Update latest checkpoint for run 15940d1d-7d8c-4253-8810-2ea934bedee4 at step 950
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +1304 -4
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 29529752
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:48b94cd8e622ebe3562c1ea9fa75f6468afc735c44a6e72504298526587d48b2
|
| 3 |
size 29529752
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 59228491
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:251f0a4a035e02d9e4cf95f08d80e5e54486d715ef25d979334aba0c383c50fc
|
| 3 |
size 59228491
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14645
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67cc8782dca6a2b07ec2b12d381573b2d3e7495fb1ce4a789f64f1147416fa5b
|
| 3 |
size 14645
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1465
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8f5a26ca81b45073ff90f001b3afe6c706ea419b3b9905df91151ee7ee1fc8a
|
| 3 |
size 1465
|
trainer_state.json
CHANGED
|
@@ -2,9 +2,9 @@
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
-
"epoch": 1.
|
| 6 |
"eval_steps": 500,
|
| 7 |
-
"global_step":
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
@@ -23408,11 +23408,1311 @@
|
|
| 23408 |
"rewards/reward_func/mean": 0.7445999979972839,
|
| 23409 |
"rewards/reward_func/std": 0.45967379212379456,
|
| 23410 |
"step": 900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23411 |
}
|
| 23412 |
],
|
| 23413 |
"logging_steps": 1,
|
| 23414 |
"max_steps": 950,
|
| 23415 |
-
"num_input_tokens_seen":
|
| 23416 |
"num_train_epochs": 2,
|
| 23417 |
"save_steps": 50,
|
| 23418 |
"stateful_callbacks": {
|
|
@@ -23422,7 +24722,7 @@
|
|
| 23422 |
"should_evaluate": false,
|
| 23423 |
"should_log": false,
|
| 23424 |
"should_save": true,
|
| 23425 |
-
"should_training_stop":
|
| 23426 |
},
|
| 23427 |
"attributes": {}
|
| 23428 |
}
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.85546875,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 950,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 23408 |
"rewards/reward_func/mean": 0.7445999979972839,
|
| 23409 |
"rewards/reward_func/std": 0.45967379212379456,
|
| 23410 |
"step": 900
|
| 23411 |
+
},
|
| 23412 |
+
{
|
| 23413 |
+
"clip_ratio/high_max": 0.0,
|
| 23414 |
+
"clip_ratio/high_mean": 0.0,
|
| 23415 |
+
"clip_ratio/low_mean": 0.0,
|
| 23416 |
+
"clip_ratio/low_min": 0.0,
|
| 23417 |
+
"clip_ratio/region_mean": 0.0,
|
| 23418 |
+
"completions/clipped_ratio": 1.0,
|
| 23419 |
+
"completions/max_length": 384.0,
|
| 23420 |
+
"completions/max_terminated_length": 0.0,
|
| 23421 |
+
"completions/mean_length": 384.0,
|
| 23422 |
+
"completions/mean_terminated_length": 0.0,
|
| 23423 |
+
"completions/min_length": 384.0,
|
| 23424 |
+
"completions/min_terminated_length": 0.0,
|
| 23425 |
+
"entropy": 0.8848395217210054,
|
| 23426 |
+
"epoch": 1.759765625,
|
| 23427 |
+
"frac_reward_zero_std": 0.0,
|
| 23428 |
+
"grad_norm": 0.2501536011695862,
|
| 23429 |
+
"learning_rate": 2.6315789473684213e-07,
|
| 23430 |
+
"loss": 0.0,
|
| 23431 |
+
"num_tokens": 4243728.0,
|
| 23432 |
+
"reward": 0.73808753490448,
|
| 23433 |
+
"reward_std": 0.4636053442955017,
|
| 23434 |
+
"rewards/reward_func/mean": 0.7380874752998352,
|
| 23435 |
+
"rewards/reward_func/std": 0.4300731420516968,
|
| 23436 |
+
"step": 901
|
| 23437 |
+
},
|
| 23438 |
+
{
|
| 23439 |
+
"clip_ratio/high_max": 0.0,
|
| 23440 |
+
"clip_ratio/high_mean": 0.0,
|
| 23441 |
+
"clip_ratio/low_mean": 0.0,
|
| 23442 |
+
"clip_ratio/low_min": 0.0,
|
| 23443 |
+
"clip_ratio/region_mean": 0.0,
|
| 23444 |
+
"completions/clipped_ratio": 1.0,
|
| 23445 |
+
"completions/max_length": 384.0,
|
| 23446 |
+
"completions/max_terminated_length": 0.0,
|
| 23447 |
+
"completions/mean_length": 384.0,
|
| 23448 |
+
"completions/mean_terminated_length": 0.0,
|
| 23449 |
+
"completions/min_length": 384.0,
|
| 23450 |
+
"completions/min_terminated_length": 0.0,
|
| 23451 |
+
"entropy": 0.5995431952178478,
|
| 23452 |
+
"epoch": 1.76171875,
|
| 23453 |
+
"frac_reward_zero_std": 0.5,
|
| 23454 |
+
"grad_norm": 0.1700550615787506,
|
| 23455 |
+
"learning_rate": 2.578947368421053e-07,
|
| 23456 |
+
"loss": -0.0,
|
| 23457 |
+
"num_tokens": 4248720.0,
|
| 23458 |
+
"reward": 0.2184000015258789,
|
| 23459 |
+
"reward_std": 0.2522551119327545,
|
| 23460 |
+
"rewards/reward_func/mean": 0.2184000015258789,
|
| 23461 |
+
"rewards/reward_func/std": 0.4044714868068695,
|
| 23462 |
+
"step": 902
|
| 23463 |
+
},
|
| 23464 |
+
{
|
| 23465 |
+
"clip_ratio/high_max": 0.0,
|
| 23466 |
+
"clip_ratio/high_mean": 0.0,
|
| 23467 |
+
"clip_ratio/low_mean": 0.0,
|
| 23468 |
+
"clip_ratio/low_min": 0.0,
|
| 23469 |
+
"clip_ratio/region_mean": 0.0,
|
| 23470 |
+
"completions/clipped_ratio": 1.0,
|
| 23471 |
+
"completions/max_length": 384.0,
|
| 23472 |
+
"completions/max_terminated_length": 0.0,
|
| 23473 |
+
"completions/mean_length": 384.0,
|
| 23474 |
+
"completions/mean_terminated_length": 0.0,
|
| 23475 |
+
"completions/min_length": 384.0,
|
| 23476 |
+
"completions/min_terminated_length": 0.0,
|
| 23477 |
+
"entropy": 0.8698443882167339,
|
| 23478 |
+
"epoch": 1.763671875,
|
| 23479 |
+
"frac_reward_zero_std": 0.0,
|
| 23480 |
+
"grad_norm": 0.2777104675769806,
|
| 23481 |
+
"learning_rate": 2.5263157894736846e-07,
|
| 23482 |
+
"loss": -0.0,
|
| 23483 |
+
"num_tokens": 4253256.0,
|
| 23484 |
+
"reward": 0.504687488079071,
|
| 23485 |
+
"reward_std": 0.5720410346984863,
|
| 23486 |
+
"rewards/reward_func/mean": 0.504687488079071,
|
| 23487 |
+
"rewards/reward_func/std": 0.5296536087989807,
|
| 23488 |
+
"step": 903
|
| 23489 |
+
},
|
| 23490 |
+
{
|
| 23491 |
+
"clip_ratio/high_max": 0.0,
|
| 23492 |
+
"clip_ratio/high_mean": 0.0,
|
| 23493 |
+
"clip_ratio/low_mean": 0.0,
|
| 23494 |
+
"clip_ratio/low_min": 0.0,
|
| 23495 |
+
"clip_ratio/region_mean": 0.0,
|
| 23496 |
+
"completions/clipped_ratio": 1.0,
|
| 23497 |
+
"completions/max_length": 384.0,
|
| 23498 |
+
"completions/max_terminated_length": 0.0,
|
| 23499 |
+
"completions/mean_length": 384.0,
|
| 23500 |
+
"completions/mean_terminated_length": 0.0,
|
| 23501 |
+
"completions/min_length": 384.0,
|
| 23502 |
+
"completions/min_terminated_length": 0.0,
|
| 23503 |
+
"entropy": 0.7046304792165756,
|
| 23504 |
+
"epoch": 1.765625,
|
| 23505 |
+
"frac_reward_zero_std": 0.0,
|
| 23506 |
+
"grad_norm": 0.23841030895709991,
|
| 23507 |
+
"learning_rate": 2.473684210526316e-07,
|
| 23508 |
+
"loss": 0.0,
|
| 23509 |
+
"num_tokens": 4258024.0,
|
| 23510 |
+
"reward": 0.6187999844551086,
|
| 23511 |
+
"reward_std": 0.5329432487487793,
|
| 23512 |
+
"rewards/reward_func/mean": 0.6187999844551086,
|
| 23513 |
+
"rewards/reward_func/std": 0.5125207901000977,
|
| 23514 |
+
"step": 904
|
| 23515 |
+
},
|
| 23516 |
+
{
|
| 23517 |
+
"clip_ratio/high_max": 0.0,
|
| 23518 |
+
"clip_ratio/high_mean": 0.0,
|
| 23519 |
+
"clip_ratio/low_mean": 0.0,
|
| 23520 |
+
"clip_ratio/low_min": 0.0,
|
| 23521 |
+
"clip_ratio/region_mean": 0.0,
|
| 23522 |
+
"completions/clipped_ratio": 1.0,
|
| 23523 |
+
"completions/max_length": 384.0,
|
| 23524 |
+
"completions/max_terminated_length": 0.0,
|
| 23525 |
+
"completions/mean_length": 384.0,
|
| 23526 |
+
"completions/mean_terminated_length": 0.0,
|
| 23527 |
+
"completions/min_length": 384.0,
|
| 23528 |
+
"completions/min_terminated_length": 0.0,
|
| 23529 |
+
"entropy": 0.7159217670559883,
|
| 23530 |
+
"epoch": 1.767578125,
|
| 23531 |
+
"frac_reward_zero_std": 0.0,
|
| 23532 |
+
"grad_norm": 0.2175939530134201,
|
| 23533 |
+
"learning_rate": 2.4210526315789473e-07,
|
| 23534 |
+
"loss": 0.0,
|
| 23535 |
+
"num_tokens": 4263024.0,
|
| 23536 |
+
"reward": 0.3587000072002411,
|
| 23537 |
+
"reward_std": 0.5176094770431519,
|
| 23538 |
+
"rewards/reward_func/mean": 0.3587000072002411,
|
| 23539 |
+
"rewards/reward_func/std": 0.4951927065849304,
|
| 23540 |
+
"step": 905
|
| 23541 |
+
},
|
| 23542 |
+
{
|
| 23543 |
+
"clip_ratio/high_max": 0.0,
|
| 23544 |
+
"clip_ratio/high_mean": 0.0,
|
| 23545 |
+
"clip_ratio/low_mean": 0.0,
|
| 23546 |
+
"clip_ratio/low_min": 0.0,
|
| 23547 |
+
"clip_ratio/region_mean": 0.0,
|
| 23548 |
+
"completions/clipped_ratio": 1.0,
|
| 23549 |
+
"completions/max_length": 384.0,
|
| 23550 |
+
"completions/max_terminated_length": 0.0,
|
| 23551 |
+
"completions/mean_length": 384.0,
|
| 23552 |
+
"completions/mean_terminated_length": 0.0,
|
| 23553 |
+
"completions/min_length": 384.0,
|
| 23554 |
+
"completions/min_terminated_length": 0.0,
|
| 23555 |
+
"entropy": 0.4197417329996824,
|
| 23556 |
+
"epoch": 1.76953125,
|
| 23557 |
+
"frac_reward_zero_std": 0.0,
|
| 23558 |
+
"grad_norm": 0.18832392990589142,
|
| 23559 |
+
"learning_rate": 2.3684210526315792e-07,
|
| 23560 |
+
"loss": 0.0,
|
| 23561 |
+
"num_tokens": 4267968.0,
|
| 23562 |
+
"reward": 0.012500000186264515,
|
| 23563 |
+
"reward_std": 0.014433757402002811,
|
| 23564 |
+
"rewards/reward_func/mean": 0.012500000186264515,
|
| 23565 |
+
"rewards/reward_func/std": 0.013363063335418701,
|
| 23566 |
+
"step": 906
|
| 23567 |
+
},
|
| 23568 |
+
{
|
| 23569 |
+
"clip_ratio/high_max": 0.0,
|
| 23570 |
+
"clip_ratio/high_mean": 0.0,
|
| 23571 |
+
"clip_ratio/low_mean": 0.0,
|
| 23572 |
+
"clip_ratio/low_min": 0.0,
|
| 23573 |
+
"clip_ratio/region_mean": 0.0,
|
| 23574 |
+
"completions/clipped_ratio": 1.0,
|
| 23575 |
+
"completions/max_length": 384.0,
|
| 23576 |
+
"completions/max_terminated_length": 0.0,
|
| 23577 |
+
"completions/mean_length": 384.0,
|
| 23578 |
+
"completions/mean_terminated_length": 0.0,
|
| 23579 |
+
"completions/min_length": 384.0,
|
| 23580 |
+
"completions/min_terminated_length": 0.0,
|
| 23581 |
+
"entropy": 0.7613006345927715,
|
| 23582 |
+
"epoch": 1.771484375,
|
| 23583 |
+
"frac_reward_zero_std": 0.5,
|
| 23584 |
+
"grad_norm": 0.17875666916370392,
|
| 23585 |
+
"learning_rate": 2.315789473684211e-07,
|
| 23586 |
+
"loss": -0.0,
|
| 23587 |
+
"num_tokens": 4272328.0,
|
| 23588 |
+
"reward": 0.9947500228881836,
|
| 23589 |
+
"reward_std": 0.010500003583729267,
|
| 23590 |
+
"rewards/reward_func/mean": 0.9947500228881836,
|
| 23591 |
+
"rewards/reward_func/std": 0.014849240891635418,
|
| 23592 |
+
"step": 907
|
| 23593 |
+
},
|
| 23594 |
+
{
|
| 23595 |
+
"clip_ratio/high_max": 0.0,
|
| 23596 |
+
"clip_ratio/high_mean": 0.0,
|
| 23597 |
+
"clip_ratio/low_mean": 0.0,
|
| 23598 |
+
"clip_ratio/low_min": 0.0,
|
| 23599 |
+
"clip_ratio/region_mean": 0.0,
|
| 23600 |
+
"completions/clipped_ratio": 1.0,
|
| 23601 |
+
"completions/max_length": 384.0,
|
| 23602 |
+
"completions/max_terminated_length": 0.0,
|
| 23603 |
+
"completions/mean_length": 384.0,
|
| 23604 |
+
"completions/mean_terminated_length": 0.0,
|
| 23605 |
+
"completions/min_length": 384.0,
|
| 23606 |
+
"completions/min_terminated_length": 0.0,
|
| 23607 |
+
"entropy": 0.7562070265412331,
|
| 23608 |
+
"epoch": 1.7734375,
|
| 23609 |
+
"frac_reward_zero_std": 0.0,
|
| 23610 |
+
"grad_norm": 0.2620117664337158,
|
| 23611 |
+
"learning_rate": 2.2631578947368425e-07,
|
| 23612 |
+
"loss": 0.0,
|
| 23613 |
+
"num_tokens": 4276768.0,
|
| 23614 |
+
"reward": 0.6265624761581421,
|
| 23615 |
+
"reward_std": 0.5368822813034058,
|
| 23616 |
+
"rewards/reward_func/mean": 0.6265624761581421,
|
| 23617 |
+
"rewards/reward_func/std": 0.515407145023346,
|
| 23618 |
+
"step": 908
|
| 23619 |
+
},
|
| 23620 |
+
{
|
| 23621 |
+
"clip_ratio/high_max": 0.0,
|
| 23622 |
+
"clip_ratio/high_mean": 0.0,
|
| 23623 |
+
"clip_ratio/low_mean": 0.0,
|
| 23624 |
+
"clip_ratio/low_min": 0.0,
|
| 23625 |
+
"clip_ratio/region_mean": 0.0,
|
| 23626 |
+
"completions/clipped_ratio": 1.0,
|
| 23627 |
+
"completions/max_length": 384.0,
|
| 23628 |
+
"completions/max_terminated_length": 0.0,
|
| 23629 |
+
"completions/mean_length": 384.0,
|
| 23630 |
+
"completions/mean_terminated_length": 0.0,
|
| 23631 |
+
"completions/min_length": 384.0,
|
| 23632 |
+
"completions/min_terminated_length": 0.0,
|
| 23633 |
+
"entropy": 0.735223077237606,
|
| 23634 |
+
"epoch": 1.775390625,
|
| 23635 |
+
"frac_reward_zero_std": 0.0,
|
| 23636 |
+
"grad_norm": 0.24371092021465302,
|
| 23637 |
+
"learning_rate": 2.2105263157894736e-07,
|
| 23638 |
+
"loss": -0.0,
|
| 23639 |
+
"num_tokens": 4281480.0,
|
| 23640 |
+
"reward": 0.9846999645233154,
|
| 23641 |
+
"reward_std": 0.03060000017285347,
|
| 23642 |
+
"rewards/reward_func/mean": 0.9846999645233154,
|
| 23643 |
+
"rewards/reward_func/std": 0.030131708830595016,
|
| 23644 |
+
"step": 909
|
| 23645 |
+
},
|
| 23646 |
+
{
|
| 23647 |
+
"clip_ratio/high_max": 0.0,
|
| 23648 |
+
"clip_ratio/high_mean": 0.0,
|
| 23649 |
+
"clip_ratio/low_mean": 0.0,
|
| 23650 |
+
"clip_ratio/low_min": 0.0,
|
| 23651 |
+
"clip_ratio/region_mean": 0.0,
|
| 23652 |
+
"completions/clipped_ratio": 1.0,
|
| 23653 |
+
"completions/max_length": 384.0,
|
| 23654 |
+
"completions/max_terminated_length": 0.0,
|
| 23655 |
+
"completions/mean_length": 384.0,
|
| 23656 |
+
"completions/mean_terminated_length": 0.0,
|
| 23657 |
+
"completions/min_length": 384.0,
|
| 23658 |
+
"completions/min_terminated_length": 0.0,
|
| 23659 |
+
"entropy": 0.888181284070015,
|
| 23660 |
+
"epoch": 1.77734375,
|
| 23661 |
+
"frac_reward_zero_std": 0.0,
|
| 23662 |
+
"grad_norm": 0.2600753605365753,
|
| 23663 |
+
"learning_rate": 2.1578947368421053e-07,
|
| 23664 |
+
"loss": 0.0,
|
| 23665 |
+
"num_tokens": 4285840.0,
|
| 23666 |
+
"reward": 0.5,
|
| 23667 |
+
"reward_std": 0.5773502588272095,
|
| 23668 |
+
"rewards/reward_func/mean": 0.5,
|
| 23669 |
+
"rewards/reward_func/std": 0.5345224738121033,
|
| 23670 |
+
"step": 910
|
| 23671 |
+
},
|
| 23672 |
+
{
|
| 23673 |
+
"clip_ratio/high_max": 0.0,
|
| 23674 |
+
"clip_ratio/high_mean": 0.0,
|
| 23675 |
+
"clip_ratio/low_mean": 0.0,
|
| 23676 |
+
"clip_ratio/low_min": 0.0,
|
| 23677 |
+
"clip_ratio/region_mean": 0.0,
|
| 23678 |
+
"completions/clipped_ratio": 1.0,
|
| 23679 |
+
"completions/max_length": 384.0,
|
| 23680 |
+
"completions/max_terminated_length": 0.0,
|
| 23681 |
+
"completions/mean_length": 384.0,
|
| 23682 |
+
"completions/mean_terminated_length": 0.0,
|
| 23683 |
+
"completions/min_length": 384.0,
|
| 23684 |
+
"completions/min_terminated_length": 0.0,
|
| 23685 |
+
"entropy": 0.6173073314130306,
|
| 23686 |
+
"epoch": 1.779296875,
|
| 23687 |
+
"frac_reward_zero_std": 0.0,
|
| 23688 |
+
"grad_norm": 0.22857871651649475,
|
| 23689 |
+
"learning_rate": 2.105263157894737e-07,
|
| 23690 |
+
"loss": -0.0,
|
| 23691 |
+
"num_tokens": 4290496.0,
|
| 23692 |
+
"reward": 0.37229999899864197,
|
| 23693 |
+
"reward_std": 0.5355914831161499,
|
| 23694 |
+
"rewards/reward_func/mean": 0.37229999899864197,
|
| 23695 |
+
"rewards/reward_func/std": 0.5138660669326782,
|
| 23696 |
+
"step": 911
|
| 23697 |
+
},
|
| 23698 |
+
{
|
| 23699 |
+
"clip_ratio/high_max": 0.0,
|
| 23700 |
+
"clip_ratio/high_mean": 0.0,
|
| 23701 |
+
"clip_ratio/low_mean": 0.0,
|
| 23702 |
+
"clip_ratio/low_min": 0.0,
|
| 23703 |
+
"clip_ratio/region_mean": 0.0,
|
| 23704 |
+
"completions/clipped_ratio": 1.0,
|
| 23705 |
+
"completions/max_length": 384.0,
|
| 23706 |
+
"completions/max_terminated_length": 0.0,
|
| 23707 |
+
"completions/mean_length": 384.0,
|
| 23708 |
+
"completions/mean_terminated_length": 0.0,
|
| 23709 |
+
"completions/min_length": 384.0,
|
| 23710 |
+
"completions/min_terminated_length": 0.0,
|
| 23711 |
+
"entropy": 0.5189631078392267,
|
| 23712 |
+
"epoch": 1.78125,
|
| 23713 |
+
"frac_reward_zero_std": 0.5,
|
| 23714 |
+
"grad_norm": 0.14599941670894623,
|
| 23715 |
+
"learning_rate": 2.0526315789473685e-07,
|
| 23716 |
+
"loss": -0.0,
|
| 23717 |
+
"num_tokens": 4295008.0,
|
| 23718 |
+
"reward": 0.8134000301361084,
|
| 23719 |
+
"reward_std": 0.23240000009536743,
|
| 23720 |
+
"rewards/reward_func/mean": 0.8134000301361084,
|
| 23721 |
+
"rewards/reward_func/std": 0.3286632299423218,
|
| 23722 |
+
"step": 912
|
| 23723 |
+
},
|
| 23724 |
+
{
|
| 23725 |
+
"clip_ratio/high_max": 0.0,
|
| 23726 |
+
"clip_ratio/high_mean": 0.0,
|
| 23727 |
+
"clip_ratio/low_mean": 0.0,
|
| 23728 |
+
"clip_ratio/low_min": 0.0,
|
| 23729 |
+
"clip_ratio/region_mean": 0.0,
|
| 23730 |
+
"completions/clipped_ratio": 1.0,
|
| 23731 |
+
"completions/max_length": 384.0,
|
| 23732 |
+
"completions/max_terminated_length": 0.0,
|
| 23733 |
+
"completions/mean_length": 384.0,
|
| 23734 |
+
"completions/mean_terminated_length": 0.0,
|
| 23735 |
+
"completions/min_length": 384.0,
|
| 23736 |
+
"completions/min_terminated_length": 0.0,
|
| 23737 |
+
"entropy": 0.8033823296427727,
|
| 23738 |
+
"epoch": 1.783203125,
|
| 23739 |
+
"frac_reward_zero_std": 0.5,
|
| 23740 |
+
"grad_norm": 0.18053551018238068,
|
| 23741 |
+
"learning_rate": 2.0000000000000002e-07,
|
| 23742 |
+
"loss": 0.0,
|
| 23743 |
+
"num_tokens": 4299768.0,
|
| 23744 |
+
"reward": 0.8362500071525574,
|
| 23745 |
+
"reward_std": 0.22793914377689362,
|
| 23746 |
+
"rewards/reward_func/mean": 0.8362500071525574,
|
| 23747 |
+
"rewards/reward_func/std": 0.3183859884738922,
|
| 23748 |
+
"step": 913
|
| 23749 |
+
},
|
| 23750 |
+
{
|
| 23751 |
+
"clip_ratio/high_max": 0.0,
|
| 23752 |
+
"clip_ratio/high_mean": 0.0,
|
| 23753 |
+
"clip_ratio/low_mean": 0.0,
|
| 23754 |
+
"clip_ratio/low_min": 0.0,
|
| 23755 |
+
"clip_ratio/region_mean": 0.0,
|
| 23756 |
+
"completions/clipped_ratio": 1.0,
|
| 23757 |
+
"completions/max_length": 384.0,
|
| 23758 |
+
"completions/max_terminated_length": 0.0,
|
| 23759 |
+
"completions/mean_length": 384.0,
|
| 23760 |
+
"completions/mean_terminated_length": 0.0,
|
| 23761 |
+
"completions/min_length": 384.0,
|
| 23762 |
+
"completions/min_terminated_length": 0.0,
|
| 23763 |
+
"entropy": 0.6970736794173717,
|
| 23764 |
+
"epoch": 1.78515625,
|
| 23765 |
+
"frac_reward_zero_std": 0.5,
|
| 23766 |
+
"grad_norm": 0.16842643916606903,
|
| 23767 |
+
"learning_rate": 1.9473684210526318e-07,
|
| 23768 |
+
"loss": -0.0,
|
| 23769 |
+
"num_tokens": 4304520.0,
|
| 23770 |
+
"reward": 0.8224999904632568,
|
| 23771 |
+
"reward_std": 0.23499999940395355,
|
| 23772 |
+
"rewards/reward_func/mean": 0.8224999904632568,
|
| 23773 |
+
"rewards/reward_func/std": 0.33234018087387085,
|
| 23774 |
+
"step": 914
|
| 23775 |
+
},
|
| 23776 |
+
{
|
| 23777 |
+
"clip_ratio/high_max": 0.0,
|
| 23778 |
+
"clip_ratio/high_mean": 0.0,
|
| 23779 |
+
"clip_ratio/low_mean": 0.0,
|
| 23780 |
+
"clip_ratio/low_min": 0.0,
|
| 23781 |
+
"clip_ratio/region_mean": 0.0,
|
| 23782 |
+
"completions/clipped_ratio": 1.0,
|
| 23783 |
+
"completions/max_length": 384.0,
|
| 23784 |
+
"completions/max_terminated_length": 0.0,
|
| 23785 |
+
"completions/mean_length": 384.0,
|
| 23786 |
+
"completions/mean_terminated_length": 0.0,
|
| 23787 |
+
"completions/min_length": 384.0,
|
| 23788 |
+
"completions/min_terminated_length": 0.0,
|
| 23789 |
+
"entropy": 0.6607658788561821,
|
| 23790 |
+
"epoch": 1.787109375,
|
| 23791 |
+
"frac_reward_zero_std": 0.0,
|
| 23792 |
+
"grad_norm": 0.22088974714279175,
|
| 23793 |
+
"learning_rate": 1.8947368421052634e-07,
|
| 23794 |
+
"loss": 0.0,
|
| 23795 |
+
"num_tokens": 4309296.0,
|
| 23796 |
+
"reward": 0.7371000051498413,
|
| 23797 |
+
"reward_std": 0.28240811824798584,
|
| 23798 |
+
"rewards/reward_func/mean": 0.7371000051498413,
|
| 23799 |
+
"rewards/reward_func/std": 0.4553808271884918,
|
| 23800 |
+
"step": 915
|
| 23801 |
+
},
|
| 23802 |
+
{
|
| 23803 |
+
"clip_ratio/high_max": 0.0,
|
| 23804 |
+
"clip_ratio/high_mean": 0.0,
|
| 23805 |
+
"clip_ratio/low_mean": 0.0,
|
| 23806 |
+
"clip_ratio/low_min": 0.0,
|
| 23807 |
+
"clip_ratio/region_mean": 0.0,
|
| 23808 |
+
"completions/clipped_ratio": 1.0,
|
| 23809 |
+
"completions/max_length": 384.0,
|
| 23810 |
+
"completions/max_terminated_length": 0.0,
|
| 23811 |
+
"completions/mean_length": 384.0,
|
| 23812 |
+
"completions/mean_terminated_length": 0.0,
|
| 23813 |
+
"completions/min_length": 384.0,
|
| 23814 |
+
"completions/min_terminated_length": 0.0,
|
| 23815 |
+
"entropy": 0.6025678031146526,
|
| 23816 |
+
"epoch": 1.7890625,
|
| 23817 |
+
"frac_reward_zero_std": 0.0,
|
| 23818 |
+
"grad_norm": 0.2665260136127472,
|
| 23819 |
+
"learning_rate": 1.8421052631578948e-07,
|
| 23820 |
+
"loss": 0.0,
|
| 23821 |
+
"num_tokens": 4314088.0,
|
| 23822 |
+
"reward": 0.8602499961853027,
|
| 23823 |
+
"reward_std": 0.24355539679527283,
|
| 23824 |
+
"rewards/reward_func/mean": 0.8602499961853027,
|
| 23825 |
+
"rewards/reward_func/std": 0.32780489325523376,
|
| 23826 |
+
"step": 916
|
| 23827 |
+
},
|
| 23828 |
+
{
|
| 23829 |
+
"clip_ratio/high_max": 0.0,
|
| 23830 |
+
"clip_ratio/high_mean": 0.0,
|
| 23831 |
+
"clip_ratio/low_mean": 0.0,
|
| 23832 |
+
"clip_ratio/low_min": 0.0,
|
| 23833 |
+
"clip_ratio/region_mean": 0.0,
|
| 23834 |
+
"completions/clipped_ratio": 1.0,
|
| 23835 |
+
"completions/max_length": 384.0,
|
| 23836 |
+
"completions/max_terminated_length": 0.0,
|
| 23837 |
+
"completions/mean_length": 384.0,
|
| 23838 |
+
"completions/mean_terminated_length": 0.0,
|
| 23839 |
+
"completions/min_length": 384.0,
|
| 23840 |
+
"completions/min_terminated_length": 0.0,
|
| 23841 |
+
"entropy": 0.421773798763752,
|
| 23842 |
+
"epoch": 1.791015625,
|
| 23843 |
+
"frac_reward_zero_std": 0.5,
|
| 23844 |
+
"grad_norm": 0.12458179891109467,
|
| 23845 |
+
"learning_rate": 1.7894736842105265e-07,
|
| 23846 |
+
"loss": 0.0,
|
| 23847 |
+
"num_tokens": 4319040.0,
|
| 23848 |
+
"reward": 0.0031250000465661287,
|
| 23849 |
+
"reward_std": 0.0062500000931322575,
|
| 23850 |
+
"rewards/reward_func/mean": 0.0031250000465661287,
|
| 23851 |
+
"rewards/reward_func/std": 0.008838835172355175,
|
| 23852 |
+
"step": 917
|
| 23853 |
+
},
|
| 23854 |
+
{
|
| 23855 |
+
"clip_ratio/high_max": 0.0,
|
| 23856 |
+
"clip_ratio/high_mean": 0.0,
|
| 23857 |
+
"clip_ratio/low_mean": 0.0,
|
| 23858 |
+
"clip_ratio/low_min": 0.0,
|
| 23859 |
+
"clip_ratio/region_mean": 0.0,
|
| 23860 |
+
"completions/clipped_ratio": 1.0,
|
| 23861 |
+
"completions/max_length": 384.0,
|
| 23862 |
+
"completions/max_terminated_length": 0.0,
|
| 23863 |
+
"completions/mean_length": 384.0,
|
| 23864 |
+
"completions/mean_terminated_length": 0.0,
|
| 23865 |
+
"completions/min_length": 384.0,
|
| 23866 |
+
"completions/min_terminated_length": 0.0,
|
| 23867 |
+
"entropy": 0.8600736558437347,
|
| 23868 |
+
"epoch": 1.79296875,
|
| 23869 |
+
"frac_reward_zero_std": 0.0,
|
| 23870 |
+
"grad_norm": 0.2437726855278015,
|
| 23871 |
+
"learning_rate": 1.7368421052631578e-07,
|
| 23872 |
+
"loss": -0.0,
|
| 23873 |
+
"num_tokens": 4323544.0,
|
| 23874 |
+
"reward": 0.5066750049591064,
|
| 23875 |
+
"reward_std": 0.4920022189617157,
|
| 23876 |
+
"rewards/reward_func/mean": 0.5066750049591064,
|
| 23877 |
+
"rewards/reward_func/std": 0.5217258930206299,
|
| 23878 |
+
"step": 918
|
| 23879 |
+
},
|
| 23880 |
+
{
|
| 23881 |
+
"clip_ratio/high_max": 0.0,
|
| 23882 |
+
"clip_ratio/high_mean": 0.0,
|
| 23883 |
+
"clip_ratio/low_mean": 0.0,
|
| 23884 |
+
"clip_ratio/low_min": 0.0,
|
| 23885 |
+
"clip_ratio/region_mean": 0.0,
|
| 23886 |
+
"completions/clipped_ratio": 1.0,
|
| 23887 |
+
"completions/max_length": 384.0,
|
| 23888 |
+
"completions/max_terminated_length": 0.0,
|
| 23889 |
+
"completions/mean_length": 384.0,
|
| 23890 |
+
"completions/mean_terminated_length": 0.0,
|
| 23891 |
+
"completions/min_length": 384.0,
|
| 23892 |
+
"completions/min_terminated_length": 0.0,
|
| 23893 |
+
"entropy": 0.7552273385226727,
|
| 23894 |
+
"epoch": 1.794921875,
|
| 23895 |
+
"frac_reward_zero_std": 0.0,
|
| 23896 |
+
"grad_norm": 0.23656296730041504,
|
| 23897 |
+
"learning_rate": 1.6842105263157895e-07,
|
| 23898 |
+
"loss": -0.0,
|
| 23899 |
+
"num_tokens": 4328408.0,
|
| 23900 |
+
"reward": 0.8660625219345093,
|
| 23901 |
+
"reward_std": 0.2540762424468994,
|
| 23902 |
+
"rewards/reward_func/mean": 0.8660625219345093,
|
| 23903 |
+
"rewards/reward_func/std": 0.34541285037994385,
|
| 23904 |
+
"step": 919
|
| 23905 |
+
},
|
| 23906 |
+
{
|
| 23907 |
+
"clip_ratio/high_max": 0.0,
|
| 23908 |
+
"clip_ratio/high_mean": 0.0,
|
| 23909 |
+
"clip_ratio/low_mean": 0.0,
|
| 23910 |
+
"clip_ratio/low_min": 0.0,
|
| 23911 |
+
"clip_ratio/region_mean": 0.0,
|
| 23912 |
+
"completions/clipped_ratio": 1.0,
|
| 23913 |
+
"completions/max_length": 384.0,
|
| 23914 |
+
"completions/max_terminated_length": 0.0,
|
| 23915 |
+
"completions/mean_length": 384.0,
|
| 23916 |
+
"completions/mean_terminated_length": 0.0,
|
| 23917 |
+
"completions/min_length": 384.0,
|
| 23918 |
+
"completions/min_terminated_length": 0.0,
|
| 23919 |
+
"entropy": 0.6802506037056446,
|
| 23920 |
+
"epoch": 1.796875,
|
| 23921 |
+
"frac_reward_zero_std": 0.5,
|
| 23922 |
+
"grad_norm": 0.1381627321243286,
|
| 23923 |
+
"learning_rate": 1.631578947368421e-07,
|
| 23924 |
+
"loss": 0.0,
|
| 23925 |
+
"num_tokens": 4333080.0,
|
| 23926 |
+
"reward": 0.875,
|
| 23927 |
+
"reward_std": 0.25,
|
| 23928 |
+
"rewards/reward_func/mean": 0.875,
|
| 23929 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 23930 |
+
"step": 920
|
| 23931 |
+
},
|
| 23932 |
+
{
|
| 23933 |
+
"clip_ratio/high_max": 0.0,
|
| 23934 |
+
"clip_ratio/high_mean": 0.0,
|
| 23935 |
+
"clip_ratio/low_mean": 0.0,
|
| 23936 |
+
"clip_ratio/low_min": 0.0,
|
| 23937 |
+
"clip_ratio/region_mean": 0.0,
|
| 23938 |
+
"completions/clipped_ratio": 1.0,
|
| 23939 |
+
"completions/max_length": 384.0,
|
| 23940 |
+
"completions/max_terminated_length": 0.0,
|
| 23941 |
+
"completions/mean_length": 384.0,
|
| 23942 |
+
"completions/mean_terminated_length": 0.0,
|
| 23943 |
+
"completions/min_length": 384.0,
|
| 23944 |
+
"completions/min_terminated_length": 0.0,
|
| 23945 |
+
"entropy": 0.19835966220125556,
|
| 23946 |
+
"epoch": 1.798828125,
|
| 23947 |
+
"frac_reward_zero_std": 1.0,
|
| 23948 |
+
"grad_norm": 0.0,
|
| 23949 |
+
"learning_rate": 1.5789473684210527e-07,
|
| 23950 |
+
"loss": 0.0,
|
| 23951 |
+
"num_tokens": 4337592.0,
|
| 23952 |
+
"reward": 0.9296000003814697,
|
| 23953 |
+
"reward_std": 0.0,
|
| 23954 |
+
"rewards/reward_func/mean": 0.9296000003814697,
|
| 23955 |
+
"rewards/reward_func/std": 0.0,
|
| 23956 |
+
"step": 921
|
| 23957 |
+
},
|
| 23958 |
+
{
|
| 23959 |
+
"clip_ratio/high_max": 0.0,
|
| 23960 |
+
"clip_ratio/high_mean": 0.0,
|
| 23961 |
+
"clip_ratio/low_mean": 0.0,
|
| 23962 |
+
"clip_ratio/low_min": 0.0,
|
| 23963 |
+
"clip_ratio/region_mean": 0.0,
|
| 23964 |
+
"completions/clipped_ratio": 1.0,
|
| 23965 |
+
"completions/max_length": 384.0,
|
| 23966 |
+
"completions/max_terminated_length": 0.0,
|
| 23967 |
+
"completions/mean_length": 384.0,
|
| 23968 |
+
"completions/mean_terminated_length": 0.0,
|
| 23969 |
+
"completions/min_length": 384.0,
|
| 23970 |
+
"completions/min_terminated_length": 0.0,
|
| 23971 |
+
"entropy": 0.8462485000491142,
|
| 23972 |
+
"epoch": 1.80078125,
|
| 23973 |
+
"frac_reward_zero_std": 0.0,
|
| 23974 |
+
"grad_norm": 0.24740563333034515,
|
| 23975 |
+
"learning_rate": 1.5263157894736844e-07,
|
| 23976 |
+
"loss": 0.0,
|
| 23977 |
+
"num_tokens": 4342608.0,
|
| 23978 |
+
"reward": 0.47290000319480896,
|
| 23979 |
+
"reward_std": 0.5461318492889404,
|
| 23980 |
+
"rewards/reward_func/mean": 0.47290000319480896,
|
| 23981 |
+
"rewards/reward_func/std": 0.5058882236480713,
|
| 23982 |
+
"step": 922
|
| 23983 |
+
},
|
| 23984 |
+
{
|
| 23985 |
+
"clip_ratio/high_max": 0.0,
|
| 23986 |
+
"clip_ratio/high_mean": 0.0,
|
| 23987 |
+
"clip_ratio/low_mean": 0.0,
|
| 23988 |
+
"clip_ratio/low_min": 0.0,
|
| 23989 |
+
"clip_ratio/region_mean": 0.0,
|
| 23990 |
+
"completions/clipped_ratio": 1.0,
|
| 23991 |
+
"completions/max_length": 384.0,
|
| 23992 |
+
"completions/max_terminated_length": 0.0,
|
| 23993 |
+
"completions/mean_length": 384.0,
|
| 23994 |
+
"completions/mean_terminated_length": 0.0,
|
| 23995 |
+
"completions/min_length": 384.0,
|
| 23996 |
+
"completions/min_terminated_length": 0.0,
|
| 23997 |
+
"entropy": 0.4742927271872759,
|
| 23998 |
+
"epoch": 1.802734375,
|
| 23999 |
+
"frac_reward_zero_std": 0.0,
|
| 24000 |
+
"grad_norm": 1.2252196073532104,
|
| 24001 |
+
"learning_rate": 1.4736842105263158e-07,
|
| 24002 |
+
"loss": 0.0,
|
| 24003 |
+
"num_tokens": 4347248.0,
|
| 24004 |
+
"reward": 0.3668999969959259,
|
| 24005 |
+
"reward_std": 0.5270397663116455,
|
| 24006 |
+
"rewards/reward_func/mean": 0.3668999969959259,
|
| 24007 |
+
"rewards/reward_func/std": 0.5063701272010803,
|
| 24008 |
+
"step": 923
|
| 24009 |
+
},
|
| 24010 |
+
{
|
| 24011 |
+
"clip_ratio/high_max": 0.0,
|
| 24012 |
+
"clip_ratio/high_mean": 0.0,
|
| 24013 |
+
"clip_ratio/low_mean": 0.0,
|
| 24014 |
+
"clip_ratio/low_min": 0.0,
|
| 24015 |
+
"clip_ratio/region_mean": 0.0,
|
| 24016 |
+
"completions/clipped_ratio": 1.0,
|
| 24017 |
+
"completions/max_length": 384.0,
|
| 24018 |
+
"completions/max_terminated_length": 0.0,
|
| 24019 |
+
"completions/mean_length": 384.0,
|
| 24020 |
+
"completions/mean_terminated_length": 0.0,
|
| 24021 |
+
"completions/min_length": 384.0,
|
| 24022 |
+
"completions/min_terminated_length": 0.0,
|
| 24023 |
+
"entropy": 0.6519879586994648,
|
| 24024 |
+
"epoch": 1.8046875,
|
| 24025 |
+
"frac_reward_zero_std": 0.0,
|
| 24026 |
+
"grad_norm": 0.25893595814704895,
|
| 24027 |
+
"learning_rate": 1.4210526315789474e-07,
|
| 24028 |
+
"loss": -0.0,
|
| 24029 |
+
"num_tokens": 4352264.0,
|
| 24030 |
+
"reward": 0.23256249725818634,
|
| 24031 |
+
"reward_std": 0.4609765410423279,
|
| 24032 |
+
"rewards/reward_func/mean": 0.23256249725818634,
|
| 24033 |
+
"rewards/reward_func/std": 0.4271462559700012,
|
| 24034 |
+
"step": 924
|
| 24035 |
+
},
|
| 24036 |
+
{
|
| 24037 |
+
"clip_ratio/high_max": 0.0,
|
| 24038 |
+
"clip_ratio/high_mean": 0.0,
|
| 24039 |
+
"clip_ratio/low_mean": 0.0,
|
| 24040 |
+
"clip_ratio/low_min": 0.0,
|
| 24041 |
+
"clip_ratio/region_mean": 0.0,
|
| 24042 |
+
"completions/clipped_ratio": 1.0,
|
| 24043 |
+
"completions/max_length": 384.0,
|
| 24044 |
+
"completions/max_terminated_length": 0.0,
|
| 24045 |
+
"completions/mean_length": 384.0,
|
| 24046 |
+
"completions/mean_terminated_length": 0.0,
|
| 24047 |
+
"completions/min_length": 384.0,
|
| 24048 |
+
"completions/min_terminated_length": 0.0,
|
| 24049 |
+
"entropy": 0.8685908392071724,
|
| 24050 |
+
"epoch": 1.806640625,
|
| 24051 |
+
"frac_reward_zero_std": 0.0,
|
| 24052 |
+
"grad_norm": 0.24174730479717255,
|
| 24053 |
+
"learning_rate": 1.368421052631579e-07,
|
| 24054 |
+
"loss": -0.0,
|
| 24055 |
+
"num_tokens": 4356792.0,
|
| 24056 |
+
"reward": 0.6343749761581421,
|
| 24057 |
+
"reward_std": 0.5239908695220947,
|
| 24058 |
+
"rewards/reward_func/mean": 0.6343749761581421,
|
| 24059 |
+
"rewards/reward_func/std": 0.5047431588172913,
|
| 24060 |
+
"step": 925
|
| 24061 |
+
},
|
| 24062 |
+
{
|
| 24063 |
+
"clip_ratio/high_max": 0.0,
|
| 24064 |
+
"clip_ratio/high_mean": 0.0,
|
| 24065 |
+
"clip_ratio/low_mean": 0.0,
|
| 24066 |
+
"clip_ratio/low_min": 0.0,
|
| 24067 |
+
"clip_ratio/region_mean": 0.0,
|
| 24068 |
+
"completions/clipped_ratio": 1.0,
|
| 24069 |
+
"completions/max_length": 384.0,
|
| 24070 |
+
"completions/max_terminated_length": 0.0,
|
| 24071 |
+
"completions/mean_length": 384.0,
|
| 24072 |
+
"completions/mean_terminated_length": 0.0,
|
| 24073 |
+
"completions/min_length": 384.0,
|
| 24074 |
+
"completions/min_terminated_length": 0.0,
|
| 24075 |
+
"entropy": 0.584613062441349,
|
| 24076 |
+
"epoch": 1.80859375,
|
| 24077 |
+
"frac_reward_zero_std": 0.5,
|
| 24078 |
+
"grad_norm": 0.15689092874526978,
|
| 24079 |
+
"learning_rate": 1.3157894736842107e-07,
|
| 24080 |
+
"loss": 0.0,
|
| 24081 |
+
"num_tokens": 4361448.0,
|
| 24082 |
+
"reward": 0.760937511920929,
|
| 24083 |
+
"reward_std": 0.27662280201911926,
|
| 24084 |
+
"rewards/reward_func/mean": 0.760937511920929,
|
| 24085 |
+
"rewards/reward_func/std": 0.4432750344276428,
|
| 24086 |
+
"step": 926
|
| 24087 |
+
},
|
| 24088 |
+
{
|
| 24089 |
+
"clip_ratio/high_max": 0.0,
|
| 24090 |
+
"clip_ratio/high_mean": 0.0,
|
| 24091 |
+
"clip_ratio/low_mean": 0.0,
|
| 24092 |
+
"clip_ratio/low_min": 0.0,
|
| 24093 |
+
"clip_ratio/region_mean": 0.0,
|
| 24094 |
+
"completions/clipped_ratio": 1.0,
|
| 24095 |
+
"completions/max_length": 384.0,
|
| 24096 |
+
"completions/max_terminated_length": 0.0,
|
| 24097 |
+
"completions/mean_length": 384.0,
|
| 24098 |
+
"completions/mean_terminated_length": 0.0,
|
| 24099 |
+
"completions/min_length": 384.0,
|
| 24100 |
+
"completions/min_terminated_length": 0.0,
|
| 24101 |
+
"entropy": 0.6101187616586685,
|
| 24102 |
+
"epoch": 1.810546875,
|
| 24103 |
+
"frac_reward_zero_std": 0.5,
|
| 24104 |
+
"grad_norm": 0.17513932287693024,
|
| 24105 |
+
"learning_rate": 1.2631578947368423e-07,
|
| 24106 |
+
"loss": 0.0,
|
| 24107 |
+
"num_tokens": 4366136.0,
|
| 24108 |
+
"reward": 0.75,
|
| 24109 |
+
"reward_std": 0.28867512941360474,
|
| 24110 |
+
"rewards/reward_func/mean": 0.75,
|
| 24111 |
+
"rewards/reward_func/std": 0.4629100561141968,
|
| 24112 |
+
"step": 927
|
| 24113 |
+
},
|
| 24114 |
+
{
|
| 24115 |
+
"clip_ratio/high_max": 0.0,
|
| 24116 |
+
"clip_ratio/high_mean": 0.0,
|
| 24117 |
+
"clip_ratio/low_mean": 0.0,
|
| 24118 |
+
"clip_ratio/low_min": 0.0,
|
| 24119 |
+
"clip_ratio/region_mean": 0.0,
|
| 24120 |
+
"completions/clipped_ratio": 1.0,
|
| 24121 |
+
"completions/max_length": 384.0,
|
| 24122 |
+
"completions/max_terminated_length": 0.0,
|
| 24123 |
+
"completions/mean_length": 384.0,
|
| 24124 |
+
"completions/mean_terminated_length": 0.0,
|
| 24125 |
+
"completions/min_length": 384.0,
|
| 24126 |
+
"completions/min_terminated_length": 0.0,
|
| 24127 |
+
"entropy": 0.6578428782522678,
|
| 24128 |
+
"epoch": 1.8125,
|
| 24129 |
+
"frac_reward_zero_std": 0.0,
|
| 24130 |
+
"grad_norm": 0.19013863801956177,
|
| 24131 |
+
"learning_rate": 1.2105263157894737e-07,
|
| 24132 |
+
"loss": 0.0,
|
| 24133 |
+
"num_tokens": 4371112.0,
|
| 24134 |
+
"reward": 0.21320000290870667,
|
| 24135 |
+
"reward_std": 0.42640000581741333,
|
| 24136 |
+
"rewards/reward_func/mean": 0.21320000290870667,
|
| 24137 |
+
"rewards/reward_func/std": 0.3952178359031677,
|
| 24138 |
+
"step": 928
|
| 24139 |
+
},
|
| 24140 |
+
{
|
| 24141 |
+
"clip_ratio/high_max": 0.0,
|
| 24142 |
+
"clip_ratio/high_mean": 0.0,
|
| 24143 |
+
"clip_ratio/low_mean": 0.0,
|
| 24144 |
+
"clip_ratio/low_min": 0.0,
|
| 24145 |
+
"clip_ratio/region_mean": 0.0,
|
| 24146 |
+
"completions/clipped_ratio": 1.0,
|
| 24147 |
+
"completions/max_length": 384.0,
|
| 24148 |
+
"completions/max_terminated_length": 0.0,
|
| 24149 |
+
"completions/mean_length": 384.0,
|
| 24150 |
+
"completions/mean_terminated_length": 0.0,
|
| 24151 |
+
"completions/min_length": 384.0,
|
| 24152 |
+
"completions/min_terminated_length": 0.0,
|
| 24153 |
+
"entropy": 0.505165308713913,
|
| 24154 |
+
"epoch": 1.814453125,
|
| 24155 |
+
"frac_reward_zero_std": 0.0,
|
| 24156 |
+
"grad_norm": 0.9573839902877808,
|
| 24157 |
+
"learning_rate": 1.1578947368421054e-07,
|
| 24158 |
+
"loss": -0.0,
|
| 24159 |
+
"num_tokens": 4375760.0,
|
| 24160 |
+
"reward": 0.6168624758720398,
|
| 24161 |
+
"reward_std": 0.5261497497558594,
|
| 24162 |
+
"rewards/reward_func/mean": 0.6168625354766846,
|
| 24163 |
+
"rewards/reward_func/std": 0.5074918866157532,
|
| 24164 |
+
"step": 929
|
| 24165 |
+
},
|
| 24166 |
+
{
|
| 24167 |
+
"clip_ratio/high_max": 0.0,
|
| 24168 |
+
"clip_ratio/high_mean": 0.0,
|
| 24169 |
+
"clip_ratio/low_mean": 0.0,
|
| 24170 |
+
"clip_ratio/low_min": 0.0,
|
| 24171 |
+
"clip_ratio/region_mean": 0.0,
|
| 24172 |
+
"completions/clipped_ratio": 1.0,
|
| 24173 |
+
"completions/max_length": 384.0,
|
| 24174 |
+
"completions/max_terminated_length": 0.0,
|
| 24175 |
+
"completions/mean_length": 384.0,
|
| 24176 |
+
"completions/mean_terminated_length": 0.0,
|
| 24177 |
+
"completions/min_length": 384.0,
|
| 24178 |
+
"completions/min_terminated_length": 0.0,
|
| 24179 |
+
"entropy": 0.874191090464592,
|
| 24180 |
+
"epoch": 1.81640625,
|
| 24181 |
+
"frac_reward_zero_std": 0.0,
|
| 24182 |
+
"grad_norm": 0.2154332995414734,
|
| 24183 |
+
"learning_rate": 1.1052631578947368e-07,
|
| 24184 |
+
"loss": 0.0,
|
| 24185 |
+
"num_tokens": 4380264.0,
|
| 24186 |
+
"reward": 0.6285499930381775,
|
| 24187 |
+
"reward_std": 0.2513039708137512,
|
| 24188 |
+
"rewards/reward_func/mean": 0.6285499930381775,
|
| 24189 |
+
"rewards/reward_func/std": 0.5067988634109497,
|
| 24190 |
+
"step": 930
|
| 24191 |
+
},
|
| 24192 |
+
{
|
| 24193 |
+
"clip_ratio/high_max": 0.0,
|
| 24194 |
+
"clip_ratio/high_mean": 0.0,
|
| 24195 |
+
"clip_ratio/low_mean": 0.0,
|
| 24196 |
+
"clip_ratio/low_min": 0.0,
|
| 24197 |
+
"clip_ratio/region_mean": 0.0,
|
| 24198 |
+
"completions/clipped_ratio": 1.0,
|
| 24199 |
+
"completions/max_length": 384.0,
|
| 24200 |
+
"completions/max_terminated_length": 0.0,
|
| 24201 |
+
"completions/mean_length": 384.0,
|
| 24202 |
+
"completions/mean_terminated_length": 0.0,
|
| 24203 |
+
"completions/min_length": 384.0,
|
| 24204 |
+
"completions/min_terminated_length": 0.0,
|
| 24205 |
+
"entropy": 0.9102950617671013,
|
| 24206 |
+
"epoch": 1.818359375,
|
| 24207 |
+
"frac_reward_zero_std": 0.5,
|
| 24208 |
+
"grad_norm": 0.1660175621509552,
|
| 24209 |
+
"learning_rate": 1.0526315789473685e-07,
|
| 24210 |
+
"loss": 0.0,
|
| 24211 |
+
"num_tokens": 4384872.0,
|
| 24212 |
+
"reward": 0.8723000288009644,
|
| 24213 |
+
"reward_std": 0.24825221300125122,
|
| 24214 |
+
"rewards/reward_func/mean": 0.8723000288009644,
|
| 24215 |
+
"rewards/reward_func/std": 0.3525434732437134,
|
| 24216 |
+
"step": 931
|
| 24217 |
+
},
|
| 24218 |
+
{
|
| 24219 |
+
"clip_ratio/high_max": 0.0,
|
| 24220 |
+
"clip_ratio/high_mean": 0.0,
|
| 24221 |
+
"clip_ratio/low_mean": 0.0,
|
| 24222 |
+
"clip_ratio/low_min": 0.0,
|
| 24223 |
+
"clip_ratio/region_mean": 0.0,
|
| 24224 |
+
"completions/clipped_ratio": 1.0,
|
| 24225 |
+
"completions/max_length": 384.0,
|
| 24226 |
+
"completions/max_terminated_length": 0.0,
|
| 24227 |
+
"completions/mean_length": 384.0,
|
| 24228 |
+
"completions/mean_terminated_length": 0.0,
|
| 24229 |
+
"completions/min_length": 384.0,
|
| 24230 |
+
"completions/min_terminated_length": 0.0,
|
| 24231 |
+
"entropy": 0.564951941370964,
|
| 24232 |
+
"epoch": 1.8203125,
|
| 24233 |
+
"frac_reward_zero_std": 0.0,
|
| 24234 |
+
"grad_norm": 0.23926304280757904,
|
| 24235 |
+
"learning_rate": 1.0000000000000001e-07,
|
| 24236 |
+
"loss": -0.0,
|
| 24237 |
+
"num_tokens": 4389696.0,
|
| 24238 |
+
"reward": 0.1210624948143959,
|
| 24239 |
+
"reward_std": 0.2421249896287918,
|
| 24240 |
+
"rewards/reward_func/mean": 0.1210624948143959,
|
| 24241 |
+
"rewards/reward_func/std": 0.33739402890205383,
|
| 24242 |
+
"step": 932
|
| 24243 |
+
},
|
| 24244 |
+
{
|
| 24245 |
+
"clip_ratio/high_max": 0.0,
|
| 24246 |
+
"clip_ratio/high_mean": 0.0,
|
| 24247 |
+
"clip_ratio/low_mean": 0.0,
|
| 24248 |
+
"clip_ratio/low_min": 0.0,
|
| 24249 |
+
"clip_ratio/region_mean": 0.0,
|
| 24250 |
+
"completions/clipped_ratio": 1.0,
|
| 24251 |
+
"completions/max_length": 384.0,
|
| 24252 |
+
"completions/max_terminated_length": 0.0,
|
| 24253 |
+
"completions/mean_length": 384.0,
|
| 24254 |
+
"completions/mean_terminated_length": 0.0,
|
| 24255 |
+
"completions/min_length": 384.0,
|
| 24256 |
+
"completions/min_terminated_length": 0.0,
|
| 24257 |
+
"entropy": 0.4994359500706196,
|
| 24258 |
+
"epoch": 1.822265625,
|
| 24259 |
+
"frac_reward_zero_std": 0.0,
|
| 24260 |
+
"grad_norm": 0.25481706857681274,
|
| 24261 |
+
"learning_rate": 9.473684210526317e-08,
|
| 24262 |
+
"loss": 0.0,
|
| 24263 |
+
"num_tokens": 4394352.0,
|
| 24264 |
+
"reward": 0.629212498664856,
|
| 24265 |
+
"reward_std": 0.49713975191116333,
|
| 24266 |
+
"rewards/reward_func/mean": 0.629212498664856,
|
| 24267 |
+
"rewards/reward_func/std": 0.47718602418899536,
|
| 24268 |
+
"step": 933
|
| 24269 |
+
},
|
| 24270 |
+
{
|
| 24271 |
+
"clip_ratio/high_max": 0.0,
|
| 24272 |
+
"clip_ratio/high_mean": 0.0,
|
| 24273 |
+
"clip_ratio/low_mean": 0.0,
|
| 24274 |
+
"clip_ratio/low_min": 0.0,
|
| 24275 |
+
"clip_ratio/region_mean": 0.0,
|
| 24276 |
+
"completions/clipped_ratio": 1.0,
|
| 24277 |
+
"completions/max_length": 384.0,
|
| 24278 |
+
"completions/max_terminated_length": 0.0,
|
| 24279 |
+
"completions/mean_length": 384.0,
|
| 24280 |
+
"completions/mean_terminated_length": 0.0,
|
| 24281 |
+
"completions/min_length": 384.0,
|
| 24282 |
+
"completions/min_terminated_length": 0.0,
|
| 24283 |
+
"entropy": 0.4711499195545912,
|
| 24284 |
+
"epoch": 1.82421875,
|
| 24285 |
+
"frac_reward_zero_std": 0.0,
|
| 24286 |
+
"grad_norm": 0.19426177442073822,
|
| 24287 |
+
"learning_rate": 8.947368421052632e-08,
|
| 24288 |
+
"loss": -0.0,
|
| 24289 |
+
"num_tokens": 4399000.0,
|
| 24290 |
+
"reward": 0.629349946975708,
|
| 24291 |
+
"reward_std": 0.5124804973602295,
|
| 24292 |
+
"rewards/reward_func/mean": 0.6293500065803528,
|
| 24293 |
+
"rewards/reward_func/std": 0.4940544664859772,
|
| 24294 |
+
"step": 934
|
| 24295 |
+
},
|
| 24296 |
+
{
|
| 24297 |
+
"clip_ratio/high_max": 0.0,
|
| 24298 |
+
"clip_ratio/high_mean": 0.0,
|
| 24299 |
+
"clip_ratio/low_mean": 0.0,
|
| 24300 |
+
"clip_ratio/low_min": 0.0,
|
| 24301 |
+
"clip_ratio/region_mean": 0.0,
|
| 24302 |
+
"completions/clipped_ratio": 1.0,
|
| 24303 |
+
"completions/max_length": 384.0,
|
| 24304 |
+
"completions/max_terminated_length": 0.0,
|
| 24305 |
+
"completions/mean_length": 384.0,
|
| 24306 |
+
"completions/mean_terminated_length": 0.0,
|
| 24307 |
+
"completions/min_length": 384.0,
|
| 24308 |
+
"completions/min_terminated_length": 0.0,
|
| 24309 |
+
"entropy": 0.8083347305655479,
|
| 24310 |
+
"epoch": 1.826171875,
|
| 24311 |
+
"frac_reward_zero_std": 0.5,
|
| 24312 |
+
"grad_norm": 0.17153169214725494,
|
| 24313 |
+
"learning_rate": 8.421052631578947e-08,
|
| 24314 |
+
"loss": 0.0,
|
| 24315 |
+
"num_tokens": 4403608.0,
|
| 24316 |
+
"reward": 0.7593749761581421,
|
| 24317 |
+
"reward_std": 0.27827125787734985,
|
| 24318 |
+
"rewards/reward_func/mean": 0.7593749761581421,
|
| 24319 |
+
"rewards/reward_func/std": 0.4460016191005707,
|
| 24320 |
+
"step": 935
|
| 24321 |
+
},
|
| 24322 |
+
{
|
| 24323 |
+
"clip_ratio/high_max": 0.0,
|
| 24324 |
+
"clip_ratio/high_mean": 0.0,
|
| 24325 |
+
"clip_ratio/low_mean": 0.0,
|
| 24326 |
+
"clip_ratio/low_min": 0.0,
|
| 24327 |
+
"clip_ratio/region_mean": 0.0,
|
| 24328 |
+
"completions/clipped_ratio": 1.0,
|
| 24329 |
+
"completions/max_length": 384.0,
|
| 24330 |
+
"completions/max_terminated_length": 0.0,
|
| 24331 |
+
"completions/mean_length": 384.0,
|
| 24332 |
+
"completions/mean_terminated_length": 0.0,
|
| 24333 |
+
"completions/min_length": 384.0,
|
| 24334 |
+
"completions/min_terminated_length": 0.0,
|
| 24335 |
+
"entropy": 0.6882787756621838,
|
| 24336 |
+
"epoch": 1.828125,
|
| 24337 |
+
"frac_reward_zero_std": 0.0,
|
| 24338 |
+
"grad_norm": 0.26333948969841003,
|
| 24339 |
+
"learning_rate": 7.894736842105264e-08,
|
| 24340 |
+
"loss": 0.0,
|
| 24341 |
+
"num_tokens": 4408400.0,
|
| 24342 |
+
"reward": 0.5950000286102295,
|
| 24343 |
+
"reward_std": 0.5117709040641785,
|
| 24344 |
+
"rewards/reward_func/mean": 0.5950000286102295,
|
| 24345 |
+
"rewards/reward_func/std": 0.4931241571903229,
|
| 24346 |
+
"step": 936
|
| 24347 |
+
},
|
| 24348 |
+
{
|
| 24349 |
+
"clip_ratio/high_max": 0.0,
|
| 24350 |
+
"clip_ratio/high_mean": 0.0,
|
| 24351 |
+
"clip_ratio/low_mean": 0.0,
|
| 24352 |
+
"clip_ratio/low_min": 0.0,
|
| 24353 |
+
"clip_ratio/region_mean": 0.0,
|
| 24354 |
+
"completions/clipped_ratio": 1.0,
|
| 24355 |
+
"completions/max_length": 384.0,
|
| 24356 |
+
"completions/max_terminated_length": 0.0,
|
| 24357 |
+
"completions/mean_length": 384.0,
|
| 24358 |
+
"completions/mean_terminated_length": 0.0,
|
| 24359 |
+
"completions/min_length": 384.0,
|
| 24360 |
+
"completions/min_terminated_length": 0.0,
|
| 24361 |
+
"entropy": 0.8129791170358658,
|
| 24362 |
+
"epoch": 1.830078125,
|
| 24363 |
+
"frac_reward_zero_std": 1.0,
|
| 24364 |
+
"grad_norm": 0.0,
|
| 24365 |
+
"learning_rate": 7.368421052631579e-08,
|
| 24366 |
+
"loss": 0.0,
|
| 24367 |
+
"num_tokens": 4412896.0,
|
| 24368 |
+
"reward": 1.0,
|
| 24369 |
+
"reward_std": 0.0,
|
| 24370 |
+
"rewards/reward_func/mean": 1.0,
|
| 24371 |
+
"rewards/reward_func/std": 0.0,
|
| 24372 |
+
"step": 937
|
| 24373 |
+
},
|
| 24374 |
+
{
|
| 24375 |
+
"clip_ratio/high_max": 0.0,
|
| 24376 |
+
"clip_ratio/high_mean": 0.0,
|
| 24377 |
+
"clip_ratio/low_mean": 0.0,
|
| 24378 |
+
"clip_ratio/low_min": 0.0,
|
| 24379 |
+
"clip_ratio/region_mean": 0.0,
|
| 24380 |
+
"completions/clipped_ratio": 1.0,
|
| 24381 |
+
"completions/max_length": 384.0,
|
| 24382 |
+
"completions/max_terminated_length": 0.0,
|
| 24383 |
+
"completions/mean_length": 384.0,
|
| 24384 |
+
"completions/mean_terminated_length": 0.0,
|
| 24385 |
+
"completions/min_length": 384.0,
|
| 24386 |
+
"completions/min_terminated_length": 0.0,
|
| 24387 |
+
"entropy": 0.45138827711343765,
|
| 24388 |
+
"epoch": 1.83203125,
|
| 24389 |
+
"frac_reward_zero_std": 0.0,
|
| 24390 |
+
"grad_norm": 0.18060848116874695,
|
| 24391 |
+
"learning_rate": 6.842105263157895e-08,
|
| 24392 |
+
"loss": 0.0,
|
| 24393 |
+
"num_tokens": 4417680.0,
|
| 24394 |
+
"reward": 0.3540624976158142,
|
| 24395 |
+
"reward_std": 0.5053315162658691,
|
| 24396 |
+
"rewards/reward_func/mean": 0.3540624976158142,
|
| 24397 |
+
"rewards/reward_func/std": 0.4852207601070404,
|
| 24398 |
+
"step": 938
|
| 24399 |
+
},
|
| 24400 |
+
{
|
| 24401 |
+
"clip_ratio/high_max": 0.0,
|
| 24402 |
+
"clip_ratio/high_mean": 0.0,
|
| 24403 |
+
"clip_ratio/low_mean": 0.0,
|
| 24404 |
+
"clip_ratio/low_min": 0.0,
|
| 24405 |
+
"clip_ratio/region_mean": 0.0,
|
| 24406 |
+
"completions/clipped_ratio": 1.0,
|
| 24407 |
+
"completions/max_length": 384.0,
|
| 24408 |
+
"completions/max_terminated_length": 0.0,
|
| 24409 |
+
"completions/mean_length": 384.0,
|
| 24410 |
+
"completions/mean_terminated_length": 0.0,
|
| 24411 |
+
"completions/min_length": 384.0,
|
| 24412 |
+
"completions/min_terminated_length": 0.0,
|
| 24413 |
+
"entropy": 0.6235154792666435,
|
| 24414 |
+
"epoch": 1.833984375,
|
| 24415 |
+
"frac_reward_zero_std": 0.0,
|
| 24416 |
+
"grad_norm": 0.16618505120277405,
|
| 24417 |
+
"learning_rate": 6.315789473684211e-08,
|
| 24418 |
+
"loss": -0.0,
|
| 24419 |
+
"num_tokens": 4422320.0,
|
| 24420 |
+
"reward": 0.7527874708175659,
|
| 24421 |
+
"reward_std": 0.4729631543159485,
|
| 24422 |
+
"rewards/reward_func/mean": 0.7527874708175659,
|
| 24423 |
+
"rewards/reward_func/std": 0.4385221302509308,
|
| 24424 |
+
"step": 939
|
| 24425 |
+
},
|
| 24426 |
+
{
|
| 24427 |
+
"clip_ratio/high_max": 0.0,
|
| 24428 |
+
"clip_ratio/high_mean": 0.0,
|
| 24429 |
+
"clip_ratio/low_mean": 0.0,
|
| 24430 |
+
"clip_ratio/low_min": 0.0,
|
| 24431 |
+
"clip_ratio/region_mean": 0.0,
|
| 24432 |
+
"completions/clipped_ratio": 1.0,
|
| 24433 |
+
"completions/max_length": 384.0,
|
| 24434 |
+
"completions/max_terminated_length": 0.0,
|
| 24435 |
+
"completions/mean_length": 384.0,
|
| 24436 |
+
"completions/mean_terminated_length": 0.0,
|
| 24437 |
+
"completions/min_length": 384.0,
|
| 24438 |
+
"completions/min_terminated_length": 0.0,
|
| 24439 |
+
"entropy": 0.32566132955253124,
|
| 24440 |
+
"epoch": 1.8359375,
|
| 24441 |
+
"frac_reward_zero_std": 0.0,
|
| 24442 |
+
"grad_norm": 0.1724817156791687,
|
| 24443 |
+
"learning_rate": 5.789473684210527e-08,
|
| 24444 |
+
"loss": 0.0,
|
| 24445 |
+
"num_tokens": 4426960.0,
|
| 24446 |
+
"reward": 0.38749998807907104,
|
| 24447 |
+
"reward_std": 0.5304675102233887,
|
| 24448 |
+
"rewards/reward_func/mean": 0.38750001788139343,
|
| 24449 |
+
"rewards/reward_func/std": 0.5074885487556458,
|
| 24450 |
+
"step": 940
|
| 24451 |
+
},
|
| 24452 |
+
{
|
| 24453 |
+
"clip_ratio/high_max": 0.0,
|
| 24454 |
+
"clip_ratio/high_mean": 0.0,
|
| 24455 |
+
"clip_ratio/low_mean": 0.0,
|
| 24456 |
+
"clip_ratio/low_min": 0.0,
|
| 24457 |
+
"clip_ratio/region_mean": 0.0,
|
| 24458 |
+
"completions/clipped_ratio": 1.0,
|
| 24459 |
+
"completions/max_length": 384.0,
|
| 24460 |
+
"completions/max_terminated_length": 0.0,
|
| 24461 |
+
"completions/mean_length": 384.0,
|
| 24462 |
+
"completions/mean_terminated_length": 0.0,
|
| 24463 |
+
"completions/min_length": 384.0,
|
| 24464 |
+
"completions/min_terminated_length": 0.0,
|
| 24465 |
+
"entropy": 0.3569133039563894,
|
| 24466 |
+
"epoch": 1.837890625,
|
| 24467 |
+
"frac_reward_zero_std": 0.0,
|
| 24468 |
+
"grad_norm": 0.17535947263240814,
|
| 24469 |
+
"learning_rate": 5.263157894736842e-08,
|
| 24470 |
+
"loss": 0.0,
|
| 24471 |
+
"num_tokens": 4431448.0,
|
| 24472 |
+
"reward": 0.6987625360488892,
|
| 24473 |
+
"reward_std": 0.4616749882698059,
|
| 24474 |
+
"rewards/reward_func/mean": 0.6987625360488892,
|
| 24475 |
+
"rewards/reward_func/std": 0.4274410605430603,
|
| 24476 |
+
"step": 941
|
| 24477 |
+
},
|
| 24478 |
+
{
|
| 24479 |
+
"clip_ratio/high_max": 0.0,
|
| 24480 |
+
"clip_ratio/high_mean": 0.0,
|
| 24481 |
+
"clip_ratio/low_mean": 0.0,
|
| 24482 |
+
"clip_ratio/low_min": 0.0,
|
| 24483 |
+
"clip_ratio/region_mean": 0.0,
|
| 24484 |
+
"completions/clipped_ratio": 1.0,
|
| 24485 |
+
"completions/max_length": 384.0,
|
| 24486 |
+
"completions/max_terminated_length": 0.0,
|
| 24487 |
+
"completions/mean_length": 384.0,
|
| 24488 |
+
"completions/mean_terminated_length": 0.0,
|
| 24489 |
+
"completions/min_length": 384.0,
|
| 24490 |
+
"completions/min_terminated_length": 0.0,
|
| 24491 |
+
"entropy": 0.5731943100690842,
|
| 24492 |
+
"epoch": 1.83984375,
|
| 24493 |
+
"frac_reward_zero_std": 0.0,
|
| 24494 |
+
"grad_norm": 0.19585168361663818,
|
| 24495 |
+
"learning_rate": 4.7368421052631586e-08,
|
| 24496 |
+
"loss": 0.0,
|
| 24497 |
+
"num_tokens": 4436104.0,
|
| 24498 |
+
"reward": 0.3721874952316284,
|
| 24499 |
+
"reward_std": 0.5249817371368408,
|
| 24500 |
+
"rewards/reward_func/mean": 0.3721874952316284,
|
| 24501 |
+
"rewards/reward_func/std": 0.5038166046142578,
|
| 24502 |
+
"step": 942
|
| 24503 |
+
},
|
| 24504 |
+
{
|
| 24505 |
+
"clip_ratio/high_max": 0.0,
|
| 24506 |
+
"clip_ratio/high_mean": 0.0,
|
| 24507 |
+
"clip_ratio/low_mean": 0.0,
|
| 24508 |
+
"clip_ratio/low_min": 0.0,
|
| 24509 |
+
"clip_ratio/region_mean": 0.0,
|
| 24510 |
+
"completions/clipped_ratio": 1.0,
|
| 24511 |
+
"completions/max_length": 384.0,
|
| 24512 |
+
"completions/max_terminated_length": 0.0,
|
| 24513 |
+
"completions/mean_length": 384.0,
|
| 24514 |
+
"completions/mean_terminated_length": 0.0,
|
| 24515 |
+
"completions/min_length": 384.0,
|
| 24516 |
+
"completions/min_terminated_length": 0.0,
|
| 24517 |
+
"entropy": 0.5804979503154755,
|
| 24518 |
+
"epoch": 1.841796875,
|
| 24519 |
+
"frac_reward_zero_std": 0.5,
|
| 24520 |
+
"grad_norm": 0.15701980888843536,
|
| 24521 |
+
"learning_rate": 4.2105263157894737e-08,
|
| 24522 |
+
"loss": 0.0,
|
| 24523 |
+
"num_tokens": 4441096.0,
|
| 24524 |
+
"reward": 0.22850000858306885,
|
| 24525 |
+
"reward_std": 0.26406246423721313,
|
| 24526 |
+
"rewards/reward_func/mean": 0.22850000858306885,
|
| 24527 |
+
"rewards/reward_func/std": 0.4233279824256897,
|
| 24528 |
+
"step": 943
|
| 24529 |
+
},
|
| 24530 |
+
{
|
| 24531 |
+
"clip_ratio/high_max": 0.0,
|
| 24532 |
+
"clip_ratio/high_mean": 0.0,
|
| 24533 |
+
"clip_ratio/low_mean": 0.0,
|
| 24534 |
+
"clip_ratio/low_min": 0.0,
|
| 24535 |
+
"clip_ratio/region_mean": 0.0,
|
| 24536 |
+
"completions/clipped_ratio": 1.0,
|
| 24537 |
+
"completions/max_length": 384.0,
|
| 24538 |
+
"completions/max_terminated_length": 0.0,
|
| 24539 |
+
"completions/mean_length": 384.0,
|
| 24540 |
+
"completions/mean_terminated_length": 0.0,
|
| 24541 |
+
"completions/min_length": 384.0,
|
| 24542 |
+
"completions/min_terminated_length": 0.0,
|
| 24543 |
+
"entropy": 0.8702232874929905,
|
| 24544 |
+
"epoch": 1.84375,
|
| 24545 |
+
"frac_reward_zero_std": 0.0,
|
| 24546 |
+
"grad_norm": 0.30385154485702515,
|
| 24547 |
+
"learning_rate": 3.6842105263157894e-08,
|
| 24548 |
+
"loss": 0.0,
|
| 24549 |
+
"num_tokens": 4445896.0,
|
| 24550 |
+
"reward": 0.5129749774932861,
|
| 24551 |
+
"reward_std": 0.47286224365234375,
|
| 24552 |
+
"rewards/reward_func/mean": 0.5129749774932861,
|
| 24553 |
+
"rewards/reward_func/std": 0.5019313097000122,
|
| 24554 |
+
"step": 944
|
| 24555 |
+
},
|
| 24556 |
+
{
|
| 24557 |
+
"clip_ratio/high_max": 0.0,
|
| 24558 |
+
"clip_ratio/high_mean": 0.0,
|
| 24559 |
+
"clip_ratio/low_mean": 0.0,
|
| 24560 |
+
"clip_ratio/low_min": 0.0,
|
| 24561 |
+
"clip_ratio/region_mean": 0.0,
|
| 24562 |
+
"completions/clipped_ratio": 1.0,
|
| 24563 |
+
"completions/max_length": 384.0,
|
| 24564 |
+
"completions/max_terminated_length": 0.0,
|
| 24565 |
+
"completions/mean_length": 384.0,
|
| 24566 |
+
"completions/mean_terminated_length": 0.0,
|
| 24567 |
+
"completions/min_length": 384.0,
|
| 24568 |
+
"completions/min_terminated_length": 0.0,
|
| 24569 |
+
"entropy": 0.6899589747190475,
|
| 24570 |
+
"epoch": 1.845703125,
|
| 24571 |
+
"frac_reward_zero_std": 0.0,
|
| 24572 |
+
"grad_norm": 0.23425976932048798,
|
| 24573 |
+
"learning_rate": 3.157894736842106e-08,
|
| 24574 |
+
"loss": 0.0,
|
| 24575 |
+
"num_tokens": 4450384.0,
|
| 24576 |
+
"reward": 0.8695999979972839,
|
| 24577 |
+
"reward_std": 0.25623539090156555,
|
| 24578 |
+
"rewards/reward_func/mean": 0.8695999979972839,
|
| 24579 |
+
"rewards/reward_func/std": 0.35150691866874695,
|
| 24580 |
+
"step": 945
|
| 24581 |
+
},
|
| 24582 |
+
{
|
| 24583 |
+
"clip_ratio/high_max": 0.0,
|
| 24584 |
+
"clip_ratio/high_mean": 0.0,
|
| 24585 |
+
"clip_ratio/low_mean": 0.0,
|
| 24586 |
+
"clip_ratio/low_min": 0.0,
|
| 24587 |
+
"clip_ratio/region_mean": 0.0,
|
| 24588 |
+
"completions/clipped_ratio": 1.0,
|
| 24589 |
+
"completions/max_length": 384.0,
|
| 24590 |
+
"completions/max_terminated_length": 0.0,
|
| 24591 |
+
"completions/mean_length": 384.0,
|
| 24592 |
+
"completions/mean_terminated_length": 0.0,
|
| 24593 |
+
"completions/min_length": 384.0,
|
| 24594 |
+
"completions/min_terminated_length": 0.0,
|
| 24595 |
+
"entropy": 0.7242371030151844,
|
| 24596 |
+
"epoch": 1.84765625,
|
| 24597 |
+
"frac_reward_zero_std": 0.5,
|
| 24598 |
+
"grad_norm": 0.18696576356887817,
|
| 24599 |
+
"learning_rate": 2.631578947368421e-08,
|
| 24600 |
+
"loss": 0.0,
|
| 24601 |
+
"num_tokens": 4454728.0,
|
| 24602 |
+
"reward": 0.8765624761581421,
|
| 24603 |
+
"reward_std": 0.24687500298023224,
|
| 24604 |
+
"rewards/reward_func/mean": 0.8765624761581421,
|
| 24605 |
+
"rewards/reward_func/std": 0.34913399815559387,
|
| 24606 |
+
"step": 946
|
| 24607 |
+
},
|
| 24608 |
+
{
|
| 24609 |
+
"clip_ratio/high_max": 0.0,
|
| 24610 |
+
"clip_ratio/high_mean": 0.0,
|
| 24611 |
+
"clip_ratio/low_mean": 0.0,
|
| 24612 |
+
"clip_ratio/low_min": 0.0,
|
| 24613 |
+
"clip_ratio/region_mean": 0.0,
|
| 24614 |
+
"completions/clipped_ratio": 1.0,
|
| 24615 |
+
"completions/max_length": 384.0,
|
| 24616 |
+
"completions/max_terminated_length": 0.0,
|
| 24617 |
+
"completions/mean_length": 384.0,
|
| 24618 |
+
"completions/mean_terminated_length": 0.0,
|
| 24619 |
+
"completions/min_length": 384.0,
|
| 24620 |
+
"completions/min_terminated_length": 0.0,
|
| 24621 |
+
"entropy": 0.6600115075707436,
|
| 24622 |
+
"epoch": 1.849609375,
|
| 24623 |
+
"frac_reward_zero_std": 0.0,
|
| 24624 |
+
"grad_norm": 0.26843783259391785,
|
| 24625 |
+
"learning_rate": 2.1052631578947368e-08,
|
| 24626 |
+
"loss": -0.0,
|
| 24627 |
+
"num_tokens": 4459744.0,
|
| 24628 |
+
"reward": 0.350600004196167,
|
| 24629 |
+
"reward_std": 0.4967568516731262,
|
| 24630 |
+
"rewards/reward_func/mean": 0.350600004196167,
|
| 24631 |
+
"rewards/reward_func/std": 0.4848525822162628,
|
| 24632 |
+
"step": 947
|
| 24633 |
+
},
|
| 24634 |
+
{
|
| 24635 |
+
"clip_ratio/high_max": 0.0,
|
| 24636 |
+
"clip_ratio/high_mean": 0.0,
|
| 24637 |
+
"clip_ratio/low_mean": 0.0,
|
| 24638 |
+
"clip_ratio/low_min": 0.0,
|
| 24639 |
+
"clip_ratio/region_mean": 0.0,
|
| 24640 |
+
"completions/clipped_ratio": 1.0,
|
| 24641 |
+
"completions/max_length": 384.0,
|
| 24642 |
+
"completions/max_terminated_length": 0.0,
|
| 24643 |
+
"completions/mean_length": 384.0,
|
| 24644 |
+
"completions/mean_terminated_length": 0.0,
|
| 24645 |
+
"completions/min_length": 384.0,
|
| 24646 |
+
"completions/min_terminated_length": 0.0,
|
| 24647 |
+
"entropy": 0.6009881878271699,
|
| 24648 |
+
"epoch": 1.8515625,
|
| 24649 |
+
"frac_reward_zero_std": 0.0,
|
| 24650 |
+
"grad_norm": 0.2861350476741791,
|
| 24651 |
+
"learning_rate": 1.578947368421053e-08,
|
| 24652 |
+
"loss": -0.0,
|
| 24653 |
+
"num_tokens": 4464248.0,
|
| 24654 |
+
"reward": 0.5809999704360962,
|
| 24655 |
+
"reward_std": 0.5007524490356445,
|
| 24656 |
+
"rewards/reward_func/mean": 0.5809999704360962,
|
| 24657 |
+
"rewards/reward_func/std": 0.4811137020587921,
|
| 24658 |
+
"step": 948
|
| 24659 |
+
},
|
| 24660 |
+
{
|
| 24661 |
+
"clip_ratio/high_max": 0.0,
|
| 24662 |
+
"clip_ratio/high_mean": 0.0,
|
| 24663 |
+
"clip_ratio/low_mean": 0.0,
|
| 24664 |
+
"clip_ratio/low_min": 0.0,
|
| 24665 |
+
"clip_ratio/region_mean": 0.0,
|
| 24666 |
+
"completions/clipped_ratio": 1.0,
|
| 24667 |
+
"completions/max_length": 384.0,
|
| 24668 |
+
"completions/max_terminated_length": 0.0,
|
| 24669 |
+
"completions/mean_length": 384.0,
|
| 24670 |
+
"completions/mean_terminated_length": 0.0,
|
| 24671 |
+
"completions/min_length": 384.0,
|
| 24672 |
+
"completions/min_terminated_length": 0.0,
|
| 24673 |
+
"entropy": 0.4016810180619359,
|
| 24674 |
+
"epoch": 1.853515625,
|
| 24675 |
+
"frac_reward_zero_std": 0.0,
|
| 24676 |
+
"grad_norm": 0.19072452187538147,
|
| 24677 |
+
"learning_rate": 1.0526315789473684e-08,
|
| 24678 |
+
"loss": -0.0,
|
| 24679 |
+
"num_tokens": 4468744.0,
|
| 24680 |
+
"reward": 0.5809999704360962,
|
| 24681 |
+
"reward_std": 0.5007524490356445,
|
| 24682 |
+
"rewards/reward_func/mean": 0.5809999704360962,
|
| 24683 |
+
"rewards/reward_func/std": 0.4811137020587921,
|
| 24684 |
+
"step": 949
|
| 24685 |
+
},
|
| 24686 |
+
{
|
| 24687 |
+
"clip_ratio/high_max": 0.0,
|
| 24688 |
+
"clip_ratio/high_mean": 0.0,
|
| 24689 |
+
"clip_ratio/low_mean": 0.0,
|
| 24690 |
+
"clip_ratio/low_min": 0.0,
|
| 24691 |
+
"clip_ratio/region_mean": 0.0,
|
| 24692 |
+
"completions/clipped_ratio": 1.0,
|
| 24693 |
+
"completions/max_length": 384.0,
|
| 24694 |
+
"completions/max_terminated_length": 0.0,
|
| 24695 |
+
"completions/mean_length": 384.0,
|
| 24696 |
+
"completions/mean_terminated_length": 0.0,
|
| 24697 |
+
"completions/min_length": 384.0,
|
| 24698 |
+
"completions/min_terminated_length": 0.0,
|
| 24699 |
+
"entropy": 0.92277492582798,
|
| 24700 |
+
"epoch": 1.85546875,
|
| 24701 |
+
"frac_reward_zero_std": 0.0,
|
| 24702 |
+
"grad_norm": 0.2709886431694031,
|
| 24703 |
+
"learning_rate": 5.263157894736842e-09,
|
| 24704 |
+
"loss": 0.0,
|
| 24705 |
+
"num_tokens": 4473248.0,
|
| 24706 |
+
"reward": 0.7473000288009644,
|
| 24707 |
+
"reward_std": 0.4982522130012512,
|
| 24708 |
+
"rewards/reward_func/mean": 0.7473000288009644,
|
| 24709 |
+
"rewards/reward_func/std": 0.4613037705421448,
|
| 24710 |
+
"step": 950
|
| 24711 |
}
|
| 24712 |
],
|
| 24713 |
"logging_steps": 1,
|
| 24714 |
"max_steps": 950,
|
| 24715 |
+
"num_input_tokens_seen": 4473248,
|
| 24716 |
"num_train_epochs": 2,
|
| 24717 |
"save_steps": 50,
|
| 24718 |
"stateful_callbacks": {
|
|
|
|
| 24722 |
"should_evaluate": false,
|
| 24723 |
"should_log": false,
|
| 24724 |
"should_save": true,
|
| 24725 |
+
"should_training_stop": true
|
| 24726 |
},
|
| 24727 |
"attributes": {}
|
| 24728 |
}
|