Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +1906 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,1906 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9981378026070763,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 134,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"grad_norm": 20.656896651231627,
|
| 14 |
+
"learning_rate": 3.5714285714285716e-07,
|
| 15 |
+
"loss": 3.6136,
|
| 16 |
+
"step": 1,
|
| 17 |
+
"trainloss/critic_chosen": 1.459133505821228,
|
| 18 |
+
"trainloss/critic_rejected": 1.468864917755127,
|
| 19 |
+
"trainloss/reward": 1.459133505821228,
|
| 20 |
+
"trainrewards/accuracies": 0.5833333134651184,
|
| 21 |
+
"trainrewards/chosen": 0.3359375,
|
| 22 |
+
"trainrewards/margins": 0.0308837890625,
|
| 23 |
+
"trainrewards/rejected": 0.3046875
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"grad_norm": 20.747827742934472,
|
| 28 |
+
"learning_rate": 7.142857142857143e-07,
|
| 29 |
+
"loss": 3.6381,
|
| 30 |
+
"step": 2,
|
| 31 |
+
"trainloss/critic_chosen": 1.4447739124298096,
|
| 32 |
+
"trainloss/critic_rejected": 1.4999535083770752,
|
| 33 |
+
"trainloss/reward": 1.4447739124298096,
|
| 34 |
+
"trainrewards/accuracies": 0.5104166865348816,
|
| 35 |
+
"trainrewards/chosen": 0.314453125,
|
| 36 |
+
"trainrewards/margins": 0.01531982421875,
|
| 37 |
+
"trainrewards/rejected": 0.298828125
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"grad_norm": 19.429850511676147,
|
| 42 |
+
"learning_rate": 1.0714285714285714e-06,
|
| 43 |
+
"loss": 3.6713,
|
| 44 |
+
"step": 3,
|
| 45 |
+
"trainloss/critic_chosen": 1.4738179445266724,
|
| 46 |
+
"trainloss/critic_rejected": 1.505049228668213,
|
| 47 |
+
"trainloss/reward": 1.4738179445266724,
|
| 48 |
+
"trainrewards/accuracies": 0.5364583134651184,
|
| 49 |
+
"trainrewards/chosen": 0.302734375,
|
| 50 |
+
"trainrewards/margins": 0.015869140625,
|
| 51 |
+
"trainrewards/rejected": 0.287109375
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"grad_norm": 19.353394477551088,
|
| 56 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 57 |
+
"loss": 3.6414,
|
| 58 |
+
"step": 4,
|
| 59 |
+
"trainloss/critic_chosen": 1.4593632221221924,
|
| 60 |
+
"trainloss/critic_rejected": 1.4944238662719727,
|
| 61 |
+
"trainloss/reward": 1.4593632221221924,
|
| 62 |
+
"trainrewards/accuracies": 0.5572916865348816,
|
| 63 |
+
"trainrewards/chosen": 0.357421875,
|
| 64 |
+
"trainrewards/margins": 0.0286865234375,
|
| 65 |
+
"trainrewards/rejected": 0.328125
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.04,
|
| 69 |
+
"grad_norm": 19.314515274756275,
|
| 70 |
+
"learning_rate": 1.7857142857142859e-06,
|
| 71 |
+
"loss": 3.6014,
|
| 72 |
+
"step": 5,
|
| 73 |
+
"trainloss/critic_chosen": 1.4528993368148804,
|
| 74 |
+
"trainloss/critic_rejected": 1.5066261291503906,
|
| 75 |
+
"trainloss/reward": 1.4528993368148804,
|
| 76 |
+
"trainrewards/accuracies": 0.7760417461395264,
|
| 77 |
+
"trainrewards/chosen": 0.494140625,
|
| 78 |
+
"trainrewards/margins": 0.130859375,
|
| 79 |
+
"trainrewards/rejected": 0.36328125
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.04,
|
| 83 |
+
"grad_norm": 17.65488892011503,
|
| 84 |
+
"learning_rate": 2.1428571428571427e-06,
|
| 85 |
+
"loss": 3.5132,
|
| 86 |
+
"step": 6,
|
| 87 |
+
"trainloss/critic_chosen": 1.4163868427276611,
|
| 88 |
+
"trainloss/critic_rejected": 1.4701489210128784,
|
| 89 |
+
"trainloss/reward": 1.4163868427276611,
|
| 90 |
+
"trainrewards/accuracies": 0.7864583730697632,
|
| 91 |
+
"trainrewards/chosen": 0.54296875,
|
| 92 |
+
"trainrewards/margins": 0.169921875,
|
| 93 |
+
"trainrewards/rejected": 0.373046875
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.05,
|
| 97 |
+
"grad_norm": 14.922878013160528,
|
| 98 |
+
"learning_rate": 2.5e-06,
|
| 99 |
+
"loss": 3.4609,
|
| 100 |
+
"step": 7,
|
| 101 |
+
"trainloss/critic_chosen": 1.4025382995605469,
|
| 102 |
+
"trainloss/critic_rejected": 1.4922353029251099,
|
| 103 |
+
"trainloss/reward": 1.4025382995605469,
|
| 104 |
+
"trainrewards/accuracies": 0.8281250596046448,
|
| 105 |
+
"trainrewards/chosen": 0.9453125,
|
| 106 |
+
"trainrewards/margins": 0.376953125,
|
| 107 |
+
"trainrewards/rejected": 0.56640625
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.06,
|
| 111 |
+
"grad_norm": 14.232594419094823,
|
| 112 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 113 |
+
"loss": 3.3368,
|
| 114 |
+
"step": 8,
|
| 115 |
+
"trainloss/critic_chosen": 1.3685592412948608,
|
| 116 |
+
"trainloss/critic_rejected": 1.4283447265625,
|
| 117 |
+
"trainloss/reward": 1.3685592412948608,
|
| 118 |
+
"trainrewards/accuracies": 0.8645833730697632,
|
| 119 |
+
"trainrewards/chosen": 1.0078125,
|
| 120 |
+
"trainrewards/margins": 0.47265625,
|
| 121 |
+
"trainrewards/rejected": 0.53515625
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.07,
|
| 125 |
+
"grad_norm": 12.480541507697524,
|
| 126 |
+
"learning_rate": 3.2142857142857147e-06,
|
| 127 |
+
"loss": 3.0686,
|
| 128 |
+
"step": 9,
|
| 129 |
+
"trainloss/critic_chosen": 1.3232171535491943,
|
| 130 |
+
"trainloss/critic_rejected": 1.3606585264205933,
|
| 131 |
+
"trainloss/reward": 1.3232171535491943,
|
| 132 |
+
"trainrewards/accuracies": 0.9114583134651184,
|
| 133 |
+
"trainrewards/chosen": 1.671875,
|
| 134 |
+
"trainrewards/margins": 1.3359375,
|
| 135 |
+
"trainrewards/rejected": 0.333984375
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.07,
|
| 139 |
+
"grad_norm": 8.25153148096521,
|
| 140 |
+
"learning_rate": 3.5714285714285718e-06,
|
| 141 |
+
"loss": 3.0194,
|
| 142 |
+
"step": 10,
|
| 143 |
+
"trainloss/critic_chosen": 1.289527177810669,
|
| 144 |
+
"trainloss/critic_rejected": 1.3465213775634766,
|
| 145 |
+
"trainloss/reward": 1.289527177810669,
|
| 146 |
+
"trainrewards/accuracies": 0.9010416865348816,
|
| 147 |
+
"trainrewards/chosen": 1.3125,
|
| 148 |
+
"trainrewards/margins": 1.4375,
|
| 149 |
+
"trainrewards/rejected": -0.1220703125
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.08,
|
| 153 |
+
"grad_norm": 15.573626174864643,
|
| 154 |
+
"learning_rate": 3.928571428571429e-06,
|
| 155 |
+
"loss": 2.9476,
|
| 156 |
+
"step": 11,
|
| 157 |
+
"trainloss/critic_chosen": 1.266632318496704,
|
| 158 |
+
"trainloss/critic_rejected": 1.3359544277191162,
|
| 159 |
+
"trainloss/reward": 1.266632318496704,
|
| 160 |
+
"trainrewards/accuracies": 0.9010417461395264,
|
| 161 |
+
"trainrewards/chosen": 0.208984375,
|
| 162 |
+
"trainrewards/margins": 2.25,
|
| 163 |
+
"trainrewards/rejected": -2.046875
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.09,
|
| 167 |
+
"grad_norm": 8.798738351150583,
|
| 168 |
+
"learning_rate": 4.2857142857142855e-06,
|
| 169 |
+
"loss": 2.9208,
|
| 170 |
+
"step": 12,
|
| 171 |
+
"trainloss/critic_chosen": 1.285549521446228,
|
| 172 |
+
"trainloss/critic_rejected": 1.336460828781128,
|
| 173 |
+
"trainloss/reward": 1.285549521446228,
|
| 174 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 175 |
+
"trainrewards/chosen": 1.390625,
|
| 176 |
+
"trainrewards/margins": 2.453125,
|
| 177 |
+
"trainrewards/rejected": -1.0625
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.1,
|
| 181 |
+
"grad_norm": 10.920314695532706,
|
| 182 |
+
"learning_rate": 4.642857142857144e-06,
|
| 183 |
+
"loss": 2.9737,
|
| 184 |
+
"step": 13,
|
| 185 |
+
"trainloss/critic_chosen": 1.3031879663467407,
|
| 186 |
+
"trainloss/critic_rejected": 1.3538345098495483,
|
| 187 |
+
"trainloss/reward": 1.3031879663467407,
|
| 188 |
+
"trainrewards/accuracies": 0.9583333134651184,
|
| 189 |
+
"trainrewards/chosen": 1.9453125,
|
| 190 |
+
"trainrewards/margins": 2.203125,
|
| 191 |
+
"trainrewards/rejected": -0.255859375
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1,
|
| 195 |
+
"grad_norm": 7.902411216600844,
|
| 196 |
+
"learning_rate": 5e-06,
|
| 197 |
+
"loss": 2.8568,
|
| 198 |
+
"step": 14,
|
| 199 |
+
"trainloss/critic_chosen": 1.2445811033248901,
|
| 200 |
+
"trainloss/critic_rejected": 1.3195788860321045,
|
| 201 |
+
"trainloss/reward": 1.2445811033248901,
|
| 202 |
+
"trainrewards/accuracies": 0.90625,
|
| 203 |
+
"trainrewards/chosen": 1.5078125,
|
| 204 |
+
"trainrewards/margins": 2.09375,
|
| 205 |
+
"trainrewards/rejected": -0.5859375
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.11,
|
| 209 |
+
"grad_norm": 11.497405609399582,
|
| 210 |
+
"learning_rate": 4.999143312438893e-06,
|
| 211 |
+
"loss": 2.8485,
|
| 212 |
+
"step": 15,
|
| 213 |
+
"trainloss/critic_chosen": 1.2575112581253052,
|
| 214 |
+
"trainloss/critic_rejected": 1.303661823272705,
|
| 215 |
+
"trainloss/reward": 1.2575112581253052,
|
| 216 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 217 |
+
"trainrewards/chosen": 0.55078125,
|
| 218 |
+
"trainrewards/margins": 1.84375,
|
| 219 |
+
"trainrewards/rejected": -1.2890625
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.12,
|
| 223 |
+
"grad_norm": 8.767557307270321,
|
| 224 |
+
"learning_rate": 4.9965738368864345e-06,
|
| 225 |
+
"loss": 2.8737,
|
| 226 |
+
"step": 16,
|
| 227 |
+
"trainloss/critic_chosen": 1.2387288808822632,
|
| 228 |
+
"trainloss/critic_rejected": 1.3038721084594727,
|
| 229 |
+
"trainloss/reward": 1.2387288808822632,
|
| 230 |
+
"trainrewards/accuracies": 0.9010416865348816,
|
| 231 |
+
"trainrewards/chosen": 1.578125,
|
| 232 |
+
"trainrewards/margins": 2.59375,
|
| 233 |
+
"trainrewards/rejected": -1.015625
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.13,
|
| 237 |
+
"grad_norm": 8.749611939075479,
|
| 238 |
+
"learning_rate": 4.992293334332821e-06,
|
| 239 |
+
"loss": 2.8681,
|
| 240 |
+
"step": 17,
|
| 241 |
+
"trainloss/critic_chosen": 1.2373957633972168,
|
| 242 |
+
"trainloss/critic_rejected": 1.302640676498413,
|
| 243 |
+
"trainloss/reward": 1.2373957633972168,
|
| 244 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 245 |
+
"trainrewards/chosen": 1.5859375,
|
| 246 |
+
"trainrewards/margins": 2.203125,
|
| 247 |
+
"trainrewards/rejected": -0.61328125
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.13,
|
| 251 |
+
"grad_norm": 8.484573172931489,
|
| 252 |
+
"learning_rate": 4.986304738420684e-06,
|
| 253 |
+
"loss": 2.8305,
|
| 254 |
+
"step": 18,
|
| 255 |
+
"trainloss/critic_chosen": 1.23964262008667,
|
| 256 |
+
"trainloss/critic_rejected": 1.297165870666504,
|
| 257 |
+
"trainloss/reward": 1.23964262008667,
|
| 258 |
+
"trainrewards/accuracies": 0.9166666865348816,
|
| 259 |
+
"trainrewards/chosen": 0.60546875,
|
| 260 |
+
"trainrewards/margins": 1.765625,
|
| 261 |
+
"trainrewards/rejected": -1.15625
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.14,
|
| 265 |
+
"grad_norm": 5.43676917020596,
|
| 266 |
+
"learning_rate": 4.978612153434527e-06,
|
| 267 |
+
"loss": 2.7193,
|
| 268 |
+
"step": 19,
|
| 269 |
+
"trainloss/critic_chosen": 1.2180960178375244,
|
| 270 |
+
"trainloss/critic_rejected": 1.2327347993850708,
|
| 271 |
+
"trainloss/reward": 1.2180960178375244,
|
| 272 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 273 |
+
"trainrewards/chosen": 1.5546875,
|
| 274 |
+
"trainrewards/margins": 2.296875,
|
| 275 |
+
"trainrewards/rejected": -0.7421875
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.15,
|
| 279 |
+
"grad_norm": 5.651490595569425,
|
| 280 |
+
"learning_rate": 4.9692208514878445e-06,
|
| 281 |
+
"loss": 2.8528,
|
| 282 |
+
"step": 20,
|
| 283 |
+
"trainloss/critic_chosen": 1.2103852033615112,
|
| 284 |
+
"trainloss/critic_rejected": 1.3035379648208618,
|
| 285 |
+
"trainloss/reward": 1.2103852033615112,
|
| 286 |
+
"trainrewards/accuracies": 0.9062500596046448,
|
| 287 |
+
"trainrewards/chosen": 1.75,
|
| 288 |
+
"trainrewards/margins": 2.6875,
|
| 289 |
+
"trainrewards/rejected": -0.94921875
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.16,
|
| 293 |
+
"grad_norm": 5.1975082682744524,
|
| 294 |
+
"learning_rate": 4.958137268909887e-06,
|
| 295 |
+
"loss": 2.7287,
|
| 296 |
+
"step": 21,
|
| 297 |
+
"trainloss/critic_chosen": 1.1857198476791382,
|
| 298 |
+
"trainloss/critic_rejected": 1.2193048000335693,
|
| 299 |
+
"trainloss/reward": 1.1857198476791382,
|
| 300 |
+
"trainrewards/accuracies": 0.9114583730697632,
|
| 301 |
+
"trainrewards/chosen": 1.4296875,
|
| 302 |
+
"trainrewards/margins": 2.21875,
|
| 303 |
+
"trainrewards/rejected": -0.7890625
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.16,
|
| 307 |
+
"grad_norm": 5.466958288822169,
|
| 308 |
+
"learning_rate": 4.9453690018345144e-06,
|
| 309 |
+
"loss": 2.7514,
|
| 310 |
+
"step": 22,
|
| 311 |
+
"trainloss/critic_chosen": 1.1868751049041748,
|
| 312 |
+
"trainloss/critic_rejected": 1.256333827972412,
|
| 313 |
+
"trainloss/reward": 1.1868751049041748,
|
| 314 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 315 |
+
"trainrewards/chosen": 0.953125,
|
| 316 |
+
"trainrewards/margins": 1.71875,
|
| 317 |
+
"trainrewards/rejected": -0.76171875
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.17,
|
| 321 |
+
"grad_norm": 4.795005616591591,
|
| 322 |
+
"learning_rate": 4.930924800994192e-06,
|
| 323 |
+
"loss": 2.7025,
|
| 324 |
+
"step": 23,
|
| 325 |
+
"trainloss/critic_chosen": 1.1841645240783691,
|
| 326 |
+
"trainloss/critic_rejected": 1.2626478672027588,
|
| 327 |
+
"trainloss/reward": 1.1841645240783691,
|
| 328 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 329 |
+
"trainrewards/chosen": 1.0625,
|
| 330 |
+
"trainrewards/margins": 2.09375,
|
| 331 |
+
"trainrewards/rejected": -1.0390625
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.18,
|
| 335 |
+
"grad_norm": 7.216599935232891,
|
| 336 |
+
"learning_rate": 4.914814565722671e-06,
|
| 337 |
+
"loss": 2.7024,
|
| 338 |
+
"step": 24,
|
| 339 |
+
"trainloss/critic_chosen": 1.1600149869918823,
|
| 340 |
+
"trainloss/critic_rejected": 1.216670036315918,
|
| 341 |
+
"trainloss/reward": 1.1600149869918823,
|
| 342 |
+
"trainrewards/accuracies": 0.90625,
|
| 343 |
+
"trainrewards/chosen": 1.953125,
|
| 344 |
+
"trainrewards/margins": 2.53125,
|
| 345 |
+
"trainrewards/rejected": -0.578125
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.19,
|
| 349 |
+
"grad_norm": 5.574536341669933,
|
| 350 |
+
"learning_rate": 4.897049337170483e-06,
|
| 351 |
+
"loss": 2.6825,
|
| 352 |
+
"step": 25,
|
| 353 |
+
"trainloss/critic_chosen": 1.17496919631958,
|
| 354 |
+
"trainloss/critic_rejected": 1.2430278062820435,
|
| 355 |
+
"trainloss/reward": 1.17496919631958,
|
| 356 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 357 |
+
"trainrewards/chosen": 1.84375,
|
| 358 |
+
"trainrewards/margins": 2.71875,
|
| 359 |
+
"trainrewards/rejected": -0.87109375
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.19,
|
| 363 |
+
"grad_norm": 8.130831144336229,
|
| 364 |
+
"learning_rate": 4.8776412907378845e-06,
|
| 365 |
+
"loss": 2.7403,
|
| 366 |
+
"step": 26,
|
| 367 |
+
"trainloss/critic_chosen": 1.1843974590301514,
|
| 368 |
+
"trainloss/critic_rejected": 1.2316250801086426,
|
| 369 |
+
"trainloss/reward": 1.1843974590301514,
|
| 370 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 371 |
+
"trainrewards/chosen": 0.322265625,
|
| 372 |
+
"trainrewards/margins": 2.09375,
|
| 373 |
+
"trainrewards/rejected": -1.78125
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.2,
|
| 377 |
+
"grad_norm": 4.106462749941039,
|
| 378 |
+
"learning_rate": 4.856603727730446e-06,
|
| 379 |
+
"loss": 2.6318,
|
| 380 |
+
"step": 27,
|
| 381 |
+
"trainloss/critic_chosen": 1.1325989961624146,
|
| 382 |
+
"trainloss/critic_rejected": 1.2125966548919678,
|
| 383 |
+
"trainloss/reward": 1.1325989961624146,
|
| 384 |
+
"trainrewards/accuracies": 0.9375,
|
| 385 |
+
"trainrewards/chosen": 0.98828125,
|
| 386 |
+
"trainrewards/margins": 1.859375,
|
| 387 |
+
"trainrewards/rejected": -0.875
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.21,
|
| 391 |
+
"grad_norm": 7.501840024960186,
|
| 392 |
+
"learning_rate": 4.833951066243004e-06,
|
| 393 |
+
"loss": 2.7439,
|
| 394 |
+
"step": 28,
|
| 395 |
+
"trainloss/critic_chosen": 1.156808853149414,
|
| 396 |
+
"trainloss/critic_rejected": 1.218095064163208,
|
| 397 |
+
"trainloss/reward": 1.156808853149414,
|
| 398 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 399 |
+
"trainrewards/chosen": 2.03125,
|
| 400 |
+
"trainrewards/margins": 2.0,
|
| 401 |
+
"trainrewards/rejected": 0.021240234375
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.22,
|
| 405 |
+
"grad_norm": 10.542645887143404,
|
| 406 |
+
"learning_rate": 4.809698831278217e-06,
|
| 407 |
+
"loss": 2.6949,
|
| 408 |
+
"step": 29,
|
| 409 |
+
"trainloss/critic_chosen": 1.146854043006897,
|
| 410 |
+
"trainloss/critic_rejected": 1.2151950597763062,
|
| 411 |
+
"trainloss/reward": 1.146854043006897,
|
| 412 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 413 |
+
"trainrewards/chosen": 2.546875,
|
| 414 |
+
"trainrewards/margins": 2.375,
|
| 415 |
+
"trainrewards/rejected": 0.169921875
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.22,
|
| 419 |
+
"grad_norm": 5.112451478263716,
|
| 420 |
+
"learning_rate": 4.783863644106502e-06,
|
| 421 |
+
"loss": 2.6784,
|
| 422 |
+
"step": 30,
|
| 423 |
+
"trainloss/critic_chosen": 1.1554011106491089,
|
| 424 |
+
"trainloss/critic_rejected": 1.2330732345581055,
|
| 425 |
+
"trainloss/reward": 1.1554011106491089,
|
| 426 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 427 |
+
"trainrewards/chosen": 1.6875,
|
| 428 |
+
"trainrewards/margins": 2.390625,
|
| 429 |
+
"trainrewards/rejected": -0.703125
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.23,
|
| 433 |
+
"grad_norm": 5.406168864359725,
|
| 434 |
+
"learning_rate": 4.7564632108746524e-06,
|
| 435 |
+
"loss": 2.716,
|
| 436 |
+
"step": 31,
|
| 437 |
+
"trainloss/critic_chosen": 1.162062168121338,
|
| 438 |
+
"trainloss/critic_rejected": 1.2383043766021729,
|
| 439 |
+
"trainloss/reward": 1.162062168121338,
|
| 440 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 441 |
+
"trainrewards/chosen": 0.609375,
|
| 442 |
+
"trainrewards/margins": 1.9765625,
|
| 443 |
+
"trainrewards/rejected": -1.3671875
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.24,
|
| 447 |
+
"grad_norm": 5.067065444400827,
|
| 448 |
+
"learning_rate": 4.72751631047092e-06,
|
| 449 |
+
"loss": 2.6516,
|
| 450 |
+
"step": 32,
|
| 451 |
+
"trainloss/critic_chosen": 1.1599905490875244,
|
| 452 |
+
"trainloss/critic_rejected": 1.2166763544082642,
|
| 453 |
+
"trainloss/reward": 1.1599905490875244,
|
| 454 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 455 |
+
"trainrewards/chosen": 0.7265625,
|
| 456 |
+
"trainrewards/margins": 2.21875,
|
| 457 |
+
"trainrewards/rejected": -1.4921875
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.25,
|
| 461 |
+
"grad_norm": 6.759001147106114,
|
| 462 |
+
"learning_rate": 4.697042781654913e-06,
|
| 463 |
+
"loss": 2.6586,
|
| 464 |
+
"step": 33,
|
| 465 |
+
"trainloss/critic_chosen": 1.1513490676879883,
|
| 466 |
+
"trainloss/critic_rejected": 1.1805285215377808,
|
| 467 |
+
"trainloss/reward": 1.1513490676879883,
|
| 468 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 469 |
+
"trainrewards/chosen": 1.8203125,
|
| 470 |
+
"trainrewards/margins": 2.234375,
|
| 471 |
+
"trainrewards/rejected": -0.408203125
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.25,
|
| 475 |
+
"grad_norm": 7.671545596305826,
|
| 476 |
+
"learning_rate": 4.665063509461098e-06,
|
| 477 |
+
"loss": 2.6397,
|
| 478 |
+
"step": 34,
|
| 479 |
+
"trainloss/critic_chosen": 1.1355525255203247,
|
| 480 |
+
"trainloss/critic_rejected": 1.1824406385421753,
|
| 481 |
+
"trainloss/reward": 1.1355525255203247,
|
| 482 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 483 |
+
"trainrewards/chosen": 2.15625,
|
| 484 |
+
"trainrewards/margins": 2.34375,
|
| 485 |
+
"trainrewards/rejected": -0.1923828125
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.26,
|
| 489 |
+
"grad_norm": 4.120967770831028,
|
| 490 |
+
"learning_rate": 4.631600410885231e-06,
|
| 491 |
+
"loss": 2.6941,
|
| 492 |
+
"step": 35,
|
| 493 |
+
"trainloss/critic_chosen": 1.1876243352890015,
|
| 494 |
+
"trainloss/critic_rejected": 1.2462928295135498,
|
| 495 |
+
"trainloss/reward": 1.1876243352890015,
|
| 496 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 497 |
+
"trainrewards/chosen": 1.6640625,
|
| 498 |
+
"trainrewards/margins": 2.453125,
|
| 499 |
+
"trainrewards/rejected": -0.78125
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.27,
|
| 503 |
+
"grad_norm": 4.873851901547121,
|
| 504 |
+
"learning_rate": 4.596676419863561e-06,
|
| 505 |
+
"loss": 2.5644,
|
| 506 |
+
"step": 36,
|
| 507 |
+
"trainloss/critic_chosen": 1.1080451011657715,
|
| 508 |
+
"trainloss/critic_rejected": 1.1967337131500244,
|
| 509 |
+
"trainloss/reward": 1.1080451011657715,
|
| 510 |
+
"trainrewards/accuracies": 0.96875,
|
| 511 |
+
"trainrewards/chosen": 0.80078125,
|
| 512 |
+
"trainrewards/margins": 2.125,
|
| 513 |
+
"trainrewards/rejected": -1.328125
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.28,
|
| 517 |
+
"grad_norm": 3.9620697117121004,
|
| 518 |
+
"learning_rate": 4.560315471555039e-06,
|
| 519 |
+
"loss": 2.5956,
|
| 520 |
+
"step": 37,
|
| 521 |
+
"trainloss/critic_chosen": 1.1373480558395386,
|
| 522 |
+
"trainloss/critic_rejected": 1.2142869234085083,
|
| 523 |
+
"trainloss/reward": 1.1373480558395386,
|
| 524 |
+
"trainrewards/accuracies": 0.9375000596046448,
|
| 525 |
+
"trainrewards/chosen": 1.0234375,
|
| 526 |
+
"trainrewards/margins": 2.40625,
|
| 527 |
+
"trainrewards/rejected": -1.390625
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.28,
|
| 531 |
+
"grad_norm": 5.768390511994523,
|
| 532 |
+
"learning_rate": 4.522542485937369e-06,
|
| 533 |
+
"loss": 2.6888,
|
| 534 |
+
"step": 38,
|
| 535 |
+
"trainloss/critic_chosen": 1.1469529867172241,
|
| 536 |
+
"trainloss/critic_rejected": 1.2045722007751465,
|
| 537 |
+
"trainloss/reward": 1.1469529867172241,
|
| 538 |
+
"trainrewards/accuracies": 0.9114583730697632,
|
| 539 |
+
"trainrewards/chosen": 1.8046875,
|
| 540 |
+
"trainrewards/margins": 2.484375,
|
| 541 |
+
"trainrewards/rejected": -0.6875
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.29,
|
| 545 |
+
"grad_norm": 4.874609149891752,
|
| 546 |
+
"learning_rate": 4.4833833507280884e-06,
|
| 547 |
+
"loss": 2.5543,
|
| 548 |
+
"step": 39,
|
| 549 |
+
"trainloss/critic_chosen": 1.1098886728286743,
|
| 550 |
+
"trainloss/critic_rejected": 1.1714903116226196,
|
| 551 |
+
"trainloss/reward": 1.1098886728286743,
|
| 552 |
+
"trainrewards/accuracies": 0.958333432674408,
|
| 553 |
+
"trainrewards/chosen": 1.8984375,
|
| 554 |
+
"trainrewards/margins": 2.625,
|
| 555 |
+
"trainrewards/rejected": -0.7265625
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3,
|
| 559 |
+
"grad_norm": 3.396284173013532,
|
| 560 |
+
"learning_rate": 4.442864903642428e-06,
|
| 561 |
+
"loss": 2.6564,
|
| 562 |
+
"step": 40,
|
| 563 |
+
"trainloss/critic_chosen": 1.1380069255828857,
|
| 564 |
+
"trainloss/critic_rejected": 1.2159972190856934,
|
| 565 |
+
"trainloss/reward": 1.1380069255828857,
|
| 566 |
+
"trainrewards/accuracies": 0.9427083134651184,
|
| 567 |
+
"trainrewards/chosen": 0.9296875,
|
| 568 |
+
"trainrewards/margins": 1.9375,
|
| 569 |
+
"trainrewards/rejected": -1.0078125
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.31,
|
| 573 |
+
"grad_norm": 3.8092153086439087,
|
| 574 |
+
"learning_rate": 4.401014914000078e-06,
|
| 575 |
+
"loss": 2.5515,
|
| 576 |
+
"step": 41,
|
| 577 |
+
"trainloss/critic_chosen": 1.123491883277893,
|
| 578 |
+
"trainloss/critic_rejected": 1.1983730792999268,
|
| 579 |
+
"trainloss/reward": 1.123491883277893,
|
| 580 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 581 |
+
"trainrewards/chosen": 1.015625,
|
| 582 |
+
"trainrewards/margins": 2.125,
|
| 583 |
+
"trainrewards/rejected": -1.1015625
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.31,
|
| 587 |
+
"grad_norm": 3.292224209377405,
|
| 588 |
+
"learning_rate": 4.357862063693486e-06,
|
| 589 |
+
"loss": 2.6296,
|
| 590 |
+
"step": 42,
|
| 591 |
+
"trainloss/critic_chosen": 1.1296896934509277,
|
| 592 |
+
"trainloss/critic_rejected": 1.193892002105713,
|
| 593 |
+
"trainloss/reward": 1.1296896934509277,
|
| 594 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 595 |
+
"trainrewards/chosen": 1.3671875,
|
| 596 |
+
"trainrewards/margins": 2.234375,
|
| 597 |
+
"trainrewards/rejected": -0.859375
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.32,
|
| 601 |
+
"grad_norm": 4.97304962284229,
|
| 602 |
+
"learning_rate": 4.313435927530719e-06,
|
| 603 |
+
"loss": 2.5984,
|
| 604 |
+
"step": 43,
|
| 605 |
+
"trainloss/critic_chosen": 1.106866478919983,
|
| 606 |
+
"trainloss/critic_rejected": 1.1839522123336792,
|
| 607 |
+
"trainloss/reward": 1.106866478919983,
|
| 608 |
+
"trainrewards/accuracies": 0.9166666865348816,
|
| 609 |
+
"trainrewards/chosen": 1.859375,
|
| 610 |
+
"trainrewards/margins": 2.515625,
|
| 611 |
+
"trainrewards/rejected": -0.6640625
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.33,
|
| 615 |
+
"grad_norm": 3.1437931542625615,
|
| 616 |
+
"learning_rate": 4.267766952966369e-06,
|
| 617 |
+
"loss": 2.6053,
|
| 618 |
+
"step": 44,
|
| 619 |
+
"trainloss/critic_chosen": 1.141026496887207,
|
| 620 |
+
"trainloss/critic_rejected": 1.1881659030914307,
|
| 621 |
+
"trainloss/reward": 1.141026496887207,
|
| 622 |
+
"trainrewards/accuracies": 0.9375,
|
| 623 |
+
"trainrewards/chosen": 1.484375,
|
| 624 |
+
"trainrewards/margins": 2.5,
|
| 625 |
+
"trainrewards/rejected": -1.015625
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.34,
|
| 629 |
+
"grad_norm": 3.0260425195721727,
|
| 630 |
+
"learning_rate": 4.220886439234385e-06,
|
| 631 |
+
"loss": 2.6162,
|
| 632 |
+
"step": 45,
|
| 633 |
+
"trainloss/critic_chosen": 1.1437909603118896,
|
| 634 |
+
"trainloss/critic_rejected": 1.1694350242614746,
|
| 635 |
+
"trainloss/reward": 1.1437909603118896,
|
| 636 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 637 |
+
"trainrewards/chosen": 1.3359375,
|
| 638 |
+
"trainrewards/margins": 2.265625,
|
| 639 |
+
"trainrewards/rejected": -0.93359375
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.34,
|
| 643 |
+
"grad_norm": 3.9421991947992803,
|
| 644 |
+
"learning_rate": 4.172826515897146e-06,
|
| 645 |
+
"loss": 2.559,
|
| 646 |
+
"step": 46,
|
| 647 |
+
"trainloss/critic_chosen": 1.1193464994430542,
|
| 648 |
+
"trainloss/critic_rejected": 1.1624045372009277,
|
| 649 |
+
"trainloss/reward": 1.1193464994430542,
|
| 650 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 651 |
+
"trainrewards/chosen": 1.2109375,
|
| 652 |
+
"trainrewards/margins": 1.96875,
|
| 653 |
+
"trainrewards/rejected": -0.75
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.35,
|
| 657 |
+
"grad_norm": 4.76800798471375,
|
| 658 |
+
"learning_rate": 4.123620120825459e-06,
|
| 659 |
+
"loss": 2.5633,
|
| 660 |
+
"step": 47,
|
| 661 |
+
"trainloss/critic_chosen": 1.1039447784423828,
|
| 662 |
+
"trainloss/critic_rejected": 1.1683855056762695,
|
| 663 |
+
"trainloss/reward": 1.1039447784423828,
|
| 664 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 665 |
+
"trainrewards/chosen": 1.5,
|
| 666 |
+
"trainrewards/margins": 1.8515625,
|
| 667 |
+
"trainrewards/rejected": -0.357421875
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.36,
|
| 671 |
+
"grad_norm": 4.677899041279874,
|
| 672 |
+
"learning_rate": 4.073300977624594e-06,
|
| 673 |
+
"loss": 2.6104,
|
| 674 |
+
"step": 48,
|
| 675 |
+
"trainloss/critic_chosen": 1.1293267011642456,
|
| 676 |
+
"trainloss/critic_rejected": 1.173600435256958,
|
| 677 |
+
"trainloss/reward": 1.1293267011642456,
|
| 678 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 679 |
+
"trainrewards/chosen": 1.5625,
|
| 680 |
+
"trainrewards/margins": 1.953125,
|
| 681 |
+
"trainrewards/rejected": -0.38671875
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.36,
|
| 685 |
+
"grad_norm": 2.8973280324668815,
|
| 686 |
+
"learning_rate": 4.021903572521802e-06,
|
| 687 |
+
"loss": 2.5884,
|
| 688 |
+
"step": 49,
|
| 689 |
+
"trainloss/critic_chosen": 1.1289738416671753,
|
| 690 |
+
"trainloss/critic_rejected": 1.169731855392456,
|
| 691 |
+
"trainloss/reward": 1.1289738416671753,
|
| 692 |
+
"trainrewards/accuracies": 0.9375000596046448,
|
| 693 |
+
"trainrewards/chosen": 1.3125,
|
| 694 |
+
"trainrewards/margins": 2.515625,
|
| 695 |
+
"trainrewards/rejected": -1.203125
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.37,
|
| 699 |
+
"grad_norm": 2.9211772383175685,
|
| 700 |
+
"learning_rate": 3.969463130731183e-06,
|
| 701 |
+
"loss": 2.5868,
|
| 702 |
+
"step": 50,
|
| 703 |
+
"trainloss/critic_chosen": 1.1367411613464355,
|
| 704 |
+
"trainloss/critic_rejected": 1.1725157499313354,
|
| 705 |
+
"trainloss/reward": 1.1367411613464355,
|
| 706 |
+
"trainrewards/accuracies": 0.9114583730697632,
|
| 707 |
+
"trainrewards/chosen": 1.265625,
|
| 708 |
+
"trainrewards/margins": 2.484375,
|
| 709 |
+
"trainrewards/rejected": -1.21875
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.38,
|
| 713 |
+
"grad_norm": 3.4239805909008627,
|
| 714 |
+
"learning_rate": 3.916015592312083e-06,
|
| 715 |
+
"loss": 2.5442,
|
| 716 |
+
"step": 51,
|
| 717 |
+
"trainloss/critic_chosen": 1.101191759109497,
|
| 718 |
+
"trainloss/critic_rejected": 1.2063257694244385,
|
| 719 |
+
"trainloss/reward": 1.101191759109497,
|
| 720 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 721 |
+
"trainrewards/chosen": 1.6171875,
|
| 722 |
+
"trainrewards/margins": 2.546875,
|
| 723 |
+
"trainrewards/rejected": -0.9296875
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.39,
|
| 727 |
+
"grad_norm": 3.3449791279382155,
|
| 728 |
+
"learning_rate": 3.861597587537568e-06,
|
| 729 |
+
"loss": 2.5532,
|
| 730 |
+
"step": 52,
|
| 731 |
+
"trainloss/critic_chosen": 1.1064534187316895,
|
| 732 |
+
"trainloss/critic_rejected": 1.1979490518569946,
|
| 733 |
+
"trainloss/reward": 1.1064534187316895,
|
| 734 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 735 |
+
"trainrewards/chosen": 1.6015625,
|
| 736 |
+
"trainrewards/margins": 2.46875,
|
| 737 |
+
"trainrewards/rejected": -0.875
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.39,
|
| 741 |
+
"grad_norm": 3.9419484082989724,
|
| 742 |
+
"learning_rate": 3.806246411789872e-06,
|
| 743 |
+
"loss": 2.6147,
|
| 744 |
+
"step": 53,
|
| 745 |
+
"trainloss/critic_chosen": 1.138405680656433,
|
| 746 |
+
"trainloss/critic_rejected": 1.1973538398742676,
|
| 747 |
+
"trainloss/reward": 1.138405680656433,
|
| 748 |
+
"trainrewards/accuracies": 0.9375,
|
| 749 |
+
"trainrewards/chosen": 1.25,
|
| 750 |
+
"trainrewards/margins": 2.609375,
|
| 751 |
+
"trainrewards/rejected": -1.3671875
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.4,
|
| 755 |
+
"grad_norm": 3.4980254593155413,
|
| 756 |
+
"learning_rate": 3.7500000000000005e-06,
|
| 757 |
+
"loss": 2.5364,
|
| 758 |
+
"step": 54,
|
| 759 |
+
"trainloss/critic_chosen": 1.0845965147018433,
|
| 760 |
+
"trainloss/critic_rejected": 1.1989306211471558,
|
| 761 |
+
"trainloss/reward": 1.0845965147018433,
|
| 762 |
+
"trainrewards/accuracies": 0.9635417461395264,
|
| 763 |
+
"trainrewards/chosen": 1.6953125,
|
| 764 |
+
"trainrewards/margins": 2.59375,
|
| 765 |
+
"trainrewards/rejected": -0.89453125
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.41,
|
| 769 |
+
"grad_norm": 3.7684432316267347,
|
| 770 |
+
"learning_rate": 3.6928969006490212e-06,
|
| 771 |
+
"loss": 2.5578,
|
| 772 |
+
"step": 55,
|
| 773 |
+
"trainloss/critic_chosen": 1.105364441871643,
|
| 774 |
+
"trainloss/critic_rejected": 1.1862692832946777,
|
| 775 |
+
"trainloss/reward": 1.105364441871643,
|
| 776 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 777 |
+
"trainrewards/chosen": 1.8046875,
|
| 778 |
+
"trainrewards/margins": 2.65625,
|
| 779 |
+
"trainrewards/rejected": -0.86328125
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.42,
|
| 783 |
+
"grad_norm": 2.733004985886796,
|
| 784 |
+
"learning_rate": 3.634976249348867e-06,
|
| 785 |
+
"loss": 2.5665,
|
| 786 |
+
"step": 56,
|
| 787 |
+
"trainloss/critic_chosen": 1.1256849765777588,
|
| 788 |
+
"trainloss/critic_rejected": 1.1650742292404175,
|
| 789 |
+
"trainloss/reward": 1.1256849765777588,
|
| 790 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 791 |
+
"trainrewards/chosen": 1.328125,
|
| 792 |
+
"trainrewards/margins": 2.390625,
|
| 793 |
+
"trainrewards/rejected": -1.0546875
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.42,
|
| 797 |
+
"grad_norm": 3.063205301802556,
|
| 798 |
+
"learning_rate": 3.5762777420207382e-06,
|
| 799 |
+
"loss": 2.5733,
|
| 800 |
+
"step": 57,
|
| 801 |
+
"trainloss/critic_chosen": 1.1022924184799194,
|
| 802 |
+
"trainloss/critic_rejected": 1.1559257507324219,
|
| 803 |
+
"trainloss/reward": 1.1022924184799194,
|
| 804 |
+
"trainrewards/accuracies": 0.9166666865348816,
|
| 805 |
+
"trainrewards/chosen": 1.40625,
|
| 806 |
+
"trainrewards/margins": 2.28125,
|
| 807 |
+
"trainrewards/rejected": -0.875
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.43,
|
| 811 |
+
"grad_norm": 3.2936675397250985,
|
| 812 |
+
"learning_rate": 3.516841607689501e-06,
|
| 813 |
+
"loss": 2.529,
|
| 814 |
+
"step": 58,
|
| 815 |
+
"trainloss/critic_chosen": 1.0919924974441528,
|
| 816 |
+
"trainloss/critic_rejected": 1.1807957887649536,
|
| 817 |
+
"trainloss/reward": 1.0919924974441528,
|
| 818 |
+
"trainrewards/accuracies": 0.9375000596046448,
|
| 819 |
+
"trainrewards/chosen": 1.0703125,
|
| 820 |
+
"trainrewards/margins": 2.0625,
|
| 821 |
+
"trainrewards/rejected": -1.0
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.44,
|
| 825 |
+
"grad_norm": 2.9687788874925505,
|
| 826 |
+
"learning_rate": 3.4567085809127247e-06,
|
| 827 |
+
"loss": 2.5538,
|
| 828 |
+
"step": 59,
|
| 829 |
+
"trainloss/critic_chosen": 1.152530312538147,
|
| 830 |
+
"trainloss/critic_rejected": 1.128198504447937,
|
| 831 |
+
"trainloss/reward": 1.152530312538147,
|
| 832 |
+
"trainrewards/accuracies": 0.9375,
|
| 833 |
+
"trainrewards/chosen": 1.3125,
|
| 834 |
+
"trainrewards/margins": 2.171875,
|
| 835 |
+
"trainrewards/rejected": -0.86328125
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.45,
|
| 839 |
+
"grad_norm": 2.5189366946202374,
|
| 840 |
+
"learning_rate": 3.39591987386325e-06,
|
| 841 |
+
"loss": 2.4931,
|
| 842 |
+
"step": 60,
|
| 843 |
+
"trainloss/critic_chosen": 1.0971665382385254,
|
| 844 |
+
"trainloss/critic_rejected": 1.189927339553833,
|
| 845 |
+
"trainloss/reward": 1.0971665382385254,
|
| 846 |
+
"trainrewards/accuracies": 0.96875,
|
| 847 |
+
"trainrewards/chosen": 1.3828125,
|
| 848 |
+
"trainrewards/margins": 2.671875,
|
| 849 |
+
"trainrewards/rejected": -1.2890625
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.45,
|
| 853 |
+
"grad_norm": 4.707774798123127,
|
| 854 |
+
"learning_rate": 3.3345171480844275e-06,
|
| 855 |
+
"loss": 2.4995,
|
| 856 |
+
"step": 61,
|
| 857 |
+
"trainloss/critic_chosen": 1.1144541501998901,
|
| 858 |
+
"trainloss/critic_rejected": 1.1472208499908447,
|
| 859 |
+
"trainloss/reward": 1.1144541501998901,
|
| 860 |
+
"trainrewards/accuracies": 0.9739583730697632,
|
| 861 |
+
"trainrewards/chosen": 1.9921875,
|
| 862 |
+
"trainrewards/margins": 2.765625,
|
| 863 |
+
"trainrewards/rejected": -0.7734375
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.46,
|
| 867 |
+
"grad_norm": 3.621977923726089,
|
| 868 |
+
"learning_rate": 3.272542485937369e-06,
|
| 869 |
+
"loss": 2.5767,
|
| 870 |
+
"step": 62,
|
| 871 |
+
"trainloss/critic_chosen": 1.1388683319091797,
|
| 872 |
+
"trainloss/critic_rejected": 1.1852062940597534,
|
| 873 |
+
"trainloss/reward": 1.1388683319091797,
|
| 874 |
+
"trainrewards/accuracies": 0.9479167461395264,
|
| 875 |
+
"trainrewards/chosen": 1.8203125,
|
| 876 |
+
"trainrewards/margins": 3.09375,
|
| 877 |
+
"trainrewards/rejected": -1.265625
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.47,
|
| 881 |
+
"grad_norm": 4.340502288849219,
|
| 882 |
+
"learning_rate": 3.2100383617598075e-06,
|
| 883 |
+
"loss": 2.5008,
|
| 884 |
+
"step": 63,
|
| 885 |
+
"trainloss/critic_chosen": 1.0960522890090942,
|
| 886 |
+
"trainloss/critic_rejected": 1.1389869451522827,
|
| 887 |
+
"trainloss/reward": 1.0960522890090942,
|
| 888 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 889 |
+
"trainrewards/chosen": 1.25,
|
| 890 |
+
"trainrewards/margins": 2.8125,
|
| 891 |
+
"trainrewards/rejected": -1.5703125
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.48,
|
| 895 |
+
"grad_norm": 3.2652013602478087,
|
| 896 |
+
"learning_rate": 3.147047612756302e-06,
|
| 897 |
+
"loss": 2.4784,
|
| 898 |
+
"step": 64,
|
| 899 |
+
"trainloss/critic_chosen": 1.1066646575927734,
|
| 900 |
+
"trainloss/critic_rejected": 1.1423835754394531,
|
| 901 |
+
"trainloss/reward": 1.1066646575927734,
|
| 902 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 903 |
+
"trainrewards/chosen": 1.2265625,
|
| 904 |
+
"trainrewards/margins": 2.859375,
|
| 905 |
+
"trainrewards/rejected": -1.6328125
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.48,
|
| 909 |
+
"grad_norm": 4.460312181878758,
|
| 910 |
+
"learning_rate": 3.0836134096397642e-06,
|
| 911 |
+
"loss": 2.5315,
|
| 912 |
+
"step": 65,
|
| 913 |
+
"trainloss/critic_chosen": 1.097680926322937,
|
| 914 |
+
"trainloss/critic_rejected": 1.1829330921173096,
|
| 915 |
+
"trainloss/reward": 1.097680926322937,
|
| 916 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 917 |
+
"trainrewards/chosen": 1.71875,
|
| 918 |
+
"trainrewards/margins": 2.375,
|
| 919 |
+
"trainrewards/rejected": -0.66015625
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.49,
|
| 923 |
+
"grad_norm": 5.398290397831798,
|
| 924 |
+
"learning_rate": 3.019779227044398e-06,
|
| 925 |
+
"loss": 2.4912,
|
| 926 |
+
"step": 66,
|
| 927 |
+
"trainloss/critic_chosen": 1.0728169679641724,
|
| 928 |
+
"trainloss/critic_rejected": 1.1528609991073608,
|
| 929 |
+
"trainloss/reward": 1.0728169679641724,
|
| 930 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 931 |
+
"trainrewards/chosen": 1.75,
|
| 932 |
+
"trainrewards/margins": 2.1875,
|
| 933 |
+
"trainrewards/rejected": -0.44140625
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.5,
|
| 937 |
+
"grad_norm": 4.530365049006353,
|
| 938 |
+
"learning_rate": 2.9555888137303695e-06,
|
| 939 |
+
"loss": 2.4768,
|
| 940 |
+
"step": 67,
|
| 941 |
+
"trainloss/critic_chosen": 1.0978233814239502,
|
| 942 |
+
"trainloss/critic_rejected": 1.1454182863235474,
|
| 943 |
+
"trainloss/reward": 1.0978233814239502,
|
| 944 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 945 |
+
"trainrewards/chosen": 1.515625,
|
| 946 |
+
"trainrewards/margins": 2.1875,
|
| 947 |
+
"trainrewards/rejected": -0.66015625
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.51,
|
| 951 |
+
"grad_norm": 3.090064735833262,
|
| 952 |
+
"learning_rate": 2.8910861626005774e-06,
|
| 953 |
+
"loss": 2.5542,
|
| 954 |
+
"step": 68,
|
| 955 |
+
"trainloss/critic_chosen": 1.1045993566513062,
|
| 956 |
+
"trainloss/critic_rejected": 1.1823933124542236,
|
| 957 |
+
"trainloss/reward": 1.1045993566513062,
|
| 958 |
+
"trainrewards/accuracies": 0.9166666865348816,
|
| 959 |
+
"trainrewards/chosen": 1.296875,
|
| 960 |
+
"trainrewards/margins": 2.296875,
|
| 961 |
+
"trainrewards/rejected": -1.0
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.51,
|
| 965 |
+
"grad_norm": 2.801294778929006,
|
| 966 |
+
"learning_rate": 2.82631548055013e-06,
|
| 967 |
+
"loss": 2.4752,
|
| 968 |
+
"step": 69,
|
| 969 |
+
"trainloss/critic_chosen": 1.0862737894058228,
|
| 970 |
+
"trainloss/critic_rejected": 1.1638906002044678,
|
| 971 |
+
"trainloss/reward": 1.0862737894058228,
|
| 972 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 973 |
+
"trainrewards/chosen": 1.46875,
|
| 974 |
+
"trainrewards/margins": 2.8125,
|
| 975 |
+
"trainrewards/rejected": -1.359375
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.52,
|
| 979 |
+
"grad_norm": 3.5888770327583503,
|
| 980 |
+
"learning_rate": 2.761321158169134e-06,
|
| 981 |
+
"loss": 2.5502,
|
| 982 |
+
"step": 70,
|
| 983 |
+
"trainloss/critic_chosen": 1.1130059957504272,
|
| 984 |
+
"trainloss/critic_rejected": 1.1747164726257324,
|
| 985 |
+
"trainloss/reward": 1.1130059957504272,
|
| 986 |
+
"trainrewards/accuracies": 0.9583333134651184,
|
| 987 |
+
"trainrewards/chosen": 1.75,
|
| 988 |
+
"trainrewards/margins": 2.953125,
|
| 989 |
+
"trainrewards/rejected": -1.203125
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.53,
|
| 993 |
+
"grad_norm": 3.553005435624982,
|
| 994 |
+
"learning_rate": 2.696147739319613e-06,
|
| 995 |
+
"loss": 2.4735,
|
| 996 |
+
"step": 71,
|
| 997 |
+
"trainloss/critic_chosen": 1.1133400201797485,
|
| 998 |
+
"trainloss/critic_rejected": 1.1409944295883179,
|
| 999 |
+
"trainloss/reward": 1.1133400201797485,
|
| 1000 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1001 |
+
"trainrewards/chosen": 1.96875,
|
| 1002 |
+
"trainrewards/margins": 3.375,
|
| 1003 |
+
"trainrewards/rejected": -1.40625
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.54,
|
| 1007 |
+
"grad_norm": 2.7088469336528145,
|
| 1008 |
+
"learning_rate": 2.6308398906073603e-06,
|
| 1009 |
+
"loss": 2.4512,
|
| 1010 |
+
"step": 72,
|
| 1011 |
+
"trainloss/critic_chosen": 1.1119564771652222,
|
| 1012 |
+
"trainloss/critic_rejected": 1.1244186162948608,
|
| 1013 |
+
"trainloss/reward": 1.1119564771652222,
|
| 1014 |
+
"trainrewards/accuracies": 0.96875,
|
| 1015 |
+
"trainrewards/chosen": 1.5703125,
|
| 1016 |
+
"trainrewards/margins": 3.03125,
|
| 1017 |
+
"trainrewards/rejected": -1.4609375
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.54,
|
| 1021 |
+
"grad_norm": 3.938561115333166,
|
| 1022 |
+
"learning_rate": 2.5654423707696834e-06,
|
| 1023 |
+
"loss": 2.4921,
|
| 1024 |
+
"step": 73,
|
| 1025 |
+
"trainloss/critic_chosen": 1.0844348669052124,
|
| 1026 |
+
"trainloss/critic_rejected": 1.163710355758667,
|
| 1027 |
+
"trainloss/reward": 1.0844348669052124,
|
| 1028 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1029 |
+
"trainrewards/chosen": 1.0703125,
|
| 1030 |
+
"trainrewards/margins": 2.734375,
|
| 1031 |
+
"trainrewards/rejected": -1.6640625
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.55,
|
| 1035 |
+
"grad_norm": 3.7076560975513293,
|
| 1036 |
+
"learning_rate": 2.5e-06,
|
| 1037 |
+
"loss": 2.4702,
|
| 1038 |
+
"step": 74,
|
| 1039 |
+
"trainloss/critic_chosen": 1.105428695678711,
|
| 1040 |
+
"trainloss/critic_rejected": 1.1134750843048096,
|
| 1041 |
+
"trainloss/reward": 1.105428695678711,
|
| 1042 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1043 |
+
"trainrewards/chosen": 1.1015625,
|
| 1044 |
+
"trainrewards/margins": 2.4375,
|
| 1045 |
+
"trainrewards/rejected": -1.328125
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.56,
|
| 1049 |
+
"grad_norm": 4.584325275815331,
|
| 1050 |
+
"learning_rate": 2.434557629230318e-06,
|
| 1051 |
+
"loss": 2.5531,
|
| 1052 |
+
"step": 75,
|
| 1053 |
+
"trainloss/critic_chosen": 1.1023496389389038,
|
| 1054 |
+
"trainloss/critic_rejected": 1.1693300008773804,
|
| 1055 |
+
"trainloss/reward": 1.1023496389389038,
|
| 1056 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1057 |
+
"trainrewards/chosen": 1.6953125,
|
| 1058 |
+
"trainrewards/margins": 2.265625,
|
| 1059 |
+
"trainrewards/rejected": -0.5703125
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.57,
|
| 1063 |
+
"grad_norm": 5.707921133643401,
|
| 1064 |
+
"learning_rate": 2.3691601093926406e-06,
|
| 1065 |
+
"loss": 2.512,
|
| 1066 |
+
"step": 76,
|
| 1067 |
+
"trainloss/critic_chosen": 1.0742218494415283,
|
| 1068 |
+
"trainloss/critic_rejected": 1.1473562717437744,
|
| 1069 |
+
"trainloss/reward": 1.0742218494415283,
|
| 1070 |
+
"trainrewards/accuracies": 0.9375000596046448,
|
| 1071 |
+
"trainrewards/chosen": 1.984375,
|
| 1072 |
+
"trainrewards/margins": 2.359375,
|
| 1073 |
+
"trainrewards/rejected": -0.380859375
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.57,
|
| 1077 |
+
"grad_norm": 5.052893345106084,
|
| 1078 |
+
"learning_rate": 2.3038522606803882e-06,
|
| 1079 |
+
"loss": 2.5495,
|
| 1080 |
+
"step": 77,
|
| 1081 |
+
"trainloss/critic_chosen": 1.09754478931427,
|
| 1082 |
+
"trainloss/critic_rejected": 1.175227165222168,
|
| 1083 |
+
"trainloss/reward": 1.09754478931427,
|
| 1084 |
+
"trainrewards/accuracies": 0.9218751192092896,
|
| 1085 |
+
"trainrewards/chosen": 1.8671875,
|
| 1086 |
+
"trainrewards/margins": 2.359375,
|
| 1087 |
+
"trainrewards/rejected": -0.490234375
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.58,
|
| 1091 |
+
"grad_norm": 3.505818483136781,
|
| 1092 |
+
"learning_rate": 2.238678841830867e-06,
|
| 1093 |
+
"loss": 2.5073,
|
| 1094 |
+
"step": 78,
|
| 1095 |
+
"trainloss/critic_chosen": 1.100816249847412,
|
| 1096 |
+
"trainloss/critic_rejected": 1.1553771495819092,
|
| 1097 |
+
"trainloss/reward": 1.100816249847412,
|
| 1098 |
+
"trainrewards/accuracies": 0.9375000596046448,
|
| 1099 |
+
"trainrewards/chosen": 1.4375,
|
| 1100 |
+
"trainrewards/margins": 2.1875,
|
| 1101 |
+
"trainrewards/rejected": -0.75
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.59,
|
| 1105 |
+
"grad_norm": 4.2251215117971475,
|
| 1106 |
+
"learning_rate": 2.173684519449872e-06,
|
| 1107 |
+
"loss": 2.5035,
|
| 1108 |
+
"step": 79,
|
| 1109 |
+
"trainloss/critic_chosen": 1.093074083328247,
|
| 1110 |
+
"trainloss/critic_rejected": 1.163825511932373,
|
| 1111 |
+
"trainloss/reward": 1.093074083328247,
|
| 1112 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1113 |
+
"trainrewards/chosen": 0.91796875,
|
| 1114 |
+
"trainrewards/margins": 2.140625,
|
| 1115 |
+
"trainrewards/rejected": -1.21875
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.6,
|
| 1119 |
+
"grad_norm": 4.171916933286059,
|
| 1120 |
+
"learning_rate": 2.1089138373994226e-06,
|
| 1121 |
+
"loss": 2.4726,
|
| 1122 |
+
"step": 80,
|
| 1123 |
+
"trainloss/critic_chosen": 1.0706841945648193,
|
| 1124 |
+
"trainloss/critic_rejected": 1.160952091217041,
|
| 1125 |
+
"trainloss/reward": 1.0706841945648193,
|
| 1126 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1127 |
+
"trainrewards/chosen": 0.98046875,
|
| 1128 |
+
"trainrewards/margins": 2.375,
|
| 1129 |
+
"trainrewards/rejected": -1.390625
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.6,
|
| 1133 |
+
"grad_norm": 2.7690433360924085,
|
| 1134 |
+
"learning_rate": 2.0444111862696313e-06,
|
| 1135 |
+
"loss": 2.4269,
|
| 1136 |
+
"step": 81,
|
| 1137 |
+
"trainloss/critic_chosen": 1.0752573013305664,
|
| 1138 |
+
"trainloss/critic_rejected": 1.1339901685714722,
|
| 1139 |
+
"trainloss/reward": 1.0752573013305664,
|
| 1140 |
+
"trainrewards/accuracies": 0.9739583730697632,
|
| 1141 |
+
"trainrewards/chosen": 1.484375,
|
| 1142 |
+
"trainrewards/margins": 2.578125,
|
| 1143 |
+
"trainrewards/rejected": -1.09375
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.61,
|
| 1147 |
+
"grad_norm": 3.358268001716196,
|
| 1148 |
+
"learning_rate": 1.9802207729556023e-06,
|
| 1149 |
+
"loss": 2.461,
|
| 1150 |
+
"step": 82,
|
| 1151 |
+
"trainloss/critic_chosen": 1.1075457334518433,
|
| 1152 |
+
"trainloss/critic_rejected": 1.1157523393630981,
|
| 1153 |
+
"trainloss/reward": 1.1075457334518433,
|
| 1154 |
+
"trainrewards/accuracies": 0.953125,
|
| 1155 |
+
"trainrewards/chosen": 1.8828125,
|
| 1156 |
+
"trainrewards/margins": 2.90625,
|
| 1157 |
+
"trainrewards/rejected": -1.0234375
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.62,
|
| 1161 |
+
"grad_norm": 4.328068525423629,
|
| 1162 |
+
"learning_rate": 1.9163865903602374e-06,
|
| 1163 |
+
"loss": 2.5352,
|
| 1164 |
+
"step": 83,
|
| 1165 |
+
"trainloss/critic_chosen": 1.1028249263763428,
|
| 1166 |
+
"trainloss/critic_rejected": 1.1644842624664307,
|
| 1167 |
+
"trainloss/reward": 1.1028249263763428,
|
| 1168 |
+
"trainrewards/accuracies": 0.96875,
|
| 1169 |
+
"trainrewards/chosen": 1.8671875,
|
| 1170 |
+
"trainrewards/margins": 2.921875,
|
| 1171 |
+
"trainrewards/rejected": -1.0625
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.63,
|
| 1175 |
+
"grad_norm": 3.266438978334478,
|
| 1176 |
+
"learning_rate": 1.852952387243698e-06,
|
| 1177 |
+
"loss": 2.4134,
|
| 1178 |
+
"step": 84,
|
| 1179 |
+
"trainloss/critic_chosen": 1.0756388902664185,
|
| 1180 |
+
"trainloss/critic_rejected": 1.1303694248199463,
|
| 1181 |
+
"trainloss/reward": 1.0756388902664185,
|
| 1182 |
+
"trainrewards/accuracies": 0.9687500596046448,
|
| 1183 |
+
"trainrewards/chosen": 1.9140625,
|
| 1184 |
+
"trainrewards/margins": 3.25,
|
| 1185 |
+
"trainrewards/rejected": -1.328125
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.63,
|
| 1189 |
+
"grad_norm": 2.386641393706194,
|
| 1190 |
+
"learning_rate": 1.7899616382401935e-06,
|
| 1191 |
+
"loss": 2.401,
|
| 1192 |
+
"step": 85,
|
| 1193 |
+
"trainloss/critic_chosen": 1.0511287450790405,
|
| 1194 |
+
"trainloss/critic_rejected": 1.128703236579895,
|
| 1195 |
+
"trainloss/reward": 1.0511287450790405,
|
| 1196 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1197 |
+
"trainrewards/chosen": 1.6015625,
|
| 1198 |
+
"trainrewards/margins": 2.953125,
|
| 1199 |
+
"trainrewards/rejected": -1.359375
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.64,
|
| 1203 |
+
"grad_norm": 3.7161933000807403,
|
| 1204 |
+
"learning_rate": 1.7274575140626318e-06,
|
| 1205 |
+
"loss": 2.4732,
|
| 1206 |
+
"step": 86,
|
| 1207 |
+
"trainloss/critic_chosen": 1.0827696323394775,
|
| 1208 |
+
"trainloss/critic_rejected": 1.1439146995544434,
|
| 1209 |
+
"trainloss/reward": 1.0827696323394775,
|
| 1210 |
+
"trainrewards/accuracies": 0.9583333134651184,
|
| 1211 |
+
"trainrewards/chosen": 1.0,
|
| 1212 |
+
"trainrewards/margins": 2.734375,
|
| 1213 |
+
"trainrewards/rejected": -1.734375
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.65,
|
| 1217 |
+
"grad_norm": 3.4186216283012754,
|
| 1218 |
+
"learning_rate": 1.665482851915573e-06,
|
| 1219 |
+
"loss": 2.5064,
|
| 1220 |
+
"step": 87,
|
| 1221 |
+
"trainloss/critic_chosen": 1.093652367591858,
|
| 1222 |
+
"trainloss/critic_rejected": 1.1373913288116455,
|
| 1223 |
+
"trainloss/reward": 1.093652367591858,
|
| 1224 |
+
"trainrewards/accuracies": 0.927083432674408,
|
| 1225 |
+
"trainrewards/chosen": 1.09375,
|
| 1226 |
+
"trainrewards/margins": 2.5625,
|
| 1227 |
+
"trainrewards/rejected": -1.46875
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.66,
|
| 1231 |
+
"grad_norm": 2.4263959266567996,
|
| 1232 |
+
"learning_rate": 1.6040801261367494e-06,
|
| 1233 |
+
"loss": 2.5409,
|
| 1234 |
+
"step": 88,
|
| 1235 |
+
"trainloss/critic_chosen": 1.1319228410720825,
|
| 1236 |
+
"trainloss/critic_rejected": 1.1887366771697998,
|
| 1237 |
+
"trainloss/reward": 1.1319228410720825,
|
| 1238 |
+
"trainrewards/accuracies": 0.9687501192092896,
|
| 1239 |
+
"trainrewards/chosen": 1.3125,
|
| 1240 |
+
"trainrewards/margins": 2.6875,
|
| 1241 |
+
"trainrewards/rejected": -1.375
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.66,
|
| 1245 |
+
"grad_norm": 4.091003192293857,
|
| 1246 |
+
"learning_rate": 1.5432914190872757e-06,
|
| 1247 |
+
"loss": 2.5386,
|
| 1248 |
+
"step": 89,
|
| 1249 |
+
"trainloss/critic_chosen": 1.1037051677703857,
|
| 1250 |
+
"trainloss/critic_rejected": 1.1342533826828003,
|
| 1251 |
+
"trainloss/reward": 1.1037051677703857,
|
| 1252 |
+
"trainrewards/accuracies": 0.9427083134651184,
|
| 1253 |
+
"trainrewards/chosen": 1.640625,
|
| 1254 |
+
"trainrewards/margins": 2.375,
|
| 1255 |
+
"trainrewards/rejected": -0.734375
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.67,
|
| 1259 |
+
"grad_norm": 4.356596196020246,
|
| 1260 |
+
"learning_rate": 1.4831583923105e-06,
|
| 1261 |
+
"loss": 2.4845,
|
| 1262 |
+
"step": 90,
|
| 1263 |
+
"trainloss/critic_chosen": 1.0889127254486084,
|
| 1264 |
+
"trainloss/critic_rejected": 1.1599314212799072,
|
| 1265 |
+
"trainloss/reward": 1.0889127254486084,
|
| 1266 |
+
"trainrewards/accuracies": 0.9583333134651184,
|
| 1267 |
+
"trainrewards/chosen": 1.875,
|
| 1268 |
+
"trainrewards/margins": 2.59375,
|
| 1269 |
+
"trainrewards/rejected": -0.7265625
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.68,
|
| 1273 |
+
"grad_norm": 3.484859150407605,
|
| 1274 |
+
"learning_rate": 1.4237222579792618e-06,
|
| 1275 |
+
"loss": 2.504,
|
| 1276 |
+
"step": 91,
|
| 1277 |
+
"trainloss/critic_chosen": 1.1031081676483154,
|
| 1278 |
+
"trainloss/critic_rejected": 1.1596983671188354,
|
| 1279 |
+
"trainloss/reward": 1.1031081676483154,
|
| 1280 |
+
"trainrewards/accuracies": 0.953125,
|
| 1281 |
+
"trainrewards/chosen": 1.7265625,
|
| 1282 |
+
"trainrewards/margins": 2.5,
|
| 1283 |
+
"trainrewards/rejected": -0.765625
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.69,
|
| 1287 |
+
"grad_norm": 3.5906077474254046,
|
| 1288 |
+
"learning_rate": 1.3650237506511333e-06,
|
| 1289 |
+
"loss": 2.497,
|
| 1290 |
+
"step": 92,
|
| 1291 |
+
"trainloss/critic_chosen": 1.1017568111419678,
|
| 1292 |
+
"trainloss/critic_rejected": 1.1597734689712524,
|
| 1293 |
+
"trainloss/reward": 1.1017568111419678,
|
| 1294 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 1295 |
+
"trainrewards/chosen": 1.734375,
|
| 1296 |
+
"trainrewards/margins": 2.609375,
|
| 1297 |
+
"trainrewards/rejected": -0.87109375
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.69,
|
| 1301 |
+
"grad_norm": 3.883326754801315,
|
| 1302 |
+
"learning_rate": 1.307103099350979e-06,
|
| 1303 |
+
"loss": 2.4881,
|
| 1304 |
+
"step": 93,
|
| 1305 |
+
"trainloss/critic_chosen": 1.1008602380752563,
|
| 1306 |
+
"trainloss/critic_rejected": 1.1622505187988281,
|
| 1307 |
+
"trainloss/reward": 1.1008602380752563,
|
| 1308 |
+
"trainrewards/accuracies": 0.9374999403953552,
|
| 1309 |
+
"trainrewards/chosen": 1.8359375,
|
| 1310 |
+
"trainrewards/margins": 2.65625,
|
| 1311 |
+
"trainrewards/rejected": -0.81640625
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.7,
|
| 1315 |
+
"grad_norm": 3.106807473961497,
|
| 1316 |
+
"learning_rate": 1.2500000000000007e-06,
|
| 1317 |
+
"loss": 2.5239,
|
| 1318 |
+
"step": 94,
|
| 1319 |
+
"trainloss/critic_chosen": 1.1186132431030273,
|
| 1320 |
+
"trainloss/critic_rejected": 1.1955691576004028,
|
| 1321 |
+
"trainloss/reward": 1.1186132431030273,
|
| 1322 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 1323 |
+
"trainrewards/chosen": 1.4453125,
|
| 1324 |
+
"trainrewards/margins": 2.78125,
|
| 1325 |
+
"trainrewards/rejected": -1.34375
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.71,
|
| 1329 |
+
"grad_norm": 3.0694983477589237,
|
| 1330 |
+
"learning_rate": 1.193753588210128e-06,
|
| 1331 |
+
"loss": 2.4975,
|
| 1332 |
+
"step": 95,
|
| 1333 |
+
"trainloss/critic_chosen": 1.089274287223816,
|
| 1334 |
+
"trainloss/critic_rejected": 1.1611120700836182,
|
| 1335 |
+
"trainloss/reward": 1.089274287223816,
|
| 1336 |
+
"trainrewards/accuracies": 0.9166667461395264,
|
| 1337 |
+
"trainrewards/chosen": 1.21875,
|
| 1338 |
+
"trainrewards/margins": 2.625,
|
| 1339 |
+
"trainrewards/rejected": -1.4140625
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.72,
|
| 1343 |
+
"grad_norm": 2.647849041797858,
|
| 1344 |
+
"learning_rate": 1.1384024124624324e-06,
|
| 1345 |
+
"loss": 2.4533,
|
| 1346 |
+
"step": 96,
|
| 1347 |
+
"trainloss/critic_chosen": 1.0731533765792847,
|
| 1348 |
+
"trainloss/critic_rejected": 1.1588420867919922,
|
| 1349 |
+
"trainloss/reward": 1.0731533765792847,
|
| 1350 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1351 |
+
"trainrewards/chosen": 1.2578125,
|
| 1352 |
+
"trainrewards/margins": 2.671875,
|
| 1353 |
+
"trainrewards/rejected": -1.40625
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.72,
|
| 1357 |
+
"grad_norm": 3.0284092019206743,
|
| 1358 |
+
"learning_rate": 1.0839844076879186e-06,
|
| 1359 |
+
"loss": 2.52,
|
| 1360 |
+
"step": 97,
|
| 1361 |
+
"trainloss/critic_chosen": 1.1046061515808105,
|
| 1362 |
+
"trainloss/critic_rejected": 1.1355366706848145,
|
| 1363 |
+
"trainloss/reward": 1.1046061515808105,
|
| 1364 |
+
"trainrewards/accuracies": 0.9114583134651184,
|
| 1365 |
+
"trainrewards/chosen": 1.5234375,
|
| 1366 |
+
"trainrewards/margins": 2.515625,
|
| 1367 |
+
"trainrewards/rejected": -1.0
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.73,
|
| 1371 |
+
"grad_norm": 3.239930376341791,
|
| 1372 |
+
"learning_rate": 1.0305368692688175e-06,
|
| 1373 |
+
"loss": 2.3829,
|
| 1374 |
+
"step": 98,
|
| 1375 |
+
"trainloss/critic_chosen": 1.0649518966674805,
|
| 1376 |
+
"trainloss/critic_rejected": 1.1148179769515991,
|
| 1377 |
+
"trainloss/reward": 1.0649518966674805,
|
| 1378 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1379 |
+
"trainrewards/chosen": 1.828125,
|
| 1380 |
+
"trainrewards/margins": 2.921875,
|
| 1381 |
+
"trainrewards/rejected": -1.0859375
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.74,
|
| 1385 |
+
"grad_norm": 2.844057334430093,
|
| 1386 |
+
"learning_rate": 9.780964274781984e-07,
|
| 1387 |
+
"loss": 2.4761,
|
| 1388 |
+
"step": 99,
|
| 1389 |
+
"trainloss/critic_chosen": 1.0876730680465698,
|
| 1390 |
+
"trainloss/critic_rejected": 1.1597809791564941,
|
| 1391 |
+
"trainloss/reward": 1.0876730680465698,
|
| 1392 |
+
"trainrewards/accuracies": 0.9583333134651184,
|
| 1393 |
+
"trainrewards/chosen": 1.65625,
|
| 1394 |
+
"trainrewards/margins": 2.703125,
|
| 1395 |
+
"trainrewards/rejected": -1.046875
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.74,
|
| 1399 |
+
"grad_norm": 2.523374395571222,
|
| 1400 |
+
"learning_rate": 9.266990223754069e-07,
|
| 1401 |
+
"loss": 2.4511,
|
| 1402 |
+
"step": 100,
|
| 1403 |
+
"trainloss/critic_chosen": 1.0983867645263672,
|
| 1404 |
+
"trainloss/critic_rejected": 1.1455085277557373,
|
| 1405 |
+
"trainloss/reward": 1.0983867645263672,
|
| 1406 |
+
"trainrewards/accuracies": 0.9791666865348816,
|
| 1407 |
+
"trainrewards/chosen": 1.5546875,
|
| 1408 |
+
"trainrewards/margins": 2.78125,
|
| 1409 |
+
"trainrewards/rejected": -1.21875
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.75,
|
| 1413 |
+
"grad_norm": 3.3175745581436917,
|
| 1414 |
+
"learning_rate": 8.763798791745413e-07,
|
| 1415 |
+
"loss": 2.453,
|
| 1416 |
+
"step": 101,
|
| 1417 |
+
"trainloss/critic_chosen": 1.094862699508667,
|
| 1418 |
+
"trainloss/critic_rejected": 1.1401726007461548,
|
| 1419 |
+
"trainloss/reward": 1.094862699508667,
|
| 1420 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1421 |
+
"trainrewards/chosen": 1.625,
|
| 1422 |
+
"trainrewards/margins": 2.78125,
|
| 1423 |
+
"trainrewards/rejected": -1.15625
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.76,
|
| 1427 |
+
"grad_norm": 2.8366252126004596,
|
| 1428 |
+
"learning_rate": 8.271734841028553e-07,
|
| 1429 |
+
"loss": 2.5483,
|
| 1430 |
+
"step": 102,
|
| 1431 |
+
"trainloss/critic_chosen": 1.0930631160736084,
|
| 1432 |
+
"trainloss/critic_rejected": 1.173164963722229,
|
| 1433 |
+
"trainloss/reward": 1.0930631160736084,
|
| 1434 |
+
"trainrewards/accuracies": 0.8958333134651184,
|
| 1435 |
+
"trainrewards/chosen": 1.390625,
|
| 1436 |
+
"trainrewards/margins": 2.515625,
|
| 1437 |
+
"trainrewards/rejected": -1.125
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.77,
|
| 1441 |
+
"grad_norm": 2.9707853433568023,
|
| 1442 |
+
"learning_rate": 7.791135607656147e-07,
|
| 1443 |
+
"loss": 2.3986,
|
| 1444 |
+
"step": 103,
|
| 1445 |
+
"trainloss/critic_chosen": 1.0701719522476196,
|
| 1446 |
+
"trainloss/critic_rejected": 1.1288470029830933,
|
| 1447 |
+
"trainloss/reward": 1.0701719522476196,
|
| 1448 |
+
"trainrewards/accuracies": 0.9791667461395264,
|
| 1449 |
+
"trainrewards/chosen": 1.6328125,
|
| 1450 |
+
"trainrewards/margins": 2.765625,
|
| 1451 |
+
"trainrewards/rejected": -1.1328125
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.77,
|
| 1455 |
+
"grad_norm": 3.309757754883857,
|
| 1456 |
+
"learning_rate": 7.322330470336314e-07,
|
| 1457 |
+
"loss": 2.429,
|
| 1458 |
+
"step": 104,
|
| 1459 |
+
"trainloss/critic_chosen": 1.0845046043395996,
|
| 1460 |
+
"trainloss/critic_rejected": 1.1261361837387085,
|
| 1461 |
+
"trainloss/reward": 1.0845046043395996,
|
| 1462 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1463 |
+
"trainrewards/chosen": 1.6796875,
|
| 1464 |
+
"trainrewards/margins": 2.71875,
|
| 1465 |
+
"trainrewards/rejected": -1.0234375
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.78,
|
| 1469 |
+
"grad_norm": 2.4159819079210556,
|
| 1470 |
+
"learning_rate": 6.865640724692815e-07,
|
| 1471 |
+
"loss": 2.3868,
|
| 1472 |
+
"step": 105,
|
| 1473 |
+
"trainloss/critic_chosen": 1.0498684644699097,
|
| 1474 |
+
"trainloss/critic_rejected": 1.131639003753662,
|
| 1475 |
+
"trainloss/reward": 1.0498684644699097,
|
| 1476 |
+
"trainrewards/accuracies": 0.9687500596046448,
|
| 1477 |
+
"trainrewards/chosen": 1.5,
|
| 1478 |
+
"trainrewards/margins": 2.90625,
|
| 1479 |
+
"trainrewards/rejected": -1.3984375
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.79,
|
| 1483 |
+
"grad_norm": 2.630877225161229,
|
| 1484 |
+
"learning_rate": 6.421379363065142e-07,
|
| 1485 |
+
"loss": 2.4745,
|
| 1486 |
+
"step": 106,
|
| 1487 |
+
"trainloss/critic_chosen": 1.0781538486480713,
|
| 1488 |
+
"trainloss/critic_rejected": 1.1649324893951416,
|
| 1489 |
+
"trainloss/reward": 1.0781538486480713,
|
| 1490 |
+
"trainrewards/accuracies": 0.9375,
|
| 1491 |
+
"trainrewards/chosen": 1.5,
|
| 1492 |
+
"trainrewards/margins": 2.78125,
|
| 1493 |
+
"trainrewards/rejected": -1.28125
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.8,
|
| 1497 |
+
"grad_norm": 2.71300394057213,
|
| 1498 |
+
"learning_rate": 5.989850859999227e-07,
|
| 1499 |
+
"loss": 2.4433,
|
| 1500 |
+
"step": 107,
|
| 1501 |
+
"trainloss/critic_chosen": 1.0875132083892822,
|
| 1502 |
+
"trainloss/critic_rejected": 1.1300991773605347,
|
| 1503 |
+
"trainloss/reward": 1.0875132083892822,
|
| 1504 |
+
"trainrewards/accuracies": 0.9635416865348816,
|
| 1505 |
+
"trainrewards/chosen": 1.4140625,
|
| 1506 |
+
"trainrewards/margins": 3.109375,
|
| 1507 |
+
"trainrewards/rejected": -1.703125
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.8,
|
| 1511 |
+
"grad_norm": 2.722489376200587,
|
| 1512 |
+
"learning_rate": 5.571350963575728e-07,
|
| 1513 |
+
"loss": 2.467,
|
| 1514 |
+
"step": 108,
|
| 1515 |
+
"trainloss/critic_chosen": 1.0709630250930786,
|
| 1516 |
+
"trainloss/critic_rejected": 1.154178500175476,
|
| 1517 |
+
"trainloss/reward": 1.0709630250930786,
|
| 1518 |
+
"trainrewards/accuracies": 0.9479166865348816,
|
| 1519 |
+
"trainrewards/chosen": 1.359375,
|
| 1520 |
+
"trainrewards/margins": 2.859375,
|
| 1521 |
+
"trainrewards/rejected": -1.5
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.81,
|
| 1525 |
+
"grad_norm": 3.255161744830997,
|
| 1526 |
+
"learning_rate": 5.166166492719124e-07,
|
| 1527 |
+
"loss": 2.4854,
|
| 1528 |
+
"step": 109,
|
| 1529 |
+
"trainloss/critic_chosen": 1.1081310510635376,
|
| 1530 |
+
"trainloss/critic_rejected": 1.152534008026123,
|
| 1531 |
+
"trainloss/reward": 1.1081310510635376,
|
| 1532 |
+
"trainrewards/accuracies": 0.973958432674408,
|
| 1533 |
+
"trainrewards/chosen": 1.34375,
|
| 1534 |
+
"trainrewards/margins": 2.96875,
|
| 1535 |
+
"trainrewards/rejected": -1.6328125
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.82,
|
| 1539 |
+
"grad_norm": 2.762498507683836,
|
| 1540 |
+
"learning_rate": 4.774575140626317e-07,
|
| 1541 |
+
"loss": 2.4388,
|
| 1542 |
+
"step": 110,
|
| 1543 |
+
"trainloss/critic_chosen": 1.065203070640564,
|
| 1544 |
+
"trainloss/critic_rejected": 1.0969582796096802,
|
| 1545 |
+
"trainloss/reward": 1.065203070640564,
|
| 1546 |
+
"trainrewards/accuracies": 0.9635416269302368,
|
| 1547 |
+
"trainrewards/chosen": 1.3046875,
|
| 1548 |
+
"trainrewards/margins": 2.609375,
|
| 1549 |
+
"trainrewards/rejected": -1.3125
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.83,
|
| 1553 |
+
"grad_norm": 2.780757216314426,
|
| 1554 |
+
"learning_rate": 4.396845284449608e-07,
|
| 1555 |
+
"loss": 2.4319,
|
| 1556 |
+
"step": 111,
|
| 1557 |
+
"trainloss/critic_chosen": 1.083713173866272,
|
| 1558 |
+
"trainloss/critic_rejected": 1.119750738143921,
|
| 1559 |
+
"trainloss/reward": 1.083713173866272,
|
| 1560 |
+
"trainrewards/accuracies": 0.9687500596046448,
|
| 1561 |
+
"trainrewards/chosen": 1.7421875,
|
| 1562 |
+
"trainrewards/margins": 3.03125,
|
| 1563 |
+
"trainrewards/rejected": -1.296875
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.83,
|
| 1567 |
+
"grad_norm": 3.7107544004289323,
|
| 1568 |
+
"learning_rate": 4.033235801364402e-07,
|
| 1569 |
+
"loss": 2.4846,
|
| 1570 |
+
"step": 112,
|
| 1571 |
+
"trainloss/critic_chosen": 1.106475830078125,
|
| 1572 |
+
"trainloss/critic_rejected": 1.1211233139038086,
|
| 1573 |
+
"trainloss/reward": 1.106475830078125,
|
| 1574 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1575 |
+
"trainrewards/chosen": 1.7421875,
|
| 1576 |
+
"trainrewards/margins": 2.703125,
|
| 1577 |
+
"trainrewards/rejected": -0.96484375
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.84,
|
| 1581 |
+
"grad_norm": 3.3073751739512787,
|
| 1582 |
+
"learning_rate": 3.683995891147696e-07,
|
| 1583 |
+
"loss": 2.4629,
|
| 1584 |
+
"step": 113,
|
| 1585 |
+
"trainloss/critic_chosen": 1.0521959066390991,
|
| 1586 |
+
"trainloss/critic_rejected": 1.173767328262329,
|
| 1587 |
+
"trainloss/reward": 1.0521959066390991,
|
| 1588 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1589 |
+
"trainrewards/chosen": 1.828125,
|
| 1590 |
+
"trainrewards/margins": 2.921875,
|
| 1591 |
+
"trainrewards/rejected": -1.0859375
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.85,
|
| 1595 |
+
"grad_norm": 2.99774406823753,
|
| 1596 |
+
"learning_rate": 3.3493649053890325e-07,
|
| 1597 |
+
"loss": 2.536,
|
| 1598 |
+
"step": 114,
|
| 1599 |
+
"trainloss/critic_chosen": 1.1110682487487793,
|
| 1600 |
+
"trainloss/critic_rejected": 1.155356526374817,
|
| 1601 |
+
"trainloss/reward": 1.1110682487487793,
|
| 1602 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 1603 |
+
"trainrewards/chosen": 1.5859375,
|
| 1604 |
+
"trainrewards/margins": 2.75,
|
| 1605 |
+
"trainrewards/rejected": -1.1640625
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.86,
|
| 1609 |
+
"grad_norm": 3.513330205624271,
|
| 1610 |
+
"learning_rate": 3.0295721834508686e-07,
|
| 1611 |
+
"loss": 2.4707,
|
| 1612 |
+
"step": 115,
|
| 1613 |
+
"trainloss/critic_chosen": 1.0783016681671143,
|
| 1614 |
+
"trainloss/critic_rejected": 1.1236062049865723,
|
| 1615 |
+
"trainloss/reward": 1.0783016681671143,
|
| 1616 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 1617 |
+
"trainrewards/chosen": 1.703125,
|
| 1618 |
+
"trainrewards/margins": 2.671875,
|
| 1619 |
+
"trainrewards/rejected": -0.97265625
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.86,
|
| 1623 |
+
"grad_norm": 2.800803642232422,
|
| 1624 |
+
"learning_rate": 2.7248368952908055e-07,
|
| 1625 |
+
"loss": 2.4803,
|
| 1626 |
+
"step": 116,
|
| 1627 |
+
"trainloss/critic_chosen": 1.080200433731079,
|
| 1628 |
+
"trainloss/critic_rejected": 1.1515780687332153,
|
| 1629 |
+
"trainloss/reward": 1.080200433731079,
|
| 1630 |
+
"trainrewards/accuracies": 0.9375,
|
| 1631 |
+
"trainrewards/chosen": 1.5546875,
|
| 1632 |
+
"trainrewards/margins": 2.5625,
|
| 1633 |
+
"trainrewards/rejected": -1.0
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.87,
|
| 1637 |
+
"grad_norm": 2.7889069352140585,
|
| 1638 |
+
"learning_rate": 2.43536789125349e-07,
|
| 1639 |
+
"loss": 2.4905,
|
| 1640 |
+
"step": 117,
|
| 1641 |
+
"trainloss/critic_chosen": 1.088797688484192,
|
| 1642 |
+
"trainloss/critic_rejected": 1.1520254611968994,
|
| 1643 |
+
"trainloss/reward": 1.088797688484192,
|
| 1644 |
+
"trainrewards/accuracies": 0.9375,
|
| 1645 |
+
"trainrewards/chosen": 1.5,
|
| 1646 |
+
"trainrewards/margins": 2.515625,
|
| 1647 |
+
"trainrewards/rejected": -1.0078125
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.88,
|
| 1651 |
+
"grad_norm": 2.931939214492335,
|
| 1652 |
+
"learning_rate": 2.1613635589349756e-07,
|
| 1653 |
+
"loss": 2.3937,
|
| 1654 |
+
"step": 118,
|
| 1655 |
+
"trainloss/critic_chosen": 1.0556377172470093,
|
| 1656 |
+
"trainloss/critic_rejected": 1.1278637647628784,
|
| 1657 |
+
"trainloss/reward": 1.0556377172470093,
|
| 1658 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1659 |
+
"trainrewards/chosen": 1.4375,
|
| 1660 |
+
"trainrewards/margins": 2.46875,
|
| 1661 |
+
"trainrewards/rejected": -1.03125
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.89,
|
| 1665 |
+
"grad_norm": 2.954190958849344,
|
| 1666 |
+
"learning_rate": 1.9030116872178317e-07,
|
| 1667 |
+
"loss": 2.418,
|
| 1668 |
+
"step": 119,
|
| 1669 |
+
"trainloss/critic_chosen": 1.0978080034255981,
|
| 1670 |
+
"trainloss/critic_rejected": 1.1148779392242432,
|
| 1671 |
+
"trainloss/reward": 1.0978080034255981,
|
| 1672 |
+
"trainrewards/accuracies": 0.9687500596046448,
|
| 1673 |
+
"trainrewards/chosen": 1.4453125,
|
| 1674 |
+
"trainrewards/margins": 2.46875,
|
| 1675 |
+
"trainrewards/rejected": -1.03125
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.89,
|
| 1679 |
+
"grad_norm": 2.9483773523832353,
|
| 1680 |
+
"learning_rate": 1.6604893375699594e-07,
|
| 1681 |
+
"loss": 2.4694,
|
| 1682 |
+
"step": 120,
|
| 1683 |
+
"trainloss/critic_chosen": 1.1018942594528198,
|
| 1684 |
+
"trainloss/critic_rejected": 1.1370322704315186,
|
| 1685 |
+
"trainloss/reward": 1.1018942594528198,
|
| 1686 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1687 |
+
"trainrewards/chosen": 1.3515625,
|
| 1688 |
+
"trainrewards/margins": 2.40625,
|
| 1689 |
+
"trainrewards/rejected": -1.0546875
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.9,
|
| 1693 |
+
"grad_norm": 2.9215978058037764,
|
| 1694 |
+
"learning_rate": 1.4339627226955394e-07,
|
| 1695 |
+
"loss": 2.4822,
|
| 1696 |
+
"step": 121,
|
| 1697 |
+
"trainloss/critic_chosen": 1.1052017211914062,
|
| 1698 |
+
"trainloss/critic_rejected": 1.148177981376648,
|
| 1699 |
+
"trainloss/reward": 1.1052017211914062,
|
| 1700 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1701 |
+
"trainrewards/chosen": 1.3515625,
|
| 1702 |
+
"trainrewards/margins": 2.515625,
|
| 1703 |
+
"trainrewards/rejected": -1.1640625
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.91,
|
| 1707 |
+
"grad_norm": 2.8843923021301667,
|
| 1708 |
+
"learning_rate": 1.223587092621162e-07,
|
| 1709 |
+
"loss": 2.4942,
|
| 1710 |
+
"step": 122,
|
| 1711 |
+
"trainloss/critic_chosen": 1.0736005306243896,
|
| 1712 |
+
"trainloss/critic_rejected": 1.168089509010315,
|
| 1713 |
+
"trainloss/reward": 1.0736005306243896,
|
| 1714 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1715 |
+
"trainrewards/chosen": 1.3359375,
|
| 1716 |
+
"trainrewards/margins": 2.328125,
|
| 1717 |
+
"trainrewards/rejected": -0.99609375
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.92,
|
| 1721 |
+
"grad_norm": 2.8724683106941193,
|
| 1722 |
+
"learning_rate": 1.0295066282951738e-07,
|
| 1723 |
+
"loss": 2.4881,
|
| 1724 |
+
"step": 123,
|
| 1725 |
+
"trainloss/critic_chosen": 1.09504234790802,
|
| 1726 |
+
"trainloss/critic_rejected": 1.1352362632751465,
|
| 1727 |
+
"trainloss/reward": 1.09504234790802,
|
| 1728 |
+
"trainrewards/accuracies": 0.9322916865348816,
|
| 1729 |
+
"trainrewards/chosen": 1.4375,
|
| 1730 |
+
"trainrewards/margins": 2.3125,
|
| 1731 |
+
"trainrewards/rejected": -0.87109375
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.92,
|
| 1735 |
+
"grad_norm": 3.0064917280475045,
|
| 1736 |
+
"learning_rate": 8.518543427732951e-08,
|
| 1737 |
+
"loss": 2.5066,
|
| 1738 |
+
"step": 124,
|
| 1739 |
+
"trainloss/critic_chosen": 1.0965893268585205,
|
| 1740 |
+
"trainloss/critic_rejected": 1.12990403175354,
|
| 1741 |
+
"trainloss/reward": 1.0965893268585205,
|
| 1742 |
+
"trainrewards/accuracies": 0.9166667461395264,
|
| 1743 |
+
"trainrewards/chosen": 1.4375,
|
| 1744 |
+
"trainrewards/margins": 2.3125,
|
| 1745 |
+
"trainrewards/rejected": -0.8828125
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.93,
|
| 1749 |
+
"grad_norm": 2.6882210161223425,
|
| 1750 |
+
"learning_rate": 6.907519900580862e-08,
|
| 1751 |
+
"loss": 2.3973,
|
| 1752 |
+
"step": 125,
|
| 1753 |
+
"trainloss/critic_chosen": 1.0724809169769287,
|
| 1754 |
+
"trainloss/critic_rejected": 1.1239736080169678,
|
| 1755 |
+
"trainloss/reward": 1.0724809169769287,
|
| 1756 |
+
"trainrewards/accuracies": 0.9687500596046448,
|
| 1757 |
+
"trainrewards/chosen": 1.546875,
|
| 1758 |
+
"trainrewards/margins": 2.5625,
|
| 1759 |
+
"trainrewards/rejected": -1.015625
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.94,
|
| 1763 |
+
"grad_norm": 3.2130299463812233,
|
| 1764 |
+
"learning_rate": 5.463099816548578e-08,
|
| 1765 |
+
"loss": 2.4583,
|
| 1766 |
+
"step": 126,
|
| 1767 |
+
"trainloss/critic_chosen": 1.053167700767517,
|
| 1768 |
+
"trainloss/critic_rejected": 1.1157554388046265,
|
| 1769 |
+
"trainloss/reward": 1.053167700767517,
|
| 1770 |
+
"trainrewards/accuracies": 0.9270833730697632,
|
| 1771 |
+
"trainrewards/chosen": 1.390625,
|
| 1772 |
+
"trainrewards/margins": 2.171875,
|
| 1773 |
+
"trainrewards/rejected": -0.7890625
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.95,
|
| 1777 |
+
"grad_norm": 2.5537231163007004,
|
| 1778 |
+
"learning_rate": 4.186273109011374e-08,
|
| 1779 |
+
"loss": 2.5432,
|
| 1780 |
+
"step": 127,
|
| 1781 |
+
"trainloss/critic_chosen": 1.1048160791397095,
|
| 1782 |
+
"trainloss/critic_rejected": 1.1709802150726318,
|
| 1783 |
+
"trainloss/reward": 1.1048160791397095,
|
| 1784 |
+
"trainrewards/accuracies": 0.9270833134651184,
|
| 1785 |
+
"trainrewards/chosen": 1.234375,
|
| 1786 |
+
"trainrewards/margins": 2.296875,
|
| 1787 |
+
"trainrewards/rejected": -1.0546875
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.95,
|
| 1791 |
+
"grad_norm": 3.455571318987563,
|
| 1792 |
+
"learning_rate": 3.077914851215585e-08,
|
| 1793 |
+
"loss": 2.4356,
|
| 1794 |
+
"step": 128,
|
| 1795 |
+
"trainloss/critic_chosen": 1.0750889778137207,
|
| 1796 |
+
"trainloss/critic_rejected": 1.1610357761383057,
|
| 1797 |
+
"trainloss/reward": 1.0750889778137207,
|
| 1798 |
+
"trainrewards/accuracies": 0.9635416865348816,
|
| 1799 |
+
"trainrewards/chosen": 1.734375,
|
| 1800 |
+
"trainrewards/margins": 2.625,
|
| 1801 |
+
"trainrewards/rejected": -0.89453125
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.96,
|
| 1805 |
+
"grad_norm": 3.095008653826813,
|
| 1806 |
+
"learning_rate": 2.1387846565474047e-08,
|
| 1807 |
+
"loss": 2.4339,
|
| 1808 |
+
"step": 129,
|
| 1809 |
+
"trainloss/critic_chosen": 1.0719571113586426,
|
| 1810 |
+
"trainloss/critic_rejected": 1.1504600048065186,
|
| 1811 |
+
"trainloss/reward": 1.0719571113586426,
|
| 1812 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1813 |
+
"trainrewards/chosen": 1.6328125,
|
| 1814 |
+
"trainrewards/margins": 2.59375,
|
| 1815 |
+
"trainrewards/rejected": -0.96484375
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.97,
|
| 1819 |
+
"grad_norm": 3.014103190303491,
|
| 1820 |
+
"learning_rate": 1.3695261579316776e-08,
|
| 1821 |
+
"loss": 2.4359,
|
| 1822 |
+
"step": 130,
|
| 1823 |
+
"trainloss/critic_chosen": 1.0608913898468018,
|
| 1824 |
+
"trainloss/critic_rejected": 1.1623928546905518,
|
| 1825 |
+
"trainloss/reward": 1.0608913898468018,
|
| 1826 |
+
"trainrewards/accuracies": 0.9791667461395264,
|
| 1827 |
+
"trainrewards/chosen": 1.5078125,
|
| 1828 |
+
"trainrewards/margins": 2.453125,
|
| 1829 |
+
"trainrewards/rejected": -0.94921875
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.98,
|
| 1833 |
+
"grad_norm": 3.5824668969231825,
|
| 1834 |
+
"learning_rate": 7.70666566718009e-09,
|
| 1835 |
+
"loss": 2.457,
|
| 1836 |
+
"step": 131,
|
| 1837 |
+
"trainloss/critic_chosen": 1.0644171237945557,
|
| 1838 |
+
"trainloss/critic_rejected": 1.1539928913116455,
|
| 1839 |
+
"trainloss/reward": 1.0644171237945557,
|
| 1840 |
+
"trainrewards/accuracies": 0.9583333730697632,
|
| 1841 |
+
"trainrewards/chosen": 1.578125,
|
| 1842 |
+
"trainrewards/margins": 2.5,
|
| 1843 |
+
"trainrewards/rejected": -0.91796875
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.98,
|
| 1847 |
+
"grad_norm": 3.085175163660414,
|
| 1848 |
+
"learning_rate": 3.4261631135654174e-09,
|
| 1849 |
+
"loss": 2.4695,
|
| 1850 |
+
"step": 132,
|
| 1851 |
+
"trainloss/critic_chosen": 1.0733301639556885,
|
| 1852 |
+
"trainloss/critic_rejected": 1.1289600133895874,
|
| 1853 |
+
"trainloss/reward": 1.0733301639556885,
|
| 1854 |
+
"trainrewards/accuracies": 0.9427083730697632,
|
| 1855 |
+
"trainrewards/chosen": 1.484375,
|
| 1856 |
+
"trainrewards/margins": 2.296875,
|
| 1857 |
+
"trainrewards/rejected": -0.80859375
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.99,
|
| 1861 |
+
"grad_norm": 2.6627177185366087,
|
| 1862 |
+
"learning_rate": 8.566875611068503e-10,
|
| 1863 |
+
"loss": 2.456,
|
| 1864 |
+
"step": 133,
|
| 1865 |
+
"trainloss/critic_chosen": 1.0969208478927612,
|
| 1866 |
+
"trainloss/critic_rejected": 1.1649024486541748,
|
| 1867 |
+
"trainloss/reward": 1.0969208478927612,
|
| 1868 |
+
"trainrewards/accuracies": 0.96875,
|
| 1869 |
+
"trainrewards/chosen": 1.4765625,
|
| 1870 |
+
"trainrewards/margins": 2.59375,
|
| 1871 |
+
"trainrewards/rejected": -1.1171875
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.0,
|
| 1875 |
+
"grad_norm": 2.6690531080971973,
|
| 1876 |
+
"learning_rate": 0.0,
|
| 1877 |
+
"loss": 2.4519,
|
| 1878 |
+
"step": 134,
|
| 1879 |
+
"trainloss/critic_chosen": 1.090218186378479,
|
| 1880 |
+
"trainloss/critic_rejected": 1.128463625907898,
|
| 1881 |
+
"trainloss/reward": 1.090218186378479,
|
| 1882 |
+
"trainrewards/accuracies": 0.9531250596046448,
|
| 1883 |
+
"trainrewards/chosen": 1.5078125,
|
| 1884 |
+
"trainrewards/margins": 2.53125,
|
| 1885 |
+
"trainrewards/rejected": -1.0234375
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.0,
|
| 1889 |
+
"step": 134,
|
| 1890 |
+
"total_flos": 0.0,
|
| 1891 |
+
"train_loss": 2.6233635464710976,
|
| 1892 |
+
"train_runtime": 32287.388,
|
| 1893 |
+
"train_samples_per_second": 0.799,
|
| 1894 |
+
"train_steps_per_second": 0.004
|
| 1895 |
+
}
|
| 1896 |
+
],
|
| 1897 |
+
"logging_steps": 1.0,
|
| 1898 |
+
"max_steps": 134,
|
| 1899 |
+
"num_input_tokens_seen": 0,
|
| 1900 |
+
"num_train_epochs": 1,
|
| 1901 |
+
"save_steps": 1000,
|
| 1902 |
+
"total_flos": 0.0,
|
| 1903 |
+
"train_batch_size": 1,
|
| 1904 |
+
"trial_name": null,
|
| 1905 |
+
"trial_params": null
|
| 1906 |
+
}
|