Upload folder using huggingface_hub
Browse files- adapter_model.safetensors +1 -1
- optimizer.pt +1 -1
- rng_state_0.pth +1 -1
- rng_state_2.pth +1 -1
- rng_state_3.pth +1 -1
- rng_state_4.pth +1 -1
- rng_state_5.pth +1 -1
- rng_state_6.pth +1 -1
- rng_state_7.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +963 -3
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16794200
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f49333bf8dc54c673cdb5d086454c1c2e11545428fdeab99e17f54d26dfd68b5
|
| 3 |
size 16794200
|
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 33664331
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:847f243c3b0717de9d7d84a818a5cca2780f187c631f0830647f9a4f32a025e5
|
| 3 |
size 33664331
|
rng_state_0.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16325
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:64809719cba422cef96de4920261da9523f70c17a2b1212c597d67f1f7a800be
|
| 3 |
size 16325
|
rng_state_2.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:eebd48be5f45b09ad66e61b713e78075a6ed9758b0743b2ba14d0149a9d32663
|
| 3 |
size 16389
|
rng_state_3.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ce6b5432b3dd7ce01917cc36abf764e4593e128309e145991b4505505b34ea51
|
| 3 |
size 16389
|
rng_state_4.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ccd4238203b96310504fffc3e0300303148399ae449f5a3c1223bd19096fe34f
|
| 3 |
size 16389
|
rng_state_5.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e7aad592eab42dab56e49c09184b6b56297b879407f0b6182f333c4c43d53d1a
|
| 3 |
size 16389
|
rng_state_6.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16389
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02690e58d4566e8b28fca4807f93c1b9ed552cda61c2c7266f83fc01375f9221
|
| 3 |
size 16389
|
rng_state_7.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 16325
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3f847170fd131c7017195f78afedddbaa655b33607982525c46cb880859458f
|
| 3 |
size 16325
|
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 sha256:9ad77e042001d2e933aab4e1dccf5c99a07800e969ec3c1619623eadd806ef6c
|
| 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,
|
|
@@ -2798,11 +2798,971 @@
|
|
| 2798 |
"rewards/format_reward_fn/mean": 0.9190057702362537,
|
| 2799 |
"rewards/format_reward_fn/std": 0.23810118879191577,
|
| 2800 |
"step": 1488
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2801 |
}
|
| 2802 |
],
|
| 2803 |
"logging_steps": 16,
|
| 2804 |
"max_steps": 10290,
|
| 2805 |
-
"num_input_tokens_seen":
|
| 2806 |
"num_train_epochs": 10,
|
| 2807 |
"save_steps": 500,
|
| 2808 |
"stateful_callbacks": {
|
|
|
|
| 2 |
"best_global_step": null,
|
| 3 |
"best_metric": null,
|
| 4 |
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.943634596695821,
|
| 6 |
"eval_steps": 500,
|
| 7 |
+
"global_step": 2000,
|
| 8 |
"is_hyper_param_search": false,
|
| 9 |
"is_local_process_zero": true,
|
| 10 |
"is_world_process_zero": true,
|
|
|
|
| 2798 |
"rewards/format_reward_fn/mean": 0.9190057702362537,
|
| 2799 |
"rewards/format_reward_fn/std": 0.23810118879191577,
|
| 2800 |
"step": 1488
|
| 2801 |
+
},
|
| 2802 |
+
{
|
| 2803 |
+
"clip_ratio/high_max": 0.0,
|
| 2804 |
+
"clip_ratio/high_mean": 0.0,
|
| 2805 |
+
"clip_ratio/low_mean": 0.0,
|
| 2806 |
+
"clip_ratio/low_min": 0.0,
|
| 2807 |
+
"clip_ratio/region_mean": 0.0,
|
| 2808 |
+
"completions/clipped_ratio": 0.07421875,
|
| 2809 |
+
"completions/max_length": 253.5,
|
| 2810 |
+
"completions/max_terminated_length": 244.625,
|
| 2811 |
+
"completions/mean_length": 190.771484375,
|
| 2812 |
+
"completions/mean_terminated_length": 185.8874397277832,
|
| 2813 |
+
"completions/min_length": 140.1875,
|
| 2814 |
+
"completions/min_terminated_length": 140.1875,
|
| 2815 |
+
"entropy": 0.0769493873231113,
|
| 2816 |
+
"epoch": 1.4616132167152576,
|
| 2817 |
+
"frac_reward_zero_std": 0.27734375,
|
| 2818 |
+
"grad_norm": 0.13221606612205505,
|
| 2819 |
+
"learning_rate": 5e-05,
|
| 2820 |
+
"loss": -0.0052,
|
| 2821 |
+
"num_tokens": 120878670.0,
|
| 2822 |
+
"reward": 11.72208970785141,
|
| 2823 |
+
"reward_std": 0.9891778491437435,
|
| 2824 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.865969829261303,
|
| 2825 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3025930265430361,
|
| 2826 |
+
"rewards/event_reward_fn/mean": 9.978515625,
|
| 2827 |
+
"rewards/event_reward_fn/std": 6.088510304689407,
|
| 2828 |
+
"rewards/format_reward_fn/mean": 0.8776041679084301,
|
| 2829 |
+
"rewards/format_reward_fn/std": 0.30370487459003925,
|
| 2830 |
+
"step": 1504
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"clip_ratio/high_max": 0.0,
|
| 2834 |
+
"clip_ratio/high_mean": 0.0,
|
| 2835 |
+
"clip_ratio/low_mean": 0.0,
|
| 2836 |
+
"clip_ratio/low_min": 0.0,
|
| 2837 |
+
"clip_ratio/region_mean": 0.0,
|
| 2838 |
+
"completions/clipped_ratio": 0.09375,
|
| 2839 |
+
"completions/max_length": 256.0,
|
| 2840 |
+
"completions/max_terminated_length": 250.875,
|
| 2841 |
+
"completions/mean_length": 210.828125,
|
| 2842 |
+
"completions/mean_terminated_length": 206.00187873840332,
|
| 2843 |
+
"completions/min_length": 159.9375,
|
| 2844 |
+
"completions/min_terminated_length": 159.9375,
|
| 2845 |
+
"entropy": 0.09037951100617647,
|
| 2846 |
+
"epoch": 1.4771622934888242,
|
| 2847 |
+
"frac_reward_zero_std": 0.31640625,
|
| 2848 |
+
"grad_norm": 0.303564190864563,
|
| 2849 |
+
"learning_rate": 5e-05,
|
| 2850 |
+
"loss": -0.0015,
|
| 2851 |
+
"num_tokens": 122164070.0,
|
| 2852 |
+
"reward": 11.119612038135529,
|
| 2853 |
+
"reward_std": 0.99767005443573,
|
| 2854 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8756668232381344,
|
| 2855 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.29590372927486897,
|
| 2856 |
+
"rewards/event_reward_fn/mean": 9.3515625,
|
| 2857 |
+
"rewards/event_reward_fn/std": 5.329805389046669,
|
| 2858 |
+
"rewards/format_reward_fn/mean": 0.8923828117549419,
|
| 2859 |
+
"rewards/format_reward_fn/std": 0.30003819055855274,
|
| 2860 |
+
"step": 1520
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"clip_ratio/high_max": 0.0,
|
| 2864 |
+
"clip_ratio/high_mean": 0.0,
|
| 2865 |
+
"clip_ratio/low_mean": 0.0,
|
| 2866 |
+
"clip_ratio/low_min": 0.0,
|
| 2867 |
+
"clip_ratio/region_mean": 0.0,
|
| 2868 |
+
"completions/clipped_ratio": 0.1669921875,
|
| 2869 |
+
"completions/max_length": 256.0,
|
| 2870 |
+
"completions/max_terminated_length": 253.0,
|
| 2871 |
+
"completions/mean_length": 215.0888671875,
|
| 2872 |
+
"completions/mean_terminated_length": 206.6857042312622,
|
| 2873 |
+
"completions/min_length": 164.4375,
|
| 2874 |
+
"completions/min_terminated_length": 164.4375,
|
| 2875 |
+
"entropy": 0.09090339438989758,
|
| 2876 |
+
"epoch": 1.4927113702623906,
|
| 2877 |
+
"frac_reward_zero_std": 0.3046875,
|
| 2878 |
+
"grad_norm": 0.16249267756938934,
|
| 2879 |
+
"learning_rate": 5e-05,
|
| 2880 |
+
"loss": 0.0016,
|
| 2881 |
+
"num_tokens": 123527081.0,
|
| 2882 |
+
"reward": 10.766064465045929,
|
| 2883 |
+
"reward_std": 0.8386576101183891,
|
| 2884 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.7968913167715073,
|
| 2885 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3705411199480295,
|
| 2886 |
+
"rewards/event_reward_fn/mean": 9.1552734375,
|
| 2887 |
+
"rewards/event_reward_fn/std": 5.637863516807556,
|
| 2888 |
+
"rewards/format_reward_fn/mean": 0.8138997405767441,
|
| 2889 |
+
"rewards/format_reward_fn/std": 0.3759169615805149,
|
| 2890 |
+
"step": 1536
|
| 2891 |
+
},
|
| 2892 |
+
{
|
| 2893 |
+
"clip_ratio/high_max": 0.0,
|
| 2894 |
+
"clip_ratio/high_mean": 0.0,
|
| 2895 |
+
"clip_ratio/low_mean": 0.0,
|
| 2896 |
+
"clip_ratio/low_min": 0.0,
|
| 2897 |
+
"clip_ratio/region_mean": 0.0,
|
| 2898 |
+
"completions/clipped_ratio": 0.1435546875,
|
| 2899 |
+
"completions/max_length": 255.375,
|
| 2900 |
+
"completions/max_terminated_length": 250.375,
|
| 2901 |
+
"completions/mean_length": 212.8251953125,
|
| 2902 |
+
"completions/mean_terminated_length": 205.47227001190186,
|
| 2903 |
+
"completions/min_length": 157.5625,
|
| 2904 |
+
"completions/min_terminated_length": 157.5625,
|
| 2905 |
+
"entropy": 0.10008962173014879,
|
| 2906 |
+
"epoch": 1.508260447035957,
|
| 2907 |
+
"frac_reward_zero_std": 0.265625,
|
| 2908 |
+
"grad_norm": 0.23113620281219482,
|
| 2909 |
+
"learning_rate": 5e-05,
|
| 2910 |
+
"loss": 0.004,
|
| 2911 |
+
"num_tokens": 124865830.0,
|
| 2912 |
+
"reward": 10.332128584384918,
|
| 2913 |
+
"reward_std": 1.082621719688177,
|
| 2914 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8261714465916157,
|
| 2915 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3391446927562356,
|
| 2916 |
+
"rewards/event_reward_fn/mean": 8.6630859375,
|
| 2917 |
+
"rewards/event_reward_fn/std": 5.3445031344890594,
|
| 2918 |
+
"rewards/format_reward_fn/mean": 0.8428710959851742,
|
| 2919 |
+
"rewards/format_reward_fn/std": 0.34356776159256697,
|
| 2920 |
+
"step": 1552
|
| 2921 |
+
},
|
| 2922 |
+
{
|
| 2923 |
+
"clip_ratio/high_max": 0.0,
|
| 2924 |
+
"clip_ratio/high_mean": 0.0,
|
| 2925 |
+
"clip_ratio/low_mean": 0.0,
|
| 2926 |
+
"clip_ratio/low_min": 0.0,
|
| 2927 |
+
"clip_ratio/region_mean": 0.0,
|
| 2928 |
+
"completions/clipped_ratio": 0.1103515625,
|
| 2929 |
+
"completions/max_length": 255.125,
|
| 2930 |
+
"completions/max_terminated_length": 247.875,
|
| 2931 |
+
"completions/mean_length": 204.4716796875,
|
| 2932 |
+
"completions/mean_terminated_length": 198.4902868270874,
|
| 2933 |
+
"completions/min_length": 149.9375,
|
| 2934 |
+
"completions/min_terminated_length": 149.9375,
|
| 2935 |
+
"entropy": 0.09716548025608063,
|
| 2936 |
+
"epoch": 1.5238095238095237,
|
| 2937 |
+
"frac_reward_zero_std": 0.30859375,
|
| 2938 |
+
"grad_norm": 0.13532325625419617,
|
| 2939 |
+
"learning_rate": 5e-05,
|
| 2940 |
+
"loss": 0.0014,
|
| 2941 |
+
"num_tokens": 126156049.0,
|
| 2942 |
+
"reward": 9.934103816747665,
|
| 2943 |
+
"reward_std": 0.9690110310912132,
|
| 2944 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8615453615784645,
|
| 2945 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2830730821006,
|
| 2946 |
+
"rewards/event_reward_fn/mean": 8.1953125,
|
| 2947 |
+
"rewards/event_reward_fn/std": 4.997192412614822,
|
| 2948 |
+
"rewards/format_reward_fn/mean": 0.8772460930049419,
|
| 2949 |
+
"rewards/format_reward_fn/std": 0.27953232545405626,
|
| 2950 |
+
"step": 1568
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"clip_ratio/high_max": 0.0,
|
| 2954 |
+
"clip_ratio/high_mean": 0.0,
|
| 2955 |
+
"clip_ratio/low_mean": 0.0,
|
| 2956 |
+
"clip_ratio/low_min": 0.0,
|
| 2957 |
+
"clip_ratio/region_mean": 0.0,
|
| 2958 |
+
"completions/clipped_ratio": 0.0634765625,
|
| 2959 |
+
"completions/max_length": 253.625,
|
| 2960 |
+
"completions/max_terminated_length": 241.8125,
|
| 2961 |
+
"completions/mean_length": 191.8193359375,
|
| 2962 |
+
"completions/mean_terminated_length": 187.4044713973999,
|
| 2963 |
+
"completions/min_length": 134.4375,
|
| 2964 |
+
"completions/min_terminated_length": 134.4375,
|
| 2965 |
+
"entropy": 0.08724062331020832,
|
| 2966 |
+
"epoch": 1.5393586005830904,
|
| 2967 |
+
"frac_reward_zero_std": 0.30078125,
|
| 2968 |
+
"grad_norm": 0.13813965022563934,
|
| 2969 |
+
"learning_rate": 5e-05,
|
| 2970 |
+
"loss": -0.0036,
|
| 2971 |
+
"num_tokens": 127483244.0,
|
| 2972 |
+
"reward": 11.109964549541473,
|
| 2973 |
+
"reward_std": 0.9232164584100246,
|
| 2974 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9082068763673306,
|
| 2975 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.25597723573446274,
|
| 2976 |
+
"rewards/event_reward_fn/mean": 9.279296875,
|
| 2977 |
+
"rewards/event_reward_fn/std": 5.3837059289216995,
|
| 2978 |
+
"rewards/format_reward_fn/mean": 0.9224609360098839,
|
| 2979 |
+
"rewards/format_reward_fn/std": 0.2576202508062124,
|
| 2980 |
+
"step": 1584
|
| 2981 |
+
},
|
| 2982 |
+
{
|
| 2983 |
+
"clip_ratio/high_max": 0.0,
|
| 2984 |
+
"clip_ratio/high_mean": 0.0,
|
| 2985 |
+
"clip_ratio/low_mean": 0.0,
|
| 2986 |
+
"clip_ratio/low_min": 0.0,
|
| 2987 |
+
"clip_ratio/region_mean": 0.0,
|
| 2988 |
+
"completions/clipped_ratio": 0.0458984375,
|
| 2989 |
+
"completions/max_length": 251.0,
|
| 2990 |
+
"completions/max_terminated_length": 242.375,
|
| 2991 |
+
"completions/mean_length": 193.9111328125,
|
| 2992 |
+
"completions/mean_terminated_length": 190.90945529937744,
|
| 2993 |
+
"completions/min_length": 148.625,
|
| 2994 |
+
"completions/min_terminated_length": 148.625,
|
| 2995 |
+
"entropy": 0.08152232086285949,
|
| 2996 |
+
"epoch": 1.554907677356657,
|
| 2997 |
+
"frac_reward_zero_std": 0.34375,
|
| 2998 |
+
"grad_norm": 0.35102641582489014,
|
| 2999 |
+
"learning_rate": 5e-05,
|
| 3000 |
+
"loss": -0.0013,
|
| 3001 |
+
"num_tokens": 128764293.0,
|
| 3002 |
+
"reward": 11.371211469173431,
|
| 3003 |
+
"reward_std": 0.8595849685370922,
|
| 3004 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9204303659498692,
|
| 3005 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.21888624806888402,
|
| 3006 |
+
"rewards/event_reward_fn/mean": 9.513671875,
|
| 3007 |
+
"rewards/event_reward_fn/std": 5.4597727209329605,
|
| 3008 |
+
"rewards/format_reward_fn/mean": 0.9371093735098839,
|
| 3009 |
+
"rewards/format_reward_fn/std": 0.2127007795497775,
|
| 3010 |
+
"step": 1600
|
| 3011 |
+
},
|
| 3012 |
+
{
|
| 3013 |
+
"clip_ratio/high_max": 0.0,
|
| 3014 |
+
"clip_ratio/high_mean": 0.0,
|
| 3015 |
+
"clip_ratio/low_mean": 0.0,
|
| 3016 |
+
"clip_ratio/low_min": 0.0,
|
| 3017 |
+
"clip_ratio/region_mean": 0.0,
|
| 3018 |
+
"completions/clipped_ratio": 0.091796875,
|
| 3019 |
+
"completions/max_length": 252.375,
|
| 3020 |
+
"completions/max_terminated_length": 245.375,
|
| 3021 |
+
"completions/mean_length": 200.150390625,
|
| 3022 |
+
"completions/mean_terminated_length": 194.58474922180176,
|
| 3023 |
+
"completions/min_length": 145.125,
|
| 3024 |
+
"completions/min_terminated_length": 145.125,
|
| 3025 |
+
"entropy": 0.08945442596450448,
|
| 3026 |
+
"epoch": 1.5704567541302237,
|
| 3027 |
+
"frac_reward_zero_std": 0.28125,
|
| 3028 |
+
"grad_norm": 0.11586015671491623,
|
| 3029 |
+
"learning_rate": 5e-05,
|
| 3030 |
+
"loss": -0.0005,
|
| 3031 |
+
"num_tokens": 130147811.0,
|
| 3032 |
+
"reward": 10.688360095024109,
|
| 3033 |
+
"reward_std": 0.8784848563373089,
|
| 3034 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8703912869095802,
|
| 3035 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.26367771509103477,
|
| 3036 |
+
"rewards/event_reward_fn/mean": 8.9267578125,
|
| 3037 |
+
"rewards/event_reward_fn/std": 5.635714888572693,
|
| 3038 |
+
"rewards/format_reward_fn/mean": 0.8912109360098839,
|
| 3039 |
+
"rewards/format_reward_fn/std": 0.2533010635524988,
|
| 3040 |
+
"step": 1616
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"clip_ratio/high_max": 0.0,
|
| 3044 |
+
"clip_ratio/high_mean": 0.0,
|
| 3045 |
+
"clip_ratio/low_mean": 0.0,
|
| 3046 |
+
"clip_ratio/low_min": 0.0,
|
| 3047 |
+
"clip_ratio/region_mean": 0.0,
|
| 3048 |
+
"completions/clipped_ratio": 0.07421875,
|
| 3049 |
+
"completions/max_length": 254.3125,
|
| 3050 |
+
"completions/max_terminated_length": 239.75,
|
| 3051 |
+
"completions/mean_length": 195.1494140625,
|
| 3052 |
+
"completions/mean_terminated_length": 190.32571697235107,
|
| 3053 |
+
"completions/min_length": 143.625,
|
| 3054 |
+
"completions/min_terminated_length": 143.625,
|
| 3055 |
+
"entropy": 0.08721820963546634,
|
| 3056 |
+
"epoch": 1.58600583090379,
|
| 3057 |
+
"frac_reward_zero_std": 0.328125,
|
| 3058 |
+
"grad_norm": 0.1575620472431183,
|
| 3059 |
+
"learning_rate": 5e-05,
|
| 3060 |
+
"loss": -0.0,
|
| 3061 |
+
"num_tokens": 131479512.0,
|
| 3062 |
+
"reward": 10.922975957393646,
|
| 3063 |
+
"reward_std": 0.7370323836803436,
|
| 3064 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8828910291194916,
|
| 3065 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2897054869681597,
|
| 3066 |
+
"rewards/event_reward_fn/mean": 9.146484375,
|
| 3067 |
+
"rewards/event_reward_fn/std": 5.057717680931091,
|
| 3068 |
+
"rewards/format_reward_fn/mean": 0.8936002627015114,
|
| 3069 |
+
"rewards/format_reward_fn/std": 0.28906678687781096,
|
| 3070 |
+
"step": 1632
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"clip_ratio/high_max": 0.0,
|
| 3074 |
+
"clip_ratio/high_mean": 0.0,
|
| 3075 |
+
"clip_ratio/low_mean": 0.0,
|
| 3076 |
+
"clip_ratio/low_min": 0.0,
|
| 3077 |
+
"clip_ratio/region_mean": 0.0,
|
| 3078 |
+
"completions/clipped_ratio": 0.0703125,
|
| 3079 |
+
"completions/max_length": 254.0,
|
| 3080 |
+
"completions/max_terminated_length": 241.6875,
|
| 3081 |
+
"completions/mean_length": 198.5185546875,
|
| 3082 |
+
"completions/mean_terminated_length": 194.11609935760498,
|
| 3083 |
+
"completions/min_length": 149.1875,
|
| 3084 |
+
"completions/min_terminated_length": 149.1875,
|
| 3085 |
+
"entropy": 0.08794478559866548,
|
| 3086 |
+
"epoch": 1.6015549076773565,
|
| 3087 |
+
"frac_reward_zero_std": 0.34375,
|
| 3088 |
+
"grad_norm": 0.16397124528884888,
|
| 3089 |
+
"learning_rate": 5e-05,
|
| 3090 |
+
"loss": 0.001,
|
| 3091 |
+
"num_tokens": 132797395.0,
|
| 3092 |
+
"reward": 10.608273446559906,
|
| 3093 |
+
"reward_std": 0.8345479369163513,
|
| 3094 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8834850341081619,
|
| 3095 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.27482672582846135,
|
| 3096 |
+
"rewards/event_reward_fn/mean": 8.8251953125,
|
| 3097 |
+
"rewards/event_reward_fn/std": 5.233703002333641,
|
| 3098 |
+
"rewards/format_reward_fn/mean": 0.8995930962264538,
|
| 3099 |
+
"rewards/format_reward_fn/std": 0.27503635361790657,
|
| 3100 |
+
"step": 1648
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"clip_ratio/high_max": 0.0,
|
| 3104 |
+
"clip_ratio/high_mean": 0.0,
|
| 3105 |
+
"clip_ratio/low_mean": 0.0,
|
| 3106 |
+
"clip_ratio/low_min": 0.0,
|
| 3107 |
+
"clip_ratio/region_mean": 0.0,
|
| 3108 |
+
"completions/clipped_ratio": 0.064453125,
|
| 3109 |
+
"completions/max_length": 254.6875,
|
| 3110 |
+
"completions/max_terminated_length": 241.25,
|
| 3111 |
+
"completions/mean_length": 192.1123046875,
|
| 3112 |
+
"completions/mean_terminated_length": 187.6938066482544,
|
| 3113 |
+
"completions/min_length": 137.3125,
|
| 3114 |
+
"completions/min_terminated_length": 137.3125,
|
| 3115 |
+
"entropy": 0.08001765748485923,
|
| 3116 |
+
"epoch": 1.6171039844509232,
|
| 3117 |
+
"frac_reward_zero_std": 0.33203125,
|
| 3118 |
+
"grad_norm": 0.16833443939685822,
|
| 3119 |
+
"learning_rate": 5e-05,
|
| 3120 |
+
"loss": 0.0008,
|
| 3121 |
+
"num_tokens": 134069966.0,
|
| 3122 |
+
"reward": 11.113677322864532,
|
| 3123 |
+
"reward_std": 0.9350622501224279,
|
| 3124 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9033686555922031,
|
| 3125 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2602922786027193,
|
| 3126 |
+
"rewards/event_reward_fn/mean": 9.2939453125,
|
| 3127 |
+
"rewards/event_reward_fn/std": 5.6752976179122925,
|
| 3128 |
+
"rewards/format_reward_fn/mean": 0.9163634702563286,
|
| 3129 |
+
"rewards/format_reward_fn/std": 0.2618194241076708,
|
| 3130 |
+
"step": 1664
|
| 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": 0.06640625,
|
| 3139 |
+
"completions/max_length": 249.3125,
|
| 3140 |
+
"completions/max_terminated_length": 241.75,
|
| 3141 |
+
"completions/mean_length": 196.7158203125,
|
| 3142 |
+
"completions/mean_terminated_length": 192.56449699401855,
|
| 3143 |
+
"completions/min_length": 129.6875,
|
| 3144 |
+
"completions/min_terminated_length": 129.6875,
|
| 3145 |
+
"entropy": 0.08379031391814351,
|
| 3146 |
+
"epoch": 1.6326530612244898,
|
| 3147 |
+
"frac_reward_zero_std": 0.3125,
|
| 3148 |
+
"grad_norm": 0.14574581384658813,
|
| 3149 |
+
"learning_rate": 5e-05,
|
| 3150 |
+
"loss": -0.0047,
|
| 3151 |
+
"num_tokens": 135435439.0,
|
| 3152 |
+
"reward": 11.539310336112976,
|
| 3153 |
+
"reward_std": 0.9443789459764957,
|
| 3154 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8967322260141373,
|
| 3155 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.22321847162675112,
|
| 3156 |
+
"rewards/event_reward_fn/mean": 9.7314453125,
|
| 3157 |
+
"rewards/event_reward_fn/std": 5.278485506772995,
|
| 3158 |
+
"rewards/format_reward_fn/mean": 0.9111328125,
|
| 3159 |
+
"rewards/format_reward_fn/std": 0.21248832251876593,
|
| 3160 |
+
"step": 1680
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"clip_ratio/high_max": 0.0,
|
| 3164 |
+
"clip_ratio/high_mean": 0.0,
|
| 3165 |
+
"clip_ratio/low_mean": 0.0,
|
| 3166 |
+
"clip_ratio/low_min": 0.0,
|
| 3167 |
+
"clip_ratio/region_mean": 0.0,
|
| 3168 |
+
"completions/clipped_ratio": 0.0791015625,
|
| 3169 |
+
"completions/max_length": 254.8125,
|
| 3170 |
+
"completions/max_terminated_length": 244.8125,
|
| 3171 |
+
"completions/mean_length": 202.83203125,
|
| 3172 |
+
"completions/mean_terminated_length": 198.2507667541504,
|
| 3173 |
+
"completions/min_length": 150.625,
|
| 3174 |
+
"completions/min_terminated_length": 150.625,
|
| 3175 |
+
"entropy": 0.08855495927855372,
|
| 3176 |
+
"epoch": 1.6482021379980565,
|
| 3177 |
+
"frac_reward_zero_std": 0.3046875,
|
| 3178 |
+
"grad_norm": 0.17940281331539154,
|
| 3179 |
+
"learning_rate": 5e-05,
|
| 3180 |
+
"loss": -0.0056,
|
| 3181 |
+
"num_tokens": 136778675.0,
|
| 3182 |
+
"reward": 11.134308993816376,
|
| 3183 |
+
"reward_std": 0.9293302595615387,
|
| 3184 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8743740394711494,
|
| 3185 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.26636734034400433,
|
| 3186 |
+
"rewards/event_reward_fn/mean": 9.3662109375,
|
| 3187 |
+
"rewards/event_reward_fn/std": 5.85838320851326,
|
| 3188 |
+
"rewards/format_reward_fn/mean": 0.8937239646911621,
|
| 3189 |
+
"rewards/format_reward_fn/std": 0.2652863524854183,
|
| 3190 |
+
"step": 1696
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"clip_ratio/high_max": 0.0,
|
| 3194 |
+
"clip_ratio/high_mean": 0.0,
|
| 3195 |
+
"clip_ratio/low_mean": 0.0,
|
| 3196 |
+
"clip_ratio/low_min": 0.0,
|
| 3197 |
+
"clip_ratio/region_mean": 0.0,
|
| 3198 |
+
"completions/clipped_ratio": 0.138671875,
|
| 3199 |
+
"completions/max_length": 256.0,
|
| 3200 |
+
"completions/max_terminated_length": 251.6875,
|
| 3201 |
+
"completions/mean_length": 213.9638671875,
|
| 3202 |
+
"completions/mean_terminated_length": 207.2107219696045,
|
| 3203 |
+
"completions/min_length": 162.375,
|
| 3204 |
+
"completions/min_terminated_length": 162.375,
|
| 3205 |
+
"entropy": 0.09064092021435499,
|
| 3206 |
+
"epoch": 1.663751214771623,
|
| 3207 |
+
"frac_reward_zero_std": 0.3046875,
|
| 3208 |
+
"grad_norm": 0.15384909510612488,
|
| 3209 |
+
"learning_rate": 5e-05,
|
| 3210 |
+
"loss": 0.0029,
|
| 3211 |
+
"num_tokens": 138044578.0,
|
| 3212 |
+
"reward": 11.337530732154846,
|
| 3213 |
+
"reward_std": 0.9400022551417351,
|
| 3214 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.824542474001646,
|
| 3215 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.33465168718248606,
|
| 3216 |
+
"rewards/event_reward_fn/mean": 9.6669921875,
|
| 3217 |
+
"rewards/event_reward_fn/std": 5.503222852945328,
|
| 3218 |
+
"rewards/format_reward_fn/mean": 0.8459960930049419,
|
| 3219 |
+
"rewards/format_reward_fn/std": 0.3366972776129842,
|
| 3220 |
+
"step": 1712
|
| 3221 |
+
},
|
| 3222 |
+
{
|
| 3223 |
+
"clip_ratio/high_max": 0.0,
|
| 3224 |
+
"clip_ratio/high_mean": 0.0,
|
| 3225 |
+
"clip_ratio/low_mean": 0.0,
|
| 3226 |
+
"clip_ratio/low_min": 0.0,
|
| 3227 |
+
"clip_ratio/region_mean": 0.0,
|
| 3228 |
+
"completions/clipped_ratio": 0.046875,
|
| 3229 |
+
"completions/max_length": 251.5625,
|
| 3230 |
+
"completions/max_terminated_length": 245.5625,
|
| 3231 |
+
"completions/mean_length": 203.677734375,
|
| 3232 |
+
"completions/mean_terminated_length": 201.15838241577148,
|
| 3233 |
+
"completions/min_length": 159.1875,
|
| 3234 |
+
"completions/min_terminated_length": 159.1875,
|
| 3235 |
+
"entropy": 0.08698790520429611,
|
| 3236 |
+
"epoch": 1.6793002915451893,
|
| 3237 |
+
"frac_reward_zero_std": 0.296875,
|
| 3238 |
+
"grad_norm": 0.11867301166057587,
|
| 3239 |
+
"learning_rate": 5e-05,
|
| 3240 |
+
"loss": 0.0003,
|
| 3241 |
+
"num_tokens": 139288124.0,
|
| 3242 |
+
"reward": 11.192306399345398,
|
| 3243 |
+
"reward_std": 0.9463471882045269,
|
| 3244 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9286414235830307,
|
| 3245 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.20035810582339764,
|
| 3246 |
+
"rewards/event_reward_fn/mean": 9.3193359375,
|
| 3247 |
+
"rewards/event_reward_fn/std": 5.200570702552795,
|
| 3248 |
+
"rewards/format_reward_fn/mean": 0.9443289637565613,
|
| 3249 |
+
"rewards/format_reward_fn/std": 0.19202105328440666,
|
| 3250 |
+
"step": 1728
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"clip_ratio/high_max": 0.0,
|
| 3254 |
+
"clip_ratio/high_mean": 0.0,
|
| 3255 |
+
"clip_ratio/low_mean": 0.0,
|
| 3256 |
+
"clip_ratio/low_min": 0.0,
|
| 3257 |
+
"clip_ratio/region_mean": 0.0,
|
| 3258 |
+
"completions/clipped_ratio": 0.0556640625,
|
| 3259 |
+
"completions/max_length": 252.125,
|
| 3260 |
+
"completions/max_terminated_length": 244.875,
|
| 3261 |
+
"completions/mean_length": 193.6708984375,
|
| 3262 |
+
"completions/mean_terminated_length": 189.9529905319214,
|
| 3263 |
+
"completions/min_length": 136.625,
|
| 3264 |
+
"completions/min_terminated_length": 136.625,
|
| 3265 |
+
"entropy": 0.08362232241779566,
|
| 3266 |
+
"epoch": 1.694849368318756,
|
| 3267 |
+
"frac_reward_zero_std": 0.33203125,
|
| 3268 |
+
"grad_norm": 0.11613152176141739,
|
| 3269 |
+
"learning_rate": 5e-05,
|
| 3270 |
+
"loss": -0.0019,
|
| 3271 |
+
"num_tokens": 140615163.0,
|
| 3272 |
+
"reward": 11.211718916893005,
|
| 3273 |
+
"reward_std": 0.8285622540861368,
|
| 3274 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9016408734023571,
|
| 3275 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.25260637141764164,
|
| 3276 |
+
"rewards/event_reward_fn/mean": 9.390625,
|
| 3277 |
+
"rewards/event_reward_fn/std": 5.310590535402298,
|
| 3278 |
+
"rewards/format_reward_fn/mean": 0.9194531291723251,
|
| 3279 |
+
"rewards/format_reward_fn/std": 0.251515906304121,
|
| 3280 |
+
"step": 1744
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"clip_ratio/high_max": 0.0,
|
| 3284 |
+
"clip_ratio/high_mean": 0.0,
|
| 3285 |
+
"clip_ratio/low_mean": 0.0,
|
| 3286 |
+
"clip_ratio/low_min": 0.0,
|
| 3287 |
+
"clip_ratio/region_mean": 0.0,
|
| 3288 |
+
"completions/clipped_ratio": 0.0751953125,
|
| 3289 |
+
"completions/max_length": 248.6875,
|
| 3290 |
+
"completions/max_terminated_length": 240.3125,
|
| 3291 |
+
"completions/mean_length": 194.3291015625,
|
| 3292 |
+
"completions/mean_terminated_length": 189.4942626953125,
|
| 3293 |
+
"completions/min_length": 139.25,
|
| 3294 |
+
"completions/min_terminated_length": 139.25,
|
| 3295 |
+
"entropy": 0.08920921664685011,
|
| 3296 |
+
"epoch": 1.7103984450923226,
|
| 3297 |
+
"frac_reward_zero_std": 0.265625,
|
| 3298 |
+
"grad_norm": 0.1495039016008377,
|
| 3299 |
+
"learning_rate": 5e-05,
|
| 3300 |
+
"loss": -0.0026,
|
| 3301 |
+
"num_tokens": 141995908.0,
|
| 3302 |
+
"reward": 11.331986844539642,
|
| 3303 |
+
"reward_std": 0.9946209099143744,
|
| 3304 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8805676624178886,
|
| 3305 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2605485112289898,
|
| 3306 |
+
"rewards/event_reward_fn/mean": 9.5615234375,
|
| 3307 |
+
"rewards/event_reward_fn/std": 5.626507669687271,
|
| 3308 |
+
"rewards/format_reward_fn/mean": 0.8898958377540112,
|
| 3309 |
+
"rewards/format_reward_fn/std": 0.25742682348936796,
|
| 3310 |
+
"step": 1760
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"clip_ratio/high_max": 0.0,
|
| 3314 |
+
"clip_ratio/high_mean": 0.0,
|
| 3315 |
+
"clip_ratio/low_mean": 0.0,
|
| 3316 |
+
"clip_ratio/low_min": 0.0,
|
| 3317 |
+
"clip_ratio/region_mean": 0.0,
|
| 3318 |
+
"completions/clipped_ratio": 0.0556640625,
|
| 3319 |
+
"completions/max_length": 250.625,
|
| 3320 |
+
"completions/max_terminated_length": 242.25,
|
| 3321 |
+
"completions/mean_length": 194.5751953125,
|
| 3322 |
+
"completions/mean_terminated_length": 190.86159992218018,
|
| 3323 |
+
"completions/min_length": 141.5625,
|
| 3324 |
+
"completions/min_terminated_length": 141.5625,
|
| 3325 |
+
"entropy": 0.09618484182283282,
|
| 3326 |
+
"epoch": 1.7259475218658893,
|
| 3327 |
+
"frac_reward_zero_std": 0.26953125,
|
| 3328 |
+
"grad_norm": 0.20417290925979614,
|
| 3329 |
+
"learning_rate": 5e-05,
|
| 3330 |
+
"loss": -0.0055,
|
| 3331 |
+
"num_tokens": 143301673.0,
|
| 3332 |
+
"reward": 10.538148939609528,
|
| 3333 |
+
"reward_std": 0.9361699968576431,
|
| 3334 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9052551127970219,
|
| 3335 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2257093784864992,
|
| 3336 |
+
"rewards/event_reward_fn/mean": 8.7158203125,
|
| 3337 |
+
"rewards/event_reward_fn/std": 4.607826009392738,
|
| 3338 |
+
"rewards/format_reward_fn/mean": 0.9170735664665699,
|
| 3339 |
+
"rewards/format_reward_fn/std": 0.22840105323120952,
|
| 3340 |
+
"step": 1776
|
| 3341 |
+
},
|
| 3342 |
+
{
|
| 3343 |
+
"clip_ratio/high_max": 0.0,
|
| 3344 |
+
"clip_ratio/high_mean": 0.0,
|
| 3345 |
+
"clip_ratio/low_mean": 0.0,
|
| 3346 |
+
"clip_ratio/low_min": 0.0,
|
| 3347 |
+
"clip_ratio/region_mean": 0.0,
|
| 3348 |
+
"completions/clipped_ratio": 0.0576171875,
|
| 3349 |
+
"completions/max_length": 252.8125,
|
| 3350 |
+
"completions/max_terminated_length": 244.4375,
|
| 3351 |
+
"completions/mean_length": 200.8662109375,
|
| 3352 |
+
"completions/mean_terminated_length": 197.39703178405762,
|
| 3353 |
+
"completions/min_length": 149.9375,
|
| 3354 |
+
"completions/min_terminated_length": 149.9375,
|
| 3355 |
+
"entropy": 0.08653424866497517,
|
| 3356 |
+
"epoch": 1.741496598639456,
|
| 3357 |
+
"frac_reward_zero_std": 0.3359375,
|
| 3358 |
+
"grad_norm": 0.14243784546852112,
|
| 3359 |
+
"learning_rate": 5e-05,
|
| 3360 |
+
"loss": 0.0011,
|
| 3361 |
+
"num_tokens": 144603412.0,
|
| 3362 |
+
"reward": 11.493825078010559,
|
| 3363 |
+
"reward_std": 0.8755283299833536,
|
| 3364 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9135190099477768,
|
| 3365 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24690337451465894,
|
| 3366 |
+
"rewards/event_reward_fn/mean": 9.658203125,
|
| 3367 |
+
"rewards/event_reward_fn/std": 5.445283606648445,
|
| 3368 |
+
"rewards/format_reward_fn/mean": 0.922102864831686,
|
| 3369 |
+
"rewards/format_reward_fn/std": 0.2481938637793064,
|
| 3370 |
+
"step": 1792
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"clip_ratio/high_max": 0.0,
|
| 3374 |
+
"clip_ratio/high_mean": 0.0,
|
| 3375 |
+
"clip_ratio/low_mean": 0.0,
|
| 3376 |
+
"clip_ratio/low_min": 0.0,
|
| 3377 |
+
"clip_ratio/region_mean": 0.0,
|
| 3378 |
+
"completions/clipped_ratio": 0.0556640625,
|
| 3379 |
+
"completions/max_length": 252.0625,
|
| 3380 |
+
"completions/max_terminated_length": 246.75,
|
| 3381 |
+
"completions/mean_length": 196.5146484375,
|
| 3382 |
+
"completions/mean_terminated_length": 193.12859344482422,
|
| 3383 |
+
"completions/min_length": 140.8125,
|
| 3384 |
+
"completions/min_terminated_length": 140.8125,
|
| 3385 |
+
"entropy": 0.08316960139200091,
|
| 3386 |
+
"epoch": 1.7570456754130224,
|
| 3387 |
+
"frac_reward_zero_std": 0.3125,
|
| 3388 |
+
"grad_norm": 0.109793521463871,
|
| 3389 |
+
"learning_rate": 5e-05,
|
| 3390 |
+
"loss": -0.0022,
|
| 3391 |
+
"num_tokens": 145919099.0,
|
| 3392 |
+
"reward": 11.740033328533173,
|
| 3393 |
+
"reward_std": 0.9224549978971481,
|
| 3394 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9150333367288113,
|
| 3395 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24498367216438055,
|
| 3396 |
+
"rewards/event_reward_fn/mean": 9.90234375,
|
| 3397 |
+
"rewards/event_reward_fn/std": 5.425331294536591,
|
| 3398 |
+
"rewards/format_reward_fn/mean": 0.9226562492549419,
|
| 3399 |
+
"rewards/format_reward_fn/std": 0.24409929476678371,
|
| 3400 |
+
"step": 1808
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"clip_ratio/high_max": 0.0,
|
| 3404 |
+
"clip_ratio/high_mean": 0.0,
|
| 3405 |
+
"clip_ratio/low_mean": 0.0,
|
| 3406 |
+
"clip_ratio/low_min": 0.0,
|
| 3407 |
+
"clip_ratio/region_mean": 0.0,
|
| 3408 |
+
"completions/clipped_ratio": 0.072265625,
|
| 3409 |
+
"completions/max_length": 254.75,
|
| 3410 |
+
"completions/max_terminated_length": 249.3125,
|
| 3411 |
+
"completions/mean_length": 208.4755859375,
|
| 3412 |
+
"completions/mean_terminated_length": 204.96116065979004,
|
| 3413 |
+
"completions/min_length": 160.5625,
|
| 3414 |
+
"completions/min_terminated_length": 160.5625,
|
| 3415 |
+
"entropy": 0.08932856796309352,
|
| 3416 |
+
"epoch": 1.7725947521865888,
|
| 3417 |
+
"frac_reward_zero_std": 0.3359375,
|
| 3418 |
+
"grad_norm": 0.24029314517974854,
|
| 3419 |
+
"learning_rate": 5e-05,
|
| 3420 |
+
"loss": 0.0045,
|
| 3421 |
+
"num_tokens": 147275350.0,
|
| 3422 |
+
"reward": 11.133660674095154,
|
| 3423 |
+
"reward_std": 1.0420608818531036,
|
| 3424 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8887974470853806,
|
| 3425 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2559172356268391,
|
| 3426 |
+
"rewards/event_reward_fn/mean": 9.34375,
|
| 3427 |
+
"rewards/event_reward_fn/std": 5.51551166176796,
|
| 3428 |
+
"rewards/format_reward_fn/mean": 0.9011132828891277,
|
| 3429 |
+
"rewards/format_reward_fn/std": 0.2507179146632552,
|
| 3430 |
+
"step": 1824
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"clip_ratio/high_max": 0.0,
|
| 3434 |
+
"clip_ratio/high_mean": 0.0,
|
| 3435 |
+
"clip_ratio/low_mean": 0.0,
|
| 3436 |
+
"clip_ratio/low_min": 0.0,
|
| 3437 |
+
"clip_ratio/region_mean": 0.0,
|
| 3438 |
+
"completions/clipped_ratio": 0.046875,
|
| 3439 |
+
"completions/max_length": 252.375,
|
| 3440 |
+
"completions/max_terminated_length": 240.875,
|
| 3441 |
+
"completions/mean_length": 198.576171875,
|
| 3442 |
+
"completions/mean_terminated_length": 195.7430601119995,
|
| 3443 |
+
"completions/min_length": 151.25,
|
| 3444 |
+
"completions/min_terminated_length": 151.25,
|
| 3445 |
+
"entropy": 0.08435806119814515,
|
| 3446 |
+
"epoch": 1.7881438289601554,
|
| 3447 |
+
"frac_reward_zero_std": 0.36328125,
|
| 3448 |
+
"grad_norm": 0.13869501650333405,
|
| 3449 |
+
"learning_rate": 5e-05,
|
| 3450 |
+
"loss": -0.0026,
|
| 3451 |
+
"num_tokens": 148539184.0,
|
| 3452 |
+
"reward": 11.031599402427673,
|
| 3453 |
+
"reward_std": 0.7965468689799309,
|
| 3454 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9208246804773808,
|
| 3455 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2361440734239295,
|
| 3456 |
+
"rewards/event_reward_fn/mean": 9.1796875,
|
| 3457 |
+
"rewards/event_reward_fn/std": 4.907300844788551,
|
| 3458 |
+
"rewards/format_reward_fn/mean": 0.9310872405767441,
|
| 3459 |
+
"rewards/format_reward_fn/std": 0.2394925099797547,
|
| 3460 |
+
"step": 1840
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"clip_ratio/high_max": 0.0,
|
| 3464 |
+
"clip_ratio/high_mean": 0.0,
|
| 3465 |
+
"clip_ratio/low_mean": 0.0,
|
| 3466 |
+
"clip_ratio/low_min": 0.0,
|
| 3467 |
+
"clip_ratio/region_mean": 0.0,
|
| 3468 |
+
"completions/clipped_ratio": 0.0380859375,
|
| 3469 |
+
"completions/max_length": 252.5,
|
| 3470 |
+
"completions/max_terminated_length": 250.125,
|
| 3471 |
+
"completions/mean_length": 202.90234375,
|
| 3472 |
+
"completions/mean_terminated_length": 200.88229370117188,
|
| 3473 |
+
"completions/min_length": 155.875,
|
| 3474 |
+
"completions/min_terminated_length": 155.875,
|
| 3475 |
+
"entropy": 0.08052209811285138,
|
| 3476 |
+
"epoch": 1.803692905733722,
|
| 3477 |
+
"frac_reward_zero_std": 0.35546875,
|
| 3478 |
+
"grad_norm": 0.1878909021615982,
|
| 3479 |
+
"learning_rate": 5e-05,
|
| 3480 |
+
"loss": 0.001,
|
| 3481 |
+
"num_tokens": 149840940.0,
|
| 3482 |
+
"reward": 10.959127485752106,
|
| 3483 |
+
"reward_std": 0.9578492008149624,
|
| 3484 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9300258904695511,
|
| 3485 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.18775073438882828,
|
| 3486 |
+
"rewards/event_reward_fn/mean": 9.0859375,
|
| 3487 |
+
"rewards/event_reward_fn/std": 5.149698540568352,
|
| 3488 |
+
"rewards/format_reward_fn/mean": 0.9431640617549419,
|
| 3489 |
+
"rewards/format_reward_fn/std": 0.17401384096592665,
|
| 3490 |
+
"step": 1856
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"clip_ratio/high_max": 0.0,
|
| 3494 |
+
"clip_ratio/high_mean": 0.0,
|
| 3495 |
+
"clip_ratio/low_mean": 0.0,
|
| 3496 |
+
"clip_ratio/low_min": 0.0,
|
| 3497 |
+
"clip_ratio/region_mean": 0.0,
|
| 3498 |
+
"completions/clipped_ratio": 0.09765625,
|
| 3499 |
+
"completions/max_length": 255.9375,
|
| 3500 |
+
"completions/max_terminated_length": 248.875,
|
| 3501 |
+
"completions/mean_length": 211.8564453125,
|
| 3502 |
+
"completions/mean_terminated_length": 207.1767454147339,
|
| 3503 |
+
"completions/min_length": 164.5625,
|
| 3504 |
+
"completions/min_terminated_length": 164.5625,
|
| 3505 |
+
"entropy": 0.08094025542959571,
|
| 3506 |
+
"epoch": 1.8192419825072887,
|
| 3507 |
+
"frac_reward_zero_std": 0.34765625,
|
| 3508 |
+
"grad_norm": 0.14807139337062836,
|
| 3509 |
+
"learning_rate": 5e-05,
|
| 3510 |
+
"loss": 0.0023,
|
| 3511 |
+
"num_tokens": 151221749.0,
|
| 3512 |
+
"reward": 11.752990126609802,
|
| 3513 |
+
"reward_std": 0.9537594802677631,
|
| 3514 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8724758252501488,
|
| 3515 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.29270493309013546,
|
| 3516 |
+
"rewards/event_reward_fn/mean": 9.9970703125,
|
| 3517 |
+
"rewards/event_reward_fn/std": 5.857491314411163,
|
| 3518 |
+
"rewards/format_reward_fn/mean": 0.8834440112113953,
|
| 3519 |
+
"rewards/format_reward_fn/std": 0.29244135320186615,
|
| 3520 |
+
"step": 1872
|
| 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": 0.0546875,
|
| 3529 |
+
"completions/max_length": 254.3125,
|
| 3530 |
+
"completions/max_terminated_length": 250.25,
|
| 3531 |
+
"completions/mean_length": 208.7001953125,
|
| 3532 |
+
"completions/mean_terminated_length": 206.05935287475586,
|
| 3533 |
+
"completions/min_length": 159.125,
|
| 3534 |
+
"completions/min_terminated_length": 159.125,
|
| 3535 |
+
"entropy": 0.08766834484413266,
|
| 3536 |
+
"epoch": 1.8347910592808552,
|
| 3537 |
+
"frac_reward_zero_std": 0.34375,
|
| 3538 |
+
"grad_norm": 0.17317424714565277,
|
| 3539 |
+
"learning_rate": 5e-05,
|
| 3540 |
+
"loss": -0.0004,
|
| 3541 |
+
"num_tokens": 152521294.0,
|
| 3542 |
+
"reward": 11.364756107330322,
|
| 3543 |
+
"reward_std": 0.9098326228559017,
|
| 3544 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8889735676348209,
|
| 3545 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2667266938369721,
|
| 3546 |
+
"rewards/event_reward_fn/mean": 9.568359375,
|
| 3547 |
+
"rewards/event_reward_fn/std": 5.424193903803825,
|
| 3548 |
+
"rewards/format_reward_fn/mean": 0.9074231162667274,
|
| 3549 |
+
"rewards/format_reward_fn/std": 0.2601332040503621,
|
| 3550 |
+
"step": 1888
|
| 3551 |
+
},
|
| 3552 |
+
{
|
| 3553 |
+
"clip_ratio/high_max": 0.0,
|
| 3554 |
+
"clip_ratio/high_mean": 0.0,
|
| 3555 |
+
"clip_ratio/low_mean": 0.0,
|
| 3556 |
+
"clip_ratio/low_min": 0.0,
|
| 3557 |
+
"clip_ratio/region_mean": 0.0,
|
| 3558 |
+
"completions/clipped_ratio": 0.0478515625,
|
| 3559 |
+
"completions/max_length": 253.1875,
|
| 3560 |
+
"completions/max_terminated_length": 248.625,
|
| 3561 |
+
"completions/mean_length": 205.2333984375,
|
| 3562 |
+
"completions/mean_terminated_length": 202.61692428588867,
|
| 3563 |
+
"completions/min_length": 158.0625,
|
| 3564 |
+
"completions/min_terminated_length": 158.0625,
|
| 3565 |
+
"entropy": 0.0873062857426703,
|
| 3566 |
+
"epoch": 1.8503401360544216,
|
| 3567 |
+
"frac_reward_zero_std": 0.3359375,
|
| 3568 |
+
"grad_norm": 0.16510100662708282,
|
| 3569 |
+
"learning_rate": 5e-05,
|
| 3570 |
+
"loss": 0.0008,
|
| 3571 |
+
"num_tokens": 153814765.0,
|
| 3572 |
+
"reward": 10.840591430664062,
|
| 3573 |
+
"reward_std": 0.806601133197546,
|
| 3574 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8800444230437279,
|
| 3575 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2592724412679672,
|
| 3576 |
+
"rewards/event_reward_fn/mean": 9.05078125,
|
| 3577 |
+
"rewards/event_reward_fn/std": 5.017846331000328,
|
| 3578 |
+
"rewards/format_reward_fn/mean": 0.9097656235098839,
|
| 3579 |
+
"rewards/format_reward_fn/std": 0.2350642140954733,
|
| 3580 |
+
"step": 1904
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"clip_ratio/high_max": 0.0,
|
| 3584 |
+
"clip_ratio/high_mean": 0.0,
|
| 3585 |
+
"clip_ratio/low_mean": 0.0,
|
| 3586 |
+
"clip_ratio/low_min": 0.0,
|
| 3587 |
+
"clip_ratio/region_mean": 0.0,
|
| 3588 |
+
"completions/clipped_ratio": 0.0478515625,
|
| 3589 |
+
"completions/max_length": 250.6875,
|
| 3590 |
+
"completions/max_terminated_length": 246.625,
|
| 3591 |
+
"completions/mean_length": 199.6025390625,
|
| 3592 |
+
"completions/mean_terminated_length": 196.8941469192505,
|
| 3593 |
+
"completions/min_length": 146.75,
|
| 3594 |
+
"completions/min_terminated_length": 146.75,
|
| 3595 |
+
"entropy": 0.08616631478071213,
|
| 3596 |
+
"epoch": 1.8658892128279883,
|
| 3597 |
+
"frac_reward_zero_std": 0.328125,
|
| 3598 |
+
"grad_norm": 0.10711020976305008,
|
| 3599 |
+
"learning_rate": 5e-05,
|
| 3600 |
+
"loss": 0.003,
|
| 3601 |
+
"num_tokens": 155159530.0,
|
| 3602 |
+
"reward": 11.430678129196167,
|
| 3603 |
+
"reward_std": 0.7845460455864668,
|
| 3604 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9038552716374397,
|
| 3605 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.21465440141037107,
|
| 3606 |
+
"rewards/event_reward_fn/mean": 9.5986328125,
|
| 3607 |
+
"rewards/event_reward_fn/std": 5.15682627260685,
|
| 3608 |
+
"rewards/format_reward_fn/mean": 0.928190104663372,
|
| 3609 |
+
"rewards/format_reward_fn/std": 0.19767758785746992,
|
| 3610 |
+
"step": 1920
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"clip_ratio/high_max": 0.0,
|
| 3614 |
+
"clip_ratio/high_mean": 0.0,
|
| 3615 |
+
"clip_ratio/low_mean": 0.0,
|
| 3616 |
+
"clip_ratio/low_min": 0.0,
|
| 3617 |
+
"clip_ratio/region_mean": 0.0,
|
| 3618 |
+
"completions/clipped_ratio": 0.0625,
|
| 3619 |
+
"completions/max_length": 253.1875,
|
| 3620 |
+
"completions/max_terminated_length": 246.6875,
|
| 3621 |
+
"completions/mean_length": 200.734375,
|
| 3622 |
+
"completions/mean_terminated_length": 197.25225925445557,
|
| 3623 |
+
"completions/min_length": 150.1875,
|
| 3624 |
+
"completions/min_terminated_length": 150.1875,
|
| 3625 |
+
"entropy": 0.08771243086084723,
|
| 3626 |
+
"epoch": 1.881438289601555,
|
| 3627 |
+
"frac_reward_zero_std": 0.27734375,
|
| 3628 |
+
"grad_norm": 0.119595006108284,
|
| 3629 |
+
"learning_rate": 5e-05,
|
| 3630 |
+
"loss": -0.001,
|
| 3631 |
+
"num_tokens": 156525614.0,
|
| 3632 |
+
"reward": 11.453014373779297,
|
| 3633 |
+
"reward_std": 1.111331295222044,
|
| 3634 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.8604980707168579,
|
| 3635 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2871107269311324,
|
| 3636 |
+
"rewards/event_reward_fn/mean": 9.708984375,
|
| 3637 |
+
"rewards/event_reward_fn/std": 5.215842500329018,
|
| 3638 |
+
"rewards/format_reward_fn/mean": 0.8835319019854069,
|
| 3639 |
+
"rewards/format_reward_fn/std": 0.2829501121304929,
|
| 3640 |
+
"step": 1936
|
| 3641 |
+
},
|
| 3642 |
+
{
|
| 3643 |
+
"clip_ratio/high_max": 0.0,
|
| 3644 |
+
"clip_ratio/high_mean": 0.0,
|
| 3645 |
+
"clip_ratio/low_mean": 0.0,
|
| 3646 |
+
"clip_ratio/low_min": 0.0,
|
| 3647 |
+
"clip_ratio/region_mean": 0.0,
|
| 3648 |
+
"completions/clipped_ratio": 0.021484375,
|
| 3649 |
+
"completions/max_length": 250.25,
|
| 3650 |
+
"completions/max_terminated_length": 245.8125,
|
| 3651 |
+
"completions/mean_length": 194.017578125,
|
| 3652 |
+
"completions/mean_terminated_length": 192.65657711029053,
|
| 3653 |
+
"completions/min_length": 140.6875,
|
| 3654 |
+
"completions/min_terminated_length": 140.6875,
|
| 3655 |
+
"entropy": 0.08764936728402972,
|
| 3656 |
+
"epoch": 1.8969873663751216,
|
| 3657 |
+
"frac_reward_zero_std": 0.28515625,
|
| 3658 |
+
"grad_norm": 0.1511303335428238,
|
| 3659 |
+
"learning_rate": 5e-05,
|
| 3660 |
+
"loss": 0.0032,
|
| 3661 |
+
"num_tokens": 157786400.0,
|
| 3662 |
+
"reward": 10.632731199264526,
|
| 3663 |
+
"reward_std": 0.9243863355368376,
|
| 3664 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9324383623898029,
|
| 3665 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.18536719167605042,
|
| 3666 |
+
"rewards/event_reward_fn/mean": 8.7451171875,
|
| 3667 |
+
"rewards/event_reward_fn/std": 5.235057607293129,
|
| 3668 |
+
"rewards/format_reward_fn/mean": 0.9551757834851742,
|
| 3669 |
+
"rewards/format_reward_fn/std": 0.16272677155211568,
|
| 3670 |
+
"step": 1952
|
| 3671 |
+
},
|
| 3672 |
+
{
|
| 3673 |
+
"clip_ratio/high_max": 0.0,
|
| 3674 |
+
"clip_ratio/high_mean": 0.0,
|
| 3675 |
+
"clip_ratio/low_mean": 0.0,
|
| 3676 |
+
"clip_ratio/low_min": 0.0,
|
| 3677 |
+
"clip_ratio/region_mean": 0.0,
|
| 3678 |
+
"completions/clipped_ratio": 0.046875,
|
| 3679 |
+
"completions/max_length": 254.125,
|
| 3680 |
+
"completions/max_terminated_length": 250.6875,
|
| 3681 |
+
"completions/mean_length": 200.95703125,
|
| 3682 |
+
"completions/mean_terminated_length": 198.30670166015625,
|
| 3683 |
+
"completions/min_length": 154.6875,
|
| 3684 |
+
"completions/min_terminated_length": 154.6875,
|
| 3685 |
+
"entropy": 0.08874167408794165,
|
| 3686 |
+
"epoch": 1.9125364431486882,
|
| 3687 |
+
"frac_reward_zero_std": 0.33984375,
|
| 3688 |
+
"grad_norm": 0.17749741673469543,
|
| 3689 |
+
"learning_rate": 5e-05,
|
| 3690 |
+
"loss": 0.0037,
|
| 3691 |
+
"num_tokens": 159093888.0,
|
| 3692 |
+
"reward": 11.275705397129059,
|
| 3693 |
+
"reward_std": 0.856599148362875,
|
| 3694 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9131891131401062,
|
| 3695 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.2150915495294612,
|
| 3696 |
+
"rewards/event_reward_fn/mean": 9.4345703125,
|
| 3697 |
+
"rewards/event_reward_fn/std": 5.729633465409279,
|
| 3698 |
+
"rewards/format_reward_fn/mean": 0.927945964038372,
|
| 3699 |
+
"rewards/format_reward_fn/std": 0.2096583191305399,
|
| 3700 |
+
"step": 1968
|
| 3701 |
+
},
|
| 3702 |
+
{
|
| 3703 |
+
"clip_ratio/high_max": 0.0,
|
| 3704 |
+
"clip_ratio/high_mean": 0.0,
|
| 3705 |
+
"clip_ratio/low_mean": 0.0,
|
| 3706 |
+
"clip_ratio/low_min": 0.0,
|
| 3707 |
+
"clip_ratio/region_mean": 0.0,
|
| 3708 |
+
"completions/clipped_ratio": 0.041015625,
|
| 3709 |
+
"completions/max_length": 253.875,
|
| 3710 |
+
"completions/max_terminated_length": 246.5625,
|
| 3711 |
+
"completions/mean_length": 200.9677734375,
|
| 3712 |
+
"completions/mean_terminated_length": 198.57855701446533,
|
| 3713 |
+
"completions/min_length": 142.8125,
|
| 3714 |
+
"completions/min_terminated_length": 142.8125,
|
| 3715 |
+
"entropy": 0.0850910097360611,
|
| 3716 |
+
"epoch": 1.9280855199222546,
|
| 3717 |
+
"frac_reward_zero_std": 0.375,
|
| 3718 |
+
"grad_norm": 0.12046821415424347,
|
| 3719 |
+
"learning_rate": 5e-05,
|
| 3720 |
+
"loss": 0.0007,
|
| 3721 |
+
"num_tokens": 160340479.0,
|
| 3722 |
+
"reward": 10.719317555427551,
|
| 3723 |
+
"reward_std": 0.8128865994513035,
|
| 3724 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.9061989188194275,
|
| 3725 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.24043723253998905,
|
| 3726 |
+
"rewards/event_reward_fn/mean": 8.892578125,
|
| 3727 |
+
"rewards/event_reward_fn/std": 5.485840782523155,
|
| 3728 |
+
"rewards/format_reward_fn/mean": 0.9205403625965118,
|
| 3729 |
+
"rewards/format_reward_fn/std": 0.2409290496725589,
|
| 3730 |
+
"step": 1984
|
| 3731 |
+
},
|
| 3732 |
+
{
|
| 3733 |
+
"clip_ratio/high_max": 0.0,
|
| 3734 |
+
"clip_ratio/high_mean": 0.0,
|
| 3735 |
+
"clip_ratio/low_mean": 0.0,
|
| 3736 |
+
"clip_ratio/low_min": 0.0,
|
| 3737 |
+
"clip_ratio/region_mean": 0.0,
|
| 3738 |
+
"completions/clipped_ratio": 0.12109375,
|
| 3739 |
+
"completions/max_length": 256.0,
|
| 3740 |
+
"completions/max_terminated_length": 251.125,
|
| 3741 |
+
"completions/mean_length": 213.712890625,
|
| 3742 |
+
"completions/mean_terminated_length": 207.85150337219238,
|
| 3743 |
+
"completions/min_length": 160.9375,
|
| 3744 |
+
"completions/min_terminated_length": 160.9375,
|
| 3745 |
+
"entropy": 0.08267078269273043,
|
| 3746 |
+
"epoch": 1.943634596695821,
|
| 3747 |
+
"frac_reward_zero_std": 0.32421875,
|
| 3748 |
+
"grad_norm": 0.09311431646347046,
|
| 3749 |
+
"learning_rate": 5e-05,
|
| 3750 |
+
"loss": 0.0044,
|
| 3751 |
+
"num_tokens": 161744217.0,
|
| 3752 |
+
"reward": 10.99679410457611,
|
| 3753 |
+
"reward_std": 0.9773008767515421,
|
| 3754 |
+
"rewards/bm25_retrieval_reward_fn/mean": 0.856908455491066,
|
| 3755 |
+
"rewards/bm25_retrieval_reward_fn/std": 0.3204036271199584,
|
| 3756 |
+
"rewards/event_reward_fn/mean": 9.2744140625,
|
| 3757 |
+
"rewards/event_reward_fn/std": 5.77374792098999,
|
| 3758 |
+
"rewards/format_reward_fn/mean": 0.8654715418815613,
|
| 3759 |
+
"rewards/format_reward_fn/std": 0.3263047467917204,
|
| 3760 |
+
"step": 2000
|
| 3761 |
}
|
| 3762 |
],
|
| 3763 |
"logging_steps": 16,
|
| 3764 |
"max_steps": 10290,
|
| 3765 |
+
"num_input_tokens_seen": 161744217,
|
| 3766 |
"num_train_epochs": 10,
|
| 3767 |
"save_steps": 500,
|
| 3768 |
"stateful_callbacks": {
|