f14
|
| 3 |
size 15024
|
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 sha256:3aa6d4cd0a1d119d88746df8d17b061da99249879d9cb64d05543ac4d112a2c5
|
| 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,
|
|
@@ -1680,6 +1680,1413 @@
|
|
| 1680 |
"learning_rate": 6.412692147824976e-05,
|
| 1681 |
"loss": 39.8384,
|
| 1682 |
"step": 239
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1683 |
}
|
| 1684 |
],
|
| 1685 |
"logging_steps": 1,
|
|
@@ -1694,12 +3101,12 @@
|
|
| 1694 |
"should_evaluate": false,
|
| 1695 |
"should_log": false,
|
| 1696 |
"should_save": true,
|
| 1697 |
-
"should_training_stop":
|
| 1698 |
},
|
| 1699 |
"attributes": {}
|
| 1700 |
}
|
| 1701 |
},
|
| 1702 |
-
"total_flos":
|
| 1703 |
"train_batch_size": 4,
|
| 1704 |
"trial_name": null,
|
| 1705 |
"trial_params": null
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994321408290744,
|
| 5 |
"eval_steps": 500,
|
| 6 |
+
"global_step": 440,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
|
|
|
| 1680 |
"learning_rate": 6.412692147824976e-05,
|
| 1681 |
"loss": 39.8384,
|
| 1682 |
"step": 239
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.545144804088586,
|
| 1686 |
+
"grad_norm": 3.1766092777252197,
|
| 1687 |
+
"learning_rate": 6.368314950360415e-05,
|
| 1688 |
+
"loss": 39.8386,
|
| 1689 |
+
"step": 240
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5474162407722885,
|
| 1693 |
+
"grad_norm": 3.190483808517456,
|
| 1694 |
+
"learning_rate": 6.323820930826879e-05,
|
| 1695 |
+
"loss": 39.8017,
|
| 1696 |
+
"step": 241
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5496876774559909,
|
| 1700 |
+
"grad_norm": 3.2229549884796143,
|
| 1701 |
+
"learning_rate": 6.279213887972179e-05,
|
| 1702 |
+
"loss": 39.7668,
|
| 1703 |
+
"step": 242
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5519591141396933,
|
| 1707 |
+
"grad_norm": 3.0010106563568115,
|
| 1708 |
+
"learning_rate": 6.234497630193666e-05,
|
| 1709 |
+
"loss": 39.9786,
|
| 1710 |
+
"step": 243
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5542305508233958,
|
| 1714 |
+
"grad_norm": 3.1917195320129395,
|
| 1715 |
+
"learning_rate": 6.189675975213094e-05,
|
| 1716 |
+
"loss": 39.812,
|
| 1717 |
+
"step": 244
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5565019875070982,
|
| 1721 |
+
"grad_norm": 3.4181742668151855,
|
| 1722 |
+
"learning_rate": 6.14475274975067e-05,
|
| 1723 |
+
"loss": 39.8809,
|
| 1724 |
+
"step": 245
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5587734241908007,
|
| 1728 |
+
"grad_norm": 3.112577438354492,
|
| 1729 |
+
"learning_rate": 6.099731789198344e-05,
|
| 1730 |
+
"loss": 39.8375,
|
| 1731 |
+
"step": 246
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5610448608745031,
|
| 1735 |
+
"grad_norm": 3.208275079727173,
|
| 1736 |
+
"learning_rate": 6.05461693729235e-05,
|
| 1737 |
+
"loss": 39.763,
|
| 1738 |
+
"step": 247
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5633162975582056,
|
| 1742 |
+
"grad_norm": 3.286311388015747,
|
| 1743 |
+
"learning_rate": 6.009412045785051e-05,
|
| 1744 |
+
"loss": 39.7137,
|
| 1745 |
+
"step": 248
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.565587734241908,
|
| 1749 |
+
"grad_norm": 3.2906556129455566,
|
| 1750 |
+
"learning_rate": 5.9641209741160855e-05,
|
| 1751 |
+
"loss": 39.6924,
|
| 1752 |
+
"step": 249
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5678591709256104,
|
| 1756 |
+
"grad_norm": 3.3440194129943848,
|
| 1757 |
+
"learning_rate": 5.918747589082853e-05,
|
| 1758 |
+
"loss": 39.8044,
|
| 1759 |
+
"step": 250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5701306076093129,
|
| 1763 |
+
"grad_norm": 3.2966935634613037,
|
| 1764 |
+
"learning_rate": 5.873295764510395e-05,
|
| 1765 |
+
"loss": 39.6652,
|
| 1766 |
+
"step": 251
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5724020442930153,
|
| 1770 |
+
"grad_norm": 3.207146644592285,
|
| 1771 |
+
"learning_rate": 5.82776938092065e-05,
|
| 1772 |
+
"loss": 39.7085,
|
| 1773 |
+
"step": 252
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.5746734809767178,
|
| 1777 |
+
"grad_norm": 3.005434513092041,
|
| 1778 |
+
"learning_rate": 5.782172325201155e-05,
|
| 1779 |
+
"loss": 39.7882,
|
| 1780 |
+
"step": 253
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5769449176604202,
|
| 1784 |
+
"grad_norm": 3.200674295425415,
|
| 1785 |
+
"learning_rate": 5.736508490273188e-05,
|
| 1786 |
+
"loss": 39.8991,
|
| 1787 |
+
"step": 254
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5792163543441227,
|
| 1791 |
+
"grad_norm": 3.2107930183410645,
|
| 1792 |
+
"learning_rate": 5.6907817747594116e-05,
|
| 1793 |
+
"loss": 39.6958,
|
| 1794 |
+
"step": 255
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5814877910278251,
|
| 1798 |
+
"grad_norm": 3.239307165145874,
|
| 1799 |
+
"learning_rate": 5.644996082651017e-05,
|
| 1800 |
+
"loss": 39.6615,
|
| 1801 |
+
"step": 256
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5837592277115276,
|
| 1805 |
+
"grad_norm": 3.256614923477173,
|
| 1806 |
+
"learning_rate": 5.599155322974417e-05,
|
| 1807 |
+
"loss": 39.5822,
|
| 1808 |
+
"step": 257
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.58603066439523,
|
| 1812 |
+
"grad_norm": 3.1306862831115723,
|
| 1813 |
+
"learning_rate": 5.553263409457504e-05,
|
| 1814 |
+
"loss": 39.7317,
|
| 1815 |
+
"step": 258
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.5883021010789324,
|
| 1819 |
+
"grad_norm": 3.114198923110962,
|
| 1820 |
+
"learning_rate": 5.507324260195515e-05,
|
| 1821 |
+
"loss": 39.654,
|
| 1822 |
+
"step": 259
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.5905735377626349,
|
| 1826 |
+
"grad_norm": 3.288999319076538,
|
| 1827 |
+
"learning_rate": 5.4613417973165106e-05,
|
| 1828 |
+
"loss": 39.5748,
|
| 1829 |
+
"step": 260
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.5928449744463373,
|
| 1833 |
+
"grad_norm": 3.4110300540924072,
|
| 1834 |
+
"learning_rate": 5.41531994664652e-05,
|
| 1835 |
+
"loss": 39.4919,
|
| 1836 |
+
"step": 261
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.5951164111300398,
|
| 1840 |
+
"grad_norm": 3.2709226608276367,
|
| 1841 |
+
"learning_rate": 5.3692626373743706e-05,
|
| 1842 |
+
"loss": 39.5359,
|
| 1843 |
+
"step": 262
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.5973878478137422,
|
| 1847 |
+
"grad_norm": 3.0505809783935547,
|
| 1848 |
+
"learning_rate": 5.3231738017162214e-05,
|
| 1849 |
+
"loss": 39.6559,
|
| 1850 |
+
"step": 263
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.5996592844974447,
|
| 1854 |
+
"grad_norm": 3.287243604660034,
|
| 1855 |
+
"learning_rate": 5.27705737457985e-05,
|
| 1856 |
+
"loss": 39.5388,
|
| 1857 |
+
"step": 264
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6019307211811471,
|
| 1861 |
+
"grad_norm": 3.1598408222198486,
|
| 1862 |
+
"learning_rate": 5.230917293228699e-05,
|
| 1863 |
+
"loss": 39.5886,
|
| 1864 |
+
"step": 265
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6042021578648495,
|
| 1868 |
+
"grad_norm": 3.160259962081909,
|
| 1869 |
+
"learning_rate": 5.184757496945726e-05,
|
| 1870 |
+
"loss": 39.6173,
|
| 1871 |
+
"step": 266
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.606473594548552,
|
| 1875 |
+
"grad_norm": 3.2377989292144775,
|
| 1876 |
+
"learning_rate": 5.138581926697082e-05,
|
| 1877 |
+
"loss": 39.5624,
|
| 1878 |
+
"step": 267
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.6087450312322544,
|
| 1882 |
+
"grad_norm": 3.2241992950439453,
|
| 1883 |
+
"learning_rate": 5.092394524795649e-05,
|
| 1884 |
+
"loss": 39.5781,
|
| 1885 |
+
"step": 268
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.6110164679159569,
|
| 1889 |
+
"grad_norm": 3.2245700359344482,
|
| 1890 |
+
"learning_rate": 5.046199234564455e-05,
|
| 1891 |
+
"loss": 39.544,
|
| 1892 |
+
"step": 269
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6132879045996593,
|
| 1896 |
+
"grad_norm": 3.128326177597046,
|
| 1897 |
+
"learning_rate": 5e-05,
|
| 1898 |
+
"loss": 39.6007,
|
| 1899 |
+
"step": 270
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6155593412833618,
|
| 1903 |
+
"grad_norm": 3.128420829772949,
|
| 1904 |
+
"learning_rate": 4.953800765435547e-05,
|
| 1905 |
+
"loss": 39.612,
|
| 1906 |
+
"step": 271
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6178307779670642,
|
| 1910 |
+
"grad_norm": 3.2072839736938477,
|
| 1911 |
+
"learning_rate": 4.907605475204352e-05,
|
| 1912 |
+
"loss": 39.5239,
|
| 1913 |
+
"step": 272
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.6201022146507666,
|
| 1917 |
+
"grad_norm": 3.284583568572998,
|
| 1918 |
+
"learning_rate": 4.861418073302919e-05,
|
| 1919 |
+
"loss": 39.4844,
|
| 1920 |
+
"step": 273
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6223736513344691,
|
| 1924 |
+
"grad_norm": 3.2405998706817627,
|
| 1925 |
+
"learning_rate": 4.8152425030542766e-05,
|
| 1926 |
+
"loss": 39.5103,
|
| 1927 |
+
"step": 274
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6246450880181715,
|
| 1931 |
+
"grad_norm": 3.397861957550049,
|
| 1932 |
+
"learning_rate": 4.7690827067713035e-05,
|
| 1933 |
+
"loss": 39.6217,
|
| 1934 |
+
"step": 275
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.626916524701874,
|
| 1938 |
+
"grad_norm": 3.4319255352020264,
|
| 1939 |
+
"learning_rate": 4.72294262542015e-05,
|
| 1940 |
+
"loss": 39.4299,
|
| 1941 |
+
"step": 276
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6291879613855764,
|
| 1945 |
+
"grad_norm": 3.159431219100952,
|
| 1946 |
+
"learning_rate": 4.676826198283779e-05,
|
| 1947 |
+
"loss": 39.534,
|
| 1948 |
+
"step": 277
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.6314593980692789,
|
| 1952 |
+
"grad_norm": 3.191432237625122,
|
| 1953 |
+
"learning_rate": 4.6307373626256306e-05,
|
| 1954 |
+
"loss": 39.5237,
|
| 1955 |
+
"step": 278
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6337308347529813,
|
| 1959 |
+
"grad_norm": 3.1432559490203857,
|
| 1960 |
+
"learning_rate": 4.5846800533534815e-05,
|
| 1961 |
+
"loss": 39.5283,
|
| 1962 |
+
"step": 279
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6360022714366838,
|
| 1966 |
+
"grad_norm": 3.033679723739624,
|
| 1967 |
+
"learning_rate": 4.5386582026834906e-05,
|
| 1968 |
+
"loss": 39.61,
|
| 1969 |
+
"step": 280
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6382737081203862,
|
| 1973 |
+
"grad_norm": 3.223874568939209,
|
| 1974 |
+
"learning_rate": 4.492675739804486e-05,
|
| 1975 |
+
"loss": 39.5168,
|
| 1976 |
+
"step": 281
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6405451448040886,
|
| 1980 |
+
"grad_norm": 3.3325939178466797,
|
| 1981 |
+
"learning_rate": 4.446736590542497e-05,
|
| 1982 |
+
"loss": 39.4294,
|
| 1983 |
+
"step": 282
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.6428165814877911,
|
| 1987 |
+
"grad_norm": 3.082111120223999,
|
| 1988 |
+
"learning_rate": 4.400844677025585e-05,
|
| 1989 |
+
"loss": 39.5168,
|
| 1990 |
+
"step": 283
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6450880181714934,
|
| 1994 |
+
"grad_norm": 3.315622091293335,
|
| 1995 |
+
"learning_rate": 4.3550039173489845e-05,
|
| 1996 |
+
"loss": 39.4095,
|
| 1997 |
+
"step": 284
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6473594548551959,
|
| 2001 |
+
"grad_norm": 3.269645929336548,
|
| 2002 |
+
"learning_rate": 4.30921822524059e-05,
|
| 2003 |
+
"loss": 39.4171,
|
| 2004 |
+
"step": 285
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6496308915388983,
|
| 2008 |
+
"grad_norm": 3.3954954147338867,
|
| 2009 |
+
"learning_rate": 4.2634915097268115e-05,
|
| 2010 |
+
"loss": 39.3404,
|
| 2011 |
+
"step": 286
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6519023282226007,
|
| 2015 |
+
"grad_norm": 3.350980520248413,
|
| 2016 |
+
"learning_rate": 4.2178276747988446e-05,
|
| 2017 |
+
"loss": 39.3607,
|
| 2018 |
+
"step": 287
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6541737649063032,
|
| 2022 |
+
"grad_norm": 3.196150302886963,
|
| 2023 |
+
"learning_rate": 4.1722306190793495e-05,
|
| 2024 |
+
"loss": 39.3995,
|
| 2025 |
+
"step": 288
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6564452015900056,
|
| 2029 |
+
"grad_norm": 3.216571807861328,
|
| 2030 |
+
"learning_rate": 4.1267042354896056e-05,
|
| 2031 |
+
"loss": 39.4511,
|
| 2032 |
+
"step": 289
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6587166382737081,
|
| 2036 |
+
"grad_norm": 3.1120824813842773,
|
| 2037 |
+
"learning_rate": 4.0812524109171476e-05,
|
| 2038 |
+
"loss": 39.5224,
|
| 2039 |
+
"step": 290
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6609880749574105,
|
| 2043 |
+
"grad_norm": 3.1584837436676025,
|
| 2044 |
+
"learning_rate": 4.0358790258839164e-05,
|
| 2045 |
+
"loss": 39.4652,
|
| 2046 |
+
"step": 291
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.663259511641113,
|
| 2050 |
+
"grad_norm": 3.1293344497680664,
|
| 2051 |
+
"learning_rate": 3.99058795421495e-05,
|
| 2052 |
+
"loss": 39.4972,
|
| 2053 |
+
"step": 292
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6655309483248154,
|
| 2057 |
+
"grad_norm": 3.218752145767212,
|
| 2058 |
+
"learning_rate": 3.9453830627076516e-05,
|
| 2059 |
+
"loss": 39.4812,
|
| 2060 |
+
"step": 293
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6678023850085179,
|
| 2064 |
+
"grad_norm": 3.051319122314453,
|
| 2065 |
+
"learning_rate": 3.9002682108016585e-05,
|
| 2066 |
+
"loss": 39.533,
|
| 2067 |
+
"step": 294
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6700738216922203,
|
| 2071 |
+
"grad_norm": 3.1116693019866943,
|
| 2072 |
+
"learning_rate": 3.855247250249331e-05,
|
| 2073 |
+
"loss": 39.4644,
|
| 2074 |
+
"step": 295
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6723452583759227,
|
| 2078 |
+
"grad_norm": 3.175712823867798,
|
| 2079 |
+
"learning_rate": 3.8103240247869075e-05,
|
| 2080 |
+
"loss": 39.4329,
|
| 2081 |
+
"step": 296
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6746166950596252,
|
| 2085 |
+
"grad_norm": 3.2702810764312744,
|
| 2086 |
+
"learning_rate": 3.765502369806334e-05,
|
| 2087 |
+
"loss": 39.3459,
|
| 2088 |
+
"step": 297
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6768881317433276,
|
| 2092 |
+
"grad_norm": 3.3634626865386963,
|
| 2093 |
+
"learning_rate": 3.720786112027822e-05,
|
| 2094 |
+
"loss": 39.3088,
|
| 2095 |
+
"step": 298
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6791595684270301,
|
| 2099 |
+
"grad_norm": 3.2262613773345947,
|
| 2100 |
+
"learning_rate": 3.676179069173121e-05,
|
| 2101 |
+
"loss": 39.3954,
|
| 2102 |
+
"step": 299
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6814310051107325,
|
| 2106 |
+
"grad_norm": 3.2690017223358154,
|
| 2107 |
+
"learning_rate": 3.631685049639586e-05,
|
| 2108 |
+
"loss": 39.3951,
|
| 2109 |
+
"step": 300
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.683702441794435,
|
| 2113 |
+
"grad_norm": 3.257148265838623,
|
| 2114 |
+
"learning_rate": 3.5873078521750246e-05,
|
| 2115 |
+
"loss": 39.4106,
|
| 2116 |
+
"step": 301
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.6859738784781374,
|
| 2120 |
+
"grad_norm": 3.2068793773651123,
|
| 2121 |
+
"learning_rate": 3.543051265553377e-05,
|
| 2122 |
+
"loss": 39.3841,
|
| 2123 |
+
"step": 302
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.6882453151618398,
|
| 2127 |
+
"grad_norm": 3.2364015579223633,
|
| 2128 |
+
"learning_rate": 3.498919068251237e-05,
|
| 2129 |
+
"loss": 39.3645,
|
| 2130 |
+
"step": 303
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.6905167518455423,
|
| 2134 |
+
"grad_norm": 3.2378106117248535,
|
| 2135 |
+
"learning_rate": 3.4549150281252636e-05,
|
| 2136 |
+
"loss": 39.3788,
|
| 2137 |
+
"step": 304
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.6927881885292447,
|
| 2141 |
+
"grad_norm": 3.2840211391448975,
|
| 2142 |
+
"learning_rate": 3.411042902090492e-05,
|
| 2143 |
+
"loss": 39.3334,
|
| 2144 |
+
"step": 305
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.6950596252129472,
|
| 2148 |
+
"grad_norm": 3.144357442855835,
|
| 2149 |
+
"learning_rate": 3.367306435799584e-05,
|
| 2150 |
+
"loss": 39.3654,
|
| 2151 |
+
"step": 306
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.6973310618966496,
|
| 2155 |
+
"grad_norm": 3.175121545791626,
|
| 2156 |
+
"learning_rate": 3.323709363323032e-05,
|
| 2157 |
+
"loss": 39.4134,
|
| 2158 |
+
"step": 307
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.6996024985803521,
|
| 2162 |
+
"grad_norm": 3.283843517303467,
|
| 2163 |
+
"learning_rate": 3.2802554068303596e-05,
|
| 2164 |
+
"loss": 39.3028,
|
| 2165 |
+
"step": 308
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7018739352640545,
|
| 2169 |
+
"grad_norm": 3.300076961517334,
|
| 2170 |
+
"learning_rate": 3.236948276272337e-05,
|
| 2171 |
+
"loss": 39.3181,
|
| 2172 |
+
"step": 309
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.704145371947757,
|
| 2176 |
+
"grad_norm": 3.330923318862915,
|
| 2177 |
+
"learning_rate": 3.1937916690642356e-05,
|
| 2178 |
+
"loss": 39.2996,
|
| 2179 |
+
"step": 310
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7064168086314594,
|
| 2183 |
+
"grad_norm": 3.320736885070801,
|
| 2184 |
+
"learning_rate": 3.150789269770155e-05,
|
| 2185 |
+
"loss": 39.3073,
|
| 2186 |
+
"step": 311
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7086882453151618,
|
| 2190 |
+
"grad_norm": 3.3675241470336914,
|
| 2191 |
+
"learning_rate": 3.107944749788449e-05,
|
| 2192 |
+
"loss": 39.298,
|
| 2193 |
+
"step": 312
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7109596819988643,
|
| 2197 |
+
"grad_norm": 3.191697597503662,
|
| 2198 |
+
"learning_rate": 3.065261767038275e-05,
|
| 2199 |
+
"loss": 39.3497,
|
| 2200 |
+
"step": 313
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7132311186825667,
|
| 2204 |
+
"grad_norm": 3.270242214202881,
|
| 2205 |
+
"learning_rate": 3.0227439656472877e-05,
|
| 2206 |
+
"loss": 39.3073,
|
| 2207 |
+
"step": 314
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7155025553662692,
|
| 2211 |
+
"grad_norm": 3.160560131072998,
|
| 2212 |
+
"learning_rate": 2.980394975640526e-05,
|
| 2213 |
+
"loss": 39.4016,
|
| 2214 |
+
"step": 315
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7177739920499716,
|
| 2218 |
+
"grad_norm": 3.158921003341675,
|
| 2219 |
+
"learning_rate": 2.9382184126304834e-05,
|
| 2220 |
+
"loss": 39.3934,
|
| 2221 |
+
"step": 316
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.720045428733674,
|
| 2225 |
+
"grad_norm": 3.112635374069214,
|
| 2226 |
+
"learning_rate": 2.8962178775084263e-05,
|
| 2227 |
+
"loss": 39.4403,
|
| 2228 |
+
"step": 317
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7223168654173765,
|
| 2232 |
+
"grad_norm": 3.1762988567352295,
|
| 2233 |
+
"learning_rate": 2.8543969561369556e-05,
|
| 2234 |
+
"loss": 39.3596,
|
| 2235 |
+
"step": 318
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7245883021010789,
|
| 2239 |
+
"grad_norm": 3.0977957248687744,
|
| 2240 |
+
"learning_rate": 2.8127592190438688e-05,
|
| 2241 |
+
"loss": 39.4288,
|
| 2242 |
+
"step": 319
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.7268597387847814,
|
| 2246 |
+
"grad_norm": 3.2995924949645996,
|
| 2247 |
+
"learning_rate": 2.771308221117309e-05,
|
| 2248 |
+
"loss": 39.3122,
|
| 2249 |
+
"step": 320
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7291311754684838,
|
| 2253 |
+
"grad_norm": 3.220078229904175,
|
| 2254 |
+
"learning_rate": 2.7300475013022663e-05,
|
| 2255 |
+
"loss": 39.3328,
|
| 2256 |
+
"step": 321
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7314026121521863,
|
| 2260 |
+
"grad_norm": 3.221082925796509,
|
| 2261 |
+
"learning_rate": 2.688980582298435e-05,
|
| 2262 |
+
"loss": 39.336,
|
| 2263 |
+
"step": 322
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7336740488358887,
|
| 2267 |
+
"grad_norm": 3.362293243408203,
|
| 2268 |
+
"learning_rate": 2.6481109702594542e-05,
|
| 2269 |
+
"loss": 39.2548,
|
| 2270 |
+
"step": 323
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7359454855195912,
|
| 2274 |
+
"grad_norm": 3.3790478706359863,
|
| 2275 |
+
"learning_rate": 2.607442154493568e-05,
|
| 2276 |
+
"loss": 39.2768,
|
| 2277 |
+
"step": 324
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.7382169222032936,
|
| 2281 |
+
"grad_norm": 3.2779300212860107,
|
| 2282 |
+
"learning_rate": 2.5669776071657192e-05,
|
| 2283 |
+
"loss": 39.4005,
|
| 2284 |
+
"step": 325
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.740488358886996,
|
| 2288 |
+
"grad_norm": 3.5260651111602783,
|
| 2289 |
+
"learning_rate": 2.5267207830011068e-05,
|
| 2290 |
+
"loss": 39.213,
|
| 2291 |
+
"step": 326
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7427597955706985,
|
| 2295 |
+
"grad_norm": 3.3797895908355713,
|
| 2296 |
+
"learning_rate": 2.4866751189902334e-05,
|
| 2297 |
+
"loss": 39.2461,
|
| 2298 |
+
"step": 327
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7450312322544009,
|
| 2302 |
+
"grad_norm": 3.299431085586548,
|
| 2303 |
+
"learning_rate": 2.446844034095466e-05,
|
| 2304 |
+
"loss": 39.2607,
|
| 2305 |
+
"step": 328
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7473026689381034,
|
| 2309 |
+
"grad_norm": 3.1269869804382324,
|
| 2310 |
+
"learning_rate": 2.4072309289591394e-05,
|
| 2311 |
+
"loss": 39.4012,
|
| 2312 |
+
"step": 329
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.7495741056218058,
|
| 2316 |
+
"grad_norm": 3.103790521621704,
|
| 2317 |
+
"learning_rate": 2.3678391856132204e-05,
|
| 2318 |
+
"loss": 39.4544,
|
| 2319 |
+
"step": 330
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7518455423055083,
|
| 2323 |
+
"grad_norm": 3.0655603408813477,
|
| 2324 |
+
"learning_rate": 2.328672167190558e-05,
|
| 2325 |
+
"loss": 39.3722,
|
| 2326 |
+
"step": 331
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7541169789892107,
|
| 2330 |
+
"grad_norm": 3.3145480155944824,
|
| 2331 |
+
"learning_rate": 2.2897332176377528e-05,
|
| 2332 |
+
"loss": 39.2439,
|
| 2333 |
+
"step": 332
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7563884156729132,
|
| 2337 |
+
"grad_norm": 3.251753091812134,
|
| 2338 |
+
"learning_rate": 2.251025661429664e-05,
|
| 2339 |
+
"loss": 39.2775,
|
| 2340 |
+
"step": 333
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.7586598523566156,
|
| 2344 |
+
"grad_norm": 3.268765687942505,
|
| 2345 |
+
"learning_rate": 2.2125528032855724e-05,
|
| 2346 |
+
"loss": 39.302,
|
| 2347 |
+
"step": 334
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.760931289040318,
|
| 2351 |
+
"grad_norm": 3.238736867904663,
|
| 2352 |
+
"learning_rate": 2.1743179278870407e-05,
|
| 2353 |
+
"loss": 39.2998,
|
| 2354 |
+
"step": 335
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7632027257240205,
|
| 2358 |
+
"grad_norm": 3.285243272781372,
|
| 2359 |
+
"learning_rate": 2.136324299597474e-05,
|
| 2360 |
+
"loss": 39.2785,
|
| 2361 |
+
"step": 336
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7654741624077229,
|
| 2365 |
+
"grad_norm": 3.3332467079162598,
|
| 2366 |
+
"learning_rate": 2.098575162183422e-05,
|
| 2367 |
+
"loss": 39.2638,
|
| 2368 |
+
"step": 337
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7677455990914254,
|
| 2372 |
+
"grad_norm": 3.088116407394409,
|
| 2373 |
+
"learning_rate": 2.061073738537635e-05,
|
| 2374 |
+
"loss": 39.3631,
|
| 2375 |
+
"step": 338
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7700170357751278,
|
| 2379 |
+
"grad_norm": 3.2848522663116455,
|
| 2380 |
+
"learning_rate": 2.0238232304039068e-05,
|
| 2381 |
+
"loss": 39.2917,
|
| 2382 |
+
"step": 339
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7722884724588303,
|
| 2386 |
+
"grad_norm": 3.111776113510132,
|
| 2387 |
+
"learning_rate": 1.9868268181037185e-05,
|
| 2388 |
+
"loss": 39.3767,
|
| 2389 |
+
"step": 340
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7745599091425327,
|
| 2393 |
+
"grad_norm": 3.173539161682129,
|
| 2394 |
+
"learning_rate": 1.9500876602647167e-05,
|
| 2395 |
+
"loss": 39.3018,
|
| 2396 |
+
"step": 341
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7768313458262351,
|
| 2400 |
+
"grad_norm": 2.8784914016723633,
|
| 2401 |
+
"learning_rate": 1.9136088935510362e-05,
|
| 2402 |
+
"loss": 39.5558,
|
| 2403 |
+
"step": 342
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7791027825099376,
|
| 2407 |
+
"grad_norm": 3.044440507888794,
|
| 2408 |
+
"learning_rate": 1.8773936323955054e-05,
|
| 2409 |
+
"loss": 39.4651,
|
| 2410 |
+
"step": 343
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.78137421919364,
|
| 2414 |
+
"grad_norm": 3.2132656574249268,
|
| 2415 |
+
"learning_rate": 1.8414449687337464e-05,
|
| 2416 |
+
"loss": 39.3461,
|
| 2417 |
+
"step": 344
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.7836456558773425,
|
| 2421 |
+
"grad_norm": 3.2220561504364014,
|
| 2422 |
+
"learning_rate": 1.8057659717401947e-05,
|
| 2423 |
+
"loss": 39.3125,
|
| 2424 |
+
"step": 345
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.7859170925610448,
|
| 2428 |
+
"grad_norm": 3.1892623901367188,
|
| 2429 |
+
"learning_rate": 1.7703596875660645e-05,
|
| 2430 |
+
"loss": 39.3226,
|
| 2431 |
+
"step": 346
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.7881885292447472,
|
| 2435 |
+
"grad_norm": 3.2378811836242676,
|
| 2436 |
+
"learning_rate": 1.7352291390792798e-05,
|
| 2437 |
+
"loss": 39.3087,
|
| 2438 |
+
"step": 347
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.7904599659284497,
|
| 2442 |
+
"grad_norm": 3.3155946731567383,
|
| 2443 |
+
"learning_rate": 1.700377325606388e-05,
|
| 2444 |
+
"loss": 39.2231,
|
| 2445 |
+
"step": 348
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.7927314026121521,
|
| 2449 |
+
"grad_norm": 3.3325536251068115,
|
| 2450 |
+
"learning_rate": 1.665807222676495e-05,
|
| 2451 |
+
"loss": 39.2634,
|
| 2452 |
+
"step": 349
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.7950028392958546,
|
| 2456 |
+
"grad_norm": 3.0400354862213135,
|
| 2457 |
+
"learning_rate": 1.631521781767214e-05,
|
| 2458 |
+
"loss": 39.4798,
|
| 2459 |
+
"step": 350
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.797274275979557,
|
| 2463 |
+
"grad_norm": 3.4280900955200195,
|
| 2464 |
+
"learning_rate": 1.5975239300526923e-05,
|
| 2465 |
+
"loss": 39.2302,
|
| 2466 |
+
"step": 351
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.7995457126632595,
|
| 2470 |
+
"grad_norm": 3.222968339920044,
|
| 2471 |
+
"learning_rate": 1.5638165701536868e-05,
|
| 2472 |
+
"loss": 39.2894,
|
| 2473 |
+
"step": 352
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8018171493469619,
|
| 2477 |
+
"grad_norm": 3.081446886062622,
|
| 2478 |
+
"learning_rate": 1.530402579889752e-05,
|
| 2479 |
+
"loss": 39.4086,
|
| 2480 |
+
"step": 353
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8040885860306644,
|
| 2484 |
+
"grad_norm": 3.1126420497894287,
|
| 2485 |
+
"learning_rate": 1.4972848120335453e-05,
|
| 2486 |
+
"loss": 39.3398,
|
| 2487 |
+
"step": 354
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8063600227143668,
|
| 2491 |
+
"grad_norm": 3.0955779552459717,
|
| 2492 |
+
"learning_rate": 1.4644660940672627e-05,
|
| 2493 |
+
"loss": 39.4006,
|
| 2494 |
+
"step": 355
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8086314593980692,
|
| 2498 |
+
"grad_norm": 3.705615758895874,
|
| 2499 |
+
"learning_rate": 1.4319492279412388e-05,
|
| 2500 |
+
"loss": 39.4957,
|
| 2501 |
+
"step": 356
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8109028960817717,
|
| 2505 |
+
"grad_norm": 3.2060508728027344,
|
| 2506 |
+
"learning_rate": 1.3997369898347278e-05,
|
| 2507 |
+
"loss": 39.3245,
|
| 2508 |
+
"step": 357
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8131743327654741,
|
| 2512 |
+
"grad_norm": 3.3021256923675537,
|
| 2513 |
+
"learning_rate": 1.3678321299188801e-05,
|
| 2514 |
+
"loss": 39.285,
|
| 2515 |
+
"step": 358
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8154457694491766,
|
| 2519 |
+
"grad_norm": 3.4100747108459473,
|
| 2520 |
+
"learning_rate": 1.336237372121944e-05,
|
| 2521 |
+
"loss": 39.213,
|
| 2522 |
+
"step": 359
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.817717206132879,
|
| 2526 |
+
"grad_norm": 3.3174431324005127,
|
| 2527 |
+
"learning_rate": 1.3049554138967051e-05,
|
| 2528 |
+
"loss": 39.2722,
|
| 2529 |
+
"step": 360
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8199886428165815,
|
| 2533 |
+
"grad_norm": 3.4100990295410156,
|
| 2534 |
+
"learning_rate": 1.2739889259901866e-05,
|
| 2535 |
+
"loss": 39.1835,
|
| 2536 |
+
"step": 361
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8222600795002839,
|
| 2540 |
+
"grad_norm": 3.3656208515167236,
|
| 2541 |
+
"learning_rate": 1.2433405522156332e-05,
|
| 2542 |
+
"loss": 39.2537,
|
| 2543 |
+
"step": 362
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8245315161839863,
|
| 2547 |
+
"grad_norm": 3.10040545463562,
|
| 2548 |
+
"learning_rate": 1.2130129092267862e-05,
|
| 2549 |
+
"loss": 39.323,
|
| 2550 |
+
"step": 363
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8268029528676888,
|
| 2554 |
+
"grad_norm": 3.2904765605926514,
|
| 2555 |
+
"learning_rate": 1.183008586294485e-05,
|
| 2556 |
+
"loss": 39.292,
|
| 2557 |
+
"step": 364
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8290743895513912,
|
| 2561 |
+
"grad_norm": 3.159372568130493,
|
| 2562 |
+
"learning_rate": 1.1533301450856054e-05,
|
| 2563 |
+
"loss": 39.3521,
|
| 2564 |
+
"step": 365
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8313458262350937,
|
| 2568 |
+
"grad_norm": 3.253182888031006,
|
| 2569 |
+
"learning_rate": 1.1239801194443506e-05,
|
| 2570 |
+
"loss": 39.3166,
|
| 2571 |
+
"step": 366
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8336172629187961,
|
| 2575 |
+
"grad_norm": 3.0811924934387207,
|
| 2576 |
+
"learning_rate": 1.0949610151759232e-05,
|
| 2577 |
+
"loss": 39.378,
|
| 2578 |
+
"step": 367
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8358886996024986,
|
| 2582 |
+
"grad_norm": 3.2115979194641113,
|
| 2583 |
+
"learning_rate": 1.066275309832584e-05,
|
| 2584 |
+
"loss": 39.3243,
|
| 2585 |
+
"step": 368
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.838160136286201,
|
| 2589 |
+
"grad_norm": 3.2386271953582764,
|
| 2590 |
+
"learning_rate": 1.0379254525021309e-05,
|
| 2591 |
+
"loss": 39.2904,
|
| 2592 |
+
"step": 369
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8404315729699035,
|
| 2596 |
+
"grad_norm": 3.2060539722442627,
|
| 2597 |
+
"learning_rate": 1.0099138635988026e-05,
|
| 2598 |
+
"loss": 39.3027,
|
| 2599 |
+
"step": 370
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8427030096536059,
|
| 2603 |
+
"grad_norm": 3.284043312072754,
|
| 2604 |
+
"learning_rate": 9.822429346566314e-06,
|
| 2605 |
+
"loss": 39.2547,
|
| 2606 |
+
"step": 371
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8449744463373083,
|
| 2610 |
+
"grad_norm": 3.284778594970703,
|
| 2611 |
+
"learning_rate": 9.549150281252633e-06,
|
| 2612 |
+
"loss": 39.2768,
|
| 2613 |
+
"step": 372
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8472458830210108,
|
| 2617 |
+
"grad_norm": 3.3626978397369385,
|
| 2618 |
+
"learning_rate": 9.279324771682585e-06,
|
| 2619 |
+
"loss": 39.2264,
|
| 2620 |
+
"step": 373
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8495173197047132,
|
| 2624 |
+
"grad_norm": 3.3178279399871826,
|
| 2625 |
+
"learning_rate": 9.012975854638949e-06,
|
| 2626 |
+
"loss": 39.2468,
|
| 2627 |
+
"step": 374
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8517887563884157,
|
| 2631 |
+
"grad_norm": 3.1009223461151123,
|
| 2632 |
+
"learning_rate": 8.75012627008489e-06,
|
| 2633 |
+
"loss": 39.3837,
|
| 2634 |
+
"step": 375
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8540601930721181,
|
| 2638 |
+
"grad_norm": 3.5075793266296387,
|
| 2639 |
+
"learning_rate": 8.490798459222476e-06,
|
| 2640 |
+
"loss": 39.1917,
|
| 2641 |
+
"step": 376
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8563316297558206,
|
| 2645 |
+
"grad_norm": 3.315697193145752,
|
| 2646 |
+
"learning_rate": 8.235014562576732e-06,
|
| 2647 |
+
"loss": 39.2783,
|
| 2648 |
+
"step": 377
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.858603066439523,
|
| 2652 |
+
"grad_norm": 3.2278759479522705,
|
| 2653 |
+
"learning_rate": 7.982796418105371e-06,
|
| 2654 |
+
"loss": 39.334,
|
| 2655 |
+
"step": 378
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8608745031232254,
|
| 2659 |
+
"grad_norm": 3.0192768573760986,
|
| 2660 |
+
"learning_rate": 7.734165559334328e-06,
|
| 2661 |
+
"loss": 39.4135,
|
| 2662 |
+
"step": 379
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8631459398069279,
|
| 2666 |
+
"grad_norm": 3.283160924911499,
|
| 2667 |
+
"learning_rate": 7.489143213519301e-06,
|
| 2668 |
+
"loss": 39.2609,
|
| 2669 |
+
"step": 380
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8654173764906303,
|
| 2673 |
+
"grad_norm": 3.1758711338043213,
|
| 2674 |
+
"learning_rate": 7.2477502998334505e-06,
|
| 2675 |
+
"loss": 39.2964,
|
| 2676 |
+
"step": 381
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8676888131743328,
|
| 2680 |
+
"grad_norm": 3.1060142517089844,
|
| 2681 |
+
"learning_rate": 7.010007427581378e-06,
|
| 2682 |
+
"loss": 39.3985,
|
| 2683 |
+
"step": 382
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8699602498580352,
|
| 2687 |
+
"grad_norm": 3.13398814201355,
|
| 2688 |
+
"learning_rate": 6.775934894439606e-06,
|
| 2689 |
+
"loss": 39.4141,
|
| 2690 |
+
"step": 383
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.8722316865417377,
|
| 2694 |
+
"grad_norm": 3.2694761753082275,
|
| 2695 |
+
"learning_rate": 6.5455526847235825e-06,
|
| 2696 |
+
"loss": 39.2613,
|
| 2697 |
+
"step": 384
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.8745031232254401,
|
| 2701 |
+
"grad_norm": 3.3462014198303223,
|
| 2702 |
+
"learning_rate": 6.318880467681526e-06,
|
| 2703 |
+
"loss": 39.2397,
|
| 2704 |
+
"step": 385
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.8767745599091425,
|
| 2708 |
+
"grad_norm": 3.345944881439209,
|
| 2709 |
+
"learning_rate": 6.0959375958151045e-06,
|
| 2710 |
+
"loss": 39.2334,
|
| 2711 |
+
"step": 386
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.879045996592845,
|
| 2715 |
+
"grad_norm": 3.225269079208374,
|
| 2716 |
+
"learning_rate": 5.876743103227217e-06,
|
| 2717 |
+
"loss": 39.2799,
|
| 2718 |
+
"step": 387
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8813174332765474,
|
| 2722 |
+
"grad_norm": 3.058452844619751,
|
| 2723 |
+
"learning_rate": 5.6613157039969055e-06,
|
| 2724 |
+
"loss": 39.4356,
|
| 2725 |
+
"step": 388
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8835888699602499,
|
| 2729 |
+
"grad_norm": 3.2678914070129395,
|
| 2730 |
+
"learning_rate": 5.449673790581611e-06,
|
| 2731 |
+
"loss": 39.2977,
|
| 2732 |
+
"step": 389
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8858603066439523,
|
| 2736 |
+
"grad_norm": 3.111703395843506,
|
| 2737 |
+
"learning_rate": 5.241835432246889e-06,
|
| 2738 |
+
"loss": 39.3645,
|
| 2739 |
+
"step": 390
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.8881317433276548,
|
| 2743 |
+
"grad_norm": 3.189887046813965,
|
| 2744 |
+
"learning_rate": 5.037818373523723e-06,
|
| 2745 |
+
"loss": 39.3412,
|
| 2746 |
+
"step": 391
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.8904031800113572,
|
| 2750 |
+
"grad_norm": 3.096855640411377,
|
| 2751 |
+
"learning_rate": 4.837640032693558e-06,
|
| 2752 |
+
"loss": 39.4025,
|
| 2753 |
+
"step": 392
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.8926746166950597,
|
| 2757 |
+
"grad_norm": 2.981090784072876,
|
| 2758 |
+
"learning_rate": 4.641317500301173e-06,
|
| 2759 |
+
"loss": 39.5613,
|
| 2760 |
+
"step": 393
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.8949460533787621,
|
| 2764 |
+
"grad_norm": 3.221526861190796,
|
| 2765 |
+
"learning_rate": 4.448867537695578e-06,
|
| 2766 |
+
"loss": 39.298,
|
| 2767 |
+
"step": 394
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.8972174900624645,
|
| 2771 |
+
"grad_norm": 3.0834949016571045,
|
| 2772 |
+
"learning_rate": 4.260306575598949e-06,
|
| 2773 |
+
"loss": 39.4121,
|
| 2774 |
+
"step": 395
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.899488926746167,
|
| 2778 |
+
"grad_norm": 3.3117966651916504,
|
| 2779 |
+
"learning_rate": 4.075650712703849e-06,
|
| 2780 |
+
"loss": 39.3448,
|
| 2781 |
+
"step": 396
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.9017603634298694,
|
| 2785 |
+
"grad_norm": 3.3456838130950928,
|
| 2786 |
+
"learning_rate": 3.8949157142987746e-06,
|
| 2787 |
+
"loss": 39.2115,
|
| 2788 |
+
"step": 397
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9040318001135719,
|
| 2792 |
+
"grad_norm": 3.269028902053833,
|
| 2793 |
+
"learning_rate": 3.71811701092219e-06,
|
| 2794 |
+
"loss": 39.2706,
|
| 2795 |
+
"step": 398
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9063032367972743,
|
| 2799 |
+
"grad_norm": 3.4121360778808594,
|
| 2800 |
+
"learning_rate": 3.545269697045067e-06,
|
| 2801 |
+
"loss": 39.2025,
|
| 2802 |
+
"step": 399
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9085746734809768,
|
| 2806 |
+
"grad_norm": 3.294390916824341,
|
| 2807 |
+
"learning_rate": 3.376388529782215e-06,
|
| 2808 |
+
"loss": 39.3322,
|
| 2809 |
+
"step": 400
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9108461101646792,
|
| 2813 |
+
"grad_norm": 3.429776668548584,
|
| 2814 |
+
"learning_rate": 3.2114879276323783e-06,
|
| 2815 |
+
"loss": 39.2353,
|
| 2816 |
+
"step": 401
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9131175468483816,
|
| 2820 |
+
"grad_norm": 3.1151254177093506,
|
| 2821 |
+
"learning_rate": 3.0505819692471792e-06,
|
| 2822 |
+
"loss": 39.3944,
|
| 2823 |
+
"step": 402
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9153889835320841,
|
| 2827 |
+
"grad_norm": 3.2361867427825928,
|
| 2828 |
+
"learning_rate": 2.8936843922291847e-06,
|
| 2829 |
+
"loss": 39.2727,
|
| 2830 |
+
"step": 403
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9176604202157865,
|
| 2834 |
+
"grad_norm": 3.2054972648620605,
|
| 2835 |
+
"learning_rate": 2.7408085919590264e-06,
|
| 2836 |
+
"loss": 39.316,
|
| 2837 |
+
"step": 404
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.919931856899489,
|
| 2841 |
+
"grad_norm": 3.019122362136841,
|
| 2842 |
+
"learning_rate": 2.591967620451707e-06,
|
| 2843 |
+
"loss": 39.4442,
|
| 2844 |
+
"step": 405
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9222032935831914,
|
| 2848 |
+
"grad_norm": 3.0654804706573486,
|
| 2849 |
+
"learning_rate": 2.4471741852423237e-06,
|
| 2850 |
+
"loss": 39.4069,
|
| 2851 |
+
"step": 406
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.9244747302668939,
|
| 2855 |
+
"grad_norm": 3.1281793117523193,
|
| 2856 |
+
"learning_rate": 2.306440648301095e-06,
|
| 2857 |
+
"loss": 39.3489,
|
| 2858 |
+
"step": 407
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.9267461669505963,
|
| 2862 |
+
"grad_norm": 3.2688112258911133,
|
| 2863 |
+
"learning_rate": 2.1697790249779636e-06,
|
| 2864 |
+
"loss": 39.3093,
|
| 2865 |
+
"step": 408
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9290176036342986,
|
| 2869 |
+
"grad_norm": 3.3317718505859375,
|
| 2870 |
+
"learning_rate": 2.0372009829767557e-06,
|
| 2871 |
+
"loss": 39.2383,
|
| 2872 |
+
"step": 409
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9312890403180011,
|
| 2876 |
+
"grad_norm": 3.1908419132232666,
|
| 2877 |
+
"learning_rate": 1.908717841359048e-06,
|
| 2878 |
+
"loss": 39.3652,
|
| 2879 |
+
"step": 410
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9335604770017035,
|
| 2883 |
+
"grad_norm": 3.279597043991089,
|
| 2884 |
+
"learning_rate": 1.784340569577758e-06,
|
| 2885 |
+
"loss": 39.2765,
|
| 2886 |
+
"step": 411
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.935831913685406,
|
| 2890 |
+
"grad_norm": 3.275874614715576,
|
| 2891 |
+
"learning_rate": 1.6640797865406288e-06,
|
| 2892 |
+
"loss": 39.2798,
|
| 2893 |
+
"step": 412
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.9381033503691084,
|
| 2897 |
+
"grad_norm": 3.114420175552368,
|
| 2898 |
+
"learning_rate": 1.547945759703623e-06,
|
| 2899 |
+
"loss": 39.3152,
|
| 2900 |
+
"step": 413
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9403747870528109,
|
| 2904 |
+
"grad_norm": 3.284017324447632,
|
| 2905 |
+
"learning_rate": 1.4359484041943038e-06,
|
| 2906 |
+
"loss": 39.3038,
|
| 2907 |
+
"step": 414
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9426462237365133,
|
| 2911 |
+
"grad_norm": 3.1268672943115234,
|
| 2912 |
+
"learning_rate": 1.328097281965357e-06,
|
| 2913 |
+
"loss": 39.361,
|
| 2914 |
+
"step": 415
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.9449176604202157,
|
| 2918 |
+
"grad_norm": 3.221214532852173,
|
| 2919 |
+
"learning_rate": 1.2244016009781701e-06,
|
| 2920 |
+
"loss": 39.2966,
|
| 2921 |
+
"step": 416
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.9471890971039182,
|
| 2925 |
+
"grad_norm": 3.1034584045410156,
|
| 2926 |
+
"learning_rate": 1.1248702144167122e-06,
|
| 2927 |
+
"loss": 39.4244,
|
| 2928 |
+
"step": 417
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9494605337876206,
|
| 2932 |
+
"grad_norm": 3.3153748512268066,
|
| 2933 |
+
"learning_rate": 1.0295116199317057e-06,
|
| 2934 |
+
"loss": 39.2435,
|
| 2935 |
+
"step": 418
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9517319704713231,
|
| 2939 |
+
"grad_norm": 3.1586427688598633,
|
| 2940 |
+
"learning_rate": 9.383339589150775e-07,
|
| 2941 |
+
"loss": 39.3637,
|
| 2942 |
+
"step": 419
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9540034071550255,
|
| 2946 |
+
"grad_norm": 3.3240771293640137,
|
| 2947 |
+
"learning_rate": 8.513450158049108e-07,
|
| 2948 |
+
"loss": 39.4772,
|
| 2949 |
+
"step": 420
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.956274843838728,
|
| 2953 |
+
"grad_norm": 3.2368547916412354,
|
| 2954 |
+
"learning_rate": 7.685522174208204e-07,
|
| 2955 |
+
"loss": 39.3071,
|
| 2956 |
+
"step": 421
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9585462805224304,
|
| 2960 |
+
"grad_norm": 3.236651659011841,
|
| 2961 |
+
"learning_rate": 6.899626323298713e-07,
|
| 2962 |
+
"loss": 39.3127,
|
| 2963 |
+
"step": 422
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9608177172061328,
|
| 2967 |
+
"grad_norm": 3.3495962619781494,
|
| 2968 |
+
"learning_rate": 6.15582970243117e-07,
|
| 2969 |
+
"loss": 39.2375,
|
| 2970 |
+
"step": 423
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9630891538898353,
|
| 2974 |
+
"grad_norm": 3.3962039947509766,
|
| 2975 |
+
"learning_rate": 5.454195814427021e-07,
|
| 2976 |
+
"loss": 39.2121,
|
| 2977 |
+
"step": 424
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9653605905735377,
|
| 2981 |
+
"grad_norm": 3.2395989894866943,
|
| 2982 |
+
"learning_rate": 4.794784562397458e-07,
|
| 2983 |
+
"loss": 39.3095,
|
| 2984 |
+
"step": 425
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9676320272572402,
|
| 2988 |
+
"grad_norm": 3.381061315536499,
|
| 2989 |
+
"learning_rate": 4.177652244628627e-07,
|
| 2990 |
+
"loss": 39.2333,
|
| 2991 |
+
"step": 426
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9699034639409426,
|
| 2995 |
+
"grad_norm": 3.2255821228027344,
|
| 2996 |
+
"learning_rate": 3.602851549775521e-07,
|
| 2997 |
+
"loss": 39.429,
|
| 2998 |
+
"step": 427
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9721749006246451,
|
| 3002 |
+
"grad_norm": 3.1304969787597656,
|
| 3003 |
+
"learning_rate": 3.0704315523631953e-07,
|
| 3004 |
+
"loss": 39.4258,
|
| 3005 |
+
"step": 428
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.9744463373083475,
|
| 3009 |
+
"grad_norm": 3.189854145050049,
|
| 3010 |
+
"learning_rate": 2.5804377085972274e-07,
|
| 3011 |
+
"loss": 39.324,
|
| 3012 |
+
"step": 429
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.97671777399205,
|
| 3016 |
+
"grad_norm": 3.220968246459961,
|
| 3017 |
+
"learning_rate": 2.1329118524827662e-07,
|
| 3018 |
+
"loss": 39.285,
|
| 3019 |
+
"step": 430
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.9789892106757524,
|
| 3023 |
+
"grad_norm": 3.2234549522399902,
|
| 3024 |
+
"learning_rate": 1.7278921922527225e-07,
|
| 3025 |
+
"loss": 39.3381,
|
| 3026 |
+
"step": 431
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.9812606473594548,
|
| 3030 |
+
"grad_norm": 3.1990137100219727,
|
| 3031 |
+
"learning_rate": 1.3654133071059893e-07,
|
| 3032 |
+
"loss": 39.3761,
|
| 3033 |
+
"step": 432
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.9835320840431573,
|
| 3037 |
+
"grad_norm": 3.230268716812134,
|
| 3038 |
+
"learning_rate": 1.0455061442548597e-07,
|
| 3039 |
+
"loss": 39.3554,
|
| 3040 |
+
"step": 433
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9858035207268597,
|
| 3044 |
+
"grad_norm": 3.22092866897583,
|
| 3045 |
+
"learning_rate": 7.681980162830282e-08,
|
| 3046 |
+
"loss": 39.281,
|
| 3047 |
+
"step": 434
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.9880749574105622,
|
| 3051 |
+
"grad_norm": 3.2216899394989014,
|
| 3052 |
+
"learning_rate": 5.3351259881379014e-08,
|
| 3053 |
+
"loss": 39.3009,
|
| 3054 |
+
"step": 435
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9903463940942646,
|
| 3058 |
+
"grad_norm": 3.283959150314331,
|
| 3059 |
+
"learning_rate": 3.4146992848854695e-08,
|
| 3060 |
+
"loss": 39.2686,
|
| 3061 |
+
"step": 436
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9926178307779671,
|
| 3065 |
+
"grad_norm": 3.333789825439453,
|
| 3066 |
+
"learning_rate": 1.920864012562862e-08,
|
| 3067 |
+
"loss": 39.2601,
|
| 3068 |
+
"step": 437
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9948892674616695,
|
| 3072 |
+
"grad_norm": 3.285109281539917,
|
| 3073 |
+
"learning_rate": 8.537477097364522e-09,
|
| 3074 |
+
"loss": 39.2603,
|
| 3075 |
+
"step": 438
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.997160704145372,
|
| 3079 |
+
"grad_norm": 3.0986461639404297,
|
| 3080 |
+
"learning_rate": 2.1344148316060354e-09,
|
| 3081 |
+
"loss": 39.4098,
|
| 3082 |
+
"step": 439
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.9994321408290744,
|
| 3086 |
+
"grad_norm": 3.2520036697387695,
|
| 3087 |
+
"learning_rate": 0.0,
|
| 3088 |
+
"loss": 39.3073,
|
| 3089 |
+
"step": 440
|
| 3090 |
}
|
| 3091 |
],
|
| 3092 |
"logging_steps": 1,
|
|
|
|
| 3101 |
"should_evaluate": false,
|
| 3102 |
"should_log": false,
|
| 3103 |
"should_save": true,
|
| 3104 |
+
"should_training_stop": true
|
| 3105 |
},
|
| 3106 |
"attributes": {}
|
| 3107 |
}
|
| 3108 |
},
|
| 3109 |
+
"total_flos": 811196274966528.0,
|
| 3110 |
"train_batch_size": 4,
|
| 3111 |
"trial_name": null,
|
| 3112 |
"trial_params": null
|