Update latest checkpoint for run 15940d1d-7d8c-4253-8810-2ea934bedee4 at step 150
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +1303 -3
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:b8d1d745dbe9439f29c04be8e2d651aaaaa39e28a377c0e0b5a773b86448a5b0
|
| 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 sha256:23ac62c931646c94c2b749bc90e9b82412e233b7ab3b22e10b31fba1b728704e
|
| 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 sha256:2c1d24a994c544cab55d39dcfd8e6eb16877b24e8fa992685c2dee0a2f5233a4
|
| 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 sha256:74a4817f6fde031e692aed008288960e65627dc1479b7074cf4c268d6cc68155
|
| 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": 0.
|
| 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,
|
|
@@ -2608,11 +2608,1311 @@
|
|
| 2608 |
"rewards/reward_func/mean": 0.625,
|
| 2609 |
"rewards/reward_func/std": 0.5175492167472839,
|
| 2610 |
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2611 |
}
|
| 2612 |
],
|
| 2613 |
"logging_steps": 1,
|
| 2614 |
"max_steps": 950,
|
| 2615 |
-
"num_input_tokens_seen":
|
| 2616 |
"num_train_epochs": 2,
|
| 2617 |
"save_steps": 50,
|
| 2618 |
"stateful_callbacks": {
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 0.29296875,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 150,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 2608 |
"rewards/reward_func/mean": 0.625,
|
| 2609 |
"rewards/reward_func/std": 0.5175492167472839,
|
| 2610 |
"step": 100
|
| 2611 |
+
},
|
| 2612 |
+
{
|
| 2613 |
+
"clip_ratio/high_max": 0.0,
|
| 2614 |
+
"clip_ratio/high_mean": 0.0,
|
| 2615 |
+
"clip_ratio/low_mean": 0.0,
|
| 2616 |
+
"clip_ratio/low_min": 0.0,
|
| 2617 |
+
"clip_ratio/region_mean": 0.0,
|
| 2618 |
+
"completions/clipped_ratio": 1.0,
|
| 2619 |
+
"completions/max_length": 384.0,
|
| 2620 |
+
"completions/max_terminated_length": 0.0,
|
| 2621 |
+
"completions/mean_length": 384.0,
|
| 2622 |
+
"completions/mean_terminated_length": 0.0,
|
| 2623 |
+
"completions/min_length": 384.0,
|
| 2624 |
+
"completions/min_terminated_length": 0.0,
|
| 2625 |
+
"entropy": 0.5260781534016132,
|
| 2626 |
+
"epoch": 0.197265625,
|
| 2627 |
+
"frac_reward_zero_std": 0.5,
|
| 2628 |
+
"grad_norm": 0.15311233699321747,
|
| 2629 |
+
"learning_rate": 4.473684210526316e-06,
|
| 2630 |
+
"loss": -0.0,
|
| 2631 |
+
"num_tokens": 472096.0,
|
| 2632 |
+
"reward": 0.0015625000232830644,
|
| 2633 |
+
"reward_std": 0.0031250000465661287,
|
| 2634 |
+
"rewards/reward_func/mean": 0.0015625000232830644,
|
| 2635 |
+
"rewards/reward_func/std": 0.0044194175861775875,
|
| 2636 |
+
"step": 101
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"clip_ratio/high_max": 0.0,
|
| 2640 |
+
"clip_ratio/high_mean": 0.0,
|
| 2641 |
+
"clip_ratio/low_mean": 0.0,
|
| 2642 |
+
"clip_ratio/low_min": 0.0,
|
| 2643 |
+
"clip_ratio/region_mean": 0.0,
|
| 2644 |
+
"completions/clipped_ratio": 1.0,
|
| 2645 |
+
"completions/max_length": 384.0,
|
| 2646 |
+
"completions/max_terminated_length": 0.0,
|
| 2647 |
+
"completions/mean_length": 384.0,
|
| 2648 |
+
"completions/mean_terminated_length": 0.0,
|
| 2649 |
+
"completions/min_length": 384.0,
|
| 2650 |
+
"completions/min_terminated_length": 0.0,
|
| 2651 |
+
"entropy": 0.7099028006196022,
|
| 2652 |
+
"epoch": 0.19921875,
|
| 2653 |
+
"frac_reward_zero_std": 0.0,
|
| 2654 |
+
"grad_norm": 0.23158502578735352,
|
| 2655 |
+
"learning_rate": 4.468421052631579e-06,
|
| 2656 |
+
"loss": 0.0,
|
| 2657 |
+
"num_tokens": 476832.0,
|
| 2658 |
+
"reward": 0.5016124844551086,
|
| 2659 |
+
"reward_std": 0.5615566968917847,
|
| 2660 |
+
"rewards/reward_func/mean": 0.5016124844551086,
|
| 2661 |
+
"rewards/reward_func/std": 0.5200324058532715,
|
| 2662 |
+
"step": 102
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"clip_ratio/high_max": 0.0,
|
| 2666 |
+
"clip_ratio/high_mean": 0.0,
|
| 2667 |
+
"clip_ratio/low_mean": 0.0,
|
| 2668 |
+
"clip_ratio/low_min": 0.0,
|
| 2669 |
+
"clip_ratio/region_mean": 0.0,
|
| 2670 |
+
"completions/clipped_ratio": 1.0,
|
| 2671 |
+
"completions/max_length": 384.0,
|
| 2672 |
+
"completions/max_terminated_length": 0.0,
|
| 2673 |
+
"completions/mean_length": 384.0,
|
| 2674 |
+
"completions/mean_terminated_length": 0.0,
|
| 2675 |
+
"completions/min_length": 384.0,
|
| 2676 |
+
"completions/min_terminated_length": 0.0,
|
| 2677 |
+
"entropy": 0.6568261608481407,
|
| 2678 |
+
"epoch": 0.201171875,
|
| 2679 |
+
"frac_reward_zero_std": 0.5,
|
| 2680 |
+
"grad_norm": 0.1539800614118576,
|
| 2681 |
+
"learning_rate": 4.463157894736842e-06,
|
| 2682 |
+
"loss": 0.0,
|
| 2683 |
+
"num_tokens": 481592.0,
|
| 2684 |
+
"reward": 0.8300000429153442,
|
| 2685 |
+
"reward_std": 0.24041630327701569,
|
| 2686 |
+
"rewards/reward_func/mean": 0.8300000429153442,
|
| 2687 |
+
"rewards/reward_func/std": 0.33602720499038696,
|
| 2688 |
+
"step": 103
|
| 2689 |
+
},
|
| 2690 |
+
{
|
| 2691 |
+
"clip_ratio/high_max": 0.0,
|
| 2692 |
+
"clip_ratio/high_mean": 0.0,
|
| 2693 |
+
"clip_ratio/low_mean": 0.0,
|
| 2694 |
+
"clip_ratio/low_min": 0.0,
|
| 2695 |
+
"clip_ratio/region_mean": 0.0,
|
| 2696 |
+
"completions/clipped_ratio": 1.0,
|
| 2697 |
+
"completions/max_length": 384.0,
|
| 2698 |
+
"completions/max_terminated_length": 0.0,
|
| 2699 |
+
"completions/mean_length": 384.0,
|
| 2700 |
+
"completions/mean_terminated_length": 0.0,
|
| 2701 |
+
"completions/min_length": 384.0,
|
| 2702 |
+
"completions/min_terminated_length": 0.0,
|
| 2703 |
+
"entropy": 0.48168842773884535,
|
| 2704 |
+
"epoch": 0.203125,
|
| 2705 |
+
"frac_reward_zero_std": 0.0,
|
| 2706 |
+
"grad_norm": 0.19704312086105347,
|
| 2707 |
+
"learning_rate": 4.4578947368421054e-06,
|
| 2708 |
+
"loss": -0.0,
|
| 2709 |
+
"num_tokens": 486440.0,
|
| 2710 |
+
"reward": 0.3668999969959259,
|
| 2711 |
+
"reward_std": 0.5270397663116455,
|
| 2712 |
+
"rewards/reward_func/mean": 0.3668999969959259,
|
| 2713 |
+
"rewards/reward_func/std": 0.5063701272010803,
|
| 2714 |
+
"step": 104
|
| 2715 |
+
},
|
| 2716 |
+
{
|
| 2717 |
+
"clip_ratio/high_max": 0.0,
|
| 2718 |
+
"clip_ratio/high_mean": 0.0,
|
| 2719 |
+
"clip_ratio/low_mean": 0.0,
|
| 2720 |
+
"clip_ratio/low_min": 0.0,
|
| 2721 |
+
"clip_ratio/region_mean": 0.0,
|
| 2722 |
+
"completions/clipped_ratio": 1.0,
|
| 2723 |
+
"completions/max_length": 384.0,
|
| 2724 |
+
"completions/max_terminated_length": 0.0,
|
| 2725 |
+
"completions/mean_length": 384.0,
|
| 2726 |
+
"completions/mean_terminated_length": 0.0,
|
| 2727 |
+
"completions/min_length": 384.0,
|
| 2728 |
+
"completions/min_terminated_length": 0.0,
|
| 2729 |
+
"entropy": 0.4500679522752762,
|
| 2730 |
+
"epoch": 0.205078125,
|
| 2731 |
+
"frac_reward_zero_std": 0.0,
|
| 2732 |
+
"grad_norm": 0.15499116480350494,
|
| 2733 |
+
"learning_rate": 4.452631578947369e-06,
|
| 2734 |
+
"loss": -0.0,
|
| 2735 |
+
"num_tokens": 491224.0,
|
| 2736 |
+
"reward": 0.12262499332427979,
|
| 2737 |
+
"reward_std": 0.24110156297683716,
|
| 2738 |
+
"rewards/reward_func/mean": 0.12262499332427979,
|
| 2739 |
+
"rewards/reward_func/std": 0.33678168058395386,
|
| 2740 |
+
"step": 105
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"clip_ratio/high_max": 0.0,
|
| 2744 |
+
"clip_ratio/high_mean": 0.0,
|
| 2745 |
+
"clip_ratio/low_mean": 0.0,
|
| 2746 |
+
"clip_ratio/low_min": 0.0,
|
| 2747 |
+
"clip_ratio/region_mean": 0.0,
|
| 2748 |
+
"completions/clipped_ratio": 1.0,
|
| 2749 |
+
"completions/max_length": 384.0,
|
| 2750 |
+
"completions/max_terminated_length": 0.0,
|
| 2751 |
+
"completions/mean_length": 384.0,
|
| 2752 |
+
"completions/mean_terminated_length": 0.0,
|
| 2753 |
+
"completions/min_length": 384.0,
|
| 2754 |
+
"completions/min_terminated_length": 0.0,
|
| 2755 |
+
"entropy": 0.6883682459592819,
|
| 2756 |
+
"epoch": 0.20703125,
|
| 2757 |
+
"frac_reward_zero_std": 0.5,
|
| 2758 |
+
"grad_norm": 0.12739039957523346,
|
| 2759 |
+
"learning_rate": 4.447368421052632e-06,
|
| 2760 |
+
"loss": 0.0,
|
| 2761 |
+
"num_tokens": 495864.0,
|
| 2762 |
+
"reward": 0.25,
|
| 2763 |
+
"reward_std": 0.28867512941360474,
|
| 2764 |
+
"rewards/reward_func/mean": 0.25,
|
| 2765 |
+
"rewards/reward_func/std": 0.4629100561141968,
|
| 2766 |
+
"step": 106
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"clip_ratio/high_max": 0.0,
|
| 2770 |
+
"clip_ratio/high_mean": 0.0,
|
| 2771 |
+
"clip_ratio/low_mean": 0.0,
|
| 2772 |
+
"clip_ratio/low_min": 0.0,
|
| 2773 |
+
"clip_ratio/region_mean": 0.0,
|
| 2774 |
+
"completions/clipped_ratio": 1.0,
|
| 2775 |
+
"completions/max_length": 384.0,
|
| 2776 |
+
"completions/max_terminated_length": 0.0,
|
| 2777 |
+
"completions/mean_length": 384.0,
|
| 2778 |
+
"completions/mean_terminated_length": 0.0,
|
| 2779 |
+
"completions/min_length": 384.0,
|
| 2780 |
+
"completions/min_terminated_length": 0.0,
|
| 2781 |
+
"entropy": 0.5138243455439806,
|
| 2782 |
+
"epoch": 0.208984375,
|
| 2783 |
+
"frac_reward_zero_std": 0.0,
|
| 2784 |
+
"grad_norm": 0.20175091922283173,
|
| 2785 |
+
"learning_rate": 4.442105263157896e-06,
|
| 2786 |
+
"loss": -0.0,
|
| 2787 |
+
"num_tokens": 500720.0,
|
| 2788 |
+
"reward": 0.4923250079154968,
|
| 2789 |
+
"reward_std": 0.5613177418708801,
|
| 2790 |
+
"rewards/reward_func/mean": 0.49232497811317444,
|
| 2791 |
+
"rewards/reward_func/std": 0.5197004675865173,
|
| 2792 |
+
"step": 107
|
| 2793 |
+
},
|
| 2794 |
+
{
|
| 2795 |
+
"clip_ratio/high_max": 0.0,
|
| 2796 |
+
"clip_ratio/high_mean": 0.0,
|
| 2797 |
+
"clip_ratio/low_mean": 0.0,
|
| 2798 |
+
"clip_ratio/low_min": 0.0,
|
| 2799 |
+
"clip_ratio/region_mean": 0.0,
|
| 2800 |
+
"completions/clipped_ratio": 1.0,
|
| 2801 |
+
"completions/max_length": 384.0,
|
| 2802 |
+
"completions/max_terminated_length": 0.0,
|
| 2803 |
+
"completions/mean_length": 384.0,
|
| 2804 |
+
"completions/mean_terminated_length": 0.0,
|
| 2805 |
+
"completions/min_length": 384.0,
|
| 2806 |
+
"completions/min_terminated_length": 0.0,
|
| 2807 |
+
"entropy": 0.7696716003119946,
|
| 2808 |
+
"epoch": 0.2109375,
|
| 2809 |
+
"frac_reward_zero_std": 0.5,
|
| 2810 |
+
"grad_norm": 0.19667677581310272,
|
| 2811 |
+
"learning_rate": 4.436842105263158e-06,
|
| 2812 |
+
"loss": 0.0,
|
| 2813 |
+
"num_tokens": 505320.0,
|
| 2814 |
+
"reward": 0.875,
|
| 2815 |
+
"reward_std": 0.25,
|
| 2816 |
+
"rewards/reward_func/mean": 0.875,
|
| 2817 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 2818 |
+
"step": 108
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"clip_ratio/high_max": 0.0,
|
| 2822 |
+
"clip_ratio/high_mean": 0.0,
|
| 2823 |
+
"clip_ratio/low_mean": 0.0,
|
| 2824 |
+
"clip_ratio/low_min": 0.0,
|
| 2825 |
+
"clip_ratio/region_mean": 0.0,
|
| 2826 |
+
"completions/clipped_ratio": 1.0,
|
| 2827 |
+
"completions/max_length": 384.0,
|
| 2828 |
+
"completions/max_terminated_length": 0.0,
|
| 2829 |
+
"completions/mean_length": 384.0,
|
| 2830 |
+
"completions/mean_terminated_length": 0.0,
|
| 2831 |
+
"completions/min_length": 384.0,
|
| 2832 |
+
"completions/min_terminated_length": 0.0,
|
| 2833 |
+
"entropy": 0.6834167242050171,
|
| 2834 |
+
"epoch": 0.212890625,
|
| 2835 |
+
"frac_reward_zero_std": 0.5,
|
| 2836 |
+
"grad_norm": 0.19136440753936768,
|
| 2837 |
+
"learning_rate": 4.431578947368421e-06,
|
| 2838 |
+
"loss": 0.0,
|
| 2839 |
+
"num_tokens": 509792.0,
|
| 2840 |
+
"reward": 0.875,
|
| 2841 |
+
"reward_std": 0.25,
|
| 2842 |
+
"rewards/reward_func/mean": 0.875,
|
| 2843 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 2844 |
+
"step": 109
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"clip_ratio/high_max": 0.0,
|
| 2848 |
+
"clip_ratio/high_mean": 0.0,
|
| 2849 |
+
"clip_ratio/low_mean": 0.0,
|
| 2850 |
+
"clip_ratio/low_min": 0.0,
|
| 2851 |
+
"clip_ratio/region_mean": 0.0,
|
| 2852 |
+
"completions/clipped_ratio": 1.0,
|
| 2853 |
+
"completions/max_length": 384.0,
|
| 2854 |
+
"completions/max_terminated_length": 0.0,
|
| 2855 |
+
"completions/mean_length": 384.0,
|
| 2856 |
+
"completions/mean_terminated_length": 0.0,
|
| 2857 |
+
"completions/min_length": 384.0,
|
| 2858 |
+
"completions/min_terminated_length": 0.0,
|
| 2859 |
+
"entropy": 0.5321553461253643,
|
| 2860 |
+
"epoch": 0.21484375,
|
| 2861 |
+
"frac_reward_zero_std": 0.0,
|
| 2862 |
+
"grad_norm": 0.2594543993473053,
|
| 2863 |
+
"learning_rate": 4.426315789473684e-06,
|
| 2864 |
+
"loss": -0.0,
|
| 2865 |
+
"num_tokens": 514392.0,
|
| 2866 |
+
"reward": 0.629687488079071,
|
| 2867 |
+
"reward_std": 0.5333658456802368,
|
| 2868 |
+
"rewards/reward_func/mean": 0.629687488079071,
|
| 2869 |
+
"rewards/reward_func/std": 0.5112107992172241,
|
| 2870 |
+
"step": 110
|
| 2871 |
+
},
|
| 2872 |
+
{
|
| 2873 |
+
"clip_ratio/high_max": 0.0,
|
| 2874 |
+
"clip_ratio/high_mean": 0.0,
|
| 2875 |
+
"clip_ratio/low_mean": 0.0,
|
| 2876 |
+
"clip_ratio/low_min": 0.0,
|
| 2877 |
+
"clip_ratio/region_mean": 0.0,
|
| 2878 |
+
"completions/clipped_ratio": 1.0,
|
| 2879 |
+
"completions/max_length": 384.0,
|
| 2880 |
+
"completions/max_terminated_length": 0.0,
|
| 2881 |
+
"completions/mean_length": 384.0,
|
| 2882 |
+
"completions/mean_terminated_length": 0.0,
|
| 2883 |
+
"completions/min_length": 384.0,
|
| 2884 |
+
"completions/min_terminated_length": 0.0,
|
| 2885 |
+
"entropy": 0.6744099333882332,
|
| 2886 |
+
"epoch": 0.216796875,
|
| 2887 |
+
"frac_reward_zero_std": 0.0,
|
| 2888 |
+
"grad_norm": 0.20642893016338348,
|
| 2889 |
+
"learning_rate": 4.4210526315789476e-06,
|
| 2890 |
+
"loss": -0.0,
|
| 2891 |
+
"num_tokens": 519056.0,
|
| 2892 |
+
"reward": 0.375,
|
| 2893 |
+
"reward_std": 0.5386751294136047,
|
| 2894 |
+
"rewards/reward_func/mean": 0.375,
|
| 2895 |
+
"rewards/reward_func/std": 0.5175492167472839,
|
| 2896 |
+
"step": 111
|
| 2897 |
+
},
|
| 2898 |
+
{
|
| 2899 |
+
"clip_ratio/high_max": 0.0,
|
| 2900 |
+
"clip_ratio/high_mean": 0.0,
|
| 2901 |
+
"clip_ratio/low_mean": 0.0,
|
| 2902 |
+
"clip_ratio/low_min": 0.0,
|
| 2903 |
+
"clip_ratio/region_mean": 0.0,
|
| 2904 |
+
"completions/clipped_ratio": 1.0,
|
| 2905 |
+
"completions/max_length": 384.0,
|
| 2906 |
+
"completions/max_terminated_length": 0.0,
|
| 2907 |
+
"completions/mean_length": 384.0,
|
| 2908 |
+
"completions/mean_terminated_length": 0.0,
|
| 2909 |
+
"completions/min_length": 384.0,
|
| 2910 |
+
"completions/min_terminated_length": 0.0,
|
| 2911 |
+
"entropy": 0.6416698023676872,
|
| 2912 |
+
"epoch": 0.21875,
|
| 2913 |
+
"frac_reward_zero_std": 0.0,
|
| 2914 |
+
"grad_norm": 0.22189435362815857,
|
| 2915 |
+
"learning_rate": 4.415789473684211e-06,
|
| 2916 |
+
"loss": 0.0,
|
| 2917 |
+
"num_tokens": 524064.0,
|
| 2918 |
+
"reward": 0.23100000619888306,
|
| 2919 |
+
"reward_std": 0.4619999825954437,
|
| 2920 |
+
"rewards/reward_func/mean": 0.23100000619888306,
|
| 2921 |
+
"rewards/reward_func/std": 0.42808806896209717,
|
| 2922 |
+
"step": 112
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"clip_ratio/high_max": 0.0,
|
| 2926 |
+
"clip_ratio/high_mean": 0.0,
|
| 2927 |
+
"clip_ratio/low_mean": 0.0,
|
| 2928 |
+
"clip_ratio/low_min": 0.0,
|
| 2929 |
+
"clip_ratio/region_mean": 0.0,
|
| 2930 |
+
"completions/clipped_ratio": 1.0,
|
| 2931 |
+
"completions/max_length": 384.0,
|
| 2932 |
+
"completions/max_terminated_length": 0.0,
|
| 2933 |
+
"completions/mean_length": 384.0,
|
| 2934 |
+
"completions/mean_terminated_length": 0.0,
|
| 2935 |
+
"completions/min_length": 384.0,
|
| 2936 |
+
"completions/min_terminated_length": 0.0,
|
| 2937 |
+
"entropy": 0.6506899148225784,
|
| 2938 |
+
"epoch": 0.220703125,
|
| 2939 |
+
"frac_reward_zero_std": 0.0,
|
| 2940 |
+
"grad_norm": 0.24659857153892517,
|
| 2941 |
+
"learning_rate": 4.410526315789474e-06,
|
| 2942 |
+
"loss": 0.0,
|
| 2943 |
+
"num_tokens": 529056.0,
|
| 2944 |
+
"reward": 0.3370000123977661,
|
| 2945 |
+
"reward_std": 0.483335018157959,
|
| 2946 |
+
"rewards/reward_func/mean": 0.3370000123977661,
|
| 2947 |
+
"rewards/reward_func/std": 0.46515026688575745,
|
| 2948 |
+
"step": 113
|
| 2949 |
+
},
|
| 2950 |
+
{
|
| 2951 |
+
"clip_ratio/high_max": 0.0,
|
| 2952 |
+
"clip_ratio/high_mean": 0.0,
|
| 2953 |
+
"clip_ratio/low_mean": 0.0,
|
| 2954 |
+
"clip_ratio/low_min": 0.0,
|
| 2955 |
+
"clip_ratio/region_mean": 0.0,
|
| 2956 |
+
"completions/clipped_ratio": 1.0,
|
| 2957 |
+
"completions/max_length": 384.0,
|
| 2958 |
+
"completions/max_terminated_length": 0.0,
|
| 2959 |
+
"completions/mean_length": 384.0,
|
| 2960 |
+
"completions/mean_terminated_length": 0.0,
|
| 2961 |
+
"completions/min_length": 384.0,
|
| 2962 |
+
"completions/min_terminated_length": 0.0,
|
| 2963 |
+
"entropy": 0.6117045246064663,
|
| 2964 |
+
"epoch": 0.22265625,
|
| 2965 |
+
"frac_reward_zero_std": 0.5,
|
| 2966 |
+
"grad_norm": 0.15469233691692352,
|
| 2967 |
+
"learning_rate": 4.405263157894737e-06,
|
| 2968 |
+
"loss": 0.0,
|
| 2969 |
+
"num_tokens": 533664.0,
|
| 2970 |
+
"reward": 0.875,
|
| 2971 |
+
"reward_std": 0.25,
|
| 2972 |
+
"rewards/reward_func/mean": 0.875,
|
| 2973 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 2974 |
+
"step": 114
|
| 2975 |
+
},
|
| 2976 |
+
{
|
| 2977 |
+
"clip_ratio/high_max": 0.0,
|
| 2978 |
+
"clip_ratio/high_mean": 0.0,
|
| 2979 |
+
"clip_ratio/low_mean": 0.0,
|
| 2980 |
+
"clip_ratio/low_min": 0.0,
|
| 2981 |
+
"clip_ratio/region_mean": 0.0,
|
| 2982 |
+
"completions/clipped_ratio": 1.0,
|
| 2983 |
+
"completions/max_length": 384.0,
|
| 2984 |
+
"completions/max_terminated_length": 0.0,
|
| 2985 |
+
"completions/mean_length": 384.0,
|
| 2986 |
+
"completions/mean_terminated_length": 0.0,
|
| 2987 |
+
"completions/min_length": 384.0,
|
| 2988 |
+
"completions/min_terminated_length": 0.0,
|
| 2989 |
+
"entropy": 0.681044191122055,
|
| 2990 |
+
"epoch": 0.224609375,
|
| 2991 |
+
"frac_reward_zero_std": 0.5,
|
| 2992 |
+
"grad_norm": 0.1914876252412796,
|
| 2993 |
+
"learning_rate": 4.4e-06,
|
| 2994 |
+
"loss": 0.0,
|
| 2995 |
+
"num_tokens": 538104.0,
|
| 2996 |
+
"reward": 0.875,
|
| 2997 |
+
"reward_std": 0.25,
|
| 2998 |
+
"rewards/reward_func/mean": 0.875,
|
| 2999 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 3000 |
+
"step": 115
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"clip_ratio/high_max": 0.0,
|
| 3004 |
+
"clip_ratio/high_mean": 0.0,
|
| 3005 |
+
"clip_ratio/low_mean": 0.0,
|
| 3006 |
+
"clip_ratio/low_min": 0.0,
|
| 3007 |
+
"clip_ratio/region_mean": 0.0,
|
| 3008 |
+
"completions/clipped_ratio": 1.0,
|
| 3009 |
+
"completions/max_length": 384.0,
|
| 3010 |
+
"completions/max_terminated_length": 0.0,
|
| 3011 |
+
"completions/mean_length": 384.0,
|
| 3012 |
+
"completions/mean_terminated_length": 0.0,
|
| 3013 |
+
"completions/min_length": 384.0,
|
| 3014 |
+
"completions/min_terminated_length": 0.0,
|
| 3015 |
+
"entropy": 0.6017662808299065,
|
| 3016 |
+
"epoch": 0.2265625,
|
| 3017 |
+
"frac_reward_zero_std": 0.0,
|
| 3018 |
+
"grad_norm": 0.2209053784608841,
|
| 3019 |
+
"learning_rate": 4.394736842105263e-06,
|
| 3020 |
+
"loss": 0.0,
|
| 3021 |
+
"num_tokens": 543024.0,
|
| 3022 |
+
"reward": 0.4899500012397766,
|
| 3023 |
+
"reward_std": 0.5516552925109863,
|
| 3024 |
+
"rewards/reward_func/mean": 0.4899500012397766,
|
| 3025 |
+
"rewards/reward_func/std": 0.5108809471130371,
|
| 3026 |
+
"step": 116
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"clip_ratio/high_max": 0.0,
|
| 3030 |
+
"clip_ratio/high_mean": 0.0,
|
| 3031 |
+
"clip_ratio/low_mean": 0.0,
|
| 3032 |
+
"clip_ratio/low_min": 0.0,
|
| 3033 |
+
"clip_ratio/region_mean": 0.0,
|
| 3034 |
+
"completions/clipped_ratio": 1.0,
|
| 3035 |
+
"completions/max_length": 384.0,
|
| 3036 |
+
"completions/max_terminated_length": 0.0,
|
| 3037 |
+
"completions/mean_length": 384.0,
|
| 3038 |
+
"completions/mean_terminated_length": 0.0,
|
| 3039 |
+
"completions/min_length": 384.0,
|
| 3040 |
+
"completions/min_terminated_length": 0.0,
|
| 3041 |
+
"entropy": 0.6683652102947235,
|
| 3042 |
+
"epoch": 0.228515625,
|
| 3043 |
+
"frac_reward_zero_std": 0.0,
|
| 3044 |
+
"grad_norm": 0.19945447146892548,
|
| 3045 |
+
"learning_rate": 4.3894736842105266e-06,
|
| 3046 |
+
"loss": -0.0,
|
| 3047 |
+
"num_tokens": 547656.0,
|
| 3048 |
+
"reward": 0.3683124780654907,
|
| 3049 |
+
"reward_std": 0.5250316858291626,
|
| 3050 |
+
"rewards/reward_func/mean": 0.3683124780654907,
|
| 3051 |
+
"rewards/reward_func/std": 0.5048869252204895,
|
| 3052 |
+
"step": 117
|
| 3053 |
+
},
|
| 3054 |
+
{
|
| 3055 |
+
"clip_ratio/high_max": 0.0,
|
| 3056 |
+
"clip_ratio/high_mean": 0.0,
|
| 3057 |
+
"clip_ratio/low_mean": 0.0,
|
| 3058 |
+
"clip_ratio/low_min": 0.0,
|
| 3059 |
+
"clip_ratio/region_mean": 0.0,
|
| 3060 |
+
"completions/clipped_ratio": 1.0,
|
| 3061 |
+
"completions/max_length": 384.0,
|
| 3062 |
+
"completions/max_terminated_length": 0.0,
|
| 3063 |
+
"completions/mean_length": 384.0,
|
| 3064 |
+
"completions/mean_terminated_length": 0.0,
|
| 3065 |
+
"completions/min_length": 384.0,
|
| 3066 |
+
"completions/min_terminated_length": 0.0,
|
| 3067 |
+
"entropy": 0.3754726406186819,
|
| 3068 |
+
"epoch": 0.23046875,
|
| 3069 |
+
"frac_reward_zero_std": 0.0,
|
| 3070 |
+
"grad_norm": 0.19966642558574677,
|
| 3071 |
+
"learning_rate": 4.38421052631579e-06,
|
| 3072 |
+
"loss": 0.0,
|
| 3073 |
+
"num_tokens": 552160.0,
|
| 3074 |
+
"reward": 0.47360000014305115,
|
| 3075 |
+
"reward_std": 0.47124379873275757,
|
| 3076 |
+
"rewards/reward_func/mean": 0.47360000014305115,
|
| 3077 |
+
"rewards/reward_func/std": 0.5068238377571106,
|
| 3078 |
+
"step": 118
|
| 3079 |
+
},
|
| 3080 |
+
{
|
| 3081 |
+
"clip_ratio/high_max": 0.0,
|
| 3082 |
+
"clip_ratio/high_mean": 0.0,
|
| 3083 |
+
"clip_ratio/low_mean": 0.0,
|
| 3084 |
+
"clip_ratio/low_min": 0.0,
|
| 3085 |
+
"clip_ratio/region_mean": 0.0,
|
| 3086 |
+
"completions/clipped_ratio": 1.0,
|
| 3087 |
+
"completions/max_length": 384.0,
|
| 3088 |
+
"completions/max_terminated_length": 0.0,
|
| 3089 |
+
"completions/mean_length": 384.0,
|
| 3090 |
+
"completions/mean_terminated_length": 0.0,
|
| 3091 |
+
"completions/min_length": 384.0,
|
| 3092 |
+
"completions/min_terminated_length": 0.0,
|
| 3093 |
+
"entropy": 0.4965004436671734,
|
| 3094 |
+
"epoch": 0.232421875,
|
| 3095 |
+
"frac_reward_zero_std": 0.5,
|
| 3096 |
+
"grad_norm": 0.1419457346200943,
|
| 3097 |
+
"learning_rate": 4.378947368421053e-06,
|
| 3098 |
+
"loss": 0.0,
|
| 3099 |
+
"num_tokens": 557144.0,
|
| 3100 |
+
"reward": 0.11100000143051147,
|
| 3101 |
+
"reward_std": 0.22200000286102295,
|
| 3102 |
+
"rewards/reward_func/mean": 0.11100000143051147,
|
| 3103 |
+
"rewards/reward_func/std": 0.3139553964138031,
|
| 3104 |
+
"step": 119
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"clip_ratio/high_max": 0.0,
|
| 3108 |
+
"clip_ratio/high_mean": 0.0,
|
| 3109 |
+
"clip_ratio/low_mean": 0.0,
|
| 3110 |
+
"clip_ratio/low_min": 0.0,
|
| 3111 |
+
"clip_ratio/region_mean": 0.0,
|
| 3112 |
+
"completions/clipped_ratio": 1.0,
|
| 3113 |
+
"completions/max_length": 384.0,
|
| 3114 |
+
"completions/max_terminated_length": 0.0,
|
| 3115 |
+
"completions/mean_length": 384.0,
|
| 3116 |
+
"completions/mean_terminated_length": 0.0,
|
| 3117 |
+
"completions/min_length": 384.0,
|
| 3118 |
+
"completions/min_terminated_length": 0.0,
|
| 3119 |
+
"entropy": 0.509646050632,
|
| 3120 |
+
"epoch": 0.234375,
|
| 3121 |
+
"frac_reward_zero_std": 0.0,
|
| 3122 |
+
"grad_norm": 0.18533200025558472,
|
| 3123 |
+
"learning_rate": 4.373684210526316e-06,
|
| 3124 |
+
"loss": -0.0,
|
| 3125 |
+
"num_tokens": 561792.0,
|
| 3126 |
+
"reward": 0.37892499566078186,
|
| 3127 |
+
"reward_std": 0.5219879150390625,
|
| 3128 |
+
"rewards/reward_func/mean": 0.37892499566078186,
|
| 3129 |
+
"rewards/reward_func/std": 0.5029488205909729,
|
| 3130 |
+
"step": 120
|
| 3131 |
+
},
|
| 3132 |
+
{
|
| 3133 |
+
"clip_ratio/high_max": 0.0,
|
| 3134 |
+
"clip_ratio/high_mean": 0.0,
|
| 3135 |
+
"clip_ratio/low_mean": 0.0,
|
| 3136 |
+
"clip_ratio/low_min": 0.0,
|
| 3137 |
+
"clip_ratio/region_mean": 0.0,
|
| 3138 |
+
"completions/clipped_ratio": 1.0,
|
| 3139 |
+
"completions/max_length": 384.0,
|
| 3140 |
+
"completions/max_terminated_length": 0.0,
|
| 3141 |
+
"completions/mean_length": 384.0,
|
| 3142 |
+
"completions/mean_terminated_length": 0.0,
|
| 3143 |
+
"completions/min_length": 384.0,
|
| 3144 |
+
"completions/min_terminated_length": 0.0,
|
| 3145 |
+
"entropy": 0.6187039539217949,
|
| 3146 |
+
"epoch": 0.236328125,
|
| 3147 |
+
"frac_reward_zero_std": 0.5,
|
| 3148 |
+
"grad_norm": 0.14618447422981262,
|
| 3149 |
+
"learning_rate": 4.368421052631579e-06,
|
| 3150 |
+
"loss": 0.0,
|
| 3151 |
+
"num_tokens": 566528.0,
|
| 3152 |
+
"reward": 0.9474999904632568,
|
| 3153 |
+
"reward_std": 0.01500000525265932,
|
| 3154 |
+
"rewards/reward_func/mean": 0.9474999904632568,
|
| 3155 |
+
"rewards/reward_func/std": 0.02121320553123951,
|
| 3156 |
+
"step": 121
|
| 3157 |
+
},
|
| 3158 |
+
{
|
| 3159 |
+
"clip_ratio/high_max": 0.0,
|
| 3160 |
+
"clip_ratio/high_mean": 0.0,
|
| 3161 |
+
"clip_ratio/low_mean": 0.0,
|
| 3162 |
+
"clip_ratio/low_min": 0.0,
|
| 3163 |
+
"clip_ratio/region_mean": 0.0,
|
| 3164 |
+
"completions/clipped_ratio": 1.0,
|
| 3165 |
+
"completions/max_length": 384.0,
|
| 3166 |
+
"completions/max_terminated_length": 0.0,
|
| 3167 |
+
"completions/mean_length": 384.0,
|
| 3168 |
+
"completions/mean_terminated_length": 0.0,
|
| 3169 |
+
"completions/min_length": 384.0,
|
| 3170 |
+
"completions/min_terminated_length": 0.0,
|
| 3171 |
+
"entropy": 0.4034488648176193,
|
| 3172 |
+
"epoch": 0.23828125,
|
| 3173 |
+
"frac_reward_zero_std": 0.0,
|
| 3174 |
+
"grad_norm": 0.17633076012134552,
|
| 3175 |
+
"learning_rate": 4.363157894736842e-06,
|
| 3176 |
+
"loss": 0.0,
|
| 3177 |
+
"num_tokens": 571704.0,
|
| 3178 |
+
"reward": 0.7593749761581421,
|
| 3179 |
+
"reward_std": 0.48124998807907104,
|
| 3180 |
+
"rewards/reward_func/mean": 0.7593749761581421,
|
| 3181 |
+
"rewards/reward_func/std": 0.4460016191005707,
|
| 3182 |
+
"step": 122
|
| 3183 |
+
},
|
| 3184 |
+
{
|
| 3185 |
+
"clip_ratio/high_max": 0.0,
|
| 3186 |
+
"clip_ratio/high_mean": 0.0,
|
| 3187 |
+
"clip_ratio/low_mean": 0.0,
|
| 3188 |
+
"clip_ratio/low_min": 0.0,
|
| 3189 |
+
"clip_ratio/region_mean": 0.0,
|
| 3190 |
+
"completions/clipped_ratio": 1.0,
|
| 3191 |
+
"completions/max_length": 384.0,
|
| 3192 |
+
"completions/max_terminated_length": 0.0,
|
| 3193 |
+
"completions/mean_length": 384.0,
|
| 3194 |
+
"completions/mean_terminated_length": 0.0,
|
| 3195 |
+
"completions/min_length": 384.0,
|
| 3196 |
+
"completions/min_terminated_length": 0.0,
|
| 3197 |
+
"entropy": 0.6987418737262487,
|
| 3198 |
+
"epoch": 0.240234375,
|
| 3199 |
+
"frac_reward_zero_std": 0.0,
|
| 3200 |
+
"grad_norm": 0.26086556911468506,
|
| 3201 |
+
"learning_rate": 4.3578947368421055e-06,
|
| 3202 |
+
"loss": -0.0,
|
| 3203 |
+
"num_tokens": 576560.0,
|
| 3204 |
+
"reward": 0.6266999840736389,
|
| 3205 |
+
"reward_std": 0.5093674659729004,
|
| 3206 |
+
"rewards/reward_func/mean": 0.6266999840736389,
|
| 3207 |
+
"rewards/reward_func/std": 0.4916507303714752,
|
| 3208 |
+
"step": 123
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"clip_ratio/high_max": 0.0,
|
| 3212 |
+
"clip_ratio/high_mean": 0.0,
|
| 3213 |
+
"clip_ratio/low_mean": 0.0,
|
| 3214 |
+
"clip_ratio/low_min": 0.0,
|
| 3215 |
+
"clip_ratio/region_mean": 0.0,
|
| 3216 |
+
"completions/clipped_ratio": 1.0,
|
| 3217 |
+
"completions/max_length": 384.0,
|
| 3218 |
+
"completions/max_terminated_length": 0.0,
|
| 3219 |
+
"completions/mean_length": 384.0,
|
| 3220 |
+
"completions/mean_terminated_length": 0.0,
|
| 3221 |
+
"completions/min_length": 384.0,
|
| 3222 |
+
"completions/min_terminated_length": 0.0,
|
| 3223 |
+
"entropy": 0.7009507194161415,
|
| 3224 |
+
"epoch": 0.2421875,
|
| 3225 |
+
"frac_reward_zero_std": 0.0,
|
| 3226 |
+
"grad_norm": 0.2221718728542328,
|
| 3227 |
+
"learning_rate": 4.352631578947369e-06,
|
| 3228 |
+
"loss": -0.0,
|
| 3229 |
+
"num_tokens": 581488.0,
|
| 3230 |
+
"reward": 0.49502497911453247,
|
| 3231 |
+
"reward_std": 0.5644694566726685,
|
| 3232 |
+
"rewards/reward_func/mean": 0.49502497911453247,
|
| 3233 |
+
"rewards/reward_func/std": 0.5226343870162964,
|
| 3234 |
+
"step": 124
|
| 3235 |
+
},
|
| 3236 |
+
{
|
| 3237 |
+
"clip_ratio/high_max": 0.0,
|
| 3238 |
+
"clip_ratio/high_mean": 0.0,
|
| 3239 |
+
"clip_ratio/low_mean": 0.0,
|
| 3240 |
+
"clip_ratio/low_min": 0.0,
|
| 3241 |
+
"clip_ratio/region_mean": 0.0,
|
| 3242 |
+
"completions/clipped_ratio": 1.0,
|
| 3243 |
+
"completions/max_length": 384.0,
|
| 3244 |
+
"completions/max_terminated_length": 0.0,
|
| 3245 |
+
"completions/mean_length": 384.0,
|
| 3246 |
+
"completions/mean_terminated_length": 0.0,
|
| 3247 |
+
"completions/min_length": 384.0,
|
| 3248 |
+
"completions/min_terminated_length": 0.0,
|
| 3249 |
+
"entropy": 0.3089421261101961,
|
| 3250 |
+
"epoch": 0.244140625,
|
| 3251 |
+
"frac_reward_zero_std": 0.0,
|
| 3252 |
+
"grad_norm": 0.16174942255020142,
|
| 3253 |
+
"learning_rate": 4.347368421052632e-06,
|
| 3254 |
+
"loss": 0.0,
|
| 3255 |
+
"num_tokens": 585992.0,
|
| 3256 |
+
"reward": 0.6930000185966492,
|
| 3257 |
+
"reward_std": 0.4620679020881653,
|
| 3258 |
+
"rewards/reward_func/mean": 0.6930000185966492,
|
| 3259 |
+
"rewards/reward_func/std": 0.4277917146682739,
|
| 3260 |
+
"step": 125
|
| 3261 |
+
},
|
| 3262 |
+
{
|
| 3263 |
+
"clip_ratio/high_max": 0.0,
|
| 3264 |
+
"clip_ratio/high_mean": 0.0,
|
| 3265 |
+
"clip_ratio/low_mean": 0.0,
|
| 3266 |
+
"clip_ratio/low_min": 0.0,
|
| 3267 |
+
"clip_ratio/region_mean": 0.0,
|
| 3268 |
+
"completions/clipped_ratio": 1.0,
|
| 3269 |
+
"completions/max_length": 384.0,
|
| 3270 |
+
"completions/max_terminated_length": 0.0,
|
| 3271 |
+
"completions/mean_length": 384.0,
|
| 3272 |
+
"completions/mean_terminated_length": 0.0,
|
| 3273 |
+
"completions/min_length": 384.0,
|
| 3274 |
+
"completions/min_terminated_length": 0.0,
|
| 3275 |
+
"entropy": 0.6322880983352661,
|
| 3276 |
+
"epoch": 0.24609375,
|
| 3277 |
+
"frac_reward_zero_std": 0.0,
|
| 3278 |
+
"grad_norm": 0.19183886051177979,
|
| 3279 |
+
"learning_rate": 4.342105263157895e-06,
|
| 3280 |
+
"loss": -0.0,
|
| 3281 |
+
"num_tokens": 590960.0,
|
| 3282 |
+
"reward": 0.2251249998807907,
|
| 3283 |
+
"reward_std": 0.25995194911956787,
|
| 3284 |
+
"rewards/reward_func/mean": 0.2251250147819519,
|
| 3285 |
+
"rewards/reward_func/std": 0.4091717302799225,
|
| 3286 |
+
"step": 126
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"clip_ratio/high_max": 0.0,
|
| 3290 |
+
"clip_ratio/high_mean": 0.0,
|
| 3291 |
+
"clip_ratio/low_mean": 0.0,
|
| 3292 |
+
"clip_ratio/low_min": 0.0,
|
| 3293 |
+
"clip_ratio/region_mean": 0.0,
|
| 3294 |
+
"completions/clipped_ratio": 1.0,
|
| 3295 |
+
"completions/max_length": 384.0,
|
| 3296 |
+
"completions/max_terminated_length": 0.0,
|
| 3297 |
+
"completions/mean_length": 384.0,
|
| 3298 |
+
"completions/mean_terminated_length": 0.0,
|
| 3299 |
+
"completions/min_length": 384.0,
|
| 3300 |
+
"completions/min_terminated_length": 0.0,
|
| 3301 |
+
"entropy": 0.8441327884793282,
|
| 3302 |
+
"epoch": 0.248046875,
|
| 3303 |
+
"frac_reward_zero_std": 0.0,
|
| 3304 |
+
"grad_norm": 0.22232002019882202,
|
| 3305 |
+
"learning_rate": 4.336842105263158e-06,
|
| 3306 |
+
"loss": -0.0,
|
| 3307 |
+
"num_tokens": 595496.0,
|
| 3308 |
+
"reward": 0.37229999899864197,
|
| 3309 |
+
"reward_std": 0.5355914831161499,
|
| 3310 |
+
"rewards/reward_func/mean": 0.37229999899864197,
|
| 3311 |
+
"rewards/reward_func/std": 0.5138660669326782,
|
| 3312 |
+
"step": 127
|
| 3313 |
+
},
|
| 3314 |
+
{
|
| 3315 |
+
"clip_ratio/high_max": 0.0,
|
| 3316 |
+
"clip_ratio/high_mean": 0.0,
|
| 3317 |
+
"clip_ratio/low_mean": 0.0,
|
| 3318 |
+
"clip_ratio/low_min": 0.0,
|
| 3319 |
+
"clip_ratio/region_mean": 0.0,
|
| 3320 |
+
"completions/clipped_ratio": 1.0,
|
| 3321 |
+
"completions/max_length": 384.0,
|
| 3322 |
+
"completions/max_terminated_length": 0.0,
|
| 3323 |
+
"completions/mean_length": 384.0,
|
| 3324 |
+
"completions/mean_terminated_length": 0.0,
|
| 3325 |
+
"completions/min_length": 384.0,
|
| 3326 |
+
"completions/min_terminated_length": 0.0,
|
| 3327 |
+
"entropy": 0.3662140220403671,
|
| 3328 |
+
"epoch": 0.25,
|
| 3329 |
+
"frac_reward_zero_std": 0.0,
|
| 3330 |
+
"grad_norm": 0.19537632167339325,
|
| 3331 |
+
"learning_rate": 4.331578947368421e-06,
|
| 3332 |
+
"loss": 0.0,
|
| 3333 |
+
"num_tokens": 599992.0,
|
| 3334 |
+
"reward": 0.4710499942302704,
|
| 3335 |
+
"reward_std": 0.46093741059303284,
|
| 3336 |
+
"rewards/reward_func/mean": 0.4710499942302704,
|
| 3337 |
+
"rewards/reward_func/std": 0.49048370122909546,
|
| 3338 |
+
"step": 128
|
| 3339 |
+
},
|
| 3340 |
+
{
|
| 3341 |
+
"clip_ratio/high_max": 0.0,
|
| 3342 |
+
"clip_ratio/high_mean": 0.0,
|
| 3343 |
+
"clip_ratio/low_mean": 0.0,
|
| 3344 |
+
"clip_ratio/low_min": 0.0,
|
| 3345 |
+
"clip_ratio/region_mean": 0.0,
|
| 3346 |
+
"completions/clipped_ratio": 1.0,
|
| 3347 |
+
"completions/max_length": 384.0,
|
| 3348 |
+
"completions/max_terminated_length": 0.0,
|
| 3349 |
+
"completions/mean_length": 384.0,
|
| 3350 |
+
"completions/mean_terminated_length": 0.0,
|
| 3351 |
+
"completions/min_length": 384.0,
|
| 3352 |
+
"completions/min_terminated_length": 0.0,
|
| 3353 |
+
"entropy": 0.6109810192137957,
|
| 3354 |
+
"epoch": 0.251953125,
|
| 3355 |
+
"frac_reward_zero_std": 0.0,
|
| 3356 |
+
"grad_norm": 0.20167317986488342,
|
| 3357 |
+
"learning_rate": 4.3263157894736845e-06,
|
| 3358 |
+
"loss": 0.0,
|
| 3359 |
+
"num_tokens": 604848.0,
|
| 3360 |
+
"reward": 0.25439998507499695,
|
| 3361 |
+
"reward_std": 0.47605061531066895,
|
| 3362 |
+
"rewards/reward_func/mean": 0.25439998507499695,
|
| 3363 |
+
"rewards/reward_func/std": 0.44077494740486145,
|
| 3364 |
+
"step": 129
|
| 3365 |
+
},
|
| 3366 |
+
{
|
| 3367 |
+
"clip_ratio/high_max": 0.0,
|
| 3368 |
+
"clip_ratio/high_mean": 0.0,
|
| 3369 |
+
"clip_ratio/low_mean": 0.0,
|
| 3370 |
+
"clip_ratio/low_min": 0.0,
|
| 3371 |
+
"clip_ratio/region_mean": 0.0,
|
| 3372 |
+
"completions/clipped_ratio": 1.0,
|
| 3373 |
+
"completions/max_length": 384.0,
|
| 3374 |
+
"completions/max_terminated_length": 0.0,
|
| 3375 |
+
"completions/mean_length": 384.0,
|
| 3376 |
+
"completions/mean_terminated_length": 0.0,
|
| 3377 |
+
"completions/min_length": 384.0,
|
| 3378 |
+
"completions/min_terminated_length": 0.0,
|
| 3379 |
+
"entropy": 0.765848021954298,
|
| 3380 |
+
"epoch": 0.25390625,
|
| 3381 |
+
"frac_reward_zero_std": 0.0,
|
| 3382 |
+
"grad_norm": 0.2271883189678192,
|
| 3383 |
+
"learning_rate": 4.321052631578948e-06,
|
| 3384 |
+
"loss": 0.0,
|
| 3385 |
+
"num_tokens": 609472.0,
|
| 3386 |
+
"reward": 0.49524998664855957,
|
| 3387 |
+
"reward_std": 0.48512208461761475,
|
| 3388 |
+
"rewards/reward_func/mean": 0.49524998664855957,
|
| 3389 |
+
"rewards/reward_func/std": 0.5163409113883972,
|
| 3390 |
+
"step": 130
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"clip_ratio/high_max": 0.0,
|
| 3394 |
+
"clip_ratio/high_mean": 0.0,
|
| 3395 |
+
"clip_ratio/low_mean": 0.0,
|
| 3396 |
+
"clip_ratio/low_min": 0.0,
|
| 3397 |
+
"clip_ratio/region_mean": 0.0,
|
| 3398 |
+
"completions/clipped_ratio": 1.0,
|
| 3399 |
+
"completions/max_length": 384.0,
|
| 3400 |
+
"completions/max_terminated_length": 0.0,
|
| 3401 |
+
"completions/mean_length": 384.0,
|
| 3402 |
+
"completions/mean_terminated_length": 0.0,
|
| 3403 |
+
"completions/min_length": 384.0,
|
| 3404 |
+
"completions/min_terminated_length": 0.0,
|
| 3405 |
+
"entropy": 0.6323870681226254,
|
| 3406 |
+
"epoch": 0.255859375,
|
| 3407 |
+
"frac_reward_zero_std": 0.0,
|
| 3408 |
+
"grad_norm": 2.204331636428833,
|
| 3409 |
+
"learning_rate": 4.315789473684211e-06,
|
| 3410 |
+
"loss": 0.0,
|
| 3411 |
+
"num_tokens": 613824.0,
|
| 3412 |
+
"reward": 0.7515624761581421,
|
| 3413 |
+
"reward_std": 0.49687498807907104,
|
| 3414 |
+
"rewards/reward_func/mean": 0.7515624761581421,
|
| 3415 |
+
"rewards/reward_func/std": 0.4600290060043335,
|
| 3416 |
+
"step": 131
|
| 3417 |
+
},
|
| 3418 |
+
{
|
| 3419 |
+
"clip_ratio/high_max": 0.0,
|
| 3420 |
+
"clip_ratio/high_mean": 0.0,
|
| 3421 |
+
"clip_ratio/low_mean": 0.0,
|
| 3422 |
+
"clip_ratio/low_min": 0.0,
|
| 3423 |
+
"clip_ratio/region_mean": 0.0,
|
| 3424 |
+
"completions/clipped_ratio": 1.0,
|
| 3425 |
+
"completions/max_length": 384.0,
|
| 3426 |
+
"completions/max_terminated_length": 0.0,
|
| 3427 |
+
"completions/mean_length": 384.0,
|
| 3428 |
+
"completions/mean_terminated_length": 0.0,
|
| 3429 |
+
"completions/min_length": 384.0,
|
| 3430 |
+
"completions/min_terminated_length": 0.0,
|
| 3431 |
+
"entropy": 0.6582241244614124,
|
| 3432 |
+
"epoch": 0.2578125,
|
| 3433 |
+
"frac_reward_zero_std": 0.0,
|
| 3434 |
+
"grad_norm": 0.23651830852031708,
|
| 3435 |
+
"learning_rate": 4.310526315789474e-06,
|
| 3436 |
+
"loss": -0.0,
|
| 3437 |
+
"num_tokens": 618512.0,
|
| 3438 |
+
"reward": 0.625,
|
| 3439 |
+
"reward_std": 0.5386751294136047,
|
| 3440 |
+
"rewards/reward_func/mean": 0.625,
|
| 3441 |
+
"rewards/reward_func/std": 0.5175492167472839,
|
| 3442 |
+
"step": 132
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"clip_ratio/high_max": 0.0,
|
| 3446 |
+
"clip_ratio/high_mean": 0.0,
|
| 3447 |
+
"clip_ratio/low_mean": 0.0,
|
| 3448 |
+
"clip_ratio/low_min": 0.0,
|
| 3449 |
+
"clip_ratio/region_mean": 0.0,
|
| 3450 |
+
"completions/clipped_ratio": 1.0,
|
| 3451 |
+
"completions/max_length": 384.0,
|
| 3452 |
+
"completions/max_terminated_length": 0.0,
|
| 3453 |
+
"completions/mean_length": 384.0,
|
| 3454 |
+
"completions/mean_terminated_length": 0.0,
|
| 3455 |
+
"completions/min_length": 384.0,
|
| 3456 |
+
"completions/min_terminated_length": 0.0,
|
| 3457 |
+
"entropy": 0.6236909478902817,
|
| 3458 |
+
"epoch": 0.259765625,
|
| 3459 |
+
"frac_reward_zero_std": 0.0,
|
| 3460 |
+
"grad_norm": 0.1894795149564743,
|
| 3461 |
+
"learning_rate": 4.305263157894737e-06,
|
| 3462 |
+
"loss": 0.0,
|
| 3463 |
+
"num_tokens": 623488.0,
|
| 3464 |
+
"reward": 0.2150000035762787,
|
| 3465 |
+
"reward_std": 0.4300000071525574,
|
| 3466 |
+
"rewards/reward_func/mean": 0.2150000035762787,
|
| 3467 |
+
"rewards/reward_func/std": 0.39838388562202454,
|
| 3468 |
+
"step": 133
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"clip_ratio/high_max": 0.0,
|
| 3472 |
+
"clip_ratio/high_mean": 0.0,
|
| 3473 |
+
"clip_ratio/low_mean": 0.0,
|
| 3474 |
+
"clip_ratio/low_min": 0.0,
|
| 3475 |
+
"clip_ratio/region_mean": 0.0,
|
| 3476 |
+
"completions/clipped_ratio": 1.0,
|
| 3477 |
+
"completions/max_length": 384.0,
|
| 3478 |
+
"completions/max_terminated_length": 0.0,
|
| 3479 |
+
"completions/mean_length": 384.0,
|
| 3480 |
+
"completions/mean_terminated_length": 0.0,
|
| 3481 |
+
"completions/min_length": 384.0,
|
| 3482 |
+
"completions/min_terminated_length": 0.0,
|
| 3483 |
+
"entropy": 0.23316527949646115,
|
| 3484 |
+
"epoch": 0.26171875,
|
| 3485 |
+
"frac_reward_zero_std": 0.0,
|
| 3486 |
+
"grad_norm": 0.14522254467010498,
|
| 3487 |
+
"learning_rate": 4.3e-06,
|
| 3488 |
+
"loss": 0.0,
|
| 3489 |
+
"num_tokens": 627992.0,
|
| 3490 |
+
"reward": 0.3499000072479248,
|
| 3491 |
+
"reward_std": 0.5033524036407471,
|
| 3492 |
+
"rewards/reward_func/mean": 0.3499000072479248,
|
| 3493 |
+
"rewards/reward_func/std": 0.48291856050491333,
|
| 3494 |
+
"step": 134
|
| 3495 |
+
},
|
| 3496 |
+
{
|
| 3497 |
+
"clip_ratio/high_max": 0.0,
|
| 3498 |
+
"clip_ratio/high_mean": 0.0,
|
| 3499 |
+
"clip_ratio/low_mean": 0.0,
|
| 3500 |
+
"clip_ratio/low_min": 0.0,
|
| 3501 |
+
"clip_ratio/region_mean": 0.0,
|
| 3502 |
+
"completions/clipped_ratio": 1.0,
|
| 3503 |
+
"completions/max_length": 384.0,
|
| 3504 |
+
"completions/max_terminated_length": 0.0,
|
| 3505 |
+
"completions/mean_length": 384.0,
|
| 3506 |
+
"completions/mean_terminated_length": 0.0,
|
| 3507 |
+
"completions/min_length": 384.0,
|
| 3508 |
+
"completions/min_terminated_length": 0.0,
|
| 3509 |
+
"entropy": 0.5633776243776083,
|
| 3510 |
+
"epoch": 0.263671875,
|
| 3511 |
+
"frac_reward_zero_std": 0.5,
|
| 3512 |
+
"grad_norm": 0.14741557836532593,
|
| 3513 |
+
"learning_rate": 4.2947368421052635e-06,
|
| 3514 |
+
"loss": 0.0,
|
| 3515 |
+
"num_tokens": 632352.0,
|
| 3516 |
+
"reward": 0.875,
|
| 3517 |
+
"reward_std": 0.25,
|
| 3518 |
+
"rewards/reward_func/mean": 0.875,
|
| 3519 |
+
"rewards/reward_func/std": 0.3535533845424652,
|
| 3520 |
+
"step": 135
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"clip_ratio/high_max": 0.0,
|
| 3524 |
+
"clip_ratio/high_mean": 0.0,
|
| 3525 |
+
"clip_ratio/low_mean": 0.0,
|
| 3526 |
+
"clip_ratio/low_min": 0.0,
|
| 3527 |
+
"clip_ratio/region_mean": 0.0,
|
| 3528 |
+
"completions/clipped_ratio": 1.0,
|
| 3529 |
+
"completions/max_length": 384.0,
|
| 3530 |
+
"completions/max_terminated_length": 0.0,
|
| 3531 |
+
"completions/mean_length": 384.0,
|
| 3532 |
+
"completions/mean_terminated_length": 0.0,
|
| 3533 |
+
"completions/min_length": 384.0,
|
| 3534 |
+
"completions/min_terminated_length": 0.0,
|
| 3535 |
+
"entropy": 0.5170593503862619,
|
| 3536 |
+
"epoch": 0.265625,
|
| 3537 |
+
"frac_reward_zero_std": 0.0,
|
| 3538 |
+
"grad_norm": 0.20114511251449585,
|
| 3539 |
+
"learning_rate": 4.289473684210527e-06,
|
| 3540 |
+
"loss": 0.0,
|
| 3541 |
+
"num_tokens": 637200.0,
|
| 3542 |
+
"reward": 0.2691749930381775,
|
| 3543 |
+
"reward_std": 0.28962743282318115,
|
| 3544 |
+
"rewards/reward_func/mean": 0.2691749930381775,
|
| 3545 |
+
"rewards/reward_func/std": 0.44533297419548035,
|
| 3546 |
+
"step": 136
|
| 3547 |
+
},
|
| 3548 |
+
{
|
| 3549 |
+
"clip_ratio/high_max": 0.0,
|
| 3550 |
+
"clip_ratio/high_mean": 0.0,
|
| 3551 |
+
"clip_ratio/low_mean": 0.0,
|
| 3552 |
+
"clip_ratio/low_min": 0.0,
|
| 3553 |
+
"clip_ratio/region_mean": 0.0,
|
| 3554 |
+
"completions/clipped_ratio": 1.0,
|
| 3555 |
+
"completions/max_length": 384.0,
|
| 3556 |
+
"completions/max_terminated_length": 0.0,
|
| 3557 |
+
"completions/mean_length": 384.0,
|
| 3558 |
+
"completions/mean_terminated_length": 0.0,
|
| 3559 |
+
"completions/min_length": 384.0,
|
| 3560 |
+
"completions/min_terminated_length": 0.0,
|
| 3561 |
+
"entropy": 0.35685587860643864,
|
| 3562 |
+
"epoch": 0.267578125,
|
| 3563 |
+
"frac_reward_zero_std": 0.0,
|
| 3564 |
+
"grad_norm": 0.17862118780612946,
|
| 3565 |
+
"learning_rate": 4.28421052631579e-06,
|
| 3566 |
+
"loss": -0.0,
|
| 3567 |
+
"num_tokens": 641688.0,
|
| 3568 |
+
"reward": 0.6972000002861023,
|
| 3569 |
+
"reward_std": 0.46480000019073486,
|
| 3570 |
+
"rewards/reward_func/mean": 0.6972000002861023,
|
| 3571 |
+
"rewards/reward_func/std": 0.43032118678092957,
|
| 3572 |
+
"step": 137
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"clip_ratio/high_max": 0.0,
|
| 3576 |
+
"clip_ratio/high_mean": 0.0,
|
| 3577 |
+
"clip_ratio/low_mean": 0.0,
|
| 3578 |
+
"clip_ratio/low_min": 0.0,
|
| 3579 |
+
"clip_ratio/region_mean": 0.0,
|
| 3580 |
+
"completions/clipped_ratio": 1.0,
|
| 3581 |
+
"completions/max_length": 384.0,
|
| 3582 |
+
"completions/max_terminated_length": 0.0,
|
| 3583 |
+
"completions/mean_length": 384.0,
|
| 3584 |
+
"completions/mean_terminated_length": 0.0,
|
| 3585 |
+
"completions/min_length": 384.0,
|
| 3586 |
+
"completions/min_terminated_length": 0.0,
|
| 3587 |
+
"entropy": 0.8764316439628601,
|
| 3588 |
+
"epoch": 0.26953125,
|
| 3589 |
+
"frac_reward_zero_std": 0.0,
|
| 3590 |
+
"grad_norm": 0.22691944241523743,
|
| 3591 |
+
"learning_rate": 4.278947368421053e-06,
|
| 3592 |
+
"loss": 0.0,
|
| 3593 |
+
"num_tokens": 646184.0,
|
| 3594 |
+
"reward": 0.49729999899864197,
|
| 3595 |
+
"reward_std": 0.5742666125297546,
|
| 3596 |
+
"rewards/reward_func/mean": 0.49729999899864197,
|
| 3597 |
+
"rewards/reward_func/std": 0.531683087348938,
|
| 3598 |
+
"step": 138
|
| 3599 |
+
},
|
| 3600 |
+
{
|
| 3601 |
+
"clip_ratio/high_max": 0.0,
|
| 3602 |
+
"clip_ratio/high_mean": 0.0,
|
| 3603 |
+
"clip_ratio/low_mean": 0.0,
|
| 3604 |
+
"clip_ratio/low_min": 0.0,
|
| 3605 |
+
"clip_ratio/region_mean": 0.0,
|
| 3606 |
+
"completions/clipped_ratio": 1.0,
|
| 3607 |
+
"completions/max_length": 384.0,
|
| 3608 |
+
"completions/max_terminated_length": 0.0,
|
| 3609 |
+
"completions/mean_length": 384.0,
|
| 3610 |
+
"completions/mean_terminated_length": 0.0,
|
| 3611 |
+
"completions/min_length": 384.0,
|
| 3612 |
+
"completions/min_terminated_length": 0.0,
|
| 3613 |
+
"entropy": 0.703185360878706,
|
| 3614 |
+
"epoch": 0.271484375,
|
| 3615 |
+
"frac_reward_zero_std": 0.0,
|
| 3616 |
+
"grad_norm": 0.21261753141880035,
|
| 3617 |
+
"learning_rate": 4.273684210526316e-06,
|
| 3618 |
+
"loss": 0.0,
|
| 3619 |
+
"num_tokens": 651160.0,
|
| 3620 |
+
"reward": 0.4440000057220459,
|
| 3621 |
+
"reward_std": 0.5126870274543762,
|
| 3622 |
+
"rewards/reward_func/mean": 0.4440000057220459,
|
| 3623 |
+
"rewards/reward_func/std": 0.47465598583221436,
|
| 3624 |
+
"step": 139
|
| 3625 |
+
},
|
| 3626 |
+
{
|
| 3627 |
+
"clip_ratio/high_max": 0.0,
|
| 3628 |
+
"clip_ratio/high_mean": 0.0,
|
| 3629 |
+
"clip_ratio/low_mean": 0.0,
|
| 3630 |
+
"clip_ratio/low_min": 0.0,
|
| 3631 |
+
"clip_ratio/region_mean": 0.0,
|
| 3632 |
+
"completions/clipped_ratio": 1.0,
|
| 3633 |
+
"completions/max_length": 384.0,
|
| 3634 |
+
"completions/max_terminated_length": 0.0,
|
| 3635 |
+
"completions/mean_length": 384.0,
|
| 3636 |
+
"completions/mean_terminated_length": 0.0,
|
| 3637 |
+
"completions/min_length": 384.0,
|
| 3638 |
+
"completions/min_terminated_length": 0.0,
|
| 3639 |
+
"entropy": 0.7478283271193504,
|
| 3640 |
+
"epoch": 0.2734375,
|
| 3641 |
+
"frac_reward_zero_std": 0.5,
|
| 3642 |
+
"grad_norm": 0.15642641484737396,
|
| 3643 |
+
"learning_rate": 4.268421052631579e-06,
|
| 3644 |
+
"loss": 0.0,
|
| 3645 |
+
"num_tokens": 656016.0,
|
| 3646 |
+
"reward": 0.11140000075101852,
|
| 3647 |
+
"reward_std": 0.22280000150203705,
|
| 3648 |
+
"rewards/reward_func/mean": 0.11140000075101852,
|
| 3649 |
+
"rewards/reward_func/std": 0.3150867819786072,
|
| 3650 |
+
"step": 140
|
| 3651 |
+
},
|
| 3652 |
+
{
|
| 3653 |
+
"clip_ratio/high_max": 0.0,
|
| 3654 |
+
"clip_ratio/high_mean": 0.0,
|
| 3655 |
+
"clip_ratio/low_mean": 0.0,
|
| 3656 |
+
"clip_ratio/low_min": 0.0,
|
| 3657 |
+
"clip_ratio/region_mean": 0.0,
|
| 3658 |
+
"completions/clipped_ratio": 1.0,
|
| 3659 |
+
"completions/max_length": 384.0,
|
| 3660 |
+
"completions/max_terminated_length": 0.0,
|
| 3661 |
+
"completions/mean_length": 384.0,
|
| 3662 |
+
"completions/mean_terminated_length": 0.0,
|
| 3663 |
+
"completions/min_length": 384.0,
|
| 3664 |
+
"completions/min_terminated_length": 0.0,
|
| 3665 |
+
"entropy": 0.507564508356154,
|
| 3666 |
+
"epoch": 0.275390625,
|
| 3667 |
+
"frac_reward_zero_std": 0.0,
|
| 3668 |
+
"grad_norm": 0.15584400296211243,
|
| 3669 |
+
"learning_rate": 4.2631578947368425e-06,
|
| 3670 |
+
"loss": 0.0,
|
| 3671 |
+
"num_tokens": 660816.0,
|
| 3672 |
+
"reward": 0.24137499928474426,
|
| 3673 |
+
"reward_std": 0.45809125900268555,
|
| 3674 |
+
"rewards/reward_func/mean": 0.24137499928474426,
|
| 3675 |
+
"rewards/reward_func/std": 0.42415857315063477,
|
| 3676 |
+
"step": 141
|
| 3677 |
+
},
|
| 3678 |
+
{
|
| 3679 |
+
"clip_ratio/high_max": 0.0,
|
| 3680 |
+
"clip_ratio/high_mean": 0.0,
|
| 3681 |
+
"clip_ratio/low_mean": 0.0,
|
| 3682 |
+
"clip_ratio/low_min": 0.0,
|
| 3683 |
+
"clip_ratio/region_mean": 0.0,
|
| 3684 |
+
"completions/clipped_ratio": 1.0,
|
| 3685 |
+
"completions/max_length": 384.0,
|
| 3686 |
+
"completions/max_terminated_length": 0.0,
|
| 3687 |
+
"completions/mean_length": 384.0,
|
| 3688 |
+
"completions/mean_terminated_length": 0.0,
|
| 3689 |
+
"completions/min_length": 384.0,
|
| 3690 |
+
"completions/min_terminated_length": 0.0,
|
| 3691 |
+
"entropy": 0.6022160463035107,
|
| 3692 |
+
"epoch": 0.27734375,
|
| 3693 |
+
"frac_reward_zero_std": 0.0,
|
| 3694 |
+
"grad_norm": 0.20966751873493195,
|
| 3695 |
+
"learning_rate": 4.257894736842106e-06,
|
| 3696 |
+
"loss": -0.0,
|
| 3697 |
+
"num_tokens": 665656.0,
|
| 3698 |
+
"reward": 0.7447500228881836,
|
| 3699 |
+
"reward_std": 0.2991751432418823,
|
| 3700 |
+
"rewards/reward_func/mean": 0.7447500228881836,
|
| 3701 |
+
"rewards/reward_func/std": 0.4598980247974396,
|
| 3702 |
+
"step": 142
|
| 3703 |
+
},
|
| 3704 |
+
{
|
| 3705 |
+
"clip_ratio/high_max": 0.0,
|
| 3706 |
+
"clip_ratio/high_mean": 0.0,
|
| 3707 |
+
"clip_ratio/low_mean": 0.0,
|
| 3708 |
+
"clip_ratio/low_min": 0.0,
|
| 3709 |
+
"clip_ratio/region_mean": 0.0,
|
| 3710 |
+
"completions/clipped_ratio": 1.0,
|
| 3711 |
+
"completions/max_length": 384.0,
|
| 3712 |
+
"completions/max_terminated_length": 0.0,
|
| 3713 |
+
"completions/mean_length": 384.0,
|
| 3714 |
+
"completions/mean_terminated_length": 0.0,
|
| 3715 |
+
"completions/min_length": 384.0,
|
| 3716 |
+
"completions/min_terminated_length": 0.0,
|
| 3717 |
+
"entropy": 0.9077504724264145,
|
| 3718 |
+
"epoch": 0.279296875,
|
| 3719 |
+
"frac_reward_zero_std": 0.0,
|
| 3720 |
+
"grad_norm": 0.27349957823753357,
|
| 3721 |
+
"learning_rate": 4.252631578947369e-06,
|
| 3722 |
+
"loss": 0.0,
|
| 3723 |
+
"num_tokens": 670144.0,
|
| 3724 |
+
"reward": 0.625,
|
| 3725 |
+
"reward_std": 0.5386751294136047,
|
| 3726 |
+
"rewards/reward_func/mean": 0.625,
|
| 3727 |
+
"rewards/reward_func/std": 0.5175492167472839,
|
| 3728 |
+
"step": 143
|
| 3729 |
+
},
|
| 3730 |
+
{
|
| 3731 |
+
"clip_ratio/high_max": 0.0,
|
| 3732 |
+
"clip_ratio/high_mean": 0.0,
|
| 3733 |
+
"clip_ratio/low_mean": 0.0,
|
| 3734 |
+
"clip_ratio/low_min": 0.0,
|
| 3735 |
+
"clip_ratio/region_mean": 0.0,
|
| 3736 |
+
"completions/clipped_ratio": 1.0,
|
| 3737 |
+
"completions/max_length": 384.0,
|
| 3738 |
+
"completions/max_terminated_length": 0.0,
|
| 3739 |
+
"completions/mean_length": 384.0,
|
| 3740 |
+
"completions/mean_terminated_length": 0.0,
|
| 3741 |
+
"completions/min_length": 384.0,
|
| 3742 |
+
"completions/min_terminated_length": 0.0,
|
| 3743 |
+
"entropy": 0.5460263751447201,
|
| 3744 |
+
"epoch": 0.28125,
|
| 3745 |
+
"frac_reward_zero_std": 0.0,
|
| 3746 |
+
"grad_norm": 0.17890246212482452,
|
| 3747 |
+
"learning_rate": 4.247368421052632e-06,
|
| 3748 |
+
"loss": 0.0,
|
| 3749 |
+
"num_tokens": 674832.0,
|
| 3750 |
+
"reward": 0.503125011920929,
|
| 3751 |
+
"reward_std": 0.49798667430877686,
|
| 3752 |
+
"rewards/reward_func/mean": 0.503125011920929,
|
| 3753 |
+
"rewards/reward_func/std": 0.5312447547912598,
|
| 3754 |
+
"step": 144
|
| 3755 |
+
},
|
| 3756 |
+
{
|
| 3757 |
+
"clip_ratio/high_max": 0.0,
|
| 3758 |
+
"clip_ratio/high_mean": 0.0,
|
| 3759 |
+
"clip_ratio/low_mean": 0.0,
|
| 3760 |
+
"clip_ratio/low_min": 0.0,
|
| 3761 |
+
"clip_ratio/region_mean": 0.0,
|
| 3762 |
+
"completions/clipped_ratio": 1.0,
|
| 3763 |
+
"completions/max_length": 384.0,
|
| 3764 |
+
"completions/max_terminated_length": 0.0,
|
| 3765 |
+
"completions/mean_length": 384.0,
|
| 3766 |
+
"completions/mean_terminated_length": 0.0,
|
| 3767 |
+
"completions/min_length": 384.0,
|
| 3768 |
+
"completions/min_terminated_length": 0.0,
|
| 3769 |
+
"entropy": 0.6752507574856281,
|
| 3770 |
+
"epoch": 0.283203125,
|
| 3771 |
+
"frac_reward_zero_std": 0.5,
|
| 3772 |
+
"grad_norm": 0.17742383480072021,
|
| 3773 |
+
"learning_rate": 4.242105263157895e-06,
|
| 3774 |
+
"loss": 0.0,
|
| 3775 |
+
"num_tokens": 679176.0,
|
| 3776 |
+
"reward": 0.75,
|
| 3777 |
+
"reward_std": 0.28867512941360474,
|
| 3778 |
+
"rewards/reward_func/mean": 0.75,
|
| 3779 |
+
"rewards/reward_func/std": 0.4629100561141968,
|
| 3780 |
+
"step": 145
|
| 3781 |
+
},
|
| 3782 |
+
{
|
| 3783 |
+
"clip_ratio/high_max": 0.0,
|
| 3784 |
+
"clip_ratio/high_mean": 0.0,
|
| 3785 |
+
"clip_ratio/low_mean": 0.0,
|
| 3786 |
+
"clip_ratio/low_min": 0.0,
|
| 3787 |
+
"clip_ratio/region_mean": 0.0,
|
| 3788 |
+
"completions/clipped_ratio": 1.0,
|
| 3789 |
+
"completions/max_length": 384.0,
|
| 3790 |
+
"completions/max_terminated_length": 0.0,
|
| 3791 |
+
"completions/mean_length": 384.0,
|
| 3792 |
+
"completions/mean_terminated_length": 0.0,
|
| 3793 |
+
"completions/min_length": 384.0,
|
| 3794 |
+
"completions/min_terminated_length": 0.0,
|
| 3795 |
+
"entropy": 0.6691841837018728,
|
| 3796 |
+
"epoch": 0.28515625,
|
| 3797 |
+
"frac_reward_zero_std": 0.0,
|
| 3798 |
+
"grad_norm": 0.2150551676750183,
|
| 3799 |
+
"learning_rate": 4.236842105263158e-06,
|
| 3800 |
+
"loss": -0.0,
|
| 3801 |
+
"num_tokens": 683968.0,
|
| 3802 |
+
"reward": 0.7343499660491943,
|
| 3803 |
+
"reward_std": 0.49009713530540466,
|
| 3804 |
+
"rewards/reward_func/mean": 0.7343499660491943,
|
| 3805 |
+
"rewards/reward_func/std": 0.45377910137176514,
|
| 3806 |
+
"step": 146
|
| 3807 |
+
},
|
| 3808 |
+
{
|
| 3809 |
+
"clip_ratio/high_max": 0.0,
|
| 3810 |
+
"clip_ratio/high_mean": 0.0,
|
| 3811 |
+
"clip_ratio/low_mean": 0.0,
|
| 3812 |
+
"clip_ratio/low_min": 0.0,
|
| 3813 |
+
"clip_ratio/region_mean": 0.0,
|
| 3814 |
+
"completions/clipped_ratio": 1.0,
|
| 3815 |
+
"completions/max_length": 384.0,
|
| 3816 |
+
"completions/max_terminated_length": 0.0,
|
| 3817 |
+
"completions/mean_length": 384.0,
|
| 3818 |
+
"completions/mean_terminated_length": 0.0,
|
| 3819 |
+
"completions/min_length": 384.0,
|
| 3820 |
+
"completions/min_terminated_length": 0.0,
|
| 3821 |
+
"entropy": 0.970278836786747,
|
| 3822 |
+
"epoch": 0.287109375,
|
| 3823 |
+
"frac_reward_zero_std": 0.0,
|
| 3824 |
+
"grad_norm": 0.27998873591423035,
|
| 3825 |
+
"learning_rate": 4.2315789473684215e-06,
|
| 3826 |
+
"loss": -0.0,
|
| 3827 |
+
"num_tokens": 688464.0,
|
| 3828 |
+
"reward": 0.37229999899864197,
|
| 3829 |
+
"reward_std": 0.5355914831161499,
|
| 3830 |
+
"rewards/reward_func/mean": 0.37229999899864197,
|
| 3831 |
+
"rewards/reward_func/std": 0.5138660669326782,
|
| 3832 |
+
"step": 147
|
| 3833 |
+
},
|
| 3834 |
+
{
|
| 3835 |
+
"clip_ratio/high_max": 0.0,
|
| 3836 |
+
"clip_ratio/high_mean": 0.0,
|
| 3837 |
+
"clip_ratio/low_mean": 0.0,
|
| 3838 |
+
"clip_ratio/low_min": 0.0,
|
| 3839 |
+
"clip_ratio/region_mean": 0.0,
|
| 3840 |
+
"completions/clipped_ratio": 1.0,
|
| 3841 |
+
"completions/max_length": 384.0,
|
| 3842 |
+
"completions/max_terminated_length": 0.0,
|
| 3843 |
+
"completions/mean_length": 384.0,
|
| 3844 |
+
"completions/mean_terminated_length": 0.0,
|
| 3845 |
+
"completions/min_length": 384.0,
|
| 3846 |
+
"completions/min_terminated_length": 0.0,
|
| 3847 |
+
"entropy": 0.38460469990968704,
|
| 3848 |
+
"epoch": 0.2890625,
|
| 3849 |
+
"frac_reward_zero_std": 0.0,
|
| 3850 |
+
"grad_norm": 0.24717585742473602,
|
| 3851 |
+
"learning_rate": 4.226315789473685e-06,
|
| 3852 |
+
"loss": 0.0,
|
| 3853 |
+
"num_tokens": 693256.0,
|
| 3854 |
+
"reward": 0.36281248927116394,
|
| 3855 |
+
"reward_std": 0.5174503326416016,
|
| 3856 |
+
"rewards/reward_func/mean": 0.36281248927116394,
|
| 3857 |
+
"rewards/reward_func/std": 0.4976207911968231,
|
| 3858 |
+
"step": 148
|
| 3859 |
+
},
|
| 3860 |
+
{
|
| 3861 |
+
"clip_ratio/high_max": 0.0,
|
| 3862 |
+
"clip_ratio/high_mean": 0.0,
|
| 3863 |
+
"clip_ratio/low_mean": 0.0,
|
| 3864 |
+
"clip_ratio/low_min": 0.0,
|
| 3865 |
+
"clip_ratio/region_mean": 0.0,
|
| 3866 |
+
"completions/clipped_ratio": 1.0,
|
| 3867 |
+
"completions/max_length": 384.0,
|
| 3868 |
+
"completions/max_terminated_length": 0.0,
|
| 3869 |
+
"completions/mean_length": 384.0,
|
| 3870 |
+
"completions/mean_terminated_length": 0.0,
|
| 3871 |
+
"completions/min_length": 384.0,
|
| 3872 |
+
"completions/min_terminated_length": 0.0,
|
| 3873 |
+
"entropy": 0.8782265558838844,
|
| 3874 |
+
"epoch": 0.291015625,
|
| 3875 |
+
"frac_reward_zero_std": 0.0,
|
| 3876 |
+
"grad_norm": 0.23354728519916534,
|
| 3877 |
+
"learning_rate": 4.221052631578948e-06,
|
| 3878 |
+
"loss": 0.0,
|
| 3879 |
+
"num_tokens": 697904.0,
|
| 3880 |
+
"reward": 0.3670499920845032,
|
| 3881 |
+
"reward_std": 0.5250914692878723,
|
| 3882 |
+
"rewards/reward_func/mean": 0.3670499920845032,
|
| 3883 |
+
"rewards/reward_func/std": 0.5067015290260315,
|
| 3884 |
+
"step": 149
|
| 3885 |
+
},
|
| 3886 |
+
{
|
| 3887 |
+
"clip_ratio/high_max": 0.0,
|
| 3888 |
+
"clip_ratio/high_mean": 0.0,
|
| 3889 |
+
"clip_ratio/low_mean": 0.0,
|
| 3890 |
+
"clip_ratio/low_min": 0.0,
|
| 3891 |
+
"clip_ratio/region_mean": 0.0,
|
| 3892 |
+
"completions/clipped_ratio": 1.0,
|
| 3893 |
+
"completions/max_length": 384.0,
|
| 3894 |
+
"completions/max_terminated_length": 0.0,
|
| 3895 |
+
"completions/mean_length": 384.0,
|
| 3896 |
+
"completions/mean_terminated_length": 0.0,
|
| 3897 |
+
"completions/min_length": 384.0,
|
| 3898 |
+
"completions/min_terminated_length": 0.0,
|
| 3899 |
+
"entropy": 0.6834246851503849,
|
| 3900 |
+
"epoch": 0.29296875,
|
| 3901 |
+
"frac_reward_zero_std": 0.0,
|
| 3902 |
+
"grad_norm": 0.2545972764492035,
|
| 3903 |
+
"learning_rate": 4.215789473684211e-06,
|
| 3904 |
+
"loss": -0.0,
|
| 3905 |
+
"num_tokens": 702248.0,
|
| 3906 |
+
"reward": 0.7447500228881836,
|
| 3907 |
+
"reward_std": 0.2991751432418823,
|
| 3908 |
+
"rewards/reward_func/mean": 0.7447500228881836,
|
| 3909 |
+
"rewards/reward_func/std": 0.4598980247974396,
|
| 3910 |
+
"step": 150
|
| 3911 |
}
|
| 3912 |
],
|
| 3913 |
"logging_steps": 1,
|
| 3914 |
"max_steps": 950,
|
| 3915 |
+
"num_input_tokens_seen": 702248,
|
| 3916 |
"num_train_epochs": 2,
|
| 3917 |
"save_steps": 50,
|
| 3918 |
"stateful_callbacks": {
|