Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +3641 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,3641 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5142,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.011668611435239206,
|
| 14 |
+
"grad_norm": 19.682453155517578,
|
| 15 |
+
"learning_rate": 3.488372093023256e-06,
|
| 16 |
+
"loss": 1.1685,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.023337222870478413,
|
| 21 |
+
"grad_norm": 2.1974496841430664,
|
| 22 |
+
"learning_rate": 7.364341085271319e-06,
|
| 23 |
+
"loss": 0.5269,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.03500583430571762,
|
| 28 |
+
"grad_norm": 2.614182710647583,
|
| 29 |
+
"learning_rate": 1.1240310077519382e-05,
|
| 30 |
+
"loss": 0.3282,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.046674445740956826,
|
| 35 |
+
"grad_norm": 1.4804197549819946,
|
| 36 |
+
"learning_rate": 1.5116279069767441e-05,
|
| 37 |
+
"loss": 0.2738,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.058343057176196034,
|
| 42 |
+
"grad_norm": 1.9591115713119507,
|
| 43 |
+
"learning_rate": 1.8992248062015506e-05,
|
| 44 |
+
"loss": 0.2523,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.07001166861143523,
|
| 49 |
+
"grad_norm": 2.3735320568084717,
|
| 50 |
+
"learning_rate": 2.2868217054263565e-05,
|
| 51 |
+
"loss": 0.2531,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.08168028004667445,
|
| 56 |
+
"grad_norm": 0.9114485383033752,
|
| 57 |
+
"learning_rate": 2.674418604651163e-05,
|
| 58 |
+
"loss": 0.238,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.09334889148191365,
|
| 63 |
+
"grad_norm": 1.5135276317596436,
|
| 64 |
+
"learning_rate": 3.062015503875969e-05,
|
| 65 |
+
"loss": 0.2141,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.10501750291715285,
|
| 70 |
+
"grad_norm": 1.330152153968811,
|
| 71 |
+
"learning_rate": 3.449612403100775e-05,
|
| 72 |
+
"loss": 0.2139,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.11668611435239207,
|
| 77 |
+
"grad_norm": 1.6825010776519775,
|
| 78 |
+
"learning_rate": 3.837209302325582e-05,
|
| 79 |
+
"loss": 0.1916,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.12835472578763127,
|
| 84 |
+
"grad_norm": 1.386888861656189,
|
| 85 |
+
"learning_rate": 4.2248062015503877e-05,
|
| 86 |
+
"loss": 0.1791,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.14002333722287047,
|
| 91 |
+
"grad_norm": 1.1155707836151123,
|
| 92 |
+
"learning_rate": 4.6124031007751936e-05,
|
| 93 |
+
"loss": 0.1535,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.1516919486581097,
|
| 98 |
+
"grad_norm": 1.7495779991149902,
|
| 99 |
+
"learning_rate": 5e-05,
|
| 100 |
+
"loss": 0.1359,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.1633605600933489,
|
| 105 |
+
"grad_norm": 1.5433738231658936,
|
| 106 |
+
"learning_rate": 5.387596899224806e-05,
|
| 107 |
+
"loss": 0.1008,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.1750291715285881,
|
| 112 |
+
"grad_norm": 1.1212965250015259,
|
| 113 |
+
"learning_rate": 5.775193798449613e-05,
|
| 114 |
+
"loss": 0.0815,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.1866977829638273,
|
| 119 |
+
"grad_norm": 0.9702732563018799,
|
| 120 |
+
"learning_rate": 6.162790697674418e-05,
|
| 121 |
+
"loss": 0.0685,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.1983663943990665,
|
| 126 |
+
"grad_norm": 1.0748980045318604,
|
| 127 |
+
"learning_rate": 6.550387596899226e-05,
|
| 128 |
+
"loss": 0.0561,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.2100350058343057,
|
| 133 |
+
"grad_norm": 1.2040292024612427,
|
| 134 |
+
"learning_rate": 6.937984496124031e-05,
|
| 135 |
+
"loss": 0.0638,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.22170361726954493,
|
| 140 |
+
"grad_norm": 1.2262201309204102,
|
| 141 |
+
"learning_rate": 7.325581395348837e-05,
|
| 142 |
+
"loss": 0.0643,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.23337222870478413,
|
| 147 |
+
"grad_norm": 1.1682775020599365,
|
| 148 |
+
"learning_rate": 7.713178294573645e-05,
|
| 149 |
+
"loss": 0.0598,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.24504084014002334,
|
| 154 |
+
"grad_norm": 0.9979167580604553,
|
| 155 |
+
"learning_rate": 8.10077519379845e-05,
|
| 156 |
+
"loss": 0.0547,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.25670945157526254,
|
| 161 |
+
"grad_norm": 0.7736290097236633,
|
| 162 |
+
"learning_rate": 8.488372093023255e-05,
|
| 163 |
+
"loss": 0.0472,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.26837806301050177,
|
| 168 |
+
"grad_norm": 0.6784246563911438,
|
| 169 |
+
"learning_rate": 8.875968992248063e-05,
|
| 170 |
+
"loss": 0.0492,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.28004667444574094,
|
| 175 |
+
"grad_norm": 0.9692291617393494,
|
| 176 |
+
"learning_rate": 9.263565891472868e-05,
|
| 177 |
+
"loss": 0.0489,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.29171528588098017,
|
| 182 |
+
"grad_norm": 0.7258503437042236,
|
| 183 |
+
"learning_rate": 9.651162790697675e-05,
|
| 184 |
+
"loss": 0.0446,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.3033838973162194,
|
| 189 |
+
"grad_norm": 0.8032006025314331,
|
| 190 |
+
"learning_rate": 9.9999989656002e-05,
|
| 191 |
+
"loss": 0.0435,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.31505250875145857,
|
| 196 |
+
"grad_norm": 0.5978022217750549,
|
| 197 |
+
"learning_rate": 9.999874838141888e-05,
|
| 198 |
+
"loss": 0.055,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.3267211201866978,
|
| 203 |
+
"grad_norm": 1.1720023155212402,
|
| 204 |
+
"learning_rate": 9.999543836608138e-05,
|
| 205 |
+
"loss": 0.0541,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.338389731621937,
|
| 210 |
+
"grad_norm": 1.0916252136230469,
|
| 211 |
+
"learning_rate": 9.999005974694422e-05,
|
| 212 |
+
"loss": 0.0544,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.3500583430571762,
|
| 217 |
+
"grad_norm": 0.8868196606636047,
|
| 218 |
+
"learning_rate": 9.99826127465523e-05,
|
| 219 |
+
"loss": 0.0504,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.3617269544924154,
|
| 224 |
+
"grad_norm": 0.7354416251182556,
|
| 225 |
+
"learning_rate": 9.997309767303165e-05,
|
| 226 |
+
"loss": 0.0496,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.3733955659276546,
|
| 231 |
+
"grad_norm": 0.6354748010635376,
|
| 232 |
+
"learning_rate": 9.99615149200765e-05,
|
| 233 |
+
"loss": 0.0494,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.38506417736289383,
|
| 238 |
+
"grad_norm": 0.7520553469657898,
|
| 239 |
+
"learning_rate": 9.99478649669331e-05,
|
| 240 |
+
"loss": 0.0477,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.396732788798133,
|
| 245 |
+
"grad_norm": 0.7409453392028809,
|
| 246 |
+
"learning_rate": 9.99321483783799e-05,
|
| 247 |
+
"loss": 0.0453,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.40840140023337224,
|
| 252 |
+
"grad_norm": 0.7333018183708191,
|
| 253 |
+
"learning_rate": 9.99143658047041e-05,
|
| 254 |
+
"loss": 0.049,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.4200700116686114,
|
| 259 |
+
"grad_norm": 0.6608542799949646,
|
| 260 |
+
"learning_rate": 9.989451798167485e-05,
|
| 261 |
+
"loss": 0.0477,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.43173862310385064,
|
| 266 |
+
"grad_norm": 0.6022766828536987,
|
| 267 |
+
"learning_rate": 9.987260573051269e-05,
|
| 268 |
+
"loss": 0.0426,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.44340723453908987,
|
| 273 |
+
"grad_norm": 1.0430063009262085,
|
| 274 |
+
"learning_rate": 9.984862995785564e-05,
|
| 275 |
+
"loss": 0.0399,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.45507584597432904,
|
| 280 |
+
"grad_norm": 0.9709864258766174,
|
| 281 |
+
"learning_rate": 9.982259165572172e-05,
|
| 282 |
+
"loss": 0.0439,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.46674445740956827,
|
| 287 |
+
"grad_norm": 0.6489934325218201,
|
| 288 |
+
"learning_rate": 9.979449190146783e-05,
|
| 289 |
+
"loss": 0.0483,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.47841306884480744,
|
| 294 |
+
"grad_norm": 0.47625336050987244,
|
| 295 |
+
"learning_rate": 9.976433185774521e-05,
|
| 296 |
+
"loss": 0.0337,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.49008168028004667,
|
| 301 |
+
"grad_norm": 0.6839993596076965,
|
| 302 |
+
"learning_rate": 9.973211277245132e-05,
|
| 303 |
+
"loss": 0.0384,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.5017502917152858,
|
| 308 |
+
"grad_norm": 0.8791218400001526,
|
| 309 |
+
"learning_rate": 9.969783597867823e-05,
|
| 310 |
+
"loss": 0.038,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.5134189031505251,
|
| 315 |
+
"grad_norm": 0.6247888803482056,
|
| 316 |
+
"learning_rate": 9.966150289465746e-05,
|
| 317 |
+
"loss": 0.0457,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.5250875145857643,
|
| 322 |
+
"grad_norm": 0.7907714247703552,
|
| 323 |
+
"learning_rate": 9.962311502370125e-05,
|
| 324 |
+
"loss": 0.0487,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.5367561260210035,
|
| 329 |
+
"grad_norm": 0.6320911049842834,
|
| 330 |
+
"learning_rate": 9.958267395414043e-05,
|
| 331 |
+
"loss": 0.053,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.5484247374562428,
|
| 336 |
+
"grad_norm": 0.8405010104179382,
|
| 337 |
+
"learning_rate": 9.954018135925868e-05,
|
| 338 |
+
"loss": 0.0441,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.5600933488914819,
|
| 343 |
+
"grad_norm": 0.8460413813591003,
|
| 344 |
+
"learning_rate": 9.949563899722325e-05,
|
| 345 |
+
"loss": 0.04,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.5717619603267211,
|
| 350 |
+
"grad_norm": 0.4561885595321655,
|
| 351 |
+
"learning_rate": 9.944904871101228e-05,
|
| 352 |
+
"loss": 0.0379,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.5834305717619603,
|
| 357 |
+
"grad_norm": 0.34813910722732544,
|
| 358 |
+
"learning_rate": 9.940041242833846e-05,
|
| 359 |
+
"loss": 0.0334,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.5950991831971996,
|
| 364 |
+
"grad_norm": 0.74653160572052,
|
| 365 |
+
"learning_rate": 9.934973216156944e-05,
|
| 366 |
+
"loss": 0.0327,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.6067677946324388,
|
| 371 |
+
"grad_norm": 1.3376106023788452,
|
| 372 |
+
"learning_rate": 9.929701000764431e-05,
|
| 373 |
+
"loss": 0.0377,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.6184364060676779,
|
| 378 |
+
"grad_norm": 0.8607772588729858,
|
| 379 |
+
"learning_rate": 9.924224814798709e-05,
|
| 380 |
+
"loss": 0.0416,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.6301050175029171,
|
| 385 |
+
"grad_norm": 0.7919967770576477,
|
| 386 |
+
"learning_rate": 9.918544884841627e-05,
|
| 387 |
+
"loss": 0.0417,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.6417736289381564,
|
| 392 |
+
"grad_norm": 0.5088178515434265,
|
| 393 |
+
"learning_rate": 9.912661445905125e-05,
|
| 394 |
+
"loss": 0.0429,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.6534422403733956,
|
| 399 |
+
"grad_norm": 0.7714805603027344,
|
| 400 |
+
"learning_rate": 9.906574741421491e-05,
|
| 401 |
+
"loss": 0.0486,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.6651108518086347,
|
| 406 |
+
"grad_norm": 0.5451911091804504,
|
| 407 |
+
"learning_rate": 9.900285023233304e-05,
|
| 408 |
+
"loss": 0.041,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.676779463243874,
|
| 413 |
+
"grad_norm": 0.45999377965927124,
|
| 414 |
+
"learning_rate": 9.893792551583004e-05,
|
| 415 |
+
"loss": 0.0342,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.6884480746791132,
|
| 420 |
+
"grad_norm": 0.597261369228363,
|
| 421 |
+
"learning_rate": 9.887097595102129e-05,
|
| 422 |
+
"loss": 0.0391,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.7001166861143524,
|
| 427 |
+
"grad_norm": 0.723434567451477,
|
| 428 |
+
"learning_rate": 9.8802004308002e-05,
|
| 429 |
+
"loss": 0.0349,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.7117852975495916,
|
| 434 |
+
"grad_norm": 0.5449537038803101,
|
| 435 |
+
"learning_rate": 9.873101344053261e-05,
|
| 436 |
+
"loss": 0.0299,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.7234539089848308,
|
| 441 |
+
"grad_norm": 0.6254649758338928,
|
| 442 |
+
"learning_rate": 9.865800628592061e-05,
|
| 443 |
+
"loss": 0.0383,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.73512252042007,
|
| 448 |
+
"grad_norm": 0.604762077331543,
|
| 449 |
+
"learning_rate": 9.85829858648992e-05,
|
| 450 |
+
"loss": 0.0373,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.7467911318553092,
|
| 455 |
+
"grad_norm": 0.7584523558616638,
|
| 456 |
+
"learning_rate": 9.850595528150207e-05,
|
| 457 |
+
"loss": 0.0336,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.7584597432905484,
|
| 462 |
+
"grad_norm": 0.5650172233581543,
|
| 463 |
+
"learning_rate": 9.84269177229352e-05,
|
| 464 |
+
"loss": 0.0344,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.7701283547257877,
|
| 469 |
+
"grad_norm": 0.381675124168396,
|
| 470 |
+
"learning_rate": 9.834587645944477e-05,
|
| 471 |
+
"loss": 0.0306,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.7817969661610268,
|
| 476 |
+
"grad_norm": 0.2907174527645111,
|
| 477 |
+
"learning_rate": 9.826283484418205e-05,
|
| 478 |
+
"loss": 0.0319,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.793465577596266,
|
| 483 |
+
"grad_norm": 0.6129659414291382,
|
| 484 |
+
"learning_rate": 9.817779631306451e-05,
|
| 485 |
+
"loss": 0.0376,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.8051341890315052,
|
| 490 |
+
"grad_norm": 0.7693837881088257,
|
| 491 |
+
"learning_rate": 9.809076438463372e-05,
|
| 492 |
+
"loss": 0.0319,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.8168028004667445,
|
| 497 |
+
"grad_norm": 0.7197820544242859,
|
| 498 |
+
"learning_rate": 9.800174265990978e-05,
|
| 499 |
+
"loss": 0.0391,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.8284714119019837,
|
| 504 |
+
"grad_norm": 0.5524908900260925,
|
| 505 |
+
"learning_rate": 9.79107348222423e-05,
|
| 506 |
+
"loss": 0.0317,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.8401400233372228,
|
| 511 |
+
"grad_norm": 0.6571870446205139,
|
| 512 |
+
"learning_rate": 9.781774463715797e-05,
|
| 513 |
+
"loss": 0.0299,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.851808634772462,
|
| 518 |
+
"grad_norm": 0.5467213988304138,
|
| 519 |
+
"learning_rate": 9.772277595220483e-05,
|
| 520 |
+
"loss": 0.0378,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.8634772462077013,
|
| 525 |
+
"grad_norm": 0.6869328618049622,
|
| 526 |
+
"learning_rate": 9.762583269679303e-05,
|
| 527 |
+
"loss": 0.0398,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.8751458576429405,
|
| 532 |
+
"grad_norm": 0.5763711929321289,
|
| 533 |
+
"learning_rate": 9.752691888203222e-05,
|
| 534 |
+
"loss": 0.0358,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.8868144690781797,
|
| 539 |
+
"grad_norm": 0.45884543657302856,
|
| 540 |
+
"learning_rate": 9.742603860056567e-05,
|
| 541 |
+
"loss": 0.0344,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.8984830805134189,
|
| 546 |
+
"grad_norm": 0.5863968133926392,
|
| 547 |
+
"learning_rate": 9.732319602640086e-05,
|
| 548 |
+
"loss": 0.0342,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.9101516919486581,
|
| 553 |
+
"grad_norm": 0.5996890068054199,
|
| 554 |
+
"learning_rate": 9.721839541473676e-05,
|
| 555 |
+
"loss": 0.0332,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.9218203033838973,
|
| 560 |
+
"grad_norm": 0.6107974052429199,
|
| 561 |
+
"learning_rate": 9.711164110178788e-05,
|
| 562 |
+
"loss": 0.0332,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.9334889148191365,
|
| 567 |
+
"grad_norm": 0.8212049603462219,
|
| 568 |
+
"learning_rate": 9.700293750460473e-05,
|
| 569 |
+
"loss": 0.0355,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.9451575262543758,
|
| 574 |
+
"grad_norm": 0.5446476936340332,
|
| 575 |
+
"learning_rate": 9.689228912089113e-05,
|
| 576 |
+
"loss": 0.038,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.9568261376896149,
|
| 581 |
+
"grad_norm": 0.7654525637626648,
|
| 582 |
+
"learning_rate": 9.67797005288181e-05,
|
| 583 |
+
"loss": 0.0367,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.9684947491248541,
|
| 588 |
+
"grad_norm": 0.3761046826839447,
|
| 589 |
+
"learning_rate": 9.666517638683446e-05,
|
| 590 |
+
"loss": 0.0362,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.9801633605600933,
|
| 595 |
+
"grad_norm": 0.7744839191436768,
|
| 596 |
+
"learning_rate": 9.654872143347398e-05,
|
| 597 |
+
"loss": 0.0324,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.9918319719953326,
|
| 602 |
+
"grad_norm": 0.4213011860847473,
|
| 603 |
+
"learning_rate": 9.643034048715946e-05,
|
| 604 |
+
"loss": 0.0312,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 1.0035005834305717,
|
| 609 |
+
"grad_norm": 0.7749153971672058,
|
| 610 |
+
"learning_rate": 9.631003844600328e-05,
|
| 611 |
+
"loss": 0.0383,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 1.015169194865811,
|
| 616 |
+
"grad_norm": 0.5128496289253235,
|
| 617 |
+
"learning_rate": 9.618782028760472e-05,
|
| 618 |
+
"loss": 0.0334,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.0268378063010501,
|
| 623 |
+
"grad_norm": 0.6189514398574829,
|
| 624 |
+
"learning_rate": 9.60636910688441e-05,
|
| 625 |
+
"loss": 0.0301,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.0385064177362895,
|
| 630 |
+
"grad_norm": 0.5949068665504456,
|
| 631 |
+
"learning_rate": 9.593765592567345e-05,
|
| 632 |
+
"loss": 0.0309,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 1.0501750291715286,
|
| 637 |
+
"grad_norm": 0.4859572947025299,
|
| 638 |
+
"learning_rate": 9.580972007290406e-05,
|
| 639 |
+
"loss": 0.0311,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 1.0618436406067677,
|
| 644 |
+
"grad_norm": 0.49274128675460815,
|
| 645 |
+
"learning_rate": 9.567988880399066e-05,
|
| 646 |
+
"loss": 0.0299,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 1.073512252042007,
|
| 651 |
+
"grad_norm": 0.7119307518005371,
|
| 652 |
+
"learning_rate": 9.55481674908125e-05,
|
| 653 |
+
"loss": 0.0315,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 1.0851808634772462,
|
| 658 |
+
"grad_norm": 0.4597378373146057,
|
| 659 |
+
"learning_rate": 9.541456158345094e-05,
|
| 660 |
+
"loss": 0.0326,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.0968494749124855,
|
| 665 |
+
"grad_norm": 0.6472983956336975,
|
| 666 |
+
"learning_rate": 9.527907660996413e-05,
|
| 667 |
+
"loss": 0.0323,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.1085180863477246,
|
| 672 |
+
"grad_norm": 0.8791772127151489,
|
| 673 |
+
"learning_rate": 9.514171817615809e-05,
|
| 674 |
+
"loss": 0.0336,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 1.1201866977829638,
|
| 679 |
+
"grad_norm": 0.9581364989280701,
|
| 680 |
+
"learning_rate": 9.500249196535487e-05,
|
| 681 |
+
"loss": 0.0334,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 1.131855309218203,
|
| 686 |
+
"grad_norm": 0.7137671709060669,
|
| 687 |
+
"learning_rate": 9.486140373815743e-05,
|
| 688 |
+
"loss": 0.0399,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 1.1435239206534422,
|
| 693 |
+
"grad_norm": 0.5093432068824768,
|
| 694 |
+
"learning_rate": 9.471845933221119e-05,
|
| 695 |
+
"loss": 0.0378,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 1.1551925320886816,
|
| 700 |
+
"grad_norm": 0.8102591633796692,
|
| 701 |
+
"learning_rate": 9.45736646619626e-05,
|
| 702 |
+
"loss": 0.0352,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.1668611435239207,
|
| 707 |
+
"grad_norm": 0.5707778334617615,
|
| 708 |
+
"learning_rate": 9.442702571841429e-05,
|
| 709 |
+
"loss": 0.0344,
|
| 710 |
+
"step": 1000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.1785297549591598,
|
| 714 |
+
"grad_norm": 0.5596927404403687,
|
| 715 |
+
"learning_rate": 9.427854856887734e-05,
|
| 716 |
+
"loss": 0.0369,
|
| 717 |
+
"step": 1010
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 1.1901983663943991,
|
| 721 |
+
"grad_norm": 0.8491670489311218,
|
| 722 |
+
"learning_rate": 9.412823935672014e-05,
|
| 723 |
+
"loss": 0.0355,
|
| 724 |
+
"step": 1020
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 1.2018669778296382,
|
| 728 |
+
"grad_norm": 0.8242259621620178,
|
| 729 |
+
"learning_rate": 9.397610430111424e-05,
|
| 730 |
+
"loss": 0.0279,
|
| 731 |
+
"step": 1030
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 1.2135355892648776,
|
| 735 |
+
"grad_norm": 0.5555998682975769,
|
| 736 |
+
"learning_rate": 9.382214969677696e-05,
|
| 737 |
+
"loss": 0.0303,
|
| 738 |
+
"step": 1040
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 1.2252042007001167,
|
| 742 |
+
"grad_norm": 0.5572234988212585,
|
| 743 |
+
"learning_rate": 9.366638191371107e-05,
|
| 744 |
+
"loss": 0.0283,
|
| 745 |
+
"step": 1050
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.2368728121353558,
|
| 749 |
+
"grad_norm": 0.7672564387321472,
|
| 750 |
+
"learning_rate": 9.350880739694115e-05,
|
| 751 |
+
"loss": 0.0315,
|
| 752 |
+
"step": 1060
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.2485414235705952,
|
| 756 |
+
"grad_norm": 0.5608231425285339,
|
| 757 |
+
"learning_rate": 9.334943266624687e-05,
|
| 758 |
+
"loss": 0.0306,
|
| 759 |
+
"step": 1070
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 1.2602100350058343,
|
| 763 |
+
"grad_norm": 0.47090911865234375,
|
| 764 |
+
"learning_rate": 9.318826431589331e-05,
|
| 765 |
+
"loss": 0.0338,
|
| 766 |
+
"step": 1080
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 1.2718786464410736,
|
| 770 |
+
"grad_norm": 0.463077187538147,
|
| 771 |
+
"learning_rate": 9.302530901435811e-05,
|
| 772 |
+
"loss": 0.0286,
|
| 773 |
+
"step": 1090
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 1.2835472578763127,
|
| 777 |
+
"grad_norm": 0.5523890852928162,
|
| 778 |
+
"learning_rate": 9.286057350405551e-05,
|
| 779 |
+
"loss": 0.0351,
|
| 780 |
+
"step": 1100
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 1.2952158693115519,
|
| 784 |
+
"grad_norm": 0.4732823371887207,
|
| 785 |
+
"learning_rate": 9.26940646010574e-05,
|
| 786 |
+
"loss": 0.0318,
|
| 787 |
+
"step": 1110
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 1.3068844807467912,
|
| 791 |
+
"grad_norm": 0.4477550685405731,
|
| 792 |
+
"learning_rate": 9.252578919481133e-05,
|
| 793 |
+
"loss": 0.0266,
|
| 794 |
+
"step": 1120
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.3185530921820303,
|
| 798 |
+
"grad_norm": 0.8199338316917419,
|
| 799 |
+
"learning_rate": 9.23557542478554e-05,
|
| 800 |
+
"loss": 0.031,
|
| 801 |
+
"step": 1130
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 1.3302217036172697,
|
| 805 |
+
"grad_norm": 0.7581847310066223,
|
| 806 |
+
"learning_rate": 9.218396679553018e-05,
|
| 807 |
+
"loss": 0.0347,
|
| 808 |
+
"step": 1140
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 1.3418903150525088,
|
| 812 |
+
"grad_norm": 0.399112343788147,
|
| 813 |
+
"learning_rate": 9.201043394568773e-05,
|
| 814 |
+
"loss": 0.0358,
|
| 815 |
+
"step": 1150
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 1.353558926487748,
|
| 819 |
+
"grad_norm": 0.5397793054580688,
|
| 820 |
+
"learning_rate": 9.18351628783973e-05,
|
| 821 |
+
"loss": 0.0337,
|
| 822 |
+
"step": 1160
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 1.3652275379229872,
|
| 826 |
+
"grad_norm": 0.4025781452655792,
|
| 827 |
+
"learning_rate": 9.165816084564844e-05,
|
| 828 |
+
"loss": 0.0391,
|
| 829 |
+
"step": 1170
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.3768961493582264,
|
| 833 |
+
"grad_norm": 0.6508492231369019,
|
| 834 |
+
"learning_rate": 9.147943517105088e-05,
|
| 835 |
+
"loss": 0.0334,
|
| 836 |
+
"step": 1180
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.3885647607934657,
|
| 840 |
+
"grad_norm": 0.6159842610359192,
|
| 841 |
+
"learning_rate": 9.129899324953144e-05,
|
| 842 |
+
"loss": 0.0313,
|
| 843 |
+
"step": 1190
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 1.4002333722287048,
|
| 847 |
+
"grad_norm": 0.8535964488983154,
|
| 848 |
+
"learning_rate": 9.111684254702817e-05,
|
| 849 |
+
"loss": 0.0324,
|
| 850 |
+
"step": 1200
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 1.411901983663944,
|
| 854 |
+
"grad_norm": 0.8584406971931458,
|
| 855 |
+
"learning_rate": 9.093299060018134e-05,
|
| 856 |
+
"loss": 0.0363,
|
| 857 |
+
"step": 1210
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 1.4235705950991833,
|
| 861 |
+
"grad_norm": 0.6858974695205688,
|
| 862 |
+
"learning_rate": 9.074744501602169e-05,
|
| 863 |
+
"loss": 0.0394,
|
| 864 |
+
"step": 1220
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 1.4352392065344224,
|
| 868 |
+
"grad_norm": 0.5008957386016846,
|
| 869 |
+
"learning_rate": 9.056021347165564e-05,
|
| 870 |
+
"loss": 0.0309,
|
| 871 |
+
"step": 1230
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.4469078179696617,
|
| 875 |
+
"grad_norm": 0.5117946863174438,
|
| 876 |
+
"learning_rate": 9.037130371394762e-05,
|
| 877 |
+
"loss": 0.0365,
|
| 878 |
+
"step": 1240
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.4585764294049008,
|
| 882 |
+
"grad_norm": 0.46373674273490906,
|
| 883 |
+
"learning_rate": 9.018072355919957e-05,
|
| 884 |
+
"loss": 0.0312,
|
| 885 |
+
"step": 1250
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 1.47024504084014,
|
| 889 |
+
"grad_norm": 0.7355740070343018,
|
| 890 |
+
"learning_rate": 8.99884808928276e-05,
|
| 891 |
+
"loss": 0.0317,
|
| 892 |
+
"step": 1260
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.4819136522753793,
|
| 896 |
+
"grad_norm": 0.8395172357559204,
|
| 897 |
+
"learning_rate": 8.979458366903555e-05,
|
| 898 |
+
"loss": 0.0335,
|
| 899 |
+
"step": 1270
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.4935822637106184,
|
| 903 |
+
"grad_norm": 0.31481388211250305,
|
| 904 |
+
"learning_rate": 8.959903991048604e-05,
|
| 905 |
+
"loss": 0.0261,
|
| 906 |
+
"step": 1280
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 1.5052508751458578,
|
| 910 |
+
"grad_norm": 0.4806414544582367,
|
| 911 |
+
"learning_rate": 8.940185770796843e-05,
|
| 912 |
+
"loss": 0.0293,
|
| 913 |
+
"step": 1290
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.5169194865810969,
|
| 917 |
+
"grad_norm": 0.36156100034713745,
|
| 918 |
+
"learning_rate": 8.920304522006415e-05,
|
| 919 |
+
"loss": 0.0372,
|
| 920 |
+
"step": 1300
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.528588098016336,
|
| 924 |
+
"grad_norm": 0.36228907108306885,
|
| 925 |
+
"learning_rate": 8.900261067280905e-05,
|
| 926 |
+
"loss": 0.027,
|
| 927 |
+
"step": 1310
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 1.5402567094515751,
|
| 931 |
+
"grad_norm": 0.45837661623954773,
|
| 932 |
+
"learning_rate": 8.8800562359353e-05,
|
| 933 |
+
"loss": 0.0294,
|
| 934 |
+
"step": 1320
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.5519253208868145,
|
| 938 |
+
"grad_norm": 0.5353251099586487,
|
| 939 |
+
"learning_rate": 8.859690863961692e-05,
|
| 940 |
+
"loss": 0.0369,
|
| 941 |
+
"step": 1330
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.5635939323220538,
|
| 945 |
+
"grad_norm": 0.41150084137916565,
|
| 946 |
+
"learning_rate": 8.839165793994673e-05,
|
| 947 |
+
"loss": 0.0271,
|
| 948 |
+
"step": 1340
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.575262543757293,
|
| 952 |
+
"grad_norm": 0.5629997849464417,
|
| 953 |
+
"learning_rate": 8.818481875276477e-05,
|
| 954 |
+
"loss": 0.0258,
|
| 955 |
+
"step": 1350
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.586931155192532,
|
| 959 |
+
"grad_norm": 0.701050877571106,
|
| 960 |
+
"learning_rate": 8.797639963621839e-05,
|
| 961 |
+
"loss": 0.0272,
|
| 962 |
+
"step": 1360
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 1.5985997666277711,
|
| 966 |
+
"grad_norm": 0.8036890029907227,
|
| 967 |
+
"learning_rate": 8.776640921382584e-05,
|
| 968 |
+
"loss": 0.0273,
|
| 969 |
+
"step": 1370
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.6102683780630105,
|
| 973 |
+
"grad_norm": 0.4057667851448059,
|
| 974 |
+
"learning_rate": 8.755485617411954e-05,
|
| 975 |
+
"loss": 0.0244,
|
| 976 |
+
"step": 1380
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 1.6219369894982498,
|
| 980 |
+
"grad_norm": 0.6381646990776062,
|
| 981 |
+
"learning_rate": 8.734174927028648e-05,
|
| 982 |
+
"loss": 0.0284,
|
| 983 |
+
"step": 1390
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.633605600933489,
|
| 987 |
+
"grad_norm": 0.394702672958374,
|
| 988 |
+
"learning_rate": 8.712709731980615e-05,
|
| 989 |
+
"loss": 0.0333,
|
| 990 |
+
"step": 1400
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 1.645274212368728,
|
| 994 |
+
"grad_norm": 0.5312820672988892,
|
| 995 |
+
"learning_rate": 8.691090920408562e-05,
|
| 996 |
+
"loss": 0.0242,
|
| 997 |
+
"step": 1410
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.6569428238039672,
|
| 1001 |
+
"grad_norm": 0.3682110011577606,
|
| 1002 |
+
"learning_rate": 8.669319386809209e-05,
|
| 1003 |
+
"loss": 0.0236,
|
| 1004 |
+
"step": 1420
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.6686114352392065,
|
| 1008 |
+
"grad_norm": 0.4435596168041229,
|
| 1009 |
+
"learning_rate": 8.647396031998285e-05,
|
| 1010 |
+
"loss": 0.0282,
|
| 1011 |
+
"step": 1430
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 1.6802800466744459,
|
| 1015 |
+
"grad_norm": 0.31747034192085266,
|
| 1016 |
+
"learning_rate": 8.625321763073249e-05,
|
| 1017 |
+
"loss": 0.0247,
|
| 1018 |
+
"step": 1440
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 1.691948658109685,
|
| 1022 |
+
"grad_norm": 0.4574846625328064,
|
| 1023 |
+
"learning_rate": 8.603097493375757e-05,
|
| 1024 |
+
"loss": 0.023,
|
| 1025 |
+
"step": 1450
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.703617269544924,
|
| 1029 |
+
"grad_norm": 0.4448521137237549,
|
| 1030 |
+
"learning_rate": 8.580724142453878e-05,
|
| 1031 |
+
"loss": 0.0269,
|
| 1032 |
+
"step": 1460
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 1.7152858809801632,
|
| 1036 |
+
"grad_norm": 0.47059154510498047,
|
| 1037 |
+
"learning_rate": 8.558202636024041e-05,
|
| 1038 |
+
"loss": 0.027,
|
| 1039 |
+
"step": 1470
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.7269544924154026,
|
| 1043 |
+
"grad_norm": 0.6006631851196289,
|
| 1044 |
+
"learning_rate": 8.535533905932738e-05,
|
| 1045 |
+
"loss": 0.028,
|
| 1046 |
+
"step": 1480
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 1.738623103850642,
|
| 1050 |
+
"grad_norm": 0.5054494738578796,
|
| 1051 |
+
"learning_rate": 8.512718890117962e-05,
|
| 1052 |
+
"loss": 0.0305,
|
| 1053 |
+
"step": 1490
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 1.750291715285881,
|
| 1057 |
+
"grad_norm": 0.44669726490974426,
|
| 1058 |
+
"learning_rate": 8.489758532570405e-05,
|
| 1059 |
+
"loss": 0.0282,
|
| 1060 |
+
"step": 1500
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.7619603267211201,
|
| 1064 |
+
"grad_norm": 0.44248783588409424,
|
| 1065 |
+
"learning_rate": 8.466653783294393e-05,
|
| 1066 |
+
"loss": 0.0329,
|
| 1067 |
+
"step": 1510
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.7736289381563592,
|
| 1071 |
+
"grad_norm": 0.5698758363723755,
|
| 1072 |
+
"learning_rate": 8.44340559826859e-05,
|
| 1073 |
+
"loss": 0.0298,
|
| 1074 |
+
"step": 1520
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.7852975495915986,
|
| 1078 |
+
"grad_norm": 0.4533137083053589,
|
| 1079 |
+
"learning_rate": 8.420014939406431e-05,
|
| 1080 |
+
"loss": 0.0283,
|
| 1081 |
+
"step": 1530
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.796966161026838,
|
| 1085 |
+
"grad_norm": 0.2834441363811493,
|
| 1086 |
+
"learning_rate": 8.396482774516324e-05,
|
| 1087 |
+
"loss": 0.0263,
|
| 1088 |
+
"step": 1540
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.808634772462077,
|
| 1092 |
+
"grad_norm": 0.6928737163543701,
|
| 1093 |
+
"learning_rate": 8.372810077261613e-05,
|
| 1094 |
+
"loss": 0.0255,
|
| 1095 |
+
"step": 1550
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.8203033838973162,
|
| 1099 |
+
"grad_norm": 0.45429399609565735,
|
| 1100 |
+
"learning_rate": 8.348997827120292e-05,
|
| 1101 |
+
"loss": 0.0303,
|
| 1102 |
+
"step": 1560
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.8319719953325553,
|
| 1106 |
+
"grad_norm": 0.5234264135360718,
|
| 1107 |
+
"learning_rate": 8.325047009344468e-05,
|
| 1108 |
+
"loss": 0.0343,
|
| 1109 |
+
"step": 1570
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.8436406067677946,
|
| 1113 |
+
"grad_norm": 0.5153536796569824,
|
| 1114 |
+
"learning_rate": 8.300958614919605e-05,
|
| 1115 |
+
"loss": 0.0277,
|
| 1116 |
+
"step": 1580
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.855309218203034,
|
| 1120 |
+
"grad_norm": 0.2998095452785492,
|
| 1121 |
+
"learning_rate": 8.27673364052352e-05,
|
| 1122 |
+
"loss": 0.0265,
|
| 1123 |
+
"step": 1590
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.866977829638273,
|
| 1127 |
+
"grad_norm": 0.6762024760246277,
|
| 1128 |
+
"learning_rate": 8.252373088485135e-05,
|
| 1129 |
+
"loss": 0.0286,
|
| 1130 |
+
"step": 1600
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.8786464410735122,
|
| 1134 |
+
"grad_norm": 0.7629134654998779,
|
| 1135 |
+
"learning_rate": 8.227877966743019e-05,
|
| 1136 |
+
"loss": 0.033,
|
| 1137 |
+
"step": 1610
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.8903150525087513,
|
| 1141 |
+
"grad_norm": 0.4351422190666199,
|
| 1142 |
+
"learning_rate": 8.203249288803674e-05,
|
| 1143 |
+
"loss": 0.0299,
|
| 1144 |
+
"step": 1620
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.9019836639439907,
|
| 1148 |
+
"grad_norm": 0.4076254665851593,
|
| 1149 |
+
"learning_rate": 8.178488073699609e-05,
|
| 1150 |
+
"loss": 0.0244,
|
| 1151 |
+
"step": 1630
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.91365227537923,
|
| 1155 |
+
"grad_norm": 0.639583945274353,
|
| 1156 |
+
"learning_rate": 8.153595345947162e-05,
|
| 1157 |
+
"loss": 0.0286,
|
| 1158 |
+
"step": 1640
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.9253208868144691,
|
| 1162 |
+
"grad_norm": 0.3300214111804962,
|
| 1163 |
+
"learning_rate": 8.128572135504124e-05,
|
| 1164 |
+
"loss": 0.0231,
|
| 1165 |
+
"step": 1650
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.9369894982497082,
|
| 1169 |
+
"grad_norm": 0.5126506090164185,
|
| 1170 |
+
"learning_rate": 8.103419477727117e-05,
|
| 1171 |
+
"loss": 0.0242,
|
| 1172 |
+
"step": 1660
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.9486581096849473,
|
| 1176 |
+
"grad_norm": 0.42610886693000793,
|
| 1177 |
+
"learning_rate": 8.07813841332876e-05,
|
| 1178 |
+
"loss": 0.0288,
|
| 1179 |
+
"step": 1670
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.9603267211201867,
|
| 1183 |
+
"grad_norm": 0.4349581301212311,
|
| 1184 |
+
"learning_rate": 8.052729988334598e-05,
|
| 1185 |
+
"loss": 0.0286,
|
| 1186 |
+
"step": 1680
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.971995332555426,
|
| 1190 |
+
"grad_norm": 0.6044297218322754,
|
| 1191 |
+
"learning_rate": 8.027195254039834e-05,
|
| 1192 |
+
"loss": 0.028,
|
| 1193 |
+
"step": 1690
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.9836639439906651,
|
| 1197 |
+
"grad_norm": 0.49344927072525024,
|
| 1198 |
+
"learning_rate": 8.001535266965829e-05,
|
| 1199 |
+
"loss": 0.0221,
|
| 1200 |
+
"step": 1700
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.9953325554259043,
|
| 1204 |
+
"grad_norm": 0.5056473016738892,
|
| 1205 |
+
"learning_rate": 7.975751088816375e-05,
|
| 1206 |
+
"loss": 0.0242,
|
| 1207 |
+
"step": 1710
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 2.0070011668611434,
|
| 1211 |
+
"grad_norm": 0.3485012948513031,
|
| 1212 |
+
"learning_rate": 7.949843786433784e-05,
|
| 1213 |
+
"loss": 0.0247,
|
| 1214 |
+
"step": 1720
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.018669778296383,
|
| 1218 |
+
"grad_norm": 0.38087254762649536,
|
| 1219 |
+
"learning_rate": 7.923814431754734e-05,
|
| 1220 |
+
"loss": 0.0248,
|
| 1221 |
+
"step": 1730
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 2.030338389731622,
|
| 1225 |
+
"grad_norm": 0.4125686585903168,
|
| 1226 |
+
"learning_rate": 7.89766410176592e-05,
|
| 1227 |
+
"loss": 0.034,
|
| 1228 |
+
"step": 1740
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 2.042007001166861,
|
| 1232 |
+
"grad_norm": 0.4918675720691681,
|
| 1233 |
+
"learning_rate": 7.871393878459496e-05,
|
| 1234 |
+
"loss": 0.026,
|
| 1235 |
+
"step": 1750
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 2.0536756126021003,
|
| 1239 |
+
"grad_norm": 0.7919062376022339,
|
| 1240 |
+
"learning_rate": 7.845004848788302e-05,
|
| 1241 |
+
"loss": 0.0277,
|
| 1242 |
+
"step": 1760
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 2.0653442240373394,
|
| 1246 |
+
"grad_norm": 0.4226606786251068,
|
| 1247 |
+
"learning_rate": 7.818498104620893e-05,
|
| 1248 |
+
"loss": 0.026,
|
| 1249 |
+
"step": 1770
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 2.077012835472579,
|
| 1253 |
+
"grad_norm": 0.3489656150341034,
|
| 1254 |
+
"learning_rate": 7.791874742696362e-05,
|
| 1255 |
+
"loss": 0.0277,
|
| 1256 |
+
"step": 1780
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.088681446907818,
|
| 1260 |
+
"grad_norm": 0.5135205388069153,
|
| 1261 |
+
"learning_rate": 7.76513586457896e-05,
|
| 1262 |
+
"loss": 0.0267,
|
| 1263 |
+
"step": 1790
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 2.100350058343057,
|
| 1267 |
+
"grad_norm": 0.34718888998031616,
|
| 1268 |
+
"learning_rate": 7.738282576612518e-05,
|
| 1269 |
+
"loss": 0.0237,
|
| 1270 |
+
"step": 1800
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 2.1120186697782963,
|
| 1274 |
+
"grad_norm": 0.6323973536491394,
|
| 1275 |
+
"learning_rate": 7.711315989874677e-05,
|
| 1276 |
+
"loss": 0.0231,
|
| 1277 |
+
"step": 1810
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 2.1236872812135354,
|
| 1281 |
+
"grad_norm": 0.33119362592697144,
|
| 1282 |
+
"learning_rate": 7.6842372201309e-05,
|
| 1283 |
+
"loss": 0.0248,
|
| 1284 |
+
"step": 1820
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 2.135355892648775,
|
| 1288 |
+
"grad_norm": 0.5421165823936462,
|
| 1289 |
+
"learning_rate": 7.657047387788332e-05,
|
| 1290 |
+
"loss": 0.0244,
|
| 1291 |
+
"step": 1830
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 2.147024504084014,
|
| 1295 |
+
"grad_norm": 0.44305363297462463,
|
| 1296 |
+
"learning_rate": 7.629747617849417e-05,
|
| 1297 |
+
"loss": 0.0253,
|
| 1298 |
+
"step": 1840
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 2.1586931155192532,
|
| 1302 |
+
"grad_norm": 0.5938749313354492,
|
| 1303 |
+
"learning_rate": 7.602339039865362e-05,
|
| 1304 |
+
"loss": 0.0282,
|
| 1305 |
+
"step": 1850
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 2.1703617269544924,
|
| 1309 |
+
"grad_norm": 0.4677620828151703,
|
| 1310 |
+
"learning_rate": 7.574822787889401e-05,
|
| 1311 |
+
"loss": 0.0276,
|
| 1312 |
+
"step": 1860
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 2.1820303383897315,
|
| 1316 |
+
"grad_norm": 0.7123854756355286,
|
| 1317 |
+
"learning_rate": 7.547200000429874e-05,
|
| 1318 |
+
"loss": 0.0288,
|
| 1319 |
+
"step": 1870
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.193698949824971,
|
| 1323 |
+
"grad_norm": 0.4638105630874634,
|
| 1324 |
+
"learning_rate": 7.519471820403112e-05,
|
| 1325 |
+
"loss": 0.0262,
|
| 1326 |
+
"step": 1880
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 2.20536756126021,
|
| 1330 |
+
"grad_norm": 0.6613805294036865,
|
| 1331 |
+
"learning_rate": 7.491639395086154e-05,
|
| 1332 |
+
"loss": 0.0235,
|
| 1333 |
+
"step": 1890
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 2.2170361726954493,
|
| 1337 |
+
"grad_norm": 0.45506903529167175,
|
| 1338 |
+
"learning_rate": 7.463703876069281e-05,
|
| 1339 |
+
"loss": 0.0318,
|
| 1340 |
+
"step": 1900
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 2.2287047841306884,
|
| 1344 |
+
"grad_norm": 0.6076046824455261,
|
| 1345 |
+
"learning_rate": 7.435666419208358e-05,
|
| 1346 |
+
"loss": 0.028,
|
| 1347 |
+
"step": 1910
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 2.2403733955659275,
|
| 1351 |
+
"grad_norm": 0.5596355199813843,
|
| 1352 |
+
"learning_rate": 7.407528184577019e-05,
|
| 1353 |
+
"loss": 0.0254,
|
| 1354 |
+
"step": 1920
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 2.252042007001167,
|
| 1358 |
+
"grad_norm": 0.4784010648727417,
|
| 1359 |
+
"learning_rate": 7.379290336418664e-05,
|
| 1360 |
+
"loss": 0.0245,
|
| 1361 |
+
"step": 1930
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.263710618436406,
|
| 1365 |
+
"grad_norm": 0.45856916904449463,
|
| 1366 |
+
"learning_rate": 7.350954043098283e-05,
|
| 1367 |
+
"loss": 0.0294,
|
| 1368 |
+
"step": 1940
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 2.2753792298716453,
|
| 1372 |
+
"grad_norm": 0.37216275930404663,
|
| 1373 |
+
"learning_rate": 7.322520477054123e-05,
|
| 1374 |
+
"loss": 0.0217,
|
| 1375 |
+
"step": 1950
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 2.2870478413068844,
|
| 1379 |
+
"grad_norm": 0.4601225256919861,
|
| 1380 |
+
"learning_rate": 7.293990814749168e-05,
|
| 1381 |
+
"loss": 0.0275,
|
| 1382 |
+
"step": 1960
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.2987164527421236,
|
| 1386 |
+
"grad_norm": 0.46999940276145935,
|
| 1387 |
+
"learning_rate": 7.265366236622475e-05,
|
| 1388 |
+
"loss": 0.0238,
|
| 1389 |
+
"step": 1970
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 2.310385064177363,
|
| 1393 |
+
"grad_norm": 0.4829197824001312,
|
| 1394 |
+
"learning_rate": 7.236647927040314e-05,
|
| 1395 |
+
"loss": 0.0215,
|
| 1396 |
+
"step": 1980
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.3220536756126022,
|
| 1400 |
+
"grad_norm": 0.3234507739543915,
|
| 1401 |
+
"learning_rate": 7.207837074247181e-05,
|
| 1402 |
+
"loss": 0.0231,
|
| 1403 |
+
"step": 1990
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.3337222870478413,
|
| 1407 |
+
"grad_norm": 0.4881344735622406,
|
| 1408 |
+
"learning_rate": 7.178934870316621e-05,
|
| 1409 |
+
"loss": 0.0283,
|
| 1410 |
+
"step": 2000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 2.3453908984830805,
|
| 1414 |
+
"grad_norm": 0.36917248368263245,
|
| 1415 |
+
"learning_rate": 7.149942511101913e-05,
|
| 1416 |
+
"loss": 0.0214,
|
| 1417 |
+
"step": 2010
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 2.3570595099183196,
|
| 1421 |
+
"grad_norm": 0.5211784243583679,
|
| 1422 |
+
"learning_rate": 7.120861196186585e-05,
|
| 1423 |
+
"loss": 0.0265,
|
| 1424 |
+
"step": 2020
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 2.368728121353559,
|
| 1428 |
+
"grad_norm": 0.6132432818412781,
|
| 1429 |
+
"learning_rate": 7.091692128834791e-05,
|
| 1430 |
+
"loss": 0.033,
|
| 1431 |
+
"step": 2030
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 2.3803967327887983,
|
| 1435 |
+
"grad_norm": 0.34162402153015137,
|
| 1436 |
+
"learning_rate": 7.062436515941504e-05,
|
| 1437 |
+
"loss": 0.0226,
|
| 1438 |
+
"step": 2040
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 2.3920653442240374,
|
| 1442 |
+
"grad_norm": 0.3589126169681549,
|
| 1443 |
+
"learning_rate": 7.033095567982602e-05,
|
| 1444 |
+
"loss": 0.0286,
|
| 1445 |
+
"step": 2050
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 2.4037339556592765,
|
| 1449 |
+
"grad_norm": 0.39690524339675903,
|
| 1450 |
+
"learning_rate": 7.003670498964769e-05,
|
| 1451 |
+
"loss": 0.0239,
|
| 1452 |
+
"step": 2060
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 2.4154025670945156,
|
| 1456 |
+
"grad_norm": 0.40809738636016846,
|
| 1457 |
+
"learning_rate": 6.974162526375272e-05,
|
| 1458 |
+
"loss": 0.028,
|
| 1459 |
+
"step": 2070
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 2.427071178529755,
|
| 1463 |
+
"grad_norm": 0.844285249710083,
|
| 1464 |
+
"learning_rate": 6.944572871131583e-05,
|
| 1465 |
+
"loss": 0.0256,
|
| 1466 |
+
"step": 2080
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 2.4387397899649943,
|
| 1470 |
+
"grad_norm": 0.4565018117427826,
|
| 1471 |
+
"learning_rate": 6.914902757530863e-05,
|
| 1472 |
+
"loss": 0.0292,
|
| 1473 |
+
"step": 2090
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 2.4504084014002334,
|
| 1477 |
+
"grad_norm": 0.6053803563117981,
|
| 1478 |
+
"learning_rate": 6.885153413199305e-05,
|
| 1479 |
+
"loss": 0.0295,
|
| 1480 |
+
"step": 2100
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 2.4620770128354725,
|
| 1484 |
+
"grad_norm": 0.29552745819091797,
|
| 1485 |
+
"learning_rate": 6.855326069041346e-05,
|
| 1486 |
+
"loss": 0.0237,
|
| 1487 |
+
"step": 2110
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.4737456242707117,
|
| 1491 |
+
"grad_norm": 0.3798177242279053,
|
| 1492 |
+
"learning_rate": 6.825421959188726e-05,
|
| 1493 |
+
"loss": 0.0199,
|
| 1494 |
+
"step": 2120
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 2.485414235705951,
|
| 1498 |
+
"grad_norm": 0.4293670952320099,
|
| 1499 |
+
"learning_rate": 6.795442320949431e-05,
|
| 1500 |
+
"loss": 0.0283,
|
| 1501 |
+
"step": 2130
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 2.4970828471411903,
|
| 1505 |
+
"grad_norm": 0.27908170223236084,
|
| 1506 |
+
"learning_rate": 6.765388394756504e-05,
|
| 1507 |
+
"loss": 0.0239,
|
| 1508 |
+
"step": 2140
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 2.5087514585764294,
|
| 1512 |
+
"grad_norm": 0.5253874063491821,
|
| 1513 |
+
"learning_rate": 6.73526142411671e-05,
|
| 1514 |
+
"loss": 0.0232,
|
| 1515 |
+
"step": 2150
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 2.5204200700116686,
|
| 1519 |
+
"grad_norm": 0.3897351920604706,
|
| 1520 |
+
"learning_rate": 6.705062655559093e-05,
|
| 1521 |
+
"loss": 0.0209,
|
| 1522 |
+
"step": 2160
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 2.5320886814469077,
|
| 1526 |
+
"grad_norm": 0.5366333723068237,
|
| 1527 |
+
"learning_rate": 6.674793338583396e-05,
|
| 1528 |
+
"loss": 0.0223,
|
| 1529 |
+
"step": 2170
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 2.5437572928821472,
|
| 1533 |
+
"grad_norm": 0.4374563694000244,
|
| 1534 |
+
"learning_rate": 6.64445472560836e-05,
|
| 1535 |
+
"loss": 0.0235,
|
| 1536 |
+
"step": 2180
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 2.5554259043173864,
|
| 1540 |
+
"grad_norm": 0.28929224610328674,
|
| 1541 |
+
"learning_rate": 6.614048071919912e-05,
|
| 1542 |
+
"loss": 0.0246,
|
| 1543 |
+
"step": 2190
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 2.5670945157526255,
|
| 1547 |
+
"grad_norm": 0.4833926260471344,
|
| 1548 |
+
"learning_rate": 6.58357463561922e-05,
|
| 1549 |
+
"loss": 0.0292,
|
| 1550 |
+
"step": 2200
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 2.5787631271878646,
|
| 1554 |
+
"grad_norm": 0.7238153219223022,
|
| 1555 |
+
"learning_rate": 6.553035677570637e-05,
|
| 1556 |
+
"loss": 0.0287,
|
| 1557 |
+
"step": 2210
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 2.5904317386231037,
|
| 1561 |
+
"grad_norm": 0.4303306043148041,
|
| 1562 |
+
"learning_rate": 6.522432461349536e-05,
|
| 1563 |
+
"loss": 0.0254,
|
| 1564 |
+
"step": 2220
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 2.6021003500583433,
|
| 1568 |
+
"grad_norm": 0.4957870543003082,
|
| 1569 |
+
"learning_rate": 6.491766253190025e-05,
|
| 1570 |
+
"loss": 0.0236,
|
| 1571 |
+
"step": 2230
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 2.6137689614935824,
|
| 1575 |
+
"grad_norm": 0.4810437858104706,
|
| 1576 |
+
"learning_rate": 6.461038321932558e-05,
|
| 1577 |
+
"loss": 0.0226,
|
| 1578 |
+
"step": 2240
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 2.6254375729288215,
|
| 1582 |
+
"grad_norm": 0.49355655908584595,
|
| 1583 |
+
"learning_rate": 6.430249938971438e-05,
|
| 1584 |
+
"loss": 0.0221,
|
| 1585 |
+
"step": 2250
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 2.6371061843640606,
|
| 1589 |
+
"grad_norm": 0.32932281494140625,
|
| 1590 |
+
"learning_rate": 6.3994023782022e-05,
|
| 1591 |
+
"loss": 0.0238,
|
| 1592 |
+
"step": 2260
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 2.6487747957992998,
|
| 1596 |
+
"grad_norm": 0.34265977144241333,
|
| 1597 |
+
"learning_rate": 6.368496915968922e-05,
|
| 1598 |
+
"loss": 0.021,
|
| 1599 |
+
"step": 2270
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 2.6604434072345393,
|
| 1603 |
+
"grad_norm": 0.39306285977363586,
|
| 1604 |
+
"learning_rate": 6.337534831011397e-05,
|
| 1605 |
+
"loss": 0.0222,
|
| 1606 |
+
"step": 2280
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 2.6721120186697784,
|
| 1610 |
+
"grad_norm": 0.5323277115821838,
|
| 1611 |
+
"learning_rate": 6.306517404412233e-05,
|
| 1612 |
+
"loss": 0.0279,
|
| 1613 |
+
"step": 2290
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 2.6837806301050176,
|
| 1617 |
+
"grad_norm": 0.28174087405204773,
|
| 1618 |
+
"learning_rate": 6.275445919543846e-05,
|
| 1619 |
+
"loss": 0.023,
|
| 1620 |
+
"step": 2300
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 2.6954492415402567,
|
| 1624 |
+
"grad_norm": 0.3223417103290558,
|
| 1625 |
+
"learning_rate": 6.244321662015358e-05,
|
| 1626 |
+
"loss": 0.0227,
|
| 1627 |
+
"step": 2310
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 2.707117852975496,
|
| 1631 |
+
"grad_norm": 0.482028603553772,
|
| 1632 |
+
"learning_rate": 6.213145919619408e-05,
|
| 1633 |
+
"loss": 0.0222,
|
| 1634 |
+
"step": 2320
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 2.7187864644107353,
|
| 1638 |
+
"grad_norm": 0.5487138628959656,
|
| 1639 |
+
"learning_rate": 6.181919982278859e-05,
|
| 1640 |
+
"loss": 0.0256,
|
| 1641 |
+
"step": 2330
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 2.7304550758459745,
|
| 1645 |
+
"grad_norm": 0.37184667587280273,
|
| 1646 |
+
"learning_rate": 6.150645141993438e-05,
|
| 1647 |
+
"loss": 0.024,
|
| 1648 |
+
"step": 2340
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 2.7421236872812136,
|
| 1652 |
+
"grad_norm": 0.36116233468055725,
|
| 1653 |
+
"learning_rate": 6.119322692786274e-05,
|
| 1654 |
+
"loss": 0.0273,
|
| 1655 |
+
"step": 2350
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 2.7537922987164527,
|
| 1659 |
+
"grad_norm": 0.334886759519577,
|
| 1660 |
+
"learning_rate": 6.0879539306503485e-05,
|
| 1661 |
+
"loss": 0.0209,
|
| 1662 |
+
"step": 2360
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 2.765460910151692,
|
| 1666 |
+
"grad_norm": 0.28347110748291016,
|
| 1667 |
+
"learning_rate": 6.056540153494886e-05,
|
| 1668 |
+
"loss": 0.0239,
|
| 1669 |
+
"step": 2370
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 2.7771295215869314,
|
| 1673 |
+
"grad_norm": 0.33316299319267273,
|
| 1674 |
+
"learning_rate": 6.0250826610916435e-05,
|
| 1675 |
+
"loss": 0.0228,
|
| 1676 |
+
"step": 2380
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 2.7887981330221705,
|
| 1680 |
+
"grad_norm": 0.44831910729408264,
|
| 1681 |
+
"learning_rate": 5.993582755021134e-05,
|
| 1682 |
+
"loss": 0.0232,
|
| 1683 |
+
"step": 2390
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 2.8004667444574096,
|
| 1687 |
+
"grad_norm": 0.4825405478477478,
|
| 1688 |
+
"learning_rate": 5.9620417386187734e-05,
|
| 1689 |
+
"loss": 0.0214,
|
| 1690 |
+
"step": 2400
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 2.8121353558926487,
|
| 1694 |
+
"grad_norm": 0.45923271775245667,
|
| 1695 |
+
"learning_rate": 5.93046091692095e-05,
|
| 1696 |
+
"loss": 0.029,
|
| 1697 |
+
"step": 2410
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 2.823803967327888,
|
| 1701 |
+
"grad_norm": 0.3167724609375,
|
| 1702 |
+
"learning_rate": 5.89884159661103e-05,
|
| 1703 |
+
"loss": 0.0229,
|
| 1704 |
+
"step": 2420
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 2.8354725787631274,
|
| 1708 |
+
"grad_norm": 0.4090178310871124,
|
| 1709 |
+
"learning_rate": 5.8671850859652945e-05,
|
| 1710 |
+
"loss": 0.0293,
|
| 1711 |
+
"step": 2430
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.8471411901983665,
|
| 1715 |
+
"grad_norm": 0.4324829578399658,
|
| 1716 |
+
"learning_rate": 5.835492694798808e-05,
|
| 1717 |
+
"loss": 0.0257,
|
| 1718 |
+
"step": 2440
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.8588098016336057,
|
| 1722 |
+
"grad_norm": 0.4097526967525482,
|
| 1723 |
+
"learning_rate": 5.803765734411216e-05,
|
| 1724 |
+
"loss": 0.0227,
|
| 1725 |
+
"step": 2450
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 2.8704784130688448,
|
| 1729 |
+
"grad_norm": 0.3852528929710388,
|
| 1730 |
+
"learning_rate": 5.772005517532502e-05,
|
| 1731 |
+
"loss": 0.0267,
|
| 1732 |
+
"step": 2460
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 2.882147024504084,
|
| 1736 |
+
"grad_norm": 0.2632818818092346,
|
| 1737 |
+
"learning_rate": 5.7402133582686576e-05,
|
| 1738 |
+
"loss": 0.0236,
|
| 1739 |
+
"step": 2470
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.8938156359393234,
|
| 1743 |
+
"grad_norm": 0.3686858117580414,
|
| 1744 |
+
"learning_rate": 5.708390572047326e-05,
|
| 1745 |
+
"loss": 0.0195,
|
| 1746 |
+
"step": 2480
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 2.9054842473745626,
|
| 1750 |
+
"grad_norm": 0.29035860300064087,
|
| 1751 |
+
"learning_rate": 5.6765384755633586e-05,
|
| 1752 |
+
"loss": 0.02,
|
| 1753 |
+
"step": 2490
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 2.9171528588098017,
|
| 1757 |
+
"grad_norm": 0.41614043712615967,
|
| 1758 |
+
"learning_rate": 5.644658386724351e-05,
|
| 1759 |
+
"loss": 0.0295,
|
| 1760 |
+
"step": 2500
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.928821470245041,
|
| 1764 |
+
"grad_norm": 0.4453172981739044,
|
| 1765 |
+
"learning_rate": 5.612751624596103e-05,
|
| 1766 |
+
"loss": 0.0201,
|
| 1767 |
+
"step": 2510
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 2.94049008168028,
|
| 1771 |
+
"grad_norm": 0.32595980167388916,
|
| 1772 |
+
"learning_rate": 5.580819509348045e-05,
|
| 1773 |
+
"loss": 0.024,
|
| 1774 |
+
"step": 2520
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 2.9521586931155195,
|
| 1778 |
+
"grad_norm": 0.39912399649620056,
|
| 1779 |
+
"learning_rate": 5.548863362198612e-05,
|
| 1780 |
+
"loss": 0.0206,
|
| 1781 |
+
"step": 2530
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.9638273045507586,
|
| 1785 |
+
"grad_norm": 0.762593686580658,
|
| 1786 |
+
"learning_rate": 5.516884505360581e-05,
|
| 1787 |
+
"loss": 0.0212,
|
| 1788 |
+
"step": 2540
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.9754959159859977,
|
| 1792 |
+
"grad_norm": 0.5083714723587036,
|
| 1793 |
+
"learning_rate": 5.484884261986364e-05,
|
| 1794 |
+
"loss": 0.0237,
|
| 1795 |
+
"step": 2550
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 2.987164527421237,
|
| 1799 |
+
"grad_norm": 0.491256445646286,
|
| 1800 |
+
"learning_rate": 5.4528639561132566e-05,
|
| 1801 |
+
"loss": 0.0244,
|
| 1802 |
+
"step": 2560
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.998833138856476,
|
| 1806 |
+
"grad_norm": 0.3807949423789978,
|
| 1807 |
+
"learning_rate": 5.420824912608659e-05,
|
| 1808 |
+
"loss": 0.0243,
|
| 1809 |
+
"step": 2570
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 3.010501750291715,
|
| 1813 |
+
"grad_norm": 0.5656011700630188,
|
| 1814 |
+
"learning_rate": 5.388768457115254e-05,
|
| 1815 |
+
"loss": 0.0251,
|
| 1816 |
+
"step": 2580
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 3.0221703617269546,
|
| 1820 |
+
"grad_norm": 0.5722658038139343,
|
| 1821 |
+
"learning_rate": 5.3566959159961615e-05,
|
| 1822 |
+
"loss": 0.0237,
|
| 1823 |
+
"step": 2590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 3.0338389731621938,
|
| 1827 |
+
"grad_norm": 0.4347064197063446,
|
| 1828 |
+
"learning_rate": 5.3246086162800613e-05,
|
| 1829 |
+
"loss": 0.0254,
|
| 1830 |
+
"step": 2600
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 3.045507584597433,
|
| 1834 |
+
"grad_norm": 0.3570729196071625,
|
| 1835 |
+
"learning_rate": 5.292507885606277e-05,
|
| 1836 |
+
"loss": 0.021,
|
| 1837 |
+
"step": 2610
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 3.057176196032672,
|
| 1841 |
+
"grad_norm": 0.34851503372192383,
|
| 1842 |
+
"learning_rate": 5.260395052169853e-05,
|
| 1843 |
+
"loss": 0.0228,
|
| 1844 |
+
"step": 2620
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 3.068844807467911,
|
| 1848 |
+
"grad_norm": 0.5343664288520813,
|
| 1849 |
+
"learning_rate": 5.228271444666596e-05,
|
| 1850 |
+
"loss": 0.0248,
|
| 1851 |
+
"step": 2630
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 3.0805134189031507,
|
| 1855 |
+
"grad_norm": 0.3728184998035431,
|
| 1856 |
+
"learning_rate": 5.196138392238097e-05,
|
| 1857 |
+
"loss": 0.0234,
|
| 1858 |
+
"step": 2640
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 3.09218203033839,
|
| 1862 |
+
"grad_norm": 0.2670184373855591,
|
| 1863 |
+
"learning_rate": 5.16399722441674e-05,
|
| 1864 |
+
"loss": 0.0217,
|
| 1865 |
+
"step": 2650
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 3.103850641773629,
|
| 1869 |
+
"grad_norm": 0.33567026257514954,
|
| 1870 |
+
"learning_rate": 5.13184927107069e-05,
|
| 1871 |
+
"loss": 0.0231,
|
| 1872 |
+
"step": 2660
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 3.115519253208868,
|
| 1876 |
+
"grad_norm": 0.4638327360153198,
|
| 1877 |
+
"learning_rate": 5.0996958623488686e-05,
|
| 1878 |
+
"loss": 0.0171,
|
| 1879 |
+
"step": 2670
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 3.1271878646441076,
|
| 1883 |
+
"grad_norm": 0.5138529539108276,
|
| 1884 |
+
"learning_rate": 5.0675383286259195e-05,
|
| 1885 |
+
"loss": 0.0257,
|
| 1886 |
+
"step": 2680
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 3.1388564760793467,
|
| 1890 |
+
"grad_norm": 0.43609845638275146,
|
| 1891 |
+
"learning_rate": 5.0353780004471606e-05,
|
| 1892 |
+
"loss": 0.0202,
|
| 1893 |
+
"step": 2690
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.150525087514586,
|
| 1897 |
+
"grad_norm": 0.3554588258266449,
|
| 1898 |
+
"learning_rate": 5.003216208473532e-05,
|
| 1899 |
+
"loss": 0.0217,
|
| 1900 |
+
"step": 2700
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 3.162193698949825,
|
| 1904 |
+
"grad_norm": 0.3532634973526001,
|
| 1905 |
+
"learning_rate": 4.971054283426542e-05,
|
| 1906 |
+
"loss": 0.0233,
|
| 1907 |
+
"step": 2710
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 3.173862310385064,
|
| 1911 |
+
"grad_norm": 0.3735610842704773,
|
| 1912 |
+
"learning_rate": 4.938893556033203e-05,
|
| 1913 |
+
"loss": 0.0191,
|
| 1914 |
+
"step": 2720
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 3.1855309218203036,
|
| 1918 |
+
"grad_norm": 0.3165076673030853,
|
| 1919 |
+
"learning_rate": 4.9067353569709734e-05,
|
| 1920 |
+
"loss": 0.0241,
|
| 1921 |
+
"step": 2730
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 3.1971995332555427,
|
| 1925 |
+
"grad_norm": 0.4481463134288788,
|
| 1926 |
+
"learning_rate": 4.874581016812701e-05,
|
| 1927 |
+
"loss": 0.0222,
|
| 1928 |
+
"step": 2740
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 3.208868144690782,
|
| 1932 |
+
"grad_norm": 0.5085083842277527,
|
| 1933 |
+
"learning_rate": 4.8424318659715676e-05,
|
| 1934 |
+
"loss": 0.0189,
|
| 1935 |
+
"step": 2750
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.220536756126021,
|
| 1939 |
+
"grad_norm": 0.4407035708427429,
|
| 1940 |
+
"learning_rate": 4.8102892346460424e-05,
|
| 1941 |
+
"loss": 0.0201,
|
| 1942 |
+
"step": 2760
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 3.23220536756126,
|
| 1946 |
+
"grad_norm": 0.23214414715766907,
|
| 1947 |
+
"learning_rate": 4.7781544527648424e-05,
|
| 1948 |
+
"loss": 0.0215,
|
| 1949 |
+
"step": 2770
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 3.2438739789964997,
|
| 1953 |
+
"grad_norm": 0.411751925945282,
|
| 1954 |
+
"learning_rate": 4.7460288499319116e-05,
|
| 1955 |
+
"loss": 0.0246,
|
| 1956 |
+
"step": 2780
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 3.2555425904317388,
|
| 1960 |
+
"grad_norm": 0.4447856843471527,
|
| 1961 |
+
"learning_rate": 4.7139137553714e-05,
|
| 1962 |
+
"loss": 0.0179,
|
| 1963 |
+
"step": 2790
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 3.267211201866978,
|
| 1967 |
+
"grad_norm": 0.3802412152290344,
|
| 1968 |
+
"learning_rate": 4.6818104978726685e-05,
|
| 1969 |
+
"loss": 0.02,
|
| 1970 |
+
"step": 2800
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 3.278879813302217,
|
| 1974 |
+
"grad_norm": 0.41835305094718933,
|
| 1975 |
+
"learning_rate": 4.649720405735312e-05,
|
| 1976 |
+
"loss": 0.0226,
|
| 1977 |
+
"step": 2810
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 3.290548424737456,
|
| 1981 |
+
"grad_norm": 0.33167389035224915,
|
| 1982 |
+
"learning_rate": 4.617644806714193e-05,
|
| 1983 |
+
"loss": 0.0218,
|
| 1984 |
+
"step": 2820
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 3.3022170361726957,
|
| 1988 |
+
"grad_norm": 0.29077398777008057,
|
| 1989 |
+
"learning_rate": 4.585585027964512e-05,
|
| 1990 |
+
"loss": 0.0154,
|
| 1991 |
+
"step": 2830
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 3.313885647607935,
|
| 1995 |
+
"grad_norm": 0.3237426280975342,
|
| 1996 |
+
"learning_rate": 4.553542395986894e-05,
|
| 1997 |
+
"loss": 0.0203,
|
| 1998 |
+
"step": 2840
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 3.325554259043174,
|
| 2002 |
+
"grad_norm": 0.3535657525062561,
|
| 2003 |
+
"learning_rate": 4.521518236572498e-05,
|
| 2004 |
+
"loss": 0.0211,
|
| 2005 |
+
"step": 2850
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 3.337222870478413,
|
| 2009 |
+
"grad_norm": 0.38993456959724426,
|
| 2010 |
+
"learning_rate": 4.489513874748167e-05,
|
| 2011 |
+
"loss": 0.017,
|
| 2012 |
+
"step": 2860
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 3.348891481913652,
|
| 2016 |
+
"grad_norm": 0.35377201437950134,
|
| 2017 |
+
"learning_rate": 4.457530634721601e-05,
|
| 2018 |
+
"loss": 0.0247,
|
| 2019 |
+
"step": 2870
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 3.3605600933488917,
|
| 2023 |
+
"grad_norm": 0.30412212014198303,
|
| 2024 |
+
"learning_rate": 4.42556983982657e-05,
|
| 2025 |
+
"loss": 0.0235,
|
| 2026 |
+
"step": 2880
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 3.372228704784131,
|
| 2030 |
+
"grad_norm": 0.2848266363143921,
|
| 2031 |
+
"learning_rate": 4.393632812468155e-05,
|
| 2032 |
+
"loss": 0.0206,
|
| 2033 |
+
"step": 2890
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 3.38389731621937,
|
| 2037 |
+
"grad_norm": 0.5236608982086182,
|
| 2038 |
+
"learning_rate": 4.3617208740680346e-05,
|
| 2039 |
+
"loss": 0.0241,
|
| 2040 |
+
"step": 2900
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 3.395565927654609,
|
| 2044 |
+
"grad_norm": 0.3452656865119934,
|
| 2045 |
+
"learning_rate": 4.3298353450098126e-05,
|
| 2046 |
+
"loss": 0.0197,
|
| 2047 |
+
"step": 2910
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 3.407234539089848,
|
| 2051 |
+
"grad_norm": 0.5942583680152893,
|
| 2052 |
+
"learning_rate": 4.297977544584383e-05,
|
| 2053 |
+
"loss": 0.017,
|
| 2054 |
+
"step": 2920
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 3.4189031505250878,
|
| 2058 |
+
"grad_norm": 0.2938035726547241,
|
| 2059 |
+
"learning_rate": 4.266148790935344e-05,
|
| 2060 |
+
"loss": 0.0224,
|
| 2061 |
+
"step": 2930
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 3.430571761960327,
|
| 2065 |
+
"grad_norm": 0.33962100744247437,
|
| 2066 |
+
"learning_rate": 4.234350401004457e-05,
|
| 2067 |
+
"loss": 0.0184,
|
| 2068 |
+
"step": 2940
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 3.442240373395566,
|
| 2072 |
+
"grad_norm": 0.2965751886367798,
|
| 2073 |
+
"learning_rate": 4.2025836904771604e-05,
|
| 2074 |
+
"loss": 0.0232,
|
| 2075 |
+
"step": 2950
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 3.453908984830805,
|
| 2079 |
+
"grad_norm": 0.4069346785545349,
|
| 2080 |
+
"learning_rate": 4.17084997372813e-05,
|
| 2081 |
+
"loss": 0.0182,
|
| 2082 |
+
"step": 2960
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 3.4655775962660442,
|
| 2086 |
+
"grad_norm": 0.3834659159183502,
|
| 2087 |
+
"learning_rate": 4.139150563766896e-05,
|
| 2088 |
+
"loss": 0.0196,
|
| 2089 |
+
"step": 2970
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 3.477246207701284,
|
| 2093 |
+
"grad_norm": 0.2882491946220398,
|
| 2094 |
+
"learning_rate": 4.1074867721835194e-05,
|
| 2095 |
+
"loss": 0.0191,
|
| 2096 |
+
"step": 2980
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 3.488914819136523,
|
| 2100 |
+
"grad_norm": 0.37607982754707336,
|
| 2101 |
+
"learning_rate": 4.075859909094314e-05,
|
| 2102 |
+
"loss": 0.0185,
|
| 2103 |
+
"step": 2990
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 3.500583430571762,
|
| 2107 |
+
"grad_norm": 0.31778615713119507,
|
| 2108 |
+
"learning_rate": 4.044271283087649e-05,
|
| 2109 |
+
"loss": 0.0197,
|
| 2110 |
+
"step": 3000
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 3.512252042007001,
|
| 2114 |
+
"grad_norm": 0.33435627818107605,
|
| 2115 |
+
"learning_rate": 4.012722201169803e-05,
|
| 2116 |
+
"loss": 0.0207,
|
| 2117 |
+
"step": 3010
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 3.5239206534422403,
|
| 2121 |
+
"grad_norm": 0.3664036691188812,
|
| 2122 |
+
"learning_rate": 3.9812139687108815e-05,
|
| 2123 |
+
"loss": 0.0192,
|
| 2124 |
+
"step": 3020
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 3.53558926487748,
|
| 2128 |
+
"grad_norm": 0.4649610221385956,
|
| 2129 |
+
"learning_rate": 3.949747889390812e-05,
|
| 2130 |
+
"loss": 0.0201,
|
| 2131 |
+
"step": 3030
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 3.547257876312719,
|
| 2135 |
+
"grad_norm": 0.45405086874961853,
|
| 2136 |
+
"learning_rate": 3.9183252651453985e-05,
|
| 2137 |
+
"loss": 0.0207,
|
| 2138 |
+
"step": 3040
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 3.558926487747958,
|
| 2142 |
+
"grad_norm": 0.29085540771484375,
|
| 2143 |
+
"learning_rate": 3.886947396112453e-05,
|
| 2144 |
+
"loss": 0.025,
|
| 2145 |
+
"step": 3050
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 3.570595099183197,
|
| 2149 |
+
"grad_norm": 0.4277862012386322,
|
| 2150 |
+
"learning_rate": 3.8556155805780056e-05,
|
| 2151 |
+
"loss": 0.0209,
|
| 2152 |
+
"step": 3060
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 3.5822637106184363,
|
| 2156 |
+
"grad_norm": 0.3923989236354828,
|
| 2157 |
+
"learning_rate": 3.8243311149225796e-05,
|
| 2158 |
+
"loss": 0.0233,
|
| 2159 |
+
"step": 3070
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 3.593932322053676,
|
| 2163 |
+
"grad_norm": 0.4537499248981476,
|
| 2164 |
+
"learning_rate": 3.793095293567558e-05,
|
| 2165 |
+
"loss": 0.0191,
|
| 2166 |
+
"step": 3080
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 3.605600933488915,
|
| 2170 |
+
"grad_norm": 0.5022493600845337,
|
| 2171 |
+
"learning_rate": 3.76190940892163e-05,
|
| 2172 |
+
"loss": 0.0211,
|
| 2173 |
+
"step": 3090
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 3.617269544924154,
|
| 2177 |
+
"grad_norm": 0.3455962538719177,
|
| 2178 |
+
"learning_rate": 3.730774751327304e-05,
|
| 2179 |
+
"loss": 0.0199,
|
| 2180 |
+
"step": 3100
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 3.628938156359393,
|
| 2184 |
+
"grad_norm": 0.33194699883461,
|
| 2185 |
+
"learning_rate": 3.699692609007528e-05,
|
| 2186 |
+
"loss": 0.0199,
|
| 2187 |
+
"step": 3110
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 3.6406067677946323,
|
| 2191 |
+
"grad_norm": 0.24475599825382233,
|
| 2192 |
+
"learning_rate": 3.668664268012386e-05,
|
| 2193 |
+
"loss": 0.0214,
|
| 2194 |
+
"step": 3120
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 3.652275379229872,
|
| 2198 |
+
"grad_norm": 0.3331003487110138,
|
| 2199 |
+
"learning_rate": 3.637691012165886e-05,
|
| 2200 |
+
"loss": 0.0185,
|
| 2201 |
+
"step": 3130
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 3.663943990665111,
|
| 2205 |
+
"grad_norm": 0.5968544483184814,
|
| 2206 |
+
"learning_rate": 3.606774123012843e-05,
|
| 2207 |
+
"loss": 0.0222,
|
| 2208 |
+
"step": 3140
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 3.67561260210035,
|
| 2212 |
+
"grad_norm": 0.4008450508117676,
|
| 2213 |
+
"learning_rate": 3.575914879765848e-05,
|
| 2214 |
+
"loss": 0.0181,
|
| 2215 |
+
"step": 3150
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 3.6872812135355892,
|
| 2219 |
+
"grad_norm": 0.26033294200897217,
|
| 2220 |
+
"learning_rate": 3.545114559252348e-05,
|
| 2221 |
+
"loss": 0.0237,
|
| 2222 |
+
"step": 3160
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 3.6989498249708284,
|
| 2226 |
+
"grad_norm": 0.3555871248245239,
|
| 2227 |
+
"learning_rate": 3.5143744358618084e-05,
|
| 2228 |
+
"loss": 0.0186,
|
| 2229 |
+
"step": 3170
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 3.710618436406068,
|
| 2233 |
+
"grad_norm": 0.4315120279788971,
|
| 2234 |
+
"learning_rate": 3.483695781492989e-05,
|
| 2235 |
+
"loss": 0.0202,
|
| 2236 |
+
"step": 3180
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 3.722287047841307,
|
| 2240 |
+
"grad_norm": 0.3939124345779419,
|
| 2241 |
+
"learning_rate": 3.453079865501319e-05,
|
| 2242 |
+
"loss": 0.0198,
|
| 2243 |
+
"step": 3190
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 3.733955659276546,
|
| 2247 |
+
"grad_norm": 0.4471058249473572,
|
| 2248 |
+
"learning_rate": 3.4225279546463694e-05,
|
| 2249 |
+
"loss": 0.0214,
|
| 2250 |
+
"step": 3200
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 3.7456242707117853,
|
| 2254 |
+
"grad_norm": 0.5129342079162598,
|
| 2255 |
+
"learning_rate": 3.392041313039446e-05,
|
| 2256 |
+
"loss": 0.017,
|
| 2257 |
+
"step": 3210
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 3.7572928821470244,
|
| 2261 |
+
"grad_norm": 0.3735811710357666,
|
| 2262 |
+
"learning_rate": 3.361621202091285e-05,
|
| 2263 |
+
"loss": 0.023,
|
| 2264 |
+
"step": 3220
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 3.768961493582264,
|
| 2268 |
+
"grad_norm": 0.41765424609184265,
|
| 2269 |
+
"learning_rate": 3.3312688804598566e-05,
|
| 2270 |
+
"loss": 0.0219,
|
| 2271 |
+
"step": 3230
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 3.780630105017503,
|
| 2275 |
+
"grad_norm": 0.3862574100494385,
|
| 2276 |
+
"learning_rate": 3.3009856039982954e-05,
|
| 2277 |
+
"loss": 0.0167,
|
| 2278 |
+
"step": 3240
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 3.792298716452742,
|
| 2282 |
+
"grad_norm": 0.38762903213500977,
|
| 2283 |
+
"learning_rate": 3.270772625702929e-05,
|
| 2284 |
+
"loss": 0.0239,
|
| 2285 |
+
"step": 3250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 3.8039673278879813,
|
| 2289 |
+
"grad_norm": 0.44288456439971924,
|
| 2290 |
+
"learning_rate": 3.240631195661442e-05,
|
| 2291 |
+
"loss": 0.0184,
|
| 2292 |
+
"step": 3260
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 3.8156359393232204,
|
| 2296 |
+
"grad_norm": 0.49590441584587097,
|
| 2297 |
+
"learning_rate": 3.210562561001145e-05,
|
| 2298 |
+
"loss": 0.0203,
|
| 2299 |
+
"step": 3270
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 3.82730455075846,
|
| 2303 |
+
"grad_norm": 0.3890804946422577,
|
| 2304 |
+
"learning_rate": 3.180567965837381e-05,
|
| 2305 |
+
"loss": 0.02,
|
| 2306 |
+
"step": 3280
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 3.838973162193699,
|
| 2310 |
+
"grad_norm": 0.36380621790885925,
|
| 2311 |
+
"learning_rate": 3.150648651222047e-05,
|
| 2312 |
+
"loss": 0.0149,
|
| 2313 |
+
"step": 3290
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 3.8506417736289382,
|
| 2317 |
+
"grad_norm": 0.5620595812797546,
|
| 2318 |
+
"learning_rate": 3.1208058550922394e-05,
|
| 2319 |
+
"loss": 0.0202,
|
| 2320 |
+
"step": 3300
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 3.8623103850641773,
|
| 2324 |
+
"grad_norm": 0.3390384316444397,
|
| 2325 |
+
"learning_rate": 3.0910408122190395e-05,
|
| 2326 |
+
"loss": 0.0169,
|
| 2327 |
+
"step": 3310
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 3.8739789964994165,
|
| 2331 |
+
"grad_norm": 0.4755258858203888,
|
| 2332 |
+
"learning_rate": 3.061354754156421e-05,
|
| 2333 |
+
"loss": 0.0187,
|
| 2334 |
+
"step": 3320
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 3.885647607934656,
|
| 2338 |
+
"grad_norm": 0.34263917803764343,
|
| 2339 |
+
"learning_rate": 3.0317489091902935e-05,
|
| 2340 |
+
"loss": 0.0216,
|
| 2341 |
+
"step": 3330
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 3.897316219369895,
|
| 2345 |
+
"grad_norm": 0.29937320947647095,
|
| 2346 |
+
"learning_rate": 3.002224502287682e-05,
|
| 2347 |
+
"loss": 0.0198,
|
| 2348 |
+
"step": 3340
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 3.9089848308051343,
|
| 2352 |
+
"grad_norm": 0.4662054777145386,
|
| 2353 |
+
"learning_rate": 2.972782755046043e-05,
|
| 2354 |
+
"loss": 0.0201,
|
| 2355 |
+
"step": 3350
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 3.9206534422403734,
|
| 2359 |
+
"grad_norm": 0.43561384081840515,
|
| 2360 |
+
"learning_rate": 2.943424885642717e-05,
|
| 2361 |
+
"loss": 0.0188,
|
| 2362 |
+
"step": 3360
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 3.9323220536756125,
|
| 2366 |
+
"grad_norm": 0.33296090364456177,
|
| 2367 |
+
"learning_rate": 2.914152108784528e-05,
|
| 2368 |
+
"loss": 0.0181,
|
| 2369 |
+
"step": 3370
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 3.943990665110852,
|
| 2373 |
+
"grad_norm": 0.5525170564651489,
|
| 2374 |
+
"learning_rate": 2.8849656356575262e-05,
|
| 2375 |
+
"loss": 0.0145,
|
| 2376 |
+
"step": 3380
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 3.955659276546091,
|
| 2380 |
+
"grad_norm": 0.4553203582763672,
|
| 2381 |
+
"learning_rate": 2.8558666738768715e-05,
|
| 2382 |
+
"loss": 0.0198,
|
| 2383 |
+
"step": 3390
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 3.9673278879813303,
|
| 2387 |
+
"grad_norm": 0.36274367570877075,
|
| 2388 |
+
"learning_rate": 2.8268564274368616e-05,
|
| 2389 |
+
"loss": 0.024,
|
| 2390 |
+
"step": 3400
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 3.9789964994165694,
|
| 2394 |
+
"grad_norm": 0.2932572066783905,
|
| 2395 |
+
"learning_rate": 2.797936096661129e-05,
|
| 2396 |
+
"loss": 0.019,
|
| 2397 |
+
"step": 3410
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 3.9906651108518085,
|
| 2401 |
+
"grad_norm": 0.22307224571704865,
|
| 2402 |
+
"learning_rate": 2.7691068781529622e-05,
|
| 2403 |
+
"loss": 0.0181,
|
| 2404 |
+
"step": 3420
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 4.002333722287048,
|
| 2408 |
+
"grad_norm": 0.23651975393295288,
|
| 2409 |
+
"learning_rate": 2.7403699647458097e-05,
|
| 2410 |
+
"loss": 0.018,
|
| 2411 |
+
"step": 3430
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 4.014002333722287,
|
| 2415 |
+
"grad_norm": 0.42548108100891113,
|
| 2416 |
+
"learning_rate": 2.711726545453912e-05,
|
| 2417 |
+
"loss": 0.022,
|
| 2418 |
+
"step": 3440
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 4.025670945157526,
|
| 2422 |
+
"grad_norm": 0.27161461114883423,
|
| 2423 |
+
"learning_rate": 2.68317780542311e-05,
|
| 2424 |
+
"loss": 0.0188,
|
| 2425 |
+
"step": 3450
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 4.037339556592766,
|
| 2429 |
+
"grad_norm": 0.4137495160102844,
|
| 2430 |
+
"learning_rate": 2.6547249258818164e-05,
|
| 2431 |
+
"loss": 0.0174,
|
| 2432 |
+
"step": 3460
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 4.049008168028005,
|
| 2436 |
+
"grad_norm": 0.23922982811927795,
|
| 2437 |
+
"learning_rate": 2.6263690840921272e-05,
|
| 2438 |
+
"loss": 0.0183,
|
| 2439 |
+
"step": 3470
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 4.060676779463244,
|
| 2443 |
+
"grad_norm": 0.4632745385169983,
|
| 2444 |
+
"learning_rate": 2.5981114533011263e-05,
|
| 2445 |
+
"loss": 0.018,
|
| 2446 |
+
"step": 3480
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 4.072345390898483,
|
| 2450 |
+
"grad_norm": 0.3738875389099121,
|
| 2451 |
+
"learning_rate": 2.5699532026923255e-05,
|
| 2452 |
+
"loss": 0.0206,
|
| 2453 |
+
"step": 3490
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 4.084014002333722,
|
| 2457 |
+
"grad_norm": 0.3949178159236908,
|
| 2458 |
+
"learning_rate": 2.5418954973373037e-05,
|
| 2459 |
+
"loss": 0.0202,
|
| 2460 |
+
"step": 3500
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 4.095682613768962,
|
| 2464 |
+
"grad_norm": 0.35208699107170105,
|
| 2465 |
+
"learning_rate": 2.513939498147494e-05,
|
| 2466 |
+
"loss": 0.0193,
|
| 2467 |
+
"step": 3510
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 4.107351225204201,
|
| 2471 |
+
"grad_norm": 0.37377116084098816,
|
| 2472 |
+
"learning_rate": 2.4860863618261444e-05,
|
| 2473 |
+
"loss": 0.0163,
|
| 2474 |
+
"step": 3520
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 4.11901983663944,
|
| 2478 |
+
"grad_norm": 0.31615620851516724,
|
| 2479 |
+
"learning_rate": 2.4583372408204726e-05,
|
| 2480 |
+
"loss": 0.0169,
|
| 2481 |
+
"step": 3530
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 4.130688448074679,
|
| 2485 |
+
"grad_norm": 0.36727479100227356,
|
| 2486 |
+
"learning_rate": 2.4306932832739653e-05,
|
| 2487 |
+
"loss": 0.0161,
|
| 2488 |
+
"step": 3540
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 4.142357059509918,
|
| 2492 |
+
"grad_norm": 0.43038618564605713,
|
| 2493 |
+
"learning_rate": 2.4031556329788902e-05,
|
| 2494 |
+
"loss": 0.0179,
|
| 2495 |
+
"step": 3550
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 4.154025670945158,
|
| 2499 |
+
"grad_norm": 0.46835821866989136,
|
| 2500 |
+
"learning_rate": 2.3757254293289567e-05,
|
| 2501 |
+
"loss": 0.0213,
|
| 2502 |
+
"step": 3560
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 4.165694282380397,
|
| 2506 |
+
"grad_norm": 0.3531014919281006,
|
| 2507 |
+
"learning_rate": 2.3484038072721758e-05,
|
| 2508 |
+
"loss": 0.019,
|
| 2509 |
+
"step": 3570
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 4.177362893815636,
|
| 2513 |
+
"grad_norm": 0.28978416323661804,
|
| 2514 |
+
"learning_rate": 2.3211918972639114e-05,
|
| 2515 |
+
"loss": 0.0156,
|
| 2516 |
+
"step": 3580
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 4.189031505250875,
|
| 2520 |
+
"grad_norm": 0.2763589024543762,
|
| 2521 |
+
"learning_rate": 2.2940908252200856e-05,
|
| 2522 |
+
"loss": 0.0171,
|
| 2523 |
+
"step": 3590
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 4.200700116686114,
|
| 2527 |
+
"grad_norm": 0.4522627294063568,
|
| 2528 |
+
"learning_rate": 2.26710171247062e-05,
|
| 2529 |
+
"loss": 0.0183,
|
| 2530 |
+
"step": 3600
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 4.212368728121354,
|
| 2534 |
+
"grad_norm": 0.37962812185287476,
|
| 2535 |
+
"learning_rate": 2.240225675713015e-05,
|
| 2536 |
+
"loss": 0.0172,
|
| 2537 |
+
"step": 3610
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 4.224037339556593,
|
| 2541 |
+
"grad_norm": 0.2463466078042984,
|
| 2542 |
+
"learning_rate": 2.2134638269661555e-05,
|
| 2543 |
+
"loss": 0.0167,
|
| 2544 |
+
"step": 3620
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 4.235705950991832,
|
| 2548 |
+
"grad_norm": 0.28136634826660156,
|
| 2549 |
+
"learning_rate": 2.186817273524307e-05,
|
| 2550 |
+
"loss": 0.0165,
|
| 2551 |
+
"step": 3630
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 4.247374562427071,
|
| 2555 |
+
"grad_norm": 0.32508429884910583,
|
| 2556 |
+
"learning_rate": 2.1602871179112855e-05,
|
| 2557 |
+
"loss": 0.0154,
|
| 2558 |
+
"step": 3640
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 4.2590431738623105,
|
| 2562 |
+
"grad_norm": 0.4417642056941986,
|
| 2563 |
+
"learning_rate": 2.1338744578348557e-05,
|
| 2564 |
+
"loss": 0.019,
|
| 2565 |
+
"step": 3650
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 4.27071178529755,
|
| 2569 |
+
"grad_norm": 0.4562079906463623,
|
| 2570 |
+
"learning_rate": 2.107580386141298e-05,
|
| 2571 |
+
"loss": 0.0206,
|
| 2572 |
+
"step": 3660
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 4.282380396732789,
|
| 2576 |
+
"grad_norm": 0.2779424488544464,
|
| 2577 |
+
"learning_rate": 2.0814059907702027e-05,
|
| 2578 |
+
"loss": 0.0172,
|
| 2579 |
+
"step": 3670
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 4.294049008168028,
|
| 2583 |
+
"grad_norm": 0.27644243836402893,
|
| 2584 |
+
"learning_rate": 2.0553523547094473e-05,
|
| 2585 |
+
"loss": 0.0162,
|
| 2586 |
+
"step": 3680
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 4.305717619603267,
|
| 2590 |
+
"grad_norm": 0.3589038848876953,
|
| 2591 |
+
"learning_rate": 2.029420555950387e-05,
|
| 2592 |
+
"loss": 0.0159,
|
| 2593 |
+
"step": 3690
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 4.3173862310385065,
|
| 2597 |
+
"grad_norm": 0.35192936658859253,
|
| 2598 |
+
"learning_rate": 2.0036116674432654e-05,
|
| 2599 |
+
"loss": 0.0163,
|
| 2600 |
+
"step": 3700
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 4.329054842473745,
|
| 2604 |
+
"grad_norm": 0.31670838594436646,
|
| 2605 |
+
"learning_rate": 1.9779267570527988e-05,
|
| 2606 |
+
"loss": 0.0142,
|
| 2607 |
+
"step": 3710
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 4.340723453908985,
|
| 2611 |
+
"grad_norm": 0.2387942522764206,
|
| 2612 |
+
"learning_rate": 1.9523668875140106e-05,
|
| 2613 |
+
"loss": 0.016,
|
| 2614 |
+
"step": 3720
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 4.352392065344224,
|
| 2618 |
+
"grad_norm": 0.28022316098213196,
|
| 2619 |
+
"learning_rate": 1.9269331163882488e-05,
|
| 2620 |
+
"loss": 0.0158,
|
| 2621 |
+
"step": 3730
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 4.364060676779463,
|
| 2625 |
+
"grad_norm": 0.5723684430122375,
|
| 2626 |
+
"learning_rate": 1.9016264960194314e-05,
|
| 2627 |
+
"loss": 0.0227,
|
| 2628 |
+
"step": 3740
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 4.3757292882147025,
|
| 2632 |
+
"grad_norm": 0.2613903880119324,
|
| 2633 |
+
"learning_rate": 1.876448073490511e-05,
|
| 2634 |
+
"loss": 0.0164,
|
| 2635 |
+
"step": 3750
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 4.387397899649942,
|
| 2639 |
+
"grad_norm": 0.4642601013183594,
|
| 2640 |
+
"learning_rate": 1.8513988905801387e-05,
|
| 2641 |
+
"loss": 0.0152,
|
| 2642 |
+
"step": 3760
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 4.399066511085181,
|
| 2646 |
+
"grad_norm": 0.38122987747192383,
|
| 2647 |
+
"learning_rate": 1.8264799837195733e-05,
|
| 2648 |
+
"loss": 0.0162,
|
| 2649 |
+
"step": 3770
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 4.41073512252042,
|
| 2653 |
+
"grad_norm": 0.27166861295700073,
|
| 2654 |
+
"learning_rate": 1.8016923839497847e-05,
|
| 2655 |
+
"loss": 0.0183,
|
| 2656 |
+
"step": 3780
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 4.422403733955659,
|
| 2660 |
+
"grad_norm": 0.29175394773483276,
|
| 2661 |
+
"learning_rate": 1.777037116878804e-05,
|
| 2662 |
+
"loss": 0.0151,
|
| 2663 |
+
"step": 3790
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 4.434072345390899,
|
| 2667 |
+
"grad_norm": 0.30123022198677063,
|
| 2668 |
+
"learning_rate": 1.7525152026392867e-05,
|
| 2669 |
+
"loss": 0.0132,
|
| 2670 |
+
"step": 3800
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 4.445740956826137,
|
| 2674 |
+
"grad_norm": 0.3631080090999603,
|
| 2675 |
+
"learning_rate": 1.7281276558462935e-05,
|
| 2676 |
+
"loss": 0.0197,
|
| 2677 |
+
"step": 3810
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 4.457409568261377,
|
| 2681 |
+
"grad_norm": 0.39106523990631104,
|
| 2682 |
+
"learning_rate": 1.7038754855553263e-05,
|
| 2683 |
+
"loss": 0.0211,
|
| 2684 |
+
"step": 3820
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 4.469078179696616,
|
| 2688 |
+
"grad_norm": 0.34949415922164917,
|
| 2689 |
+
"learning_rate": 1.679759695220562e-05,
|
| 2690 |
+
"loss": 0.0188,
|
| 2691 |
+
"step": 3830
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 4.480746791131855,
|
| 2695 |
+
"grad_norm": 0.506232738494873,
|
| 2696 |
+
"learning_rate": 1.655781282653347e-05,
|
| 2697 |
+
"loss": 0.0225,
|
| 2698 |
+
"step": 3840
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 4.492415402567095,
|
| 2702 |
+
"grad_norm": 0.23795835673809052,
|
| 2703 |
+
"learning_rate": 1.6319412399809015e-05,
|
| 2704 |
+
"loss": 0.0161,
|
| 2705 |
+
"step": 3850
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 4.504084014002334,
|
| 2709 |
+
"grad_norm": 0.4405321478843689,
|
| 2710 |
+
"learning_rate": 1.608240553605273e-05,
|
| 2711 |
+
"loss": 0.0186,
|
| 2712 |
+
"step": 3860
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 4.515752625437573,
|
| 2716 |
+
"grad_norm": 0.23445239663124084,
|
| 2717 |
+
"learning_rate": 1.584680204162528e-05,
|
| 2718 |
+
"loss": 0.0151,
|
| 2719 |
+
"step": 3870
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 4.527421236872812,
|
| 2723 |
+
"grad_norm": 0.2741932272911072,
|
| 2724 |
+
"learning_rate": 1.5612611664821676e-05,
|
| 2725 |
+
"loss": 0.0147,
|
| 2726 |
+
"step": 3880
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 4.539089848308051,
|
| 2730 |
+
"grad_norm": 0.3066684901714325,
|
| 2731 |
+
"learning_rate": 1.537984409546805e-05,
|
| 2732 |
+
"loss": 0.0132,
|
| 2733 |
+
"step": 3890
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 4.550758459743291,
|
| 2737 |
+
"grad_norm": 0.2490091770887375,
|
| 2738 |
+
"learning_rate": 1.5148508964520585e-05,
|
| 2739 |
+
"loss": 0.0176,
|
| 2740 |
+
"step": 3900
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 4.562427071178529,
|
| 2744 |
+
"grad_norm": 0.31570225954055786,
|
| 2745 |
+
"learning_rate": 1.4918615843667171e-05,
|
| 2746 |
+
"loss": 0.0167,
|
| 2747 |
+
"step": 3910
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 4.574095682613769,
|
| 2751 |
+
"grad_norm": 0.3133799433708191,
|
| 2752 |
+
"learning_rate": 1.4690174244931304e-05,
|
| 2753 |
+
"loss": 0.0172,
|
| 2754 |
+
"step": 3920
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 4.585764294049008,
|
| 2758 |
+
"grad_norm": 0.34936434030532837,
|
| 2759 |
+
"learning_rate": 1.4463193620278453e-05,
|
| 2760 |
+
"loss": 0.0148,
|
| 2761 |
+
"step": 3930
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 4.597432905484247,
|
| 2765 |
+
"grad_norm": 0.289408802986145,
|
| 2766 |
+
"learning_rate": 1.4237683361225118e-05,
|
| 2767 |
+
"loss": 0.0149,
|
| 2768 |
+
"step": 3940
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 4.609101516919487,
|
| 2772 |
+
"grad_norm": 0.3845432698726654,
|
| 2773 |
+
"learning_rate": 1.401365279845011e-05,
|
| 2774 |
+
"loss": 0.016,
|
| 2775 |
+
"step": 3950
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 4.620770128354726,
|
| 2779 |
+
"grad_norm": 0.3027130663394928,
|
| 2780 |
+
"learning_rate": 1.3791111201408601e-05,
|
| 2781 |
+
"loss": 0.0149,
|
| 2782 |
+
"step": 3960
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 4.632438739789965,
|
| 2786 |
+
"grad_norm": 0.29027318954467773,
|
| 2787 |
+
"learning_rate": 1.3570067777948497e-05,
|
| 2788 |
+
"loss": 0.0148,
|
| 2789 |
+
"step": 3970
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 4.6441073512252045,
|
| 2793 |
+
"grad_norm": 0.29574844241142273,
|
| 2794 |
+
"learning_rate": 1.3350531673929501e-05,
|
| 2795 |
+
"loss": 0.0142,
|
| 2796 |
+
"step": 3980
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 4.655775962660443,
|
| 2800 |
+
"grad_norm": 0.38981419801712036,
|
| 2801 |
+
"learning_rate": 1.3132511972844725e-05,
|
| 2802 |
+
"loss": 0.0127,
|
| 2803 |
+
"step": 3990
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 4.667444574095683,
|
| 2807 |
+
"grad_norm": 0.318435400724411,
|
| 2808 |
+
"learning_rate": 1.291601769544476e-05,
|
| 2809 |
+
"loss": 0.0146,
|
| 2810 |
+
"step": 4000
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 4.679113185530921,
|
| 2814 |
+
"grad_norm": 0.3407133221626282,
|
| 2815 |
+
"learning_rate": 1.2701057799364591e-05,
|
| 2816 |
+
"loss": 0.0161,
|
| 2817 |
+
"step": 4010
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 4.690781796966161,
|
| 2821 |
+
"grad_norm": 0.17900419235229492,
|
| 2822 |
+
"learning_rate": 1.2487641178752779e-05,
|
| 2823 |
+
"loss": 0.0135,
|
| 2824 |
+
"step": 4020
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 4.7024504084014005,
|
| 2828 |
+
"grad_norm": 0.4556140899658203,
|
| 2829 |
+
"learning_rate": 1.2275776663903565e-05,
|
| 2830 |
+
"loss": 0.0187,
|
| 2831 |
+
"step": 4030
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 4.714119019836639,
|
| 2835 |
+
"grad_norm": 0.3229597210884094,
|
| 2836 |
+
"learning_rate": 1.2065473020891544e-05,
|
| 2837 |
+
"loss": 0.0169,
|
| 2838 |
+
"step": 4040
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 4.725787631271879,
|
| 2842 |
+
"grad_norm": 0.4072445333003998,
|
| 2843 |
+
"learning_rate": 1.1856738951208852e-05,
|
| 2844 |
+
"loss": 0.0157,
|
| 2845 |
+
"step": 4050
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 4.737456242707118,
|
| 2849 |
+
"grad_norm": 0.32965341210365295,
|
| 2850 |
+
"learning_rate": 1.164958309140523e-05,
|
| 2851 |
+
"loss": 0.0148,
|
| 2852 |
+
"step": 4060
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 4.749124854142357,
|
| 2856 |
+
"grad_norm": 0.341792494058609,
|
| 2857 |
+
"learning_rate": 1.144401401273062e-05,
|
| 2858 |
+
"loss": 0.017,
|
| 2859 |
+
"step": 4070
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 4.7607934655775965,
|
| 2863 |
+
"grad_norm": 0.34901103377342224,
|
| 2864 |
+
"learning_rate": 1.1240040220780573e-05,
|
| 2865 |
+
"loss": 0.015,
|
| 2866 |
+
"step": 4080
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 4.772462077012835,
|
| 2870 |
+
"grad_norm": 0.48683103919029236,
|
| 2871 |
+
"learning_rate": 1.103767015514427e-05,
|
| 2872 |
+
"loss": 0.0188,
|
| 2873 |
+
"step": 4090
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 4.784130688448075,
|
| 2877 |
+
"grad_norm": 0.22785413265228271,
|
| 2878 |
+
"learning_rate": 1.083691218905532e-05,
|
| 2879 |
+
"loss": 0.0151,
|
| 2880 |
+
"step": 4100
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 4.795799299883313,
|
| 2884 |
+
"grad_norm": 0.18910996615886688,
|
| 2885 |
+
"learning_rate": 1.0637774629045444e-05,
|
| 2886 |
+
"loss": 0.0178,
|
| 2887 |
+
"step": 4110
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 4.807467911318553,
|
| 2891 |
+
"grad_norm": 0.41959840059280396,
|
| 2892 |
+
"learning_rate": 1.0440265714600572e-05,
|
| 2893 |
+
"loss": 0.0165,
|
| 2894 |
+
"step": 4120
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 4.819136522753793,
|
| 2898 |
+
"grad_norm": 0.40362274646759033,
|
| 2899 |
+
"learning_rate": 1.0244393617820113e-05,
|
| 2900 |
+
"loss": 0.0129,
|
| 2901 |
+
"step": 4130
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 4.830805134189031,
|
| 2905 |
+
"grad_norm": 0.5385197401046753,
|
| 2906 |
+
"learning_rate": 1.0050166443078713e-05,
|
| 2907 |
+
"loss": 0.0144,
|
| 2908 |
+
"step": 4140
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 4.842473745624271,
|
| 2912 |
+
"grad_norm": 0.3056287169456482,
|
| 2913 |
+
"learning_rate": 9.857592226690958e-06,
|
| 2914 |
+
"loss": 0.0109,
|
| 2915 |
+
"step": 4150
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 4.85414235705951,
|
| 2919 |
+
"grad_norm": 0.3714611530303955,
|
| 2920 |
+
"learning_rate": 9.666678936578938e-06,
|
| 2921 |
+
"loss": 0.0195,
|
| 2922 |
+
"step": 4160
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 4.865810968494749,
|
| 2926 |
+
"grad_norm": 0.36349254846572876,
|
| 2927 |
+
"learning_rate": 9.477434471942426e-06,
|
| 2928 |
+
"loss": 0.0149,
|
| 2929 |
+
"step": 4170
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 4.877479579929989,
|
| 2933 |
+
"grad_norm": 0.23277226090431213,
|
| 2934 |
+
"learning_rate": 9.289866662932189e-06,
|
| 2935 |
+
"loss": 0.0144,
|
| 2936 |
+
"step": 4180
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 4.889148191365227,
|
| 2940 |
+
"grad_norm": 0.2719990909099579,
|
| 2941 |
+
"learning_rate": 9.10398327032589e-06,
|
| 2942 |
+
"loss": 0.0158,
|
| 2943 |
+
"step": 4190
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 4.900816802800467,
|
| 2947 |
+
"grad_norm": 0.23565804958343506,
|
| 2948 |
+
"learning_rate": 8.91979198520706e-06,
|
| 2949 |
+
"loss": 0.0134,
|
| 2950 |
+
"step": 4200
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 4.9124854142357055,
|
| 2954 |
+
"grad_norm": 0.3723337650299072,
|
| 2955 |
+
"learning_rate": 8.737300428646856e-06,
|
| 2956 |
+
"loss": 0.0145,
|
| 2957 |
+
"step": 4210
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 4.924154025670945,
|
| 2961 |
+
"grad_norm": 0.20222023129463196,
|
| 2962 |
+
"learning_rate": 8.556516151388661e-06,
|
| 2963 |
+
"loss": 0.0128,
|
| 2964 |
+
"step": 4220
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 4.935822637106185,
|
| 2968 |
+
"grad_norm": 0.23124603927135468,
|
| 2969 |
+
"learning_rate": 8.377446633535795e-06,
|
| 2970 |
+
"loss": 0.0148,
|
| 2971 |
+
"step": 4230
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 4.947491248541423,
|
| 2975 |
+
"grad_norm": 0.30160167813301086,
|
| 2976 |
+
"learning_rate": 8.20009928424189e-06,
|
| 2977 |
+
"loss": 0.0134,
|
| 2978 |
+
"step": 4240
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 4.959159859976663,
|
| 2982 |
+
"grad_norm": 0.328268438577652,
|
| 2983 |
+
"learning_rate": 8.024481441404441e-06,
|
| 2984 |
+
"loss": 0.0139,
|
| 2985 |
+
"step": 4250
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 4.970828471411902,
|
| 2989 |
+
"grad_norm": 0.3776269257068634,
|
| 2990 |
+
"learning_rate": 7.8506003713611e-06,
|
| 2991 |
+
"loss": 0.0166,
|
| 2992 |
+
"step": 4260
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 4.982497082847141,
|
| 2996 |
+
"grad_norm": 0.1584804654121399,
|
| 2997 |
+
"learning_rate": 7.67846326858907e-06,
|
| 2998 |
+
"loss": 0.0152,
|
| 2999 |
+
"step": 4270
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 4.994165694282381,
|
| 3003 |
+
"grad_norm": 0.44497108459472656,
|
| 3004 |
+
"learning_rate": 7.508077255407459e-06,
|
| 3005 |
+
"loss": 0.0144,
|
| 3006 |
+
"step": 4280
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 5.005834305717619,
|
| 3010 |
+
"grad_norm": 0.24507474899291992,
|
| 3011 |
+
"learning_rate": 7.339449381682506e-06,
|
| 3012 |
+
"loss": 0.016,
|
| 3013 |
+
"step": 4290
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 5.017502917152859,
|
| 3017 |
+
"grad_norm": 0.22279092669487,
|
| 3018 |
+
"learning_rate": 7.172586624535993e-06,
|
| 3019 |
+
"loss": 0.02,
|
| 3020 |
+
"step": 4300
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 5.029171528588098,
|
| 3024 |
+
"grad_norm": 0.43602585792541504,
|
| 3025 |
+
"learning_rate": 7.007495888056432e-06,
|
| 3026 |
+
"loss": 0.0124,
|
| 3027 |
+
"step": 4310
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 5.040840140023337,
|
| 3031 |
+
"grad_norm": 0.19565297663211823,
|
| 3032 |
+
"learning_rate": 6.8441840030135205e-06,
|
| 3033 |
+
"loss": 0.0143,
|
| 3034 |
+
"step": 4320
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 5.052508751458577,
|
| 3038 |
+
"grad_norm": 0.34382426738739014,
|
| 3039 |
+
"learning_rate": 6.68265772657547e-06,
|
| 3040 |
+
"loss": 0.0183,
|
| 3041 |
+
"step": 4330
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 5.064177362893815,
|
| 3045 |
+
"grad_norm": 0.28925248980522156,
|
| 3046 |
+
"learning_rate": 6.522923742029374e-06,
|
| 3047 |
+
"loss": 0.0124,
|
| 3048 |
+
"step": 4340
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 5.075845974329055,
|
| 3052 |
+
"grad_norm": 0.28685078024864197,
|
| 3053 |
+
"learning_rate": 6.364988658504767e-06,
|
| 3054 |
+
"loss": 0.014,
|
| 3055 |
+
"step": 4350
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 5.0875145857642945,
|
| 3059 |
+
"grad_norm": 0.30919069051742554,
|
| 3060 |
+
"learning_rate": 6.208859010700091e-06,
|
| 3061 |
+
"loss": 0.0168,
|
| 3062 |
+
"step": 4360
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 5.099183197199533,
|
| 3066 |
+
"grad_norm": 0.41703978180885315,
|
| 3067 |
+
"learning_rate": 6.054541258612373e-06,
|
| 3068 |
+
"loss": 0.0132,
|
| 3069 |
+
"step": 4370
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 5.110851808634773,
|
| 3073 |
+
"grad_norm": 0.33975979685783386,
|
| 3074 |
+
"learning_rate": 5.9020417872698965e-06,
|
| 3075 |
+
"loss": 0.012,
|
| 3076 |
+
"step": 4380
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 5.122520420070011,
|
| 3080 |
+
"grad_norm": 0.5082515478134155,
|
| 3081 |
+
"learning_rate": 5.751366906468025e-06,
|
| 3082 |
+
"loss": 0.016,
|
| 3083 |
+
"step": 4390
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 5.134189031505251,
|
| 3087 |
+
"grad_norm": 0.36697059869766235,
|
| 3088 |
+
"learning_rate": 5.60252285050818e-06,
|
| 3089 |
+
"loss": 0.0161,
|
| 3090 |
+
"step": 4400
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 5.14585764294049,
|
| 3094 |
+
"grad_norm": 0.24588853120803833,
|
| 3095 |
+
"learning_rate": 5.455515777939779e-06,
|
| 3096 |
+
"loss": 0.0126,
|
| 3097 |
+
"step": 4410
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 5.157526254375729,
|
| 3101 |
+
"grad_norm": 0.1793574094772339,
|
| 3102 |
+
"learning_rate": 5.310351771305566e-06,
|
| 3103 |
+
"loss": 0.0126,
|
| 3104 |
+
"step": 4420
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 5.169194865810969,
|
| 3108 |
+
"grad_norm": 0.20527991652488708,
|
| 3109 |
+
"learning_rate": 5.167036836889811e-06,
|
| 3110 |
+
"loss": 0.013,
|
| 3111 |
+
"step": 4430
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 5.180863477246207,
|
| 3115 |
+
"grad_norm": 0.30803123116493225,
|
| 3116 |
+
"learning_rate": 5.025576904469842e-06,
|
| 3117 |
+
"loss": 0.0166,
|
| 3118 |
+
"step": 4440
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 5.192532088681447,
|
| 3122 |
+
"grad_norm": 0.31528738141059875,
|
| 3123 |
+
"learning_rate": 4.885977827070748e-06,
|
| 3124 |
+
"loss": 0.0154,
|
| 3125 |
+
"step": 4450
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 5.204200700116687,
|
| 3129 |
+
"grad_norm": 0.3894023597240448,
|
| 3130 |
+
"learning_rate": 4.748245380723104e-06,
|
| 3131 |
+
"loss": 0.0104,
|
| 3132 |
+
"step": 4460
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 5.215869311551925,
|
| 3136 |
+
"grad_norm": 0.42655327916145325,
|
| 3137 |
+
"learning_rate": 4.612385264224084e-06,
|
| 3138 |
+
"loss": 0.011,
|
| 3139 |
+
"step": 4470
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 5.227537922987165,
|
| 3143 |
+
"grad_norm": 0.28541433811187744,
|
| 3144 |
+
"learning_rate": 4.4784030989015905e-06,
|
| 3145 |
+
"loss": 0.0119,
|
| 3146 |
+
"step": 4480
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 5.2392065344224035,
|
| 3150 |
+
"grad_norm": 0.2481224685907364,
|
| 3151 |
+
"learning_rate": 4.34630442838172e-06,
|
| 3152 |
+
"loss": 0.0174,
|
| 3153 |
+
"step": 4490
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 5.250875145857643,
|
| 3157 |
+
"grad_norm": 0.2757529616355896,
|
| 3158 |
+
"learning_rate": 4.216094718359348e-06,
|
| 3159 |
+
"loss": 0.0164,
|
| 3160 |
+
"step": 4500
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 5.262543757292882,
|
| 3164 |
+
"grad_norm": 0.26155319809913635,
|
| 3165 |
+
"learning_rate": 4.087779356371996e-06,
|
| 3166 |
+
"loss": 0.0149,
|
| 3167 |
+
"step": 4510
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 5.274212368728121,
|
| 3171 |
+
"grad_norm": 0.3611695468425751,
|
| 3172 |
+
"learning_rate": 3.961363651576977e-06,
|
| 3173 |
+
"loss": 0.0147,
|
| 3174 |
+
"step": 4520
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 5.285880980163361,
|
| 3178 |
+
"grad_norm": 0.30108821392059326,
|
| 3179 |
+
"learning_rate": 3.836852834531607e-06,
|
| 3180 |
+
"loss": 0.01,
|
| 3181 |
+
"step": 4530
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 5.2975495915985995,
|
| 3185 |
+
"grad_norm": 0.43208441138267517,
|
| 3186 |
+
"learning_rate": 3.714252056976897e-06,
|
| 3187 |
+
"loss": 0.0131,
|
| 3188 |
+
"step": 4540
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 5.309218203033839,
|
| 3192 |
+
"grad_norm": 0.21732571721076965,
|
| 3193 |
+
"learning_rate": 3.5935663916243145e-06,
|
| 3194 |
+
"loss": 0.011,
|
| 3195 |
+
"step": 4550
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 5.320886814469079,
|
| 3199 |
+
"grad_norm": 0.24558159708976746,
|
| 3200 |
+
"learning_rate": 3.4748008319459455e-06,
|
| 3201 |
+
"loss": 0.015,
|
| 3202 |
+
"step": 4560
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 5.332555425904317,
|
| 3206 |
+
"grad_norm": 0.3239372968673706,
|
| 3207 |
+
"learning_rate": 3.357960291967882e-06,
|
| 3208 |
+
"loss": 0.0138,
|
| 3209 |
+
"step": 4570
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 5.344224037339557,
|
| 3213 |
+
"grad_norm": 0.4319896101951599,
|
| 3214 |
+
"learning_rate": 3.2430496060668613e-06,
|
| 3215 |
+
"loss": 0.0144,
|
| 3216 |
+
"step": 4580
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 5.3558926487747955,
|
| 3220 |
+
"grad_norm": 0.3155679404735565,
|
| 3221 |
+
"learning_rate": 3.1300735287703e-06,
|
| 3222 |
+
"loss": 0.0125,
|
| 3223 |
+
"step": 4590
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 5.367561260210035,
|
| 3227 |
+
"grad_norm": 0.3422655761241913,
|
| 3228 |
+
"learning_rate": 3.019036734559494e-06,
|
| 3229 |
+
"loss": 0.0154,
|
| 3230 |
+
"step": 4600
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 5.379229871645274,
|
| 3234 |
+
"grad_norm": 0.3783782422542572,
|
| 3235 |
+
"learning_rate": 2.9099438176762973e-06,
|
| 3236 |
+
"loss": 0.0166,
|
| 3237 |
+
"step": 4610
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 5.390898483080513,
|
| 3241 |
+
"grad_norm": 0.14136770367622375,
|
| 3242 |
+
"learning_rate": 2.8027992919329616e-06,
|
| 3243 |
+
"loss": 0.0147,
|
| 3244 |
+
"step": 4620
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 5.402567094515753,
|
| 3248 |
+
"grad_norm": 0.3206866979598999,
|
| 3249 |
+
"learning_rate": 2.6976075905253917e-06,
|
| 3250 |
+
"loss": 0.011,
|
| 3251 |
+
"step": 4630
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 5.414235705950992,
|
| 3255 |
+
"grad_norm": 0.36189404129981995,
|
| 3256 |
+
"learning_rate": 2.594373065849748e-06,
|
| 3257 |
+
"loss": 0.014,
|
| 3258 |
+
"step": 4640
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 5.425904317386231,
|
| 3262 |
+
"grad_norm": 0.41719111800193787,
|
| 3263 |
+
"learning_rate": 2.4930999893223004e-06,
|
| 3264 |
+
"loss": 0.0126,
|
| 3265 |
+
"step": 4650
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 5.437572928821471,
|
| 3269 |
+
"grad_norm": 0.5754780173301697,
|
| 3270 |
+
"learning_rate": 2.393792551202778e-06,
|
| 3271 |
+
"loss": 0.0144,
|
| 3272 |
+
"step": 4660
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 5.449241540256709,
|
| 3276 |
+
"grad_norm": 0.3934665620326996,
|
| 3277 |
+
"learning_rate": 2.2964548604209213e-06,
|
| 3278 |
+
"loss": 0.0159,
|
| 3279 |
+
"step": 4670
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 5.460910151691949,
|
| 3283 |
+
"grad_norm": 0.4623982906341553,
|
| 3284 |
+
"learning_rate": 2.2010909444064866e-06,
|
| 3285 |
+
"loss": 0.0161,
|
| 3286 |
+
"step": 4680
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 5.472578763127188,
|
| 3290 |
+
"grad_norm": 0.2888776957988739,
|
| 3291 |
+
"learning_rate": 2.107704748922662e-06,
|
| 3292 |
+
"loss": 0.0133,
|
| 3293 |
+
"step": 4690
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 5.484247374562427,
|
| 3297 |
+
"grad_norm": 0.3349481225013733,
|
| 3298 |
+
"learning_rate": 2.0163001379027157e-06,
|
| 3299 |
+
"loss": 0.0162,
|
| 3300 |
+
"step": 4700
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 5.495915985997666,
|
| 3304 |
+
"grad_norm": 0.1728476732969284,
|
| 3305 |
+
"learning_rate": 1.926880893290217e-06,
|
| 3306 |
+
"loss": 0.0107,
|
| 3307 |
+
"step": 4710
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 5.507584597432905,
|
| 3311 |
+
"grad_norm": 0.27100396156311035,
|
| 3312 |
+
"learning_rate": 1.8394507148824847e-06,
|
| 3313 |
+
"loss": 0.0143,
|
| 3314 |
+
"step": 4720
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 5.519253208868145,
|
| 3318 |
+
"grad_norm": 0.29534074664115906,
|
| 3319 |
+
"learning_rate": 1.7540132201775361e-06,
|
| 3320 |
+
"loss": 0.0114,
|
| 3321 |
+
"step": 4730
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 5.530921820303384,
|
| 3325 |
+
"grad_norm": 0.26686033606529236,
|
| 3326 |
+
"learning_rate": 1.6705719442244138e-06,
|
| 3327 |
+
"loss": 0.0119,
|
| 3328 |
+
"step": 4740
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 5.542590431738623,
|
| 3332 |
+
"grad_norm": 0.26302769780158997,
|
| 3333 |
+
"learning_rate": 1.5891303394768908e-06,
|
| 3334 |
+
"loss": 0.0136,
|
| 3335 |
+
"step": 4750
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"epoch": 5.554259043173863,
|
| 3339 |
+
"grad_norm": 0.1138404905796051,
|
| 3340 |
+
"learning_rate": 1.5096917756506734e-06,
|
| 3341 |
+
"loss": 0.0099,
|
| 3342 |
+
"step": 4760
|
| 3343 |
+
},
|
| 3344 |
+
{
|
| 3345 |
+
"epoch": 5.565927654609101,
|
| 3346 |
+
"grad_norm": 0.22708551585674286,
|
| 3347 |
+
"learning_rate": 1.4322595395839133e-06,
|
| 3348 |
+
"loss": 0.0162,
|
| 3349 |
+
"step": 4770
|
| 3350 |
+
},
|
| 3351 |
+
{
|
| 3352 |
+
"epoch": 5.577596266044341,
|
| 3353 |
+
"grad_norm": 0.460024893283844,
|
| 3354 |
+
"learning_rate": 1.3568368351012717e-06,
|
| 3355 |
+
"loss": 0.0161,
|
| 3356 |
+
"step": 4780
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 5.58926487747958,
|
| 3360 |
+
"grad_norm": 0.32277268171310425,
|
| 3361 |
+
"learning_rate": 1.2834267828813084e-06,
|
| 3362 |
+
"loss": 0.0143,
|
| 3363 |
+
"step": 4790
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 5.600933488914819,
|
| 3367 |
+
"grad_norm": 0.33768701553344727,
|
| 3368 |
+
"learning_rate": 1.212032420327397e-06,
|
| 3369 |
+
"loss": 0.0122,
|
| 3370 |
+
"step": 4800
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 5.612602100350058,
|
| 3374 |
+
"grad_norm": 0.25165292620658875,
|
| 3375 |
+
"learning_rate": 1.1426567014420297e-06,
|
| 3376 |
+
"loss": 0.0112,
|
| 3377 |
+
"step": 4810
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 5.6242707117852975,
|
| 3381 |
+
"grad_norm": 0.2219328135251999,
|
| 3382 |
+
"learning_rate": 1.0753024967045945e-06,
|
| 3383 |
+
"loss": 0.0137,
|
| 3384 |
+
"step": 4820
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 5.635939323220537,
|
| 3388 |
+
"grad_norm": 0.3845769762992859,
|
| 3389 |
+
"learning_rate": 1.0099725929526348e-06,
|
| 3390 |
+
"loss": 0.0156,
|
| 3391 |
+
"step": 4830
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 5.647607934655776,
|
| 3395 |
+
"grad_norm": 0.3202395737171173,
|
| 3396 |
+
"learning_rate": 9.466696932664998e-07,
|
| 3397 |
+
"loss": 0.0102,
|
| 3398 |
+
"step": 4840
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 5.659276546091015,
|
| 3402 |
+
"grad_norm": 0.21468676626682281,
|
| 3403 |
+
"learning_rate": 8.853964168575146e-07,
|
| 3404 |
+
"loss": 0.013,
|
| 3405 |
+
"step": 4850
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 5.670945157526255,
|
| 3409 |
+
"grad_norm": 0.30721360445022583,
|
| 3410 |
+
"learning_rate": 8.2615529895963e-07,
|
| 3411 |
+
"loss": 0.0142,
|
| 3412 |
+
"step": 4860
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 5.6826137689614935,
|
| 3416 |
+
"grad_norm": 0.3136475086212158,
|
| 3417 |
+
"learning_rate": 7.689487907244941e-07,
|
| 3418 |
+
"loss": 0.0113,
|
| 3419 |
+
"step": 4870
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 5.694282380396733,
|
| 3423 |
+
"grad_norm": 0.31257662177085876,
|
| 3424 |
+
"learning_rate": 7.137792591200676e-07,
|
| 3425 |
+
"loss": 0.0135,
|
| 3426 |
+
"step": 4880
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 5.705950991831972,
|
| 3430 |
+
"grad_norm": 0.33189886808395386,
|
| 3431 |
+
"learning_rate": 6.606489868326571e-07,
|
| 3432 |
+
"loss": 0.015,
|
| 3433 |
+
"step": 4890
|
| 3434 |
+
},
|
| 3435 |
+
{
|
| 3436 |
+
"epoch": 5.717619603267211,
|
| 3437 |
+
"grad_norm": 0.27678588032722473,
|
| 3438 |
+
"learning_rate": 6.095601721724852e-07,
|
| 3439 |
+
"loss": 0.0139,
|
| 3440 |
+
"step": 4900
|
| 3441 |
+
},
|
| 3442 |
+
{
|
| 3443 |
+
"epoch": 5.72928821470245,
|
| 3444 |
+
"grad_norm": 0.43014249205589294,
|
| 3445 |
+
"learning_rate": 5.605149289827194e-07,
|
| 3446 |
+
"loss": 0.0119,
|
| 3447 |
+
"step": 4910
|
| 3448 |
+
},
|
| 3449 |
+
{
|
| 3450 |
+
"epoch": 5.7409568261376895,
|
| 3451 |
+
"grad_norm": 0.1929558962583542,
|
| 3452 |
+
"learning_rate": 5.135152865520298e-07,
|
| 3453 |
+
"loss": 0.0136,
|
| 3454 |
+
"step": 4920
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 5.752625437572929,
|
| 3458 |
+
"grad_norm": 0.2798265218734741,
|
| 3459 |
+
"learning_rate": 4.685631895306131e-07,
|
| 3460 |
+
"loss": 0.0097,
|
| 3461 |
+
"step": 4930
|
| 3462 |
+
},
|
| 3463 |
+
{
|
| 3464 |
+
"epoch": 5.764294049008168,
|
| 3465 |
+
"grad_norm": 0.29584571719169617,
|
| 3466 |
+
"learning_rate": 4.256604978497225e-07,
|
| 3467 |
+
"loss": 0.0107,
|
| 3468 |
+
"step": 4940
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"epoch": 5.775962660443407,
|
| 3472 |
+
"grad_norm": 0.29433825612068176,
|
| 3473 |
+
"learning_rate": 3.848089866447413e-07,
|
| 3474 |
+
"loss": 0.0164,
|
| 3475 |
+
"step": 4950
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 5.787631271878647,
|
| 3479 |
+
"grad_norm": 0.24111002683639526,
|
| 3480 |
+
"learning_rate": 3.4601034618169105e-07,
|
| 3481 |
+
"loss": 0.0136,
|
| 3482 |
+
"step": 4960
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 5.799299883313886,
|
| 3486 |
+
"grad_norm": 0.236686110496521,
|
| 3487 |
+
"learning_rate": 3.0926618178733214e-07,
|
| 3488 |
+
"loss": 0.0133,
|
| 3489 |
+
"step": 4970
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 5.810968494749125,
|
| 3493 |
+
"grad_norm": 0.3044494390487671,
|
| 3494 |
+
"learning_rate": 2.745780137827336e-07,
|
| 3495 |
+
"loss": 0.0131,
|
| 3496 |
+
"step": 4980
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 5.822637106184364,
|
| 3500 |
+
"grad_norm": 0.2573210299015045,
|
| 3501 |
+
"learning_rate": 2.419472774203568e-07,
|
| 3502 |
+
"loss": 0.0136,
|
| 3503 |
+
"step": 4990
|
| 3504 |
+
},
|
| 3505 |
+
{
|
| 3506 |
+
"epoch": 5.834305717619603,
|
| 3507 |
+
"grad_norm": 0.347267746925354,
|
| 3508 |
+
"learning_rate": 2.1137532282469174e-07,
|
| 3509 |
+
"loss": 0.0147,
|
| 3510 |
+
"step": 5000
|
| 3511 |
+
},
|
| 3512 |
+
{
|
| 3513 |
+
"epoch": 5.845974329054842,
|
| 3514 |
+
"grad_norm": 0.26792603731155396,
|
| 3515 |
+
"learning_rate": 1.828634149363573e-07,
|
| 3516 |
+
"loss": 0.0127,
|
| 3517 |
+
"step": 5010
|
| 3518 |
+
},
|
| 3519 |
+
{
|
| 3520 |
+
"epoch": 5.857642940490082,
|
| 3521 |
+
"grad_norm": 0.31212228536605835,
|
| 3522 |
+
"learning_rate": 1.5641273345980977e-07,
|
| 3523 |
+
"loss": 0.0122,
|
| 3524 |
+
"step": 5020
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 5.869311551925321,
|
| 3528 |
+
"grad_norm": 0.2252100259065628,
|
| 3529 |
+
"learning_rate": 1.3202437281449321e-07,
|
| 3530 |
+
"loss": 0.0113,
|
| 3531 |
+
"step": 5030
|
| 3532 |
+
},
|
| 3533 |
+
{
|
| 3534 |
+
"epoch": 5.88098016336056,
|
| 3535 |
+
"grad_norm": 0.29899704456329346,
|
| 3536 |
+
"learning_rate": 1.0969934208958643e-07,
|
| 3537 |
+
"loss": 0.0142,
|
| 3538 |
+
"step": 5040
|
| 3539 |
+
},
|
| 3540 |
+
{
|
| 3541 |
+
"epoch": 5.892648774795799,
|
| 3542 |
+
"grad_norm": 0.30285921692848206,
|
| 3543 |
+
"learning_rate": 8.943856500223113e-08,
|
| 3544 |
+
"loss": 0.015,
|
| 3545 |
+
"step": 5050
|
| 3546 |
+
},
|
| 3547 |
+
{
|
| 3548 |
+
"epoch": 5.904317386231039,
|
| 3549 |
+
"grad_norm": 0.30103516578674316,
|
| 3550 |
+
"learning_rate": 7.124287985931233e-08,
|
| 3551 |
+
"loss": 0.0117,
|
| 3552 |
+
"step": 5060
|
| 3553 |
+
},
|
| 3554 |
+
{
|
| 3555 |
+
"epoch": 5.915985997666278,
|
| 3556 |
+
"grad_norm": 0.21845895051956177,
|
| 3557 |
+
"learning_rate": 5.511303952278612e-08,
|
| 3558 |
+
"loss": 0.0132,
|
| 3559 |
+
"step": 5070
|
| 3560 |
+
},
|
| 3561 |
+
{
|
| 3562 |
+
"epoch": 5.927654609101517,
|
| 3563 |
+
"grad_norm": 0.4260839521884918,
|
| 3564 |
+
"learning_rate": 4.104971137851577e-08,
|
| 3565 |
+
"loss": 0.0121,
|
| 3566 |
+
"step": 5080
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 5.939323220536756,
|
| 3570 |
+
"grad_norm": 0.26336410641670227,
|
| 3571 |
+
"learning_rate": 2.9053477308677022e-08,
|
| 3572 |
+
"loss": 0.0105,
|
| 3573 |
+
"step": 5090
|
| 3574 |
+
},
|
| 3575 |
+
{
|
| 3576 |
+
"epoch": 5.950991831971995,
|
| 3577 |
+
"grad_norm": 0.3341667950153351,
|
| 3578 |
+
"learning_rate": 1.9124833667655272e-08,
|
| 3579 |
+
"loss": 0.0128,
|
| 3580 |
+
"step": 5100
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"epoch": 5.962660443407234,
|
| 3584 |
+
"grad_norm": 0.2763518691062927,
|
| 3585 |
+
"learning_rate": 1.1264191261528557e-08,
|
| 3586 |
+
"loss": 0.0175,
|
| 3587 |
+
"step": 5110
|
| 3588 |
+
},
|
| 3589 |
+
{
|
| 3590 |
+
"epoch": 5.974329054842474,
|
| 3591 |
+
"grad_norm": 0.30684781074523926,
|
| 3592 |
+
"learning_rate": 5.471875331070075e-09,
|
| 3593 |
+
"loss": 0.0171,
|
| 3594 |
+
"step": 5120
|
| 3595 |
+
},
|
| 3596 |
+
{
|
| 3597 |
+
"epoch": 5.985997666277713,
|
| 3598 |
+
"grad_norm": 0.24833644926548004,
|
| 3599 |
+
"learning_rate": 1.7481255382756355e-09,
|
| 3600 |
+
"loss": 0.0119,
|
| 3601 |
+
"step": 5130
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"epoch": 5.997666277712952,
|
| 3605 |
+
"grad_norm": 0.36518436670303345,
|
| 3606 |
+
"learning_rate": 9.309595645490277e-11,
|
| 3607 |
+
"loss": 0.0099,
|
| 3608 |
+
"step": 5140
|
| 3609 |
+
},
|
| 3610 |
+
{
|
| 3611 |
+
"epoch": 6.0,
|
| 3612 |
+
"step": 5142,
|
| 3613 |
+
"total_flos": 0.0,
|
| 3614 |
+
"train_loss": 0.03185599890419107,
|
| 3615 |
+
"train_runtime": 5146.2418,
|
| 3616 |
+
"train_samples_per_second": 48.949,
|
| 3617 |
+
"train_steps_per_second": 0.999
|
| 3618 |
+
}
|
| 3619 |
+
],
|
| 3620 |
+
"logging_steps": 10,
|
| 3621 |
+
"max_steps": 5142,
|
| 3622 |
+
"num_input_tokens_seen": 0,
|
| 3623 |
+
"num_train_epochs": 6,
|
| 3624 |
+
"save_steps": 3000,
|
| 3625 |
+
"stateful_callbacks": {
|
| 3626 |
+
"TrainerControl": {
|
| 3627 |
+
"args": {
|
| 3628 |
+
"should_epoch_stop": false,
|
| 3629 |
+
"should_evaluate": false,
|
| 3630 |
+
"should_log": false,
|
| 3631 |
+
"should_save": true,
|
| 3632 |
+
"should_training_stop": true
|
| 3633 |
+
},
|
| 3634 |
+
"attributes": {}
|
| 3635 |
+
}
|
| 3636 |
+
},
|
| 3637 |
+
"total_flos": 0.0,
|
| 3638 |
+
"train_batch_size": 49,
|
| 3639 |
+
"trial_name": null,
|
| 3640 |
+
"trial_params": null
|
| 3641 |
+
}
|