Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +4186 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,4186 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 5000,
|
| 6 |
+
"global_step": 34113,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 1.997068566235746e-05,
|
| 14 |
+
"loss": 0.5682,
|
| 15 |
+
"step": 50
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 1.994137132471492e-05,
|
| 20 |
+
"loss": 0.4713,
|
| 21 |
+
"step": 100
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 1.9912056987072378e-05,
|
| 26 |
+
"loss": 0.4706,
|
| 27 |
+
"step": 150
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.02,
|
| 31 |
+
"learning_rate": 1.9882742649429837e-05,
|
| 32 |
+
"loss": 0.4316,
|
| 33 |
+
"step": 200
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
+
"learning_rate": 1.9853428311787297e-05,
|
| 38 |
+
"loss": 0.4492,
|
| 39 |
+
"step": 250
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 1.9824113974144757e-05,
|
| 44 |
+
"loss": 0.4548,
|
| 45 |
+
"step": 300
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"learning_rate": 1.9794799636502217e-05,
|
| 50 |
+
"loss": 0.4145,
|
| 51 |
+
"step": 350
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04,
|
| 55 |
+
"learning_rate": 1.9765485298859673e-05,
|
| 56 |
+
"loss": 0.4183,
|
| 57 |
+
"step": 400
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.04,
|
| 61 |
+
"learning_rate": 1.9736170961217133e-05,
|
| 62 |
+
"loss": 0.4257,
|
| 63 |
+
"step": 450
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.04,
|
| 67 |
+
"learning_rate": 1.9706856623574592e-05,
|
| 68 |
+
"loss": 0.4009,
|
| 69 |
+
"step": 500
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05,
|
| 73 |
+
"learning_rate": 1.9677542285932052e-05,
|
| 74 |
+
"loss": 0.398,
|
| 75 |
+
"step": 550
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.05,
|
| 79 |
+
"learning_rate": 1.9648227948289512e-05,
|
| 80 |
+
"loss": 0.4313,
|
| 81 |
+
"step": 600
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.06,
|
| 85 |
+
"learning_rate": 1.961891361064697e-05,
|
| 86 |
+
"loss": 0.4121,
|
| 87 |
+
"step": 650
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.06,
|
| 91 |
+
"learning_rate": 1.9589599273004428e-05,
|
| 92 |
+
"loss": 0.3895,
|
| 93 |
+
"step": 700
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 1.9560284935361888e-05,
|
| 98 |
+
"loss": 0.4008,
|
| 99 |
+
"step": 750
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.07,
|
| 103 |
+
"learning_rate": 1.9530970597719344e-05,
|
| 104 |
+
"loss": 0.4141,
|
| 105 |
+
"step": 800
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.07,
|
| 109 |
+
"learning_rate": 1.9501656260076804e-05,
|
| 110 |
+
"loss": 0.3785,
|
| 111 |
+
"step": 850
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.08,
|
| 115 |
+
"learning_rate": 1.9472341922434264e-05,
|
| 116 |
+
"loss": 0.3999,
|
| 117 |
+
"step": 900
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.08,
|
| 121 |
+
"learning_rate": 1.9443027584791724e-05,
|
| 122 |
+
"loss": 0.4068,
|
| 123 |
+
"step": 950
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.09,
|
| 127 |
+
"learning_rate": 1.9413713247149183e-05,
|
| 128 |
+
"loss": 0.3879,
|
| 129 |
+
"step": 1000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.09,
|
| 133 |
+
"learning_rate": 1.938439890950664e-05,
|
| 134 |
+
"loss": 0.3935,
|
| 135 |
+
"step": 1050
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 1.93550845718641e-05,
|
| 140 |
+
"loss": 0.4013,
|
| 141 |
+
"step": 1100
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1,
|
| 145 |
+
"learning_rate": 1.932577023422156e-05,
|
| 146 |
+
"loss": 0.3796,
|
| 147 |
+
"step": 1150
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.11,
|
| 151 |
+
"learning_rate": 1.929645589657902e-05,
|
| 152 |
+
"loss": 0.3819,
|
| 153 |
+
"step": 1200
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.11,
|
| 157 |
+
"learning_rate": 1.926714155893648e-05,
|
| 158 |
+
"loss": 0.386,
|
| 159 |
+
"step": 1250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.11,
|
| 163 |
+
"learning_rate": 1.9237827221293935e-05,
|
| 164 |
+
"loss": 0.3602,
|
| 165 |
+
"step": 1300
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.12,
|
| 169 |
+
"learning_rate": 1.9208512883651395e-05,
|
| 170 |
+
"loss": 0.3966,
|
| 171 |
+
"step": 1350
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.12,
|
| 175 |
+
"learning_rate": 1.9179198546008855e-05,
|
| 176 |
+
"loss": 0.3806,
|
| 177 |
+
"step": 1400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.13,
|
| 181 |
+
"learning_rate": 1.9149884208366314e-05,
|
| 182 |
+
"loss": 0.3692,
|
| 183 |
+
"step": 1450
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.13,
|
| 187 |
+
"learning_rate": 1.9120569870723774e-05,
|
| 188 |
+
"loss": 0.3725,
|
| 189 |
+
"step": 1500
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.14,
|
| 193 |
+
"learning_rate": 1.909125553308123e-05,
|
| 194 |
+
"loss": 0.375,
|
| 195 |
+
"step": 1550
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.14,
|
| 199 |
+
"learning_rate": 1.906194119543869e-05,
|
| 200 |
+
"loss": 0.3542,
|
| 201 |
+
"step": 1600
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.15,
|
| 205 |
+
"learning_rate": 1.903262685779615e-05,
|
| 206 |
+
"loss": 0.3608,
|
| 207 |
+
"step": 1650
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.15,
|
| 211 |
+
"learning_rate": 1.900331252015361e-05,
|
| 212 |
+
"loss": 0.3744,
|
| 213 |
+
"step": 1700
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.15,
|
| 217 |
+
"learning_rate": 1.897399818251107e-05,
|
| 218 |
+
"loss": 0.3966,
|
| 219 |
+
"step": 1750
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.16,
|
| 223 |
+
"learning_rate": 1.894468384486853e-05,
|
| 224 |
+
"loss": 0.3351,
|
| 225 |
+
"step": 1800
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.16,
|
| 229 |
+
"learning_rate": 1.8915369507225986e-05,
|
| 230 |
+
"loss": 0.3449,
|
| 231 |
+
"step": 1850
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.17,
|
| 235 |
+
"learning_rate": 1.8886055169583445e-05,
|
| 236 |
+
"loss": 0.3885,
|
| 237 |
+
"step": 1900
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.17,
|
| 241 |
+
"learning_rate": 1.8856740831940902e-05,
|
| 242 |
+
"loss": 0.3905,
|
| 243 |
+
"step": 1950
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.18,
|
| 247 |
+
"learning_rate": 1.882742649429836e-05,
|
| 248 |
+
"loss": 0.3613,
|
| 249 |
+
"step": 2000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.18,
|
| 253 |
+
"learning_rate": 1.879811215665582e-05,
|
| 254 |
+
"loss": 0.3666,
|
| 255 |
+
"step": 2050
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.18,
|
| 259 |
+
"learning_rate": 1.876879781901328e-05,
|
| 260 |
+
"loss": 0.359,
|
| 261 |
+
"step": 2100
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.19,
|
| 265 |
+
"learning_rate": 1.873948348137074e-05,
|
| 266 |
+
"loss": 0.3636,
|
| 267 |
+
"step": 2150
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.19,
|
| 271 |
+
"learning_rate": 1.8710169143728197e-05,
|
| 272 |
+
"loss": 0.3625,
|
| 273 |
+
"step": 2200
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.2,
|
| 277 |
+
"learning_rate": 1.8680854806085657e-05,
|
| 278 |
+
"loss": 0.3506,
|
| 279 |
+
"step": 2250
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.2,
|
| 283 |
+
"learning_rate": 1.8651540468443117e-05,
|
| 284 |
+
"loss": 0.3658,
|
| 285 |
+
"step": 2300
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.21,
|
| 289 |
+
"learning_rate": 1.8622226130800576e-05,
|
| 290 |
+
"loss": 0.3739,
|
| 291 |
+
"step": 2350
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.21,
|
| 295 |
+
"learning_rate": 1.8592911793158036e-05,
|
| 296 |
+
"loss": 0.3508,
|
| 297 |
+
"step": 2400
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.22,
|
| 301 |
+
"learning_rate": 1.8563597455515493e-05,
|
| 302 |
+
"loss": 0.3345,
|
| 303 |
+
"step": 2450
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.22,
|
| 307 |
+
"learning_rate": 1.8534283117872952e-05,
|
| 308 |
+
"loss": 0.3523,
|
| 309 |
+
"step": 2500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.22,
|
| 313 |
+
"learning_rate": 1.8504968780230412e-05,
|
| 314 |
+
"loss": 0.3412,
|
| 315 |
+
"step": 2550
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.23,
|
| 319 |
+
"learning_rate": 1.8475654442587872e-05,
|
| 320 |
+
"loss": 0.356,
|
| 321 |
+
"step": 2600
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.23,
|
| 325 |
+
"learning_rate": 1.844634010494533e-05,
|
| 326 |
+
"loss": 0.3635,
|
| 327 |
+
"step": 2650
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.24,
|
| 331 |
+
"learning_rate": 1.8417025767302788e-05,
|
| 332 |
+
"loss": 0.3507,
|
| 333 |
+
"step": 2700
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.24,
|
| 337 |
+
"learning_rate": 1.8387711429660248e-05,
|
| 338 |
+
"loss": 0.3429,
|
| 339 |
+
"step": 2750
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.25,
|
| 343 |
+
"learning_rate": 1.8358397092017708e-05,
|
| 344 |
+
"loss": 0.3352,
|
| 345 |
+
"step": 2800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.25,
|
| 349 |
+
"learning_rate": 1.8329082754375167e-05,
|
| 350 |
+
"loss": 0.3754,
|
| 351 |
+
"step": 2850
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.26,
|
| 355 |
+
"learning_rate": 1.8299768416732627e-05,
|
| 356 |
+
"loss": 0.3536,
|
| 357 |
+
"step": 2900
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.26,
|
| 361 |
+
"learning_rate": 1.8270454079090087e-05,
|
| 362 |
+
"loss": 0.3583,
|
| 363 |
+
"step": 2950
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.26,
|
| 367 |
+
"learning_rate": 1.8241139741447543e-05,
|
| 368 |
+
"loss": 0.3687,
|
| 369 |
+
"step": 3000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.27,
|
| 373 |
+
"learning_rate": 1.8211825403805003e-05,
|
| 374 |
+
"loss": 0.3456,
|
| 375 |
+
"step": 3050
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.27,
|
| 379 |
+
"learning_rate": 1.8182511066162463e-05,
|
| 380 |
+
"loss": 0.3417,
|
| 381 |
+
"step": 3100
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.28,
|
| 385 |
+
"learning_rate": 1.815319672851992e-05,
|
| 386 |
+
"loss": 0.3271,
|
| 387 |
+
"step": 3150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.28,
|
| 391 |
+
"learning_rate": 1.812388239087738e-05,
|
| 392 |
+
"loss": 0.3445,
|
| 393 |
+
"step": 3200
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.29,
|
| 397 |
+
"learning_rate": 1.809456805323484e-05,
|
| 398 |
+
"loss": 0.353,
|
| 399 |
+
"step": 3250
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.29,
|
| 403 |
+
"learning_rate": 1.80652537155923e-05,
|
| 404 |
+
"loss": 0.3432,
|
| 405 |
+
"step": 3300
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.29,
|
| 409 |
+
"learning_rate": 1.8035939377949755e-05,
|
| 410 |
+
"loss": 0.3309,
|
| 411 |
+
"step": 3350
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.3,
|
| 415 |
+
"learning_rate": 1.8006625040307214e-05,
|
| 416 |
+
"loss": 0.3343,
|
| 417 |
+
"step": 3400
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.3,
|
| 421 |
+
"learning_rate": 1.7977310702664674e-05,
|
| 422 |
+
"loss": 0.3458,
|
| 423 |
+
"step": 3450
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.31,
|
| 427 |
+
"learning_rate": 1.7947996365022134e-05,
|
| 428 |
+
"loss": 0.3306,
|
| 429 |
+
"step": 3500
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.31,
|
| 433 |
+
"learning_rate": 1.7918682027379594e-05,
|
| 434 |
+
"loss": 0.3384,
|
| 435 |
+
"step": 3550
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.32,
|
| 439 |
+
"learning_rate": 1.788936768973705e-05,
|
| 440 |
+
"loss": 0.3392,
|
| 441 |
+
"step": 3600
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.32,
|
| 445 |
+
"learning_rate": 1.786005335209451e-05,
|
| 446 |
+
"loss": 0.3407,
|
| 447 |
+
"step": 3650
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.33,
|
| 451 |
+
"learning_rate": 1.783073901445197e-05,
|
| 452 |
+
"loss": 0.3594,
|
| 453 |
+
"step": 3700
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.33,
|
| 457 |
+
"learning_rate": 1.780142467680943e-05,
|
| 458 |
+
"loss": 0.3516,
|
| 459 |
+
"step": 3750
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.33,
|
| 463 |
+
"learning_rate": 1.777211033916689e-05,
|
| 464 |
+
"loss": 0.3084,
|
| 465 |
+
"step": 3800
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.34,
|
| 469 |
+
"learning_rate": 1.7742796001524346e-05,
|
| 470 |
+
"loss": 0.3271,
|
| 471 |
+
"step": 3850
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.34,
|
| 475 |
+
"learning_rate": 1.7713481663881805e-05,
|
| 476 |
+
"loss": 0.3462,
|
| 477 |
+
"step": 3900
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.35,
|
| 481 |
+
"learning_rate": 1.7684167326239265e-05,
|
| 482 |
+
"loss": 0.3449,
|
| 483 |
+
"step": 3950
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.35,
|
| 487 |
+
"learning_rate": 1.7654852988596725e-05,
|
| 488 |
+
"loss": 0.3553,
|
| 489 |
+
"step": 4000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.36,
|
| 493 |
+
"learning_rate": 1.7625538650954185e-05,
|
| 494 |
+
"loss": 0.3522,
|
| 495 |
+
"step": 4050
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.36,
|
| 499 |
+
"learning_rate": 1.7596224313311644e-05,
|
| 500 |
+
"loss": 0.3519,
|
| 501 |
+
"step": 4100
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.36,
|
| 505 |
+
"learning_rate": 1.75669099756691e-05,
|
| 506 |
+
"loss": 0.3216,
|
| 507 |
+
"step": 4150
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.37,
|
| 511 |
+
"learning_rate": 1.753759563802656e-05,
|
| 512 |
+
"loss": 0.3366,
|
| 513 |
+
"step": 4200
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.37,
|
| 517 |
+
"learning_rate": 1.750828130038402e-05,
|
| 518 |
+
"loss": 0.3783,
|
| 519 |
+
"step": 4250
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.38,
|
| 523 |
+
"learning_rate": 1.747896696274148e-05,
|
| 524 |
+
"loss": 0.3379,
|
| 525 |
+
"step": 4300
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.38,
|
| 529 |
+
"learning_rate": 1.744965262509894e-05,
|
| 530 |
+
"loss": 0.3351,
|
| 531 |
+
"step": 4350
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.39,
|
| 535 |
+
"learning_rate": 1.7420338287456396e-05,
|
| 536 |
+
"loss": 0.3364,
|
| 537 |
+
"step": 4400
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.39,
|
| 541 |
+
"learning_rate": 1.7391023949813856e-05,
|
| 542 |
+
"loss": 0.3466,
|
| 543 |
+
"step": 4450
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.4,
|
| 547 |
+
"learning_rate": 1.7361709612171312e-05,
|
| 548 |
+
"loss": 0.3491,
|
| 549 |
+
"step": 4500
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.4,
|
| 553 |
+
"learning_rate": 1.7332395274528772e-05,
|
| 554 |
+
"loss": 0.2991,
|
| 555 |
+
"step": 4550
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.4,
|
| 559 |
+
"learning_rate": 1.7303080936886232e-05,
|
| 560 |
+
"loss": 0.3177,
|
| 561 |
+
"step": 4600
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.41,
|
| 565 |
+
"learning_rate": 1.727376659924369e-05,
|
| 566 |
+
"loss": 0.343,
|
| 567 |
+
"step": 4650
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.41,
|
| 571 |
+
"learning_rate": 1.724445226160115e-05,
|
| 572 |
+
"loss": 0.3361,
|
| 573 |
+
"step": 4700
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.42,
|
| 577 |
+
"learning_rate": 1.7215137923958608e-05,
|
| 578 |
+
"loss": 0.3444,
|
| 579 |
+
"step": 4750
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.42,
|
| 583 |
+
"learning_rate": 1.7185823586316067e-05,
|
| 584 |
+
"loss": 0.3498,
|
| 585 |
+
"step": 4800
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.43,
|
| 589 |
+
"learning_rate": 1.7156509248673527e-05,
|
| 590 |
+
"loss": 0.3397,
|
| 591 |
+
"step": 4850
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.43,
|
| 595 |
+
"learning_rate": 1.7127194911030987e-05,
|
| 596 |
+
"loss": 0.3221,
|
| 597 |
+
"step": 4900
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.44,
|
| 601 |
+
"learning_rate": 1.7097880573388447e-05,
|
| 602 |
+
"loss": 0.3193,
|
| 603 |
+
"step": 4950
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.44,
|
| 607 |
+
"learning_rate": 1.7068566235745903e-05,
|
| 608 |
+
"loss": 0.3417,
|
| 609 |
+
"step": 5000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.44,
|
| 613 |
+
"eval_accuracy": 0.8577541429631462,
|
| 614 |
+
"eval_combined_score": 0.831721954408897,
|
| 615 |
+
"eval_f1": 0.8056897658546476,
|
| 616 |
+
"eval_loss": 0.31978192925453186,
|
| 617 |
+
"eval_runtime": 51.8428,
|
| 618 |
+
"eval_samples_per_second": 779.857,
|
| 619 |
+
"eval_steps_per_second": 48.743,
|
| 620 |
+
"step": 5000
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.44,
|
| 624 |
+
"learning_rate": 1.7039251898103363e-05,
|
| 625 |
+
"loss": 0.3475,
|
| 626 |
+
"step": 5050
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.45,
|
| 630 |
+
"learning_rate": 1.7009937560460823e-05,
|
| 631 |
+
"loss": 0.3044,
|
| 632 |
+
"step": 5100
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.45,
|
| 636 |
+
"learning_rate": 1.6980623222818282e-05,
|
| 637 |
+
"loss": 0.3138,
|
| 638 |
+
"step": 5150
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.46,
|
| 642 |
+
"learning_rate": 1.6951308885175742e-05,
|
| 643 |
+
"loss": 0.336,
|
| 644 |
+
"step": 5200
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.46,
|
| 648 |
+
"learning_rate": 1.6921994547533202e-05,
|
| 649 |
+
"loss": 0.3363,
|
| 650 |
+
"step": 5250
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.47,
|
| 654 |
+
"learning_rate": 1.6892680209890658e-05,
|
| 655 |
+
"loss": 0.3189,
|
| 656 |
+
"step": 5300
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.47,
|
| 660 |
+
"learning_rate": 1.6863365872248118e-05,
|
| 661 |
+
"loss": 0.344,
|
| 662 |
+
"step": 5350
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.47,
|
| 666 |
+
"learning_rate": 1.6834051534605578e-05,
|
| 667 |
+
"loss": 0.3346,
|
| 668 |
+
"step": 5400
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.48,
|
| 672 |
+
"learning_rate": 1.6804737196963038e-05,
|
| 673 |
+
"loss": 0.33,
|
| 674 |
+
"step": 5450
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.48,
|
| 678 |
+
"learning_rate": 1.6775422859320497e-05,
|
| 679 |
+
"loss": 0.3098,
|
| 680 |
+
"step": 5500
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.49,
|
| 684 |
+
"learning_rate": 1.6746108521677954e-05,
|
| 685 |
+
"loss": 0.3385,
|
| 686 |
+
"step": 5550
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.49,
|
| 690 |
+
"learning_rate": 1.6716794184035413e-05,
|
| 691 |
+
"loss": 0.3568,
|
| 692 |
+
"step": 5600
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.5,
|
| 696 |
+
"learning_rate": 1.668747984639287e-05,
|
| 697 |
+
"loss": 0.3464,
|
| 698 |
+
"step": 5650
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.5,
|
| 702 |
+
"learning_rate": 1.665816550875033e-05,
|
| 703 |
+
"loss": 0.3276,
|
| 704 |
+
"step": 5700
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.51,
|
| 708 |
+
"learning_rate": 1.662885117110779e-05,
|
| 709 |
+
"loss": 0.3316,
|
| 710 |
+
"step": 5750
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.51,
|
| 714 |
+
"learning_rate": 1.659953683346525e-05,
|
| 715 |
+
"loss": 0.3165,
|
| 716 |
+
"step": 5800
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.51,
|
| 720 |
+
"learning_rate": 1.657022249582271e-05,
|
| 721 |
+
"loss": 0.327,
|
| 722 |
+
"step": 5850
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.52,
|
| 726 |
+
"learning_rate": 1.6540908158180165e-05,
|
| 727 |
+
"loss": 0.3476,
|
| 728 |
+
"step": 5900
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.52,
|
| 732 |
+
"learning_rate": 1.6511593820537625e-05,
|
| 733 |
+
"loss": 0.3278,
|
| 734 |
+
"step": 5950
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.53,
|
| 738 |
+
"learning_rate": 1.6482279482895085e-05,
|
| 739 |
+
"loss": 0.3286,
|
| 740 |
+
"step": 6000
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.53,
|
| 744 |
+
"learning_rate": 1.6452965145252545e-05,
|
| 745 |
+
"loss": 0.3077,
|
| 746 |
+
"step": 6050
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.54,
|
| 750 |
+
"learning_rate": 1.6423650807610004e-05,
|
| 751 |
+
"loss": 0.3257,
|
| 752 |
+
"step": 6100
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.54,
|
| 756 |
+
"learning_rate": 1.639433646996746e-05,
|
| 757 |
+
"loss": 0.3239,
|
| 758 |
+
"step": 6150
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.55,
|
| 762 |
+
"learning_rate": 1.636502213232492e-05,
|
| 763 |
+
"loss": 0.337,
|
| 764 |
+
"step": 6200
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.55,
|
| 768 |
+
"learning_rate": 1.633570779468238e-05,
|
| 769 |
+
"loss": 0.3101,
|
| 770 |
+
"step": 6250
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.55,
|
| 774 |
+
"learning_rate": 1.630639345703984e-05,
|
| 775 |
+
"loss": 0.2938,
|
| 776 |
+
"step": 6300
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.56,
|
| 780 |
+
"learning_rate": 1.62770791193973e-05,
|
| 781 |
+
"loss": 0.2929,
|
| 782 |
+
"step": 6350
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.56,
|
| 786 |
+
"learning_rate": 1.624776478175476e-05,
|
| 787 |
+
"loss": 0.3085,
|
| 788 |
+
"step": 6400
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.57,
|
| 792 |
+
"learning_rate": 1.6218450444112216e-05,
|
| 793 |
+
"loss": 0.319,
|
| 794 |
+
"step": 6450
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.57,
|
| 798 |
+
"learning_rate": 1.6189136106469676e-05,
|
| 799 |
+
"loss": 0.3263,
|
| 800 |
+
"step": 6500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.58,
|
| 804 |
+
"learning_rate": 1.6159821768827135e-05,
|
| 805 |
+
"loss": 0.3221,
|
| 806 |
+
"step": 6550
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.58,
|
| 810 |
+
"learning_rate": 1.6130507431184595e-05,
|
| 811 |
+
"loss": 0.3094,
|
| 812 |
+
"step": 6600
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.58,
|
| 816 |
+
"learning_rate": 1.6101193093542055e-05,
|
| 817 |
+
"loss": 0.3159,
|
| 818 |
+
"step": 6650
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.59,
|
| 822 |
+
"learning_rate": 1.607187875589951e-05,
|
| 823 |
+
"loss": 0.3263,
|
| 824 |
+
"step": 6700
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.59,
|
| 828 |
+
"learning_rate": 1.604256441825697e-05,
|
| 829 |
+
"loss": 0.332,
|
| 830 |
+
"step": 6750
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.6,
|
| 834 |
+
"learning_rate": 1.601325008061443e-05,
|
| 835 |
+
"loss": 0.3239,
|
| 836 |
+
"step": 6800
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.6,
|
| 840 |
+
"learning_rate": 1.5983935742971887e-05,
|
| 841 |
+
"loss": 0.3319,
|
| 842 |
+
"step": 6850
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.61,
|
| 846 |
+
"learning_rate": 1.5954621405329347e-05,
|
| 847 |
+
"loss": 0.3023,
|
| 848 |
+
"step": 6900
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.61,
|
| 852 |
+
"learning_rate": 1.5925307067686807e-05,
|
| 853 |
+
"loss": 0.3214,
|
| 854 |
+
"step": 6950
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.62,
|
| 858 |
+
"learning_rate": 1.5895992730044266e-05,
|
| 859 |
+
"loss": 0.3129,
|
| 860 |
+
"step": 7000
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.62,
|
| 864 |
+
"learning_rate": 1.5866678392401723e-05,
|
| 865 |
+
"loss": 0.3239,
|
| 866 |
+
"step": 7050
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.62,
|
| 870 |
+
"learning_rate": 1.5837364054759183e-05,
|
| 871 |
+
"loss": 0.3133,
|
| 872 |
+
"step": 7100
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.63,
|
| 876 |
+
"learning_rate": 1.5808049717116642e-05,
|
| 877 |
+
"loss": 0.336,
|
| 878 |
+
"step": 7150
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.63,
|
| 882 |
+
"learning_rate": 1.5778735379474102e-05,
|
| 883 |
+
"loss": 0.326,
|
| 884 |
+
"step": 7200
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.64,
|
| 888 |
+
"learning_rate": 1.5749421041831562e-05,
|
| 889 |
+
"loss": 0.2881,
|
| 890 |
+
"step": 7250
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.64,
|
| 894 |
+
"learning_rate": 1.5720106704189018e-05,
|
| 895 |
+
"loss": 0.3103,
|
| 896 |
+
"step": 7300
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.65,
|
| 900 |
+
"learning_rate": 1.5690792366546478e-05,
|
| 901 |
+
"loss": 0.3039,
|
| 902 |
+
"step": 7350
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.65,
|
| 906 |
+
"learning_rate": 1.5661478028903938e-05,
|
| 907 |
+
"loss": 0.3034,
|
| 908 |
+
"step": 7400
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.66,
|
| 912 |
+
"learning_rate": 1.5632163691261397e-05,
|
| 913 |
+
"loss": 0.3353,
|
| 914 |
+
"step": 7450
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.66,
|
| 918 |
+
"learning_rate": 1.5602849353618857e-05,
|
| 919 |
+
"loss": 0.3276,
|
| 920 |
+
"step": 7500
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.66,
|
| 924 |
+
"learning_rate": 1.5573535015976317e-05,
|
| 925 |
+
"loss": 0.3217,
|
| 926 |
+
"step": 7550
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.67,
|
| 930 |
+
"learning_rate": 1.5544220678333773e-05,
|
| 931 |
+
"loss": 0.2949,
|
| 932 |
+
"step": 7600
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.67,
|
| 936 |
+
"learning_rate": 1.5514906340691233e-05,
|
| 937 |
+
"loss": 0.3221,
|
| 938 |
+
"step": 7650
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.68,
|
| 942 |
+
"learning_rate": 1.5485592003048693e-05,
|
| 943 |
+
"loss": 0.3359,
|
| 944 |
+
"step": 7700
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.68,
|
| 948 |
+
"learning_rate": 1.5456277665406153e-05,
|
| 949 |
+
"loss": 0.3073,
|
| 950 |
+
"step": 7750
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.69,
|
| 954 |
+
"learning_rate": 1.5426963327763612e-05,
|
| 955 |
+
"loss": 0.3026,
|
| 956 |
+
"step": 7800
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.69,
|
| 960 |
+
"learning_rate": 1.539764899012107e-05,
|
| 961 |
+
"loss": 0.3191,
|
| 962 |
+
"step": 7850
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.69,
|
| 966 |
+
"learning_rate": 1.536833465247853e-05,
|
| 967 |
+
"loss": 0.3274,
|
| 968 |
+
"step": 7900
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.7,
|
| 972 |
+
"learning_rate": 1.5339020314835988e-05,
|
| 973 |
+
"loss": 0.3191,
|
| 974 |
+
"step": 7950
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.7,
|
| 978 |
+
"learning_rate": 1.5309705977193448e-05,
|
| 979 |
+
"loss": 0.3064,
|
| 980 |
+
"step": 8000
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.71,
|
| 984 |
+
"learning_rate": 1.5280391639550908e-05,
|
| 985 |
+
"loss": 0.3145,
|
| 986 |
+
"step": 8050
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.71,
|
| 990 |
+
"learning_rate": 1.5251077301908364e-05,
|
| 991 |
+
"loss": 0.2941,
|
| 992 |
+
"step": 8100
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.72,
|
| 996 |
+
"learning_rate": 1.5221762964265822e-05,
|
| 997 |
+
"loss": 0.33,
|
| 998 |
+
"step": 8150
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.72,
|
| 1002 |
+
"learning_rate": 1.5192448626623282e-05,
|
| 1003 |
+
"loss": 0.3249,
|
| 1004 |
+
"step": 8200
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.73,
|
| 1008 |
+
"learning_rate": 1.516313428898074e-05,
|
| 1009 |
+
"loss": 0.2896,
|
| 1010 |
+
"step": 8250
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.73,
|
| 1014 |
+
"learning_rate": 1.51338199513382e-05,
|
| 1015 |
+
"loss": 0.3377,
|
| 1016 |
+
"step": 8300
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.73,
|
| 1020 |
+
"learning_rate": 1.510450561369566e-05,
|
| 1021 |
+
"loss": 0.2989,
|
| 1022 |
+
"step": 8350
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.74,
|
| 1026 |
+
"learning_rate": 1.5075191276053118e-05,
|
| 1027 |
+
"loss": 0.3102,
|
| 1028 |
+
"step": 8400
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.74,
|
| 1032 |
+
"learning_rate": 1.5045876938410577e-05,
|
| 1033 |
+
"loss": 0.2975,
|
| 1034 |
+
"step": 8450
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.75,
|
| 1038 |
+
"learning_rate": 1.5016562600768037e-05,
|
| 1039 |
+
"loss": 0.2968,
|
| 1040 |
+
"step": 8500
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.75,
|
| 1044 |
+
"learning_rate": 1.4987248263125495e-05,
|
| 1045 |
+
"loss": 0.2903,
|
| 1046 |
+
"step": 8550
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.76,
|
| 1050 |
+
"learning_rate": 1.4957933925482955e-05,
|
| 1051 |
+
"loss": 0.3167,
|
| 1052 |
+
"step": 8600
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.76,
|
| 1056 |
+
"learning_rate": 1.4928619587840413e-05,
|
| 1057 |
+
"loss": 0.3162,
|
| 1058 |
+
"step": 8650
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.77,
|
| 1062 |
+
"learning_rate": 1.4899305250197873e-05,
|
| 1063 |
+
"loss": 0.314,
|
| 1064 |
+
"step": 8700
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.77,
|
| 1068 |
+
"learning_rate": 1.4869990912555333e-05,
|
| 1069 |
+
"loss": 0.2955,
|
| 1070 |
+
"step": 8750
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.77,
|
| 1074 |
+
"learning_rate": 1.484067657491279e-05,
|
| 1075 |
+
"loss": 0.3077,
|
| 1076 |
+
"step": 8800
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.78,
|
| 1080 |
+
"learning_rate": 1.481136223727025e-05,
|
| 1081 |
+
"loss": 0.3056,
|
| 1082 |
+
"step": 8850
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.78,
|
| 1086 |
+
"learning_rate": 1.478204789962771e-05,
|
| 1087 |
+
"loss": 0.3262,
|
| 1088 |
+
"step": 8900
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.79,
|
| 1092 |
+
"learning_rate": 1.4752733561985168e-05,
|
| 1093 |
+
"loss": 0.3035,
|
| 1094 |
+
"step": 8950
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.79,
|
| 1098 |
+
"learning_rate": 1.4723419224342628e-05,
|
| 1099 |
+
"loss": 0.2915,
|
| 1100 |
+
"step": 9000
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.8,
|
| 1104 |
+
"learning_rate": 1.4694104886700086e-05,
|
| 1105 |
+
"loss": 0.3258,
|
| 1106 |
+
"step": 9050
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.8,
|
| 1110 |
+
"learning_rate": 1.4664790549057546e-05,
|
| 1111 |
+
"loss": 0.3312,
|
| 1112 |
+
"step": 9100
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.8,
|
| 1116 |
+
"learning_rate": 1.4635476211415006e-05,
|
| 1117 |
+
"loss": 0.3457,
|
| 1118 |
+
"step": 9150
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.81,
|
| 1122 |
+
"learning_rate": 1.4606161873772464e-05,
|
| 1123 |
+
"loss": 0.3183,
|
| 1124 |
+
"step": 9200
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.81,
|
| 1128 |
+
"learning_rate": 1.4576847536129923e-05,
|
| 1129 |
+
"loss": 0.3034,
|
| 1130 |
+
"step": 9250
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.82,
|
| 1134 |
+
"learning_rate": 1.4547533198487383e-05,
|
| 1135 |
+
"loss": 0.2851,
|
| 1136 |
+
"step": 9300
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.82,
|
| 1140 |
+
"learning_rate": 1.451821886084484e-05,
|
| 1141 |
+
"loss": 0.3081,
|
| 1142 |
+
"step": 9350
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.83,
|
| 1146 |
+
"learning_rate": 1.4488904523202298e-05,
|
| 1147 |
+
"loss": 0.2989,
|
| 1148 |
+
"step": 9400
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.83,
|
| 1152 |
+
"learning_rate": 1.4459590185559757e-05,
|
| 1153 |
+
"loss": 0.3022,
|
| 1154 |
+
"step": 9450
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.84,
|
| 1158 |
+
"learning_rate": 1.4430275847917217e-05,
|
| 1159 |
+
"loss": 0.3055,
|
| 1160 |
+
"step": 9500
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.84,
|
| 1164 |
+
"learning_rate": 1.4400961510274675e-05,
|
| 1165 |
+
"loss": 0.2944,
|
| 1166 |
+
"step": 9550
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.84,
|
| 1170 |
+
"learning_rate": 1.4371647172632135e-05,
|
| 1171 |
+
"loss": 0.3061,
|
| 1172 |
+
"step": 9600
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.85,
|
| 1176 |
+
"learning_rate": 1.4342332834989595e-05,
|
| 1177 |
+
"loss": 0.2882,
|
| 1178 |
+
"step": 9650
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.85,
|
| 1182 |
+
"learning_rate": 1.4313018497347053e-05,
|
| 1183 |
+
"loss": 0.3045,
|
| 1184 |
+
"step": 9700
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.86,
|
| 1188 |
+
"learning_rate": 1.4283704159704513e-05,
|
| 1189 |
+
"loss": 0.2752,
|
| 1190 |
+
"step": 9750
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.86,
|
| 1194 |
+
"learning_rate": 1.425438982206197e-05,
|
| 1195 |
+
"loss": 0.3305,
|
| 1196 |
+
"step": 9800
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.87,
|
| 1200 |
+
"learning_rate": 1.422507548441943e-05,
|
| 1201 |
+
"loss": 0.3221,
|
| 1202 |
+
"step": 9850
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.87,
|
| 1206 |
+
"learning_rate": 1.419576114677689e-05,
|
| 1207 |
+
"loss": 0.3056,
|
| 1208 |
+
"step": 9900
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.88,
|
| 1212 |
+
"learning_rate": 1.4166446809134348e-05,
|
| 1213 |
+
"loss": 0.3073,
|
| 1214 |
+
"step": 9950
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.88,
|
| 1218 |
+
"learning_rate": 1.4137132471491808e-05,
|
| 1219 |
+
"loss": 0.2998,
|
| 1220 |
+
"step": 10000
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.88,
|
| 1224 |
+
"eval_accuracy": 0.8724214692060351,
|
| 1225 |
+
"eval_combined_score": 0.8488343963458012,
|
| 1226 |
+
"eval_f1": 0.8252473234855671,
|
| 1227 |
+
"eval_loss": 0.2907809317111969,
|
| 1228 |
+
"eval_runtime": 51.9541,
|
| 1229 |
+
"eval_samples_per_second": 778.187,
|
| 1230 |
+
"eval_steps_per_second": 48.639,
|
| 1231 |
+
"step": 10000
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 0.88,
|
| 1235 |
+
"learning_rate": 1.4107818133849268e-05,
|
| 1236 |
+
"loss": 0.3164,
|
| 1237 |
+
"step": 10050
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 0.89,
|
| 1241 |
+
"learning_rate": 1.4078503796206726e-05,
|
| 1242 |
+
"loss": 0.3194,
|
| 1243 |
+
"step": 10100
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 0.89,
|
| 1247 |
+
"learning_rate": 1.4049189458564186e-05,
|
| 1248 |
+
"loss": 0.3147,
|
| 1249 |
+
"step": 10150
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.9,
|
| 1253 |
+
"learning_rate": 1.4019875120921644e-05,
|
| 1254 |
+
"loss": 0.3012,
|
| 1255 |
+
"step": 10200
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.9,
|
| 1259 |
+
"learning_rate": 1.3990560783279103e-05,
|
| 1260 |
+
"loss": 0.321,
|
| 1261 |
+
"step": 10250
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 0.91,
|
| 1265 |
+
"learning_rate": 1.3961246445636563e-05,
|
| 1266 |
+
"loss": 0.2941,
|
| 1267 |
+
"step": 10300
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 0.91,
|
| 1271 |
+
"learning_rate": 1.3931932107994021e-05,
|
| 1272 |
+
"loss": 0.3123,
|
| 1273 |
+
"step": 10350
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 0.91,
|
| 1277 |
+
"learning_rate": 1.3902617770351481e-05,
|
| 1278 |
+
"loss": 0.3137,
|
| 1279 |
+
"step": 10400
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 0.92,
|
| 1283 |
+
"learning_rate": 1.387330343270894e-05,
|
| 1284 |
+
"loss": 0.3022,
|
| 1285 |
+
"step": 10450
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 0.92,
|
| 1289 |
+
"learning_rate": 1.3843989095066399e-05,
|
| 1290 |
+
"loss": 0.2877,
|
| 1291 |
+
"step": 10500
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.93,
|
| 1295 |
+
"learning_rate": 1.3814674757423859e-05,
|
| 1296 |
+
"loss": 0.2911,
|
| 1297 |
+
"step": 10550
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.93,
|
| 1301 |
+
"learning_rate": 1.3785360419781315e-05,
|
| 1302 |
+
"loss": 0.2837,
|
| 1303 |
+
"step": 10600
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 0.94,
|
| 1307 |
+
"learning_rate": 1.3756046082138775e-05,
|
| 1308 |
+
"loss": 0.297,
|
| 1309 |
+
"step": 10650
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 0.94,
|
| 1313 |
+
"learning_rate": 1.3726731744496233e-05,
|
| 1314 |
+
"loss": 0.312,
|
| 1315 |
+
"step": 10700
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 0.95,
|
| 1319 |
+
"learning_rate": 1.3697417406853692e-05,
|
| 1320 |
+
"loss": 0.2899,
|
| 1321 |
+
"step": 10750
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 0.95,
|
| 1325 |
+
"learning_rate": 1.3668103069211152e-05,
|
| 1326 |
+
"loss": 0.3167,
|
| 1327 |
+
"step": 10800
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 0.95,
|
| 1331 |
+
"learning_rate": 1.363878873156861e-05,
|
| 1332 |
+
"loss": 0.3129,
|
| 1333 |
+
"step": 10850
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.96,
|
| 1337 |
+
"learning_rate": 1.360947439392607e-05,
|
| 1338 |
+
"loss": 0.2963,
|
| 1339 |
+
"step": 10900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.96,
|
| 1343 |
+
"learning_rate": 1.3580160056283528e-05,
|
| 1344 |
+
"loss": 0.2793,
|
| 1345 |
+
"step": 10950
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 0.97,
|
| 1349 |
+
"learning_rate": 1.3550845718640988e-05,
|
| 1350 |
+
"loss": 0.2995,
|
| 1351 |
+
"step": 11000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 0.97,
|
| 1355 |
+
"learning_rate": 1.3521531380998448e-05,
|
| 1356 |
+
"loss": 0.3328,
|
| 1357 |
+
"step": 11050
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 0.98,
|
| 1361 |
+
"learning_rate": 1.3492217043355906e-05,
|
| 1362 |
+
"loss": 0.3088,
|
| 1363 |
+
"step": 11100
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 0.98,
|
| 1367 |
+
"learning_rate": 1.3462902705713365e-05,
|
| 1368 |
+
"loss": 0.3016,
|
| 1369 |
+
"step": 11150
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 0.98,
|
| 1373 |
+
"learning_rate": 1.3433588368070825e-05,
|
| 1374 |
+
"loss": 0.3053,
|
| 1375 |
+
"step": 11200
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.99,
|
| 1379 |
+
"learning_rate": 1.3404274030428283e-05,
|
| 1380 |
+
"loss": 0.3138,
|
| 1381 |
+
"step": 11250
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.99,
|
| 1385 |
+
"learning_rate": 1.3374959692785743e-05,
|
| 1386 |
+
"loss": 0.2762,
|
| 1387 |
+
"step": 11300
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.0,
|
| 1391 |
+
"learning_rate": 1.3345645355143201e-05,
|
| 1392 |
+
"loss": 0.3029,
|
| 1393 |
+
"step": 11350
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.0,
|
| 1397 |
+
"learning_rate": 1.3316331017500661e-05,
|
| 1398 |
+
"loss": 0.2657,
|
| 1399 |
+
"step": 11400
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.01,
|
| 1403 |
+
"learning_rate": 1.328701667985812e-05,
|
| 1404 |
+
"loss": 0.2511,
|
| 1405 |
+
"step": 11450
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.01,
|
| 1409 |
+
"learning_rate": 1.3257702342215579e-05,
|
| 1410 |
+
"loss": 0.2462,
|
| 1411 |
+
"step": 11500
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 1.02,
|
| 1415 |
+
"learning_rate": 1.3228388004573038e-05,
|
| 1416 |
+
"loss": 0.2443,
|
| 1417 |
+
"step": 11550
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.02,
|
| 1421 |
+
"learning_rate": 1.3199073666930498e-05,
|
| 1422 |
+
"loss": 0.2628,
|
| 1423 |
+
"step": 11600
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.02,
|
| 1427 |
+
"learning_rate": 1.3169759329287956e-05,
|
| 1428 |
+
"loss": 0.2512,
|
| 1429 |
+
"step": 11650
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.03,
|
| 1433 |
+
"learning_rate": 1.3140444991645416e-05,
|
| 1434 |
+
"loss": 0.2404,
|
| 1435 |
+
"step": 11700
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.03,
|
| 1439 |
+
"learning_rate": 1.3111130654002874e-05,
|
| 1440 |
+
"loss": 0.2554,
|
| 1441 |
+
"step": 11750
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.04,
|
| 1445 |
+
"learning_rate": 1.3081816316360334e-05,
|
| 1446 |
+
"loss": 0.2538,
|
| 1447 |
+
"step": 11800
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.04,
|
| 1451 |
+
"learning_rate": 1.305250197871779e-05,
|
| 1452 |
+
"loss": 0.2456,
|
| 1453 |
+
"step": 11850
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.05,
|
| 1457 |
+
"learning_rate": 1.302318764107525e-05,
|
| 1458 |
+
"loss": 0.2648,
|
| 1459 |
+
"step": 11900
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.05,
|
| 1463 |
+
"learning_rate": 1.299387330343271e-05,
|
| 1464 |
+
"loss": 0.2482,
|
| 1465 |
+
"step": 11950
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.06,
|
| 1469 |
+
"learning_rate": 1.2964558965790168e-05,
|
| 1470 |
+
"loss": 0.2426,
|
| 1471 |
+
"step": 12000
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.06,
|
| 1475 |
+
"learning_rate": 1.2935244628147628e-05,
|
| 1476 |
+
"loss": 0.2642,
|
| 1477 |
+
"step": 12050
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.06,
|
| 1481 |
+
"learning_rate": 1.2905930290505086e-05,
|
| 1482 |
+
"loss": 0.2584,
|
| 1483 |
+
"step": 12100
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.07,
|
| 1487 |
+
"learning_rate": 1.2876615952862545e-05,
|
| 1488 |
+
"loss": 0.2655,
|
| 1489 |
+
"step": 12150
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 1.07,
|
| 1493 |
+
"learning_rate": 1.2847301615220005e-05,
|
| 1494 |
+
"loss": 0.2561,
|
| 1495 |
+
"step": 12200
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.08,
|
| 1499 |
+
"learning_rate": 1.2817987277577463e-05,
|
| 1500 |
+
"loss": 0.2291,
|
| 1501 |
+
"step": 12250
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.08,
|
| 1505 |
+
"learning_rate": 1.2788672939934923e-05,
|
| 1506 |
+
"loss": 0.2587,
|
| 1507 |
+
"step": 12300
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.09,
|
| 1511 |
+
"learning_rate": 1.2759358602292383e-05,
|
| 1512 |
+
"loss": 0.2192,
|
| 1513 |
+
"step": 12350
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.09,
|
| 1517 |
+
"learning_rate": 1.2730044264649841e-05,
|
| 1518 |
+
"loss": 0.2523,
|
| 1519 |
+
"step": 12400
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.09,
|
| 1523 |
+
"learning_rate": 1.27007299270073e-05,
|
| 1524 |
+
"loss": 0.2478,
|
| 1525 |
+
"step": 12450
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.1,
|
| 1529 |
+
"learning_rate": 1.2671415589364759e-05,
|
| 1530 |
+
"loss": 0.2576,
|
| 1531 |
+
"step": 12500
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 1.1,
|
| 1535 |
+
"learning_rate": 1.2642101251722218e-05,
|
| 1536 |
+
"loss": 0.2397,
|
| 1537 |
+
"step": 12550
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.11,
|
| 1541 |
+
"learning_rate": 1.2612786914079678e-05,
|
| 1542 |
+
"loss": 0.2549,
|
| 1543 |
+
"step": 12600
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.11,
|
| 1547 |
+
"learning_rate": 1.2583472576437136e-05,
|
| 1548 |
+
"loss": 0.26,
|
| 1549 |
+
"step": 12650
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.12,
|
| 1553 |
+
"learning_rate": 1.2554158238794596e-05,
|
| 1554 |
+
"loss": 0.2699,
|
| 1555 |
+
"step": 12700
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.12,
|
| 1559 |
+
"learning_rate": 1.2524843901152056e-05,
|
| 1560 |
+
"loss": 0.2646,
|
| 1561 |
+
"step": 12750
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.13,
|
| 1565 |
+
"learning_rate": 1.2495529563509514e-05,
|
| 1566 |
+
"loss": 0.2444,
|
| 1567 |
+
"step": 12800
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.13,
|
| 1571 |
+
"learning_rate": 1.2466215225866974e-05,
|
| 1572 |
+
"loss": 0.243,
|
| 1573 |
+
"step": 12850
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 1.13,
|
| 1577 |
+
"learning_rate": 1.2436900888224432e-05,
|
| 1578 |
+
"loss": 0.2505,
|
| 1579 |
+
"step": 12900
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.14,
|
| 1583 |
+
"learning_rate": 1.2407586550581891e-05,
|
| 1584 |
+
"loss": 0.2527,
|
| 1585 |
+
"step": 12950
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.14,
|
| 1589 |
+
"learning_rate": 1.2378272212939351e-05,
|
| 1590 |
+
"loss": 0.2611,
|
| 1591 |
+
"step": 13000
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.15,
|
| 1595 |
+
"learning_rate": 1.234895787529681e-05,
|
| 1596 |
+
"loss": 0.2444,
|
| 1597 |
+
"step": 13050
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.15,
|
| 1601 |
+
"learning_rate": 1.2319643537654267e-05,
|
| 1602 |
+
"loss": 0.2539,
|
| 1603 |
+
"step": 13100
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.16,
|
| 1607 |
+
"learning_rate": 1.2290329200011725e-05,
|
| 1608 |
+
"loss": 0.273,
|
| 1609 |
+
"step": 13150
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 1.16,
|
| 1613 |
+
"learning_rate": 1.2261014862369185e-05,
|
| 1614 |
+
"loss": 0.2541,
|
| 1615 |
+
"step": 13200
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.17,
|
| 1619 |
+
"learning_rate": 1.2231700524726643e-05,
|
| 1620 |
+
"loss": 0.2362,
|
| 1621 |
+
"step": 13250
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.17,
|
| 1625 |
+
"learning_rate": 1.2202386187084103e-05,
|
| 1626 |
+
"loss": 0.2574,
|
| 1627 |
+
"step": 13300
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.17,
|
| 1631 |
+
"learning_rate": 1.2173071849441563e-05,
|
| 1632 |
+
"loss": 0.2428,
|
| 1633 |
+
"step": 13350
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.18,
|
| 1637 |
+
"learning_rate": 1.214375751179902e-05,
|
| 1638 |
+
"loss": 0.2654,
|
| 1639 |
+
"step": 13400
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.18,
|
| 1643 |
+
"learning_rate": 1.211444317415648e-05,
|
| 1644 |
+
"loss": 0.2323,
|
| 1645 |
+
"step": 13450
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.19,
|
| 1649 |
+
"learning_rate": 1.208512883651394e-05,
|
| 1650 |
+
"loss": 0.2563,
|
| 1651 |
+
"step": 13500
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.19,
|
| 1655 |
+
"learning_rate": 1.2055814498871398e-05,
|
| 1656 |
+
"loss": 0.2266,
|
| 1657 |
+
"step": 13550
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 1.2,
|
| 1661 |
+
"learning_rate": 1.2026500161228858e-05,
|
| 1662 |
+
"loss": 0.2728,
|
| 1663 |
+
"step": 13600
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.2,
|
| 1667 |
+
"learning_rate": 1.1997185823586316e-05,
|
| 1668 |
+
"loss": 0.2396,
|
| 1669 |
+
"step": 13650
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.2,
|
| 1673 |
+
"learning_rate": 1.1967871485943776e-05,
|
| 1674 |
+
"loss": 0.2506,
|
| 1675 |
+
"step": 13700
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.21,
|
| 1679 |
+
"learning_rate": 1.1938557148301236e-05,
|
| 1680 |
+
"loss": 0.2257,
|
| 1681 |
+
"step": 13750
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.21,
|
| 1685 |
+
"learning_rate": 1.1909242810658694e-05,
|
| 1686 |
+
"loss": 0.2312,
|
| 1687 |
+
"step": 13800
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.22,
|
| 1691 |
+
"learning_rate": 1.1879928473016154e-05,
|
| 1692 |
+
"loss": 0.2338,
|
| 1693 |
+
"step": 13850
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 1.22,
|
| 1697 |
+
"learning_rate": 1.1850614135373613e-05,
|
| 1698 |
+
"loss": 0.2559,
|
| 1699 |
+
"step": 13900
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 1.23,
|
| 1703 |
+
"learning_rate": 1.1821299797731071e-05,
|
| 1704 |
+
"loss": 0.2376,
|
| 1705 |
+
"step": 13950
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.23,
|
| 1709 |
+
"learning_rate": 1.1791985460088531e-05,
|
| 1710 |
+
"loss": 0.2564,
|
| 1711 |
+
"step": 14000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 1.24,
|
| 1715 |
+
"learning_rate": 1.176267112244599e-05,
|
| 1716 |
+
"loss": 0.2321,
|
| 1717 |
+
"step": 14050
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.24,
|
| 1721 |
+
"learning_rate": 1.1733356784803449e-05,
|
| 1722 |
+
"loss": 0.2414,
|
| 1723 |
+
"step": 14100
|
| 1724 |
+
},
|
| 1725 |
+
{
|
| 1726 |
+
"epoch": 1.24,
|
| 1727 |
+
"learning_rate": 1.1704042447160909e-05,
|
| 1728 |
+
"loss": 0.2263,
|
| 1729 |
+
"step": 14150
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 1.25,
|
| 1733 |
+
"learning_rate": 1.1674728109518367e-05,
|
| 1734 |
+
"loss": 0.2545,
|
| 1735 |
+
"step": 14200
|
| 1736 |
+
},
|
| 1737 |
+
{
|
| 1738 |
+
"epoch": 1.25,
|
| 1739 |
+
"learning_rate": 1.1645413771875827e-05,
|
| 1740 |
+
"loss": 0.2497,
|
| 1741 |
+
"step": 14250
|
| 1742 |
+
},
|
| 1743 |
+
{
|
| 1744 |
+
"epoch": 1.26,
|
| 1745 |
+
"learning_rate": 1.1616099434233283e-05,
|
| 1746 |
+
"loss": 0.2305,
|
| 1747 |
+
"step": 14300
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 1.26,
|
| 1751 |
+
"learning_rate": 1.1586785096590743e-05,
|
| 1752 |
+
"loss": 0.2425,
|
| 1753 |
+
"step": 14350
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 1.27,
|
| 1757 |
+
"learning_rate": 1.15574707589482e-05,
|
| 1758 |
+
"loss": 0.2456,
|
| 1759 |
+
"step": 14400
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 1.27,
|
| 1763 |
+
"learning_rate": 1.152815642130566e-05,
|
| 1764 |
+
"loss": 0.2333,
|
| 1765 |
+
"step": 14450
|
| 1766 |
+
},
|
| 1767 |
+
{
|
| 1768 |
+
"epoch": 1.28,
|
| 1769 |
+
"learning_rate": 1.149884208366312e-05,
|
| 1770 |
+
"loss": 0.24,
|
| 1771 |
+
"step": 14500
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 1.28,
|
| 1775 |
+
"learning_rate": 1.1469527746020578e-05,
|
| 1776 |
+
"loss": 0.2463,
|
| 1777 |
+
"step": 14550
|
| 1778 |
+
},
|
| 1779 |
+
{
|
| 1780 |
+
"epoch": 1.28,
|
| 1781 |
+
"learning_rate": 1.1440213408378038e-05,
|
| 1782 |
+
"loss": 0.2338,
|
| 1783 |
+
"step": 14600
|
| 1784 |
+
},
|
| 1785 |
+
{
|
| 1786 |
+
"epoch": 1.29,
|
| 1787 |
+
"learning_rate": 1.1410899070735498e-05,
|
| 1788 |
+
"loss": 0.2603,
|
| 1789 |
+
"step": 14650
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 1.29,
|
| 1793 |
+
"learning_rate": 1.1381584733092956e-05,
|
| 1794 |
+
"loss": 0.2406,
|
| 1795 |
+
"step": 14700
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 1.3,
|
| 1799 |
+
"learning_rate": 1.1352270395450416e-05,
|
| 1800 |
+
"loss": 0.2335,
|
| 1801 |
+
"step": 14750
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.3,
|
| 1805 |
+
"learning_rate": 1.1322956057807874e-05,
|
| 1806 |
+
"loss": 0.2454,
|
| 1807 |
+
"step": 14800
|
| 1808 |
+
},
|
| 1809 |
+
{
|
| 1810 |
+
"epoch": 1.31,
|
| 1811 |
+
"learning_rate": 1.1293641720165334e-05,
|
| 1812 |
+
"loss": 0.2415,
|
| 1813 |
+
"step": 14850
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.31,
|
| 1817 |
+
"learning_rate": 1.1264327382522793e-05,
|
| 1818 |
+
"loss": 0.2284,
|
| 1819 |
+
"step": 14900
|
| 1820 |
+
},
|
| 1821 |
+
{
|
| 1822 |
+
"epoch": 1.31,
|
| 1823 |
+
"learning_rate": 1.1235013044880251e-05,
|
| 1824 |
+
"loss": 0.2386,
|
| 1825 |
+
"step": 14950
|
| 1826 |
+
},
|
| 1827 |
+
{
|
| 1828 |
+
"epoch": 1.32,
|
| 1829 |
+
"learning_rate": 1.1205698707237711e-05,
|
| 1830 |
+
"loss": 0.2629,
|
| 1831 |
+
"step": 15000
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 1.32,
|
| 1835 |
+
"eval_accuracy": 0.8762799901063567,
|
| 1836 |
+
"eval_combined_score": 0.8531586834663621,
|
| 1837 |
+
"eval_f1": 0.8300373768263676,
|
| 1838 |
+
"eval_loss": 0.2969573140144348,
|
| 1839 |
+
"eval_runtime": 51.2924,
|
| 1840 |
+
"eval_samples_per_second": 788.226,
|
| 1841 |
+
"eval_steps_per_second": 49.267,
|
| 1842 |
+
"step": 15000
|
| 1843 |
+
},
|
| 1844 |
+
{
|
| 1845 |
+
"epoch": 1.32,
|
| 1846 |
+
"learning_rate": 1.1176384369595171e-05,
|
| 1847 |
+
"loss": 0.2755,
|
| 1848 |
+
"step": 15050
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 1.33,
|
| 1852 |
+
"learning_rate": 1.1147070031952629e-05,
|
| 1853 |
+
"loss": 0.2464,
|
| 1854 |
+
"step": 15100
|
| 1855 |
+
},
|
| 1856 |
+
{
|
| 1857 |
+
"epoch": 1.33,
|
| 1858 |
+
"learning_rate": 1.1117755694310089e-05,
|
| 1859 |
+
"loss": 0.2575,
|
| 1860 |
+
"step": 15150
|
| 1861 |
+
},
|
| 1862 |
+
{
|
| 1863 |
+
"epoch": 1.34,
|
| 1864 |
+
"learning_rate": 1.1088441356667547e-05,
|
| 1865 |
+
"loss": 0.2552,
|
| 1866 |
+
"step": 15200
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 1.34,
|
| 1870 |
+
"learning_rate": 1.1059127019025007e-05,
|
| 1871 |
+
"loss": 0.2306,
|
| 1872 |
+
"step": 15250
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 1.35,
|
| 1876 |
+
"learning_rate": 1.1029812681382466e-05,
|
| 1877 |
+
"loss": 0.234,
|
| 1878 |
+
"step": 15300
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 1.35,
|
| 1882 |
+
"learning_rate": 1.1000498343739924e-05,
|
| 1883 |
+
"loss": 0.2359,
|
| 1884 |
+
"step": 15350
|
| 1885 |
+
},
|
| 1886 |
+
{
|
| 1887 |
+
"epoch": 1.35,
|
| 1888 |
+
"learning_rate": 1.0971184006097384e-05,
|
| 1889 |
+
"loss": 0.2641,
|
| 1890 |
+
"step": 15400
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 1.36,
|
| 1894 |
+
"learning_rate": 1.0941869668454844e-05,
|
| 1895 |
+
"loss": 0.2635,
|
| 1896 |
+
"step": 15450
|
| 1897 |
+
},
|
| 1898 |
+
{
|
| 1899 |
+
"epoch": 1.36,
|
| 1900 |
+
"learning_rate": 1.0912555330812302e-05,
|
| 1901 |
+
"loss": 0.2325,
|
| 1902 |
+
"step": 15500
|
| 1903 |
+
},
|
| 1904 |
+
{
|
| 1905 |
+
"epoch": 1.37,
|
| 1906 |
+
"learning_rate": 1.0883240993169758e-05,
|
| 1907 |
+
"loss": 0.2732,
|
| 1908 |
+
"step": 15550
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 1.37,
|
| 1912 |
+
"learning_rate": 1.0853926655527218e-05,
|
| 1913 |
+
"loss": 0.2401,
|
| 1914 |
+
"step": 15600
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 1.38,
|
| 1918 |
+
"learning_rate": 1.0824612317884678e-05,
|
| 1919 |
+
"loss": 0.224,
|
| 1920 |
+
"step": 15650
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.38,
|
| 1924 |
+
"learning_rate": 1.0795297980242136e-05,
|
| 1925 |
+
"loss": 0.273,
|
| 1926 |
+
"step": 15700
|
| 1927 |
+
},
|
| 1928 |
+
{
|
| 1929 |
+
"epoch": 1.39,
|
| 1930 |
+
"learning_rate": 1.0765983642599596e-05,
|
| 1931 |
+
"loss": 0.2647,
|
| 1932 |
+
"step": 15750
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 1.39,
|
| 1936 |
+
"learning_rate": 1.0736669304957055e-05,
|
| 1937 |
+
"loss": 0.2472,
|
| 1938 |
+
"step": 15800
|
| 1939 |
+
},
|
| 1940 |
+
{
|
| 1941 |
+
"epoch": 1.39,
|
| 1942 |
+
"learning_rate": 1.0707354967314513e-05,
|
| 1943 |
+
"loss": 0.2477,
|
| 1944 |
+
"step": 15850
|
| 1945 |
+
},
|
| 1946 |
+
{
|
| 1947 |
+
"epoch": 1.4,
|
| 1948 |
+
"learning_rate": 1.0678040629671973e-05,
|
| 1949 |
+
"loss": 0.2433,
|
| 1950 |
+
"step": 15900
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.4,
|
| 1954 |
+
"learning_rate": 1.0648726292029431e-05,
|
| 1955 |
+
"loss": 0.2835,
|
| 1956 |
+
"step": 15950
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 1.41,
|
| 1960 |
+
"learning_rate": 1.0619411954386891e-05,
|
| 1961 |
+
"loss": 0.2519,
|
| 1962 |
+
"step": 16000
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 1.41,
|
| 1966 |
+
"learning_rate": 1.059009761674435e-05,
|
| 1967 |
+
"loss": 0.2284,
|
| 1968 |
+
"step": 16050
|
| 1969 |
+
},
|
| 1970 |
+
{
|
| 1971 |
+
"epoch": 1.42,
|
| 1972 |
+
"learning_rate": 1.0560783279101809e-05,
|
| 1973 |
+
"loss": 0.2785,
|
| 1974 |
+
"step": 16100
|
| 1975 |
+
},
|
| 1976 |
+
{
|
| 1977 |
+
"epoch": 1.42,
|
| 1978 |
+
"learning_rate": 1.0531468941459269e-05,
|
| 1979 |
+
"loss": 0.2448,
|
| 1980 |
+
"step": 16150
|
| 1981 |
+
},
|
| 1982 |
+
{
|
| 1983 |
+
"epoch": 1.42,
|
| 1984 |
+
"learning_rate": 1.0502154603816728e-05,
|
| 1985 |
+
"loss": 0.2402,
|
| 1986 |
+
"step": 16200
|
| 1987 |
+
},
|
| 1988 |
+
{
|
| 1989 |
+
"epoch": 1.43,
|
| 1990 |
+
"learning_rate": 1.0472840266174186e-05,
|
| 1991 |
+
"loss": 0.2627,
|
| 1992 |
+
"step": 16250
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 1.43,
|
| 1996 |
+
"learning_rate": 1.0443525928531646e-05,
|
| 1997 |
+
"loss": 0.2549,
|
| 1998 |
+
"step": 16300
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 1.44,
|
| 2002 |
+
"learning_rate": 1.0414211590889104e-05,
|
| 2003 |
+
"loss": 0.2381,
|
| 2004 |
+
"step": 16350
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 1.44,
|
| 2008 |
+
"learning_rate": 1.0384897253246564e-05,
|
| 2009 |
+
"loss": 0.2293,
|
| 2010 |
+
"step": 16400
|
| 2011 |
+
},
|
| 2012 |
+
{
|
| 2013 |
+
"epoch": 1.45,
|
| 2014 |
+
"learning_rate": 1.0355582915604024e-05,
|
| 2015 |
+
"loss": 0.2412,
|
| 2016 |
+
"step": 16450
|
| 2017 |
+
},
|
| 2018 |
+
{
|
| 2019 |
+
"epoch": 1.45,
|
| 2020 |
+
"learning_rate": 1.0326268577961482e-05,
|
| 2021 |
+
"loss": 0.2441,
|
| 2022 |
+
"step": 16500
|
| 2023 |
+
},
|
| 2024 |
+
{
|
| 2025 |
+
"epoch": 1.46,
|
| 2026 |
+
"learning_rate": 1.0296954240318942e-05,
|
| 2027 |
+
"loss": 0.2388,
|
| 2028 |
+
"step": 16550
|
| 2029 |
+
},
|
| 2030 |
+
{
|
| 2031 |
+
"epoch": 1.46,
|
| 2032 |
+
"learning_rate": 1.0267639902676401e-05,
|
| 2033 |
+
"loss": 0.2309,
|
| 2034 |
+
"step": 16600
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 1.46,
|
| 2038 |
+
"learning_rate": 1.023832556503386e-05,
|
| 2039 |
+
"loss": 0.2332,
|
| 2040 |
+
"step": 16650
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 1.47,
|
| 2044 |
+
"learning_rate": 1.020901122739132e-05,
|
| 2045 |
+
"loss": 0.2573,
|
| 2046 |
+
"step": 16700
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 1.47,
|
| 2050 |
+
"learning_rate": 1.0179696889748777e-05,
|
| 2051 |
+
"loss": 0.2337,
|
| 2052 |
+
"step": 16750
|
| 2053 |
+
},
|
| 2054 |
+
{
|
| 2055 |
+
"epoch": 1.48,
|
| 2056 |
+
"learning_rate": 1.0150382552106235e-05,
|
| 2057 |
+
"loss": 0.2245,
|
| 2058 |
+
"step": 16800
|
| 2059 |
+
},
|
| 2060 |
+
{
|
| 2061 |
+
"epoch": 1.48,
|
| 2062 |
+
"learning_rate": 1.0121068214463693e-05,
|
| 2063 |
+
"loss": 0.2336,
|
| 2064 |
+
"step": 16850
|
| 2065 |
+
},
|
| 2066 |
+
{
|
| 2067 |
+
"epoch": 1.49,
|
| 2068 |
+
"learning_rate": 1.0091753876821153e-05,
|
| 2069 |
+
"loss": 0.2487,
|
| 2070 |
+
"step": 16900
|
| 2071 |
+
},
|
| 2072 |
+
{
|
| 2073 |
+
"epoch": 1.49,
|
| 2074 |
+
"learning_rate": 1.0062439539178613e-05,
|
| 2075 |
+
"loss": 0.2219,
|
| 2076 |
+
"step": 16950
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 1.5,
|
| 2080 |
+
"learning_rate": 1.0033125201536071e-05,
|
| 2081 |
+
"loss": 0.2544,
|
| 2082 |
+
"step": 17000
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 1.5,
|
| 2086 |
+
"learning_rate": 1.000381086389353e-05,
|
| 2087 |
+
"loss": 0.2568,
|
| 2088 |
+
"step": 17050
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 1.5,
|
| 2092 |
+
"learning_rate": 9.974496526250989e-06,
|
| 2093 |
+
"loss": 0.236,
|
| 2094 |
+
"step": 17100
|
| 2095 |
+
},
|
| 2096 |
+
{
|
| 2097 |
+
"epoch": 1.51,
|
| 2098 |
+
"learning_rate": 9.945182188608449e-06,
|
| 2099 |
+
"loss": 0.2557,
|
| 2100 |
+
"step": 17150
|
| 2101 |
+
},
|
| 2102 |
+
{
|
| 2103 |
+
"epoch": 1.51,
|
| 2104 |
+
"learning_rate": 9.915867850965908e-06,
|
| 2105 |
+
"loss": 0.2252,
|
| 2106 |
+
"step": 17200
|
| 2107 |
+
},
|
| 2108 |
+
{
|
| 2109 |
+
"epoch": 1.52,
|
| 2110 |
+
"learning_rate": 9.886553513323366e-06,
|
| 2111 |
+
"loss": 0.2656,
|
| 2112 |
+
"step": 17250
|
| 2113 |
+
},
|
| 2114 |
+
{
|
| 2115 |
+
"epoch": 1.52,
|
| 2116 |
+
"learning_rate": 9.857239175680826e-06,
|
| 2117 |
+
"loss": 0.2237,
|
| 2118 |
+
"step": 17300
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 1.53,
|
| 2122 |
+
"learning_rate": 9.827924838038286e-06,
|
| 2123 |
+
"loss": 0.241,
|
| 2124 |
+
"step": 17350
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 1.53,
|
| 2128 |
+
"learning_rate": 9.798610500395744e-06,
|
| 2129 |
+
"loss": 0.2435,
|
| 2130 |
+
"step": 17400
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.53,
|
| 2134 |
+
"learning_rate": 9.769296162753204e-06,
|
| 2135 |
+
"loss": 0.2497,
|
| 2136 |
+
"step": 17450
|
| 2137 |
+
},
|
| 2138 |
+
{
|
| 2139 |
+
"epoch": 1.54,
|
| 2140 |
+
"learning_rate": 9.739981825110662e-06,
|
| 2141 |
+
"loss": 0.2487,
|
| 2142 |
+
"step": 17500
|
| 2143 |
+
},
|
| 2144 |
+
{
|
| 2145 |
+
"epoch": 1.54,
|
| 2146 |
+
"learning_rate": 9.710667487468122e-06,
|
| 2147 |
+
"loss": 0.2287,
|
| 2148 |
+
"step": 17550
|
| 2149 |
+
},
|
| 2150 |
+
{
|
| 2151 |
+
"epoch": 1.55,
|
| 2152 |
+
"learning_rate": 9.681353149825581e-06,
|
| 2153 |
+
"loss": 0.2559,
|
| 2154 |
+
"step": 17600
|
| 2155 |
+
},
|
| 2156 |
+
{
|
| 2157 |
+
"epoch": 1.55,
|
| 2158 |
+
"learning_rate": 9.65203881218304e-06,
|
| 2159 |
+
"loss": 0.2363,
|
| 2160 |
+
"step": 17650
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 1.56,
|
| 2164 |
+
"learning_rate": 9.6227244745405e-06,
|
| 2165 |
+
"loss": 0.2533,
|
| 2166 |
+
"step": 17700
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 1.56,
|
| 2170 |
+
"learning_rate": 9.593410136897957e-06,
|
| 2171 |
+
"loss": 0.2297,
|
| 2172 |
+
"step": 17750
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 1.57,
|
| 2176 |
+
"learning_rate": 9.564095799255417e-06,
|
| 2177 |
+
"loss": 0.2595,
|
| 2178 |
+
"step": 17800
|
| 2179 |
+
},
|
| 2180 |
+
{
|
| 2181 |
+
"epoch": 1.57,
|
| 2182 |
+
"learning_rate": 9.534781461612875e-06,
|
| 2183 |
+
"loss": 0.2238,
|
| 2184 |
+
"step": 17850
|
| 2185 |
+
},
|
| 2186 |
+
{
|
| 2187 |
+
"epoch": 1.57,
|
| 2188 |
+
"learning_rate": 9.505467123970335e-06,
|
| 2189 |
+
"loss": 0.2475,
|
| 2190 |
+
"step": 17900
|
| 2191 |
+
},
|
| 2192 |
+
{
|
| 2193 |
+
"epoch": 1.58,
|
| 2194 |
+
"learning_rate": 9.476152786327793e-06,
|
| 2195 |
+
"loss": 0.2523,
|
| 2196 |
+
"step": 17950
|
| 2197 |
+
},
|
| 2198 |
+
{
|
| 2199 |
+
"epoch": 1.58,
|
| 2200 |
+
"learning_rate": 9.446838448685253e-06,
|
| 2201 |
+
"loss": 0.2412,
|
| 2202 |
+
"step": 18000
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 1.59,
|
| 2206 |
+
"learning_rate": 9.417524111042712e-06,
|
| 2207 |
+
"loss": 0.2409,
|
| 2208 |
+
"step": 18050
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 1.59,
|
| 2212 |
+
"learning_rate": 9.38820977340017e-06,
|
| 2213 |
+
"loss": 0.2731,
|
| 2214 |
+
"step": 18100
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 1.6,
|
| 2218 |
+
"learning_rate": 9.35889543575763e-06,
|
| 2219 |
+
"loss": 0.2754,
|
| 2220 |
+
"step": 18150
|
| 2221 |
+
},
|
| 2222 |
+
{
|
| 2223 |
+
"epoch": 1.6,
|
| 2224 |
+
"learning_rate": 9.329581098115088e-06,
|
| 2225 |
+
"loss": 0.2206,
|
| 2226 |
+
"step": 18200
|
| 2227 |
+
},
|
| 2228 |
+
{
|
| 2229 |
+
"epoch": 1.6,
|
| 2230 |
+
"learning_rate": 9.300266760472548e-06,
|
| 2231 |
+
"loss": 0.249,
|
| 2232 |
+
"step": 18250
|
| 2233 |
+
},
|
| 2234 |
+
{
|
| 2235 |
+
"epoch": 1.61,
|
| 2236 |
+
"learning_rate": 9.270952422830008e-06,
|
| 2237 |
+
"loss": 0.2286,
|
| 2238 |
+
"step": 18300
|
| 2239 |
+
},
|
| 2240 |
+
{
|
| 2241 |
+
"epoch": 1.61,
|
| 2242 |
+
"learning_rate": 9.241638085187466e-06,
|
| 2243 |
+
"loss": 0.2532,
|
| 2244 |
+
"step": 18350
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 1.62,
|
| 2248 |
+
"learning_rate": 9.212323747544924e-06,
|
| 2249 |
+
"loss": 0.2323,
|
| 2250 |
+
"step": 18400
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 1.62,
|
| 2254 |
+
"learning_rate": 9.183009409902384e-06,
|
| 2255 |
+
"loss": 0.2399,
|
| 2256 |
+
"step": 18450
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 1.63,
|
| 2260 |
+
"learning_rate": 9.153695072259843e-06,
|
| 2261 |
+
"loss": 0.2368,
|
| 2262 |
+
"step": 18500
|
| 2263 |
+
},
|
| 2264 |
+
{
|
| 2265 |
+
"epoch": 1.63,
|
| 2266 |
+
"learning_rate": 9.124380734617302e-06,
|
| 2267 |
+
"loss": 0.2139,
|
| 2268 |
+
"step": 18550
|
| 2269 |
+
},
|
| 2270 |
+
{
|
| 2271 |
+
"epoch": 1.64,
|
| 2272 |
+
"learning_rate": 9.095066396974761e-06,
|
| 2273 |
+
"loss": 0.2403,
|
| 2274 |
+
"step": 18600
|
| 2275 |
+
},
|
| 2276 |
+
{
|
| 2277 |
+
"epoch": 1.64,
|
| 2278 |
+
"learning_rate": 9.06575205933222e-06,
|
| 2279 |
+
"loss": 0.2247,
|
| 2280 |
+
"step": 18650
|
| 2281 |
+
},
|
| 2282 |
+
{
|
| 2283 |
+
"epoch": 1.64,
|
| 2284 |
+
"learning_rate": 9.036437721689679e-06,
|
| 2285 |
+
"loss": 0.2607,
|
| 2286 |
+
"step": 18700
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 1.65,
|
| 2290 |
+
"learning_rate": 9.007123384047139e-06,
|
| 2291 |
+
"loss": 0.2357,
|
| 2292 |
+
"step": 18750
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 1.65,
|
| 2296 |
+
"learning_rate": 8.977809046404597e-06,
|
| 2297 |
+
"loss": 0.2436,
|
| 2298 |
+
"step": 18800
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 1.66,
|
| 2302 |
+
"learning_rate": 8.948494708762057e-06,
|
| 2303 |
+
"loss": 0.2424,
|
| 2304 |
+
"step": 18850
|
| 2305 |
+
},
|
| 2306 |
+
{
|
| 2307 |
+
"epoch": 1.66,
|
| 2308 |
+
"learning_rate": 8.919180371119516e-06,
|
| 2309 |
+
"loss": 0.2457,
|
| 2310 |
+
"step": 18900
|
| 2311 |
+
},
|
| 2312 |
+
{
|
| 2313 |
+
"epoch": 1.67,
|
| 2314 |
+
"learning_rate": 8.889866033476975e-06,
|
| 2315 |
+
"loss": 0.2229,
|
| 2316 |
+
"step": 18950
|
| 2317 |
+
},
|
| 2318 |
+
{
|
| 2319 |
+
"epoch": 1.67,
|
| 2320 |
+
"learning_rate": 8.860551695834433e-06,
|
| 2321 |
+
"loss": 0.2636,
|
| 2322 |
+
"step": 19000
|
| 2323 |
+
},
|
| 2324 |
+
{
|
| 2325 |
+
"epoch": 1.68,
|
| 2326 |
+
"learning_rate": 8.831237358191892e-06,
|
| 2327 |
+
"loss": 0.229,
|
| 2328 |
+
"step": 19050
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 1.68,
|
| 2332 |
+
"learning_rate": 8.80192302054935e-06,
|
| 2333 |
+
"loss": 0.2427,
|
| 2334 |
+
"step": 19100
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 1.68,
|
| 2338 |
+
"learning_rate": 8.77260868290681e-06,
|
| 2339 |
+
"loss": 0.2538,
|
| 2340 |
+
"step": 19150
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 1.69,
|
| 2344 |
+
"learning_rate": 8.74329434526427e-06,
|
| 2345 |
+
"loss": 0.247,
|
| 2346 |
+
"step": 19200
|
| 2347 |
+
},
|
| 2348 |
+
{
|
| 2349 |
+
"epoch": 1.69,
|
| 2350 |
+
"learning_rate": 8.713980007621728e-06,
|
| 2351 |
+
"loss": 0.2229,
|
| 2352 |
+
"step": 19250
|
| 2353 |
+
},
|
| 2354 |
+
{
|
| 2355 |
+
"epoch": 1.7,
|
| 2356 |
+
"learning_rate": 8.684665669979188e-06,
|
| 2357 |
+
"loss": 0.254,
|
| 2358 |
+
"step": 19300
|
| 2359 |
+
},
|
| 2360 |
+
{
|
| 2361 |
+
"epoch": 1.7,
|
| 2362 |
+
"learning_rate": 8.655351332336648e-06,
|
| 2363 |
+
"loss": 0.2495,
|
| 2364 |
+
"step": 19350
|
| 2365 |
+
},
|
| 2366 |
+
{
|
| 2367 |
+
"epoch": 1.71,
|
| 2368 |
+
"learning_rate": 8.626036994694106e-06,
|
| 2369 |
+
"loss": 0.2554,
|
| 2370 |
+
"step": 19400
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 1.71,
|
| 2374 |
+
"learning_rate": 8.596722657051565e-06,
|
| 2375 |
+
"loss": 0.2327,
|
| 2376 |
+
"step": 19450
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 1.71,
|
| 2380 |
+
"learning_rate": 8.567408319409023e-06,
|
| 2381 |
+
"loss": 0.2484,
|
| 2382 |
+
"step": 19500
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 1.72,
|
| 2386 |
+
"learning_rate": 8.538093981766483e-06,
|
| 2387 |
+
"loss": 0.2396,
|
| 2388 |
+
"step": 19550
|
| 2389 |
+
},
|
| 2390 |
+
{
|
| 2391 |
+
"epoch": 1.72,
|
| 2392 |
+
"learning_rate": 8.508779644123941e-06,
|
| 2393 |
+
"loss": 0.2321,
|
| 2394 |
+
"step": 19600
|
| 2395 |
+
},
|
| 2396 |
+
{
|
| 2397 |
+
"epoch": 1.73,
|
| 2398 |
+
"learning_rate": 8.479465306481401e-06,
|
| 2399 |
+
"loss": 0.2402,
|
| 2400 |
+
"step": 19650
|
| 2401 |
+
},
|
| 2402 |
+
{
|
| 2403 |
+
"epoch": 1.73,
|
| 2404 |
+
"learning_rate": 8.450150968838859e-06,
|
| 2405 |
+
"loss": 0.2264,
|
| 2406 |
+
"step": 19700
|
| 2407 |
+
},
|
| 2408 |
+
{
|
| 2409 |
+
"epoch": 1.74,
|
| 2410 |
+
"learning_rate": 8.420836631196319e-06,
|
| 2411 |
+
"loss": 0.2628,
|
| 2412 |
+
"step": 19750
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 1.74,
|
| 2416 |
+
"learning_rate": 8.391522293553777e-06,
|
| 2417 |
+
"loss": 0.2303,
|
| 2418 |
+
"step": 19800
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 1.75,
|
| 2422 |
+
"learning_rate": 8.362207955911237e-06,
|
| 2423 |
+
"loss": 0.2279,
|
| 2424 |
+
"step": 19850
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 1.75,
|
| 2428 |
+
"learning_rate": 8.332893618268696e-06,
|
| 2429 |
+
"loss": 0.2655,
|
| 2430 |
+
"step": 19900
|
| 2431 |
+
},
|
| 2432 |
+
{
|
| 2433 |
+
"epoch": 1.75,
|
| 2434 |
+
"learning_rate": 8.303579280626155e-06,
|
| 2435 |
+
"loss": 0.2302,
|
| 2436 |
+
"step": 19950
|
| 2437 |
+
},
|
| 2438 |
+
{
|
| 2439 |
+
"epoch": 1.76,
|
| 2440 |
+
"learning_rate": 8.274264942983614e-06,
|
| 2441 |
+
"loss": 0.2269,
|
| 2442 |
+
"step": 20000
|
| 2443 |
+
},
|
| 2444 |
+
{
|
| 2445 |
+
"epoch": 1.76,
|
| 2446 |
+
"eval_accuracy": 0.8844917140737076,
|
| 2447 |
+
"eval_combined_score": 0.862498589277291,
|
| 2448 |
+
"eval_f1": 0.8405054644808744,
|
| 2449 |
+
"eval_loss": 0.28743284940719604,
|
| 2450 |
+
"eval_runtime": 51.5779,
|
| 2451 |
+
"eval_samples_per_second": 783.863,
|
| 2452 |
+
"eval_steps_per_second": 48.994,
|
| 2453 |
+
"step": 20000
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 1.76,
|
| 2457 |
+
"learning_rate": 8.244950605341074e-06,
|
| 2458 |
+
"loss": 0.2402,
|
| 2459 |
+
"step": 20050
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 1.77,
|
| 2463 |
+
"learning_rate": 8.215636267698532e-06,
|
| 2464 |
+
"loss": 0.2583,
|
| 2465 |
+
"step": 20100
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 1.77,
|
| 2469 |
+
"learning_rate": 8.186321930055992e-06,
|
| 2470 |
+
"loss": 0.2302,
|
| 2471 |
+
"step": 20150
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 1.78,
|
| 2475 |
+
"learning_rate": 8.15700759241345e-06,
|
| 2476 |
+
"loss": 0.242,
|
| 2477 |
+
"step": 20200
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 1.78,
|
| 2481 |
+
"learning_rate": 8.127693254770908e-06,
|
| 2482 |
+
"loss": 0.2462,
|
| 2483 |
+
"step": 20250
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 1.79,
|
| 2487 |
+
"learning_rate": 8.098378917128368e-06,
|
| 2488 |
+
"loss": 0.2049,
|
| 2489 |
+
"step": 20300
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 1.79,
|
| 2493 |
+
"learning_rate": 8.069064579485828e-06,
|
| 2494 |
+
"loss": 0.2429,
|
| 2495 |
+
"step": 20350
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 1.79,
|
| 2499 |
+
"learning_rate": 8.039750241843286e-06,
|
| 2500 |
+
"loss": 0.243,
|
| 2501 |
+
"step": 20400
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 1.8,
|
| 2505 |
+
"learning_rate": 8.010435904200745e-06,
|
| 2506 |
+
"loss": 0.2478,
|
| 2507 |
+
"step": 20450
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 1.8,
|
| 2511 |
+
"learning_rate": 7.981121566558205e-06,
|
| 2512 |
+
"loss": 0.2477,
|
| 2513 |
+
"step": 20500
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 1.81,
|
| 2517 |
+
"learning_rate": 7.951807228915663e-06,
|
| 2518 |
+
"loss": 0.2414,
|
| 2519 |
+
"step": 20550
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 1.81,
|
| 2523 |
+
"learning_rate": 7.922492891273123e-06,
|
| 2524 |
+
"loss": 0.2488,
|
| 2525 |
+
"step": 20600
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 1.82,
|
| 2529 |
+
"learning_rate": 7.893178553630581e-06,
|
| 2530 |
+
"loss": 0.2498,
|
| 2531 |
+
"step": 20650
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 1.82,
|
| 2535 |
+
"learning_rate": 7.86386421598804e-06,
|
| 2536 |
+
"loss": 0.2532,
|
| 2537 |
+
"step": 20700
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 1.82,
|
| 2541 |
+
"learning_rate": 7.8345498783455e-06,
|
| 2542 |
+
"loss": 0.2199,
|
| 2543 |
+
"step": 20750
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 1.83,
|
| 2547 |
+
"learning_rate": 7.805235540702959e-06,
|
| 2548 |
+
"loss": 0.2398,
|
| 2549 |
+
"step": 20800
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 1.83,
|
| 2553 |
+
"learning_rate": 7.775921203060417e-06,
|
| 2554 |
+
"loss": 0.2197,
|
| 2555 |
+
"step": 20850
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 1.84,
|
| 2559 |
+
"learning_rate": 7.746606865417876e-06,
|
| 2560 |
+
"loss": 0.2322,
|
| 2561 |
+
"step": 20900
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 1.84,
|
| 2565 |
+
"learning_rate": 7.717292527775334e-06,
|
| 2566 |
+
"loss": 0.2081,
|
| 2567 |
+
"step": 20950
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 1.85,
|
| 2571 |
+
"learning_rate": 7.687978190132794e-06,
|
| 2572 |
+
"loss": 0.2536,
|
| 2573 |
+
"step": 21000
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 1.85,
|
| 2577 |
+
"learning_rate": 7.658663852490254e-06,
|
| 2578 |
+
"loss": 0.2343,
|
| 2579 |
+
"step": 21050
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 1.86,
|
| 2583 |
+
"learning_rate": 7.629349514847713e-06,
|
| 2584 |
+
"loss": 0.2423,
|
| 2585 |
+
"step": 21100
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 1.86,
|
| 2589 |
+
"learning_rate": 7.600035177205172e-06,
|
| 2590 |
+
"loss": 0.266,
|
| 2591 |
+
"step": 21150
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 1.86,
|
| 2595 |
+
"learning_rate": 7.570720839562631e-06,
|
| 2596 |
+
"loss": 0.2309,
|
| 2597 |
+
"step": 21200
|
| 2598 |
+
},
|
| 2599 |
+
{
|
| 2600 |
+
"epoch": 1.87,
|
| 2601 |
+
"learning_rate": 7.54140650192009e-06,
|
| 2602 |
+
"loss": 0.2472,
|
| 2603 |
+
"step": 21250
|
| 2604 |
+
},
|
| 2605 |
+
{
|
| 2606 |
+
"epoch": 1.87,
|
| 2607 |
+
"learning_rate": 7.512092164277549e-06,
|
| 2608 |
+
"loss": 0.249,
|
| 2609 |
+
"step": 21300
|
| 2610 |
+
},
|
| 2611 |
+
{
|
| 2612 |
+
"epoch": 1.88,
|
| 2613 |
+
"learning_rate": 7.482777826635008e-06,
|
| 2614 |
+
"loss": 0.2311,
|
| 2615 |
+
"step": 21350
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 1.88,
|
| 2619 |
+
"learning_rate": 7.453463488992467e-06,
|
| 2620 |
+
"loss": 0.2355,
|
| 2621 |
+
"step": 21400
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 1.89,
|
| 2625 |
+
"learning_rate": 7.424149151349925e-06,
|
| 2626 |
+
"loss": 0.2592,
|
| 2627 |
+
"step": 21450
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 1.89,
|
| 2631 |
+
"learning_rate": 7.394834813707384e-06,
|
| 2632 |
+
"loss": 0.2394,
|
| 2633 |
+
"step": 21500
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 1.9,
|
| 2637 |
+
"learning_rate": 7.365520476064844e-06,
|
| 2638 |
+
"loss": 0.2329,
|
| 2639 |
+
"step": 21550
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 1.9,
|
| 2643 |
+
"learning_rate": 7.336206138422303e-06,
|
| 2644 |
+
"loss": 0.2295,
|
| 2645 |
+
"step": 21600
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 1.9,
|
| 2649 |
+
"learning_rate": 7.306891800779762e-06,
|
| 2650 |
+
"loss": 0.2503,
|
| 2651 |
+
"step": 21650
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 1.91,
|
| 2655 |
+
"learning_rate": 7.277577463137221e-06,
|
| 2656 |
+
"loss": 0.2346,
|
| 2657 |
+
"step": 21700
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 1.91,
|
| 2661 |
+
"learning_rate": 7.24826312549468e-06,
|
| 2662 |
+
"loss": 0.2565,
|
| 2663 |
+
"step": 21750
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 1.92,
|
| 2667 |
+
"learning_rate": 7.218948787852139e-06,
|
| 2668 |
+
"loss": 0.2244,
|
| 2669 |
+
"step": 21800
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 1.92,
|
| 2673 |
+
"learning_rate": 7.189634450209598e-06,
|
| 2674 |
+
"loss": 0.2218,
|
| 2675 |
+
"step": 21850
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 1.93,
|
| 2679 |
+
"learning_rate": 7.160320112567057e-06,
|
| 2680 |
+
"loss": 0.2502,
|
| 2681 |
+
"step": 21900
|
| 2682 |
+
},
|
| 2683 |
+
{
|
| 2684 |
+
"epoch": 1.93,
|
| 2685 |
+
"learning_rate": 7.131005774924516e-06,
|
| 2686 |
+
"loss": 0.2257,
|
| 2687 |
+
"step": 21950
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 1.93,
|
| 2691 |
+
"learning_rate": 7.101691437281976e-06,
|
| 2692 |
+
"loss": 0.2244,
|
| 2693 |
+
"step": 22000
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 1.94,
|
| 2697 |
+
"learning_rate": 7.072377099639435e-06,
|
| 2698 |
+
"loss": 0.2475,
|
| 2699 |
+
"step": 22050
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 1.94,
|
| 2703 |
+
"learning_rate": 7.043062761996893e-06,
|
| 2704 |
+
"loss": 0.2452,
|
| 2705 |
+
"step": 22100
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 1.95,
|
| 2709 |
+
"learning_rate": 7.013748424354352e-06,
|
| 2710 |
+
"loss": 0.2274,
|
| 2711 |
+
"step": 22150
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 1.95,
|
| 2715 |
+
"learning_rate": 6.984434086711811e-06,
|
| 2716 |
+
"loss": 0.2314,
|
| 2717 |
+
"step": 22200
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 1.96,
|
| 2721 |
+
"learning_rate": 6.9551197490692704e-06,
|
| 2722 |
+
"loss": 0.2385,
|
| 2723 |
+
"step": 22250
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 1.96,
|
| 2727 |
+
"learning_rate": 6.925805411426729e-06,
|
| 2728 |
+
"loss": 0.2313,
|
| 2729 |
+
"step": 22300
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 1.97,
|
| 2733 |
+
"learning_rate": 6.896491073784188e-06,
|
| 2734 |
+
"loss": 0.2552,
|
| 2735 |
+
"step": 22350
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 1.97,
|
| 2739 |
+
"learning_rate": 6.867176736141647e-06,
|
| 2740 |
+
"loss": 0.2397,
|
| 2741 |
+
"step": 22400
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 1.97,
|
| 2745 |
+
"learning_rate": 6.837862398499107e-06,
|
| 2746 |
+
"loss": 0.2117,
|
| 2747 |
+
"step": 22450
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 1.98,
|
| 2751 |
+
"learning_rate": 6.808548060856566e-06,
|
| 2752 |
+
"loss": 0.2273,
|
| 2753 |
+
"step": 22500
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 1.98,
|
| 2757 |
+
"learning_rate": 6.779233723214025e-06,
|
| 2758 |
+
"loss": 0.2218,
|
| 2759 |
+
"step": 22550
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 1.99,
|
| 2763 |
+
"learning_rate": 6.749919385571484e-06,
|
| 2764 |
+
"loss": 0.2412,
|
| 2765 |
+
"step": 22600
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"epoch": 1.99,
|
| 2769 |
+
"learning_rate": 6.7206050479289434e-06,
|
| 2770 |
+
"loss": 0.2309,
|
| 2771 |
+
"step": 22650
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 2.0,
|
| 2775 |
+
"learning_rate": 6.6912907102864015e-06,
|
| 2776 |
+
"loss": 0.2383,
|
| 2777 |
+
"step": 22700
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 2.0,
|
| 2781 |
+
"learning_rate": 6.66197637264386e-06,
|
| 2782 |
+
"loss": 0.2217,
|
| 2783 |
+
"step": 22750
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.01,
|
| 2787 |
+
"learning_rate": 6.632662035001319e-06,
|
| 2788 |
+
"loss": 0.1919,
|
| 2789 |
+
"step": 22800
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 2.01,
|
| 2793 |
+
"learning_rate": 6.603347697358778e-06,
|
| 2794 |
+
"loss": 0.1773,
|
| 2795 |
+
"step": 22850
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 2.01,
|
| 2799 |
+
"learning_rate": 6.574033359716237e-06,
|
| 2800 |
+
"loss": 0.191,
|
| 2801 |
+
"step": 22900
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 2.02,
|
| 2805 |
+
"learning_rate": 6.544719022073697e-06,
|
| 2806 |
+
"loss": 0.1741,
|
| 2807 |
+
"step": 22950
|
| 2808 |
+
},
|
| 2809 |
+
{
|
| 2810 |
+
"epoch": 2.02,
|
| 2811 |
+
"learning_rate": 6.515404684431156e-06,
|
| 2812 |
+
"loss": 0.2017,
|
| 2813 |
+
"step": 23000
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 2.03,
|
| 2817 |
+
"learning_rate": 6.486090346788615e-06,
|
| 2818 |
+
"loss": 0.1802,
|
| 2819 |
+
"step": 23050
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 2.03,
|
| 2823 |
+
"learning_rate": 6.456776009146074e-06,
|
| 2824 |
+
"loss": 0.1713,
|
| 2825 |
+
"step": 23100
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.04,
|
| 2829 |
+
"learning_rate": 6.427461671503533e-06,
|
| 2830 |
+
"loss": 0.1787,
|
| 2831 |
+
"step": 23150
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 2.04,
|
| 2835 |
+
"learning_rate": 6.398147333860992e-06,
|
| 2836 |
+
"loss": 0.2174,
|
| 2837 |
+
"step": 23200
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 2.04,
|
| 2841 |
+
"learning_rate": 6.368832996218451e-06,
|
| 2842 |
+
"loss": 0.2077,
|
| 2843 |
+
"step": 23250
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 2.05,
|
| 2847 |
+
"learning_rate": 6.33951865857591e-06,
|
| 2848 |
+
"loss": 0.1675,
|
| 2849 |
+
"step": 23300
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 2.05,
|
| 2853 |
+
"learning_rate": 6.310204320933368e-06,
|
| 2854 |
+
"loss": 0.1859,
|
| 2855 |
+
"step": 23350
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 2.06,
|
| 2859 |
+
"learning_rate": 6.280889983290828e-06,
|
| 2860 |
+
"loss": 0.1706,
|
| 2861 |
+
"step": 23400
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 2.06,
|
| 2865 |
+
"learning_rate": 6.251575645648287e-06,
|
| 2866 |
+
"loss": 0.1833,
|
| 2867 |
+
"step": 23450
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 2.07,
|
| 2871 |
+
"learning_rate": 6.222261308005746e-06,
|
| 2872 |
+
"loss": 0.1649,
|
| 2873 |
+
"step": 23500
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 2.07,
|
| 2877 |
+
"learning_rate": 6.192946970363205e-06,
|
| 2878 |
+
"loss": 0.185,
|
| 2879 |
+
"step": 23550
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.08,
|
| 2883 |
+
"learning_rate": 6.1636326327206645e-06,
|
| 2884 |
+
"loss": 0.1773,
|
| 2885 |
+
"step": 23600
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 2.08,
|
| 2889 |
+
"learning_rate": 6.134318295078123e-06,
|
| 2890 |
+
"loss": 0.2054,
|
| 2891 |
+
"step": 23650
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 2.08,
|
| 2895 |
+
"learning_rate": 6.105003957435582e-06,
|
| 2896 |
+
"loss": 0.1757,
|
| 2897 |
+
"step": 23700
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 2.09,
|
| 2901 |
+
"learning_rate": 6.075689619793041e-06,
|
| 2902 |
+
"loss": 0.1938,
|
| 2903 |
+
"step": 23750
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 2.09,
|
| 2907 |
+
"learning_rate": 6.046375282150501e-06,
|
| 2908 |
+
"loss": 0.1999,
|
| 2909 |
+
"step": 23800
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 2.1,
|
| 2913 |
+
"learning_rate": 6.01706094450796e-06,
|
| 2914 |
+
"loss": 0.2059,
|
| 2915 |
+
"step": 23850
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 2.1,
|
| 2919 |
+
"learning_rate": 5.987746606865419e-06,
|
| 2920 |
+
"loss": 0.1761,
|
| 2921 |
+
"step": 23900
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.11,
|
| 2925 |
+
"learning_rate": 5.958432269222877e-06,
|
| 2926 |
+
"loss": 0.1768,
|
| 2927 |
+
"step": 23950
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 2.11,
|
| 2931 |
+
"learning_rate": 5.929117931580336e-06,
|
| 2932 |
+
"loss": 0.1949,
|
| 2933 |
+
"step": 24000
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 2.12,
|
| 2937 |
+
"learning_rate": 5.8998035939377955e-06,
|
| 2938 |
+
"loss": 0.2134,
|
| 2939 |
+
"step": 24050
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 2.12,
|
| 2943 |
+
"learning_rate": 5.8704892562952545e-06,
|
| 2944 |
+
"loss": 0.1839,
|
| 2945 |
+
"step": 24100
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 2.12,
|
| 2949 |
+
"learning_rate": 5.841174918652713e-06,
|
| 2950 |
+
"loss": 0.1592,
|
| 2951 |
+
"step": 24150
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 2.13,
|
| 2955 |
+
"learning_rate": 5.811860581010172e-06,
|
| 2956 |
+
"loss": 0.179,
|
| 2957 |
+
"step": 24200
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 2.13,
|
| 2961 |
+
"learning_rate": 5.782546243367631e-06,
|
| 2962 |
+
"loss": 0.1968,
|
| 2963 |
+
"step": 24250
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 2.14,
|
| 2967 |
+
"learning_rate": 5.753231905725091e-06,
|
| 2968 |
+
"loss": 0.1921,
|
| 2969 |
+
"step": 24300
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 2.14,
|
| 2973 |
+
"learning_rate": 5.72391756808255e-06,
|
| 2974 |
+
"loss": 0.2058,
|
| 2975 |
+
"step": 24350
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 2.15,
|
| 2979 |
+
"learning_rate": 5.694603230440009e-06,
|
| 2980 |
+
"loss": 0.1971,
|
| 2981 |
+
"step": 24400
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 2.15,
|
| 2985 |
+
"learning_rate": 5.665288892797468e-06,
|
| 2986 |
+
"loss": 0.182,
|
| 2987 |
+
"step": 24450
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 2.15,
|
| 2991 |
+
"learning_rate": 5.6359745551549274e-06,
|
| 2992 |
+
"loss": 0.1757,
|
| 2993 |
+
"step": 24500
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 2.16,
|
| 2997 |
+
"learning_rate": 5.606660217512386e-06,
|
| 2998 |
+
"loss": 0.1907,
|
| 2999 |
+
"step": 24550
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 2.16,
|
| 3003 |
+
"learning_rate": 5.5773458798698444e-06,
|
| 3004 |
+
"loss": 0.1923,
|
| 3005 |
+
"step": 24600
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 2.17,
|
| 3009 |
+
"learning_rate": 5.548031542227303e-06,
|
| 3010 |
+
"loss": 0.1792,
|
| 3011 |
+
"step": 24650
|
| 3012 |
+
},
|
| 3013 |
+
{
|
| 3014 |
+
"epoch": 2.17,
|
| 3015 |
+
"learning_rate": 5.518717204584762e-06,
|
| 3016 |
+
"loss": 0.1839,
|
| 3017 |
+
"step": 24700
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 2.18,
|
| 3021 |
+
"learning_rate": 5.489402866942222e-06,
|
| 3022 |
+
"loss": 0.1699,
|
| 3023 |
+
"step": 24750
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 2.18,
|
| 3027 |
+
"learning_rate": 5.460088529299681e-06,
|
| 3028 |
+
"loss": 0.1801,
|
| 3029 |
+
"step": 24800
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 2.19,
|
| 3033 |
+
"learning_rate": 5.43077419165714e-06,
|
| 3034 |
+
"loss": 0.1823,
|
| 3035 |
+
"step": 24850
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 2.19,
|
| 3039 |
+
"learning_rate": 5.401459854014599e-06,
|
| 3040 |
+
"loss": 0.1741,
|
| 3041 |
+
"step": 24900
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 2.19,
|
| 3045 |
+
"learning_rate": 5.3721455163720585e-06,
|
| 3046 |
+
"loss": 0.2303,
|
| 3047 |
+
"step": 24950
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 2.2,
|
| 3051 |
+
"learning_rate": 5.3428311787295174e-06,
|
| 3052 |
+
"loss": 0.1933,
|
| 3053 |
+
"step": 25000
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 2.2,
|
| 3057 |
+
"eval_accuracy": 0.8866930497155577,
|
| 3058 |
+
"eval_combined_score": 0.8668614011392094,
|
| 3059 |
+
"eval_f1": 0.8470297525628611,
|
| 3060 |
+
"eval_loss": 0.2961539924144745,
|
| 3061 |
+
"eval_runtime": 51.9998,
|
| 3062 |
+
"eval_samples_per_second": 777.503,
|
| 3063 |
+
"eval_steps_per_second": 48.596,
|
| 3064 |
+
"step": 25000
|
| 3065 |
+
},
|
| 3066 |
+
{
|
| 3067 |
+
"epoch": 2.2,
|
| 3068 |
+
"learning_rate": 5.313516841086976e-06,
|
| 3069 |
+
"loss": 0.1978,
|
| 3070 |
+
"step": 25050
|
| 3071 |
+
},
|
| 3072 |
+
{
|
| 3073 |
+
"epoch": 2.21,
|
| 3074 |
+
"learning_rate": 5.284202503444435e-06,
|
| 3075 |
+
"loss": 0.1708,
|
| 3076 |
+
"step": 25100
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 2.21,
|
| 3080 |
+
"learning_rate": 5.254888165801895e-06,
|
| 3081 |
+
"loss": 0.1674,
|
| 3082 |
+
"step": 25150
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 2.22,
|
| 3086 |
+
"learning_rate": 5.225573828159353e-06,
|
| 3087 |
+
"loss": 0.1804,
|
| 3088 |
+
"step": 25200
|
| 3089 |
+
},
|
| 3090 |
+
{
|
| 3091 |
+
"epoch": 2.22,
|
| 3092 |
+
"learning_rate": 5.196259490516812e-06,
|
| 3093 |
+
"loss": 0.2008,
|
| 3094 |
+
"step": 25250
|
| 3095 |
+
},
|
| 3096 |
+
{
|
| 3097 |
+
"epoch": 2.22,
|
| 3098 |
+
"learning_rate": 5.166945152874271e-06,
|
| 3099 |
+
"loss": 0.1639,
|
| 3100 |
+
"step": 25300
|
| 3101 |
+
},
|
| 3102 |
+
{
|
| 3103 |
+
"epoch": 2.23,
|
| 3104 |
+
"learning_rate": 5.13763081523173e-06,
|
| 3105 |
+
"loss": 0.181,
|
| 3106 |
+
"step": 25350
|
| 3107 |
+
},
|
| 3108 |
+
{
|
| 3109 |
+
"epoch": 2.23,
|
| 3110 |
+
"learning_rate": 5.108316477589189e-06,
|
| 3111 |
+
"loss": 0.1856,
|
| 3112 |
+
"step": 25400
|
| 3113 |
+
},
|
| 3114 |
+
{
|
| 3115 |
+
"epoch": 2.24,
|
| 3116 |
+
"learning_rate": 5.0790021399466485e-06,
|
| 3117 |
+
"loss": 0.1882,
|
| 3118 |
+
"step": 25450
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 2.24,
|
| 3122 |
+
"learning_rate": 5.049687802304107e-06,
|
| 3123 |
+
"loss": 0.2002,
|
| 3124 |
+
"step": 25500
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 2.25,
|
| 3128 |
+
"learning_rate": 5.020373464661566e-06,
|
| 3129 |
+
"loss": 0.1728,
|
| 3130 |
+
"step": 25550
|
| 3131 |
+
},
|
| 3132 |
+
{
|
| 3133 |
+
"epoch": 2.25,
|
| 3134 |
+
"learning_rate": 4.991059127019025e-06,
|
| 3135 |
+
"loss": 0.1722,
|
| 3136 |
+
"step": 25600
|
| 3137 |
+
},
|
| 3138 |
+
{
|
| 3139 |
+
"epoch": 2.26,
|
| 3140 |
+
"learning_rate": 4.961744789376484e-06,
|
| 3141 |
+
"loss": 0.1615,
|
| 3142 |
+
"step": 25650
|
| 3143 |
+
},
|
| 3144 |
+
{
|
| 3145 |
+
"epoch": 2.26,
|
| 3146 |
+
"learning_rate": 4.932430451733943e-06,
|
| 3147 |
+
"loss": 0.172,
|
| 3148 |
+
"step": 25700
|
| 3149 |
+
},
|
| 3150 |
+
{
|
| 3151 |
+
"epoch": 2.26,
|
| 3152 |
+
"learning_rate": 4.903116114091403e-06,
|
| 3153 |
+
"loss": 0.158,
|
| 3154 |
+
"step": 25750
|
| 3155 |
+
},
|
| 3156 |
+
{
|
| 3157 |
+
"epoch": 2.27,
|
| 3158 |
+
"learning_rate": 4.873801776448862e-06,
|
| 3159 |
+
"loss": 0.1692,
|
| 3160 |
+
"step": 25800
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 2.27,
|
| 3164 |
+
"learning_rate": 4.844487438806321e-06,
|
| 3165 |
+
"loss": 0.1603,
|
| 3166 |
+
"step": 25850
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 2.28,
|
| 3170 |
+
"learning_rate": 4.8151731011637796e-06,
|
| 3171 |
+
"loss": 0.1776,
|
| 3172 |
+
"step": 25900
|
| 3173 |
+
},
|
| 3174 |
+
{
|
| 3175 |
+
"epoch": 2.28,
|
| 3176 |
+
"learning_rate": 4.785858763521239e-06,
|
| 3177 |
+
"loss": 0.2094,
|
| 3178 |
+
"step": 25950
|
| 3179 |
+
},
|
| 3180 |
+
{
|
| 3181 |
+
"epoch": 2.29,
|
| 3182 |
+
"learning_rate": 4.756544425878697e-06,
|
| 3183 |
+
"loss": 0.2003,
|
| 3184 |
+
"step": 26000
|
| 3185 |
+
},
|
| 3186 |
+
{
|
| 3187 |
+
"epoch": 2.29,
|
| 3188 |
+
"learning_rate": 4.727230088236156e-06,
|
| 3189 |
+
"loss": 0.1595,
|
| 3190 |
+
"step": 26050
|
| 3191 |
+
},
|
| 3192 |
+
{
|
| 3193 |
+
"epoch": 2.3,
|
| 3194 |
+
"learning_rate": 4.697915750593616e-06,
|
| 3195 |
+
"loss": 0.1782,
|
| 3196 |
+
"step": 26100
|
| 3197 |
+
},
|
| 3198 |
+
{
|
| 3199 |
+
"epoch": 2.3,
|
| 3200 |
+
"learning_rate": 4.668601412951075e-06,
|
| 3201 |
+
"loss": 0.1713,
|
| 3202 |
+
"step": 26150
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 2.3,
|
| 3206 |
+
"learning_rate": 4.639287075308534e-06,
|
| 3207 |
+
"loss": 0.1746,
|
| 3208 |
+
"step": 26200
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 2.31,
|
| 3212 |
+
"learning_rate": 4.609972737665993e-06,
|
| 3213 |
+
"loss": 0.1829,
|
| 3214 |
+
"step": 26250
|
| 3215 |
+
},
|
| 3216 |
+
{
|
| 3217 |
+
"epoch": 2.31,
|
| 3218 |
+
"learning_rate": 4.580658400023452e-06,
|
| 3219 |
+
"loss": 0.192,
|
| 3220 |
+
"step": 26300
|
| 3221 |
+
},
|
| 3222 |
+
{
|
| 3223 |
+
"epoch": 2.32,
|
| 3224 |
+
"learning_rate": 4.551344062380911e-06,
|
| 3225 |
+
"loss": 0.1756,
|
| 3226 |
+
"step": 26350
|
| 3227 |
+
},
|
| 3228 |
+
{
|
| 3229 |
+
"epoch": 2.32,
|
| 3230 |
+
"learning_rate": 4.5220297247383695e-06,
|
| 3231 |
+
"loss": 0.1954,
|
| 3232 |
+
"step": 26400
|
| 3233 |
+
},
|
| 3234 |
+
{
|
| 3235 |
+
"epoch": 2.33,
|
| 3236 |
+
"learning_rate": 4.492715387095829e-06,
|
| 3237 |
+
"loss": 0.1786,
|
| 3238 |
+
"step": 26450
|
| 3239 |
+
},
|
| 3240 |
+
{
|
| 3241 |
+
"epoch": 2.33,
|
| 3242 |
+
"learning_rate": 4.463401049453288e-06,
|
| 3243 |
+
"loss": 0.1837,
|
| 3244 |
+
"step": 26500
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 2.33,
|
| 3248 |
+
"learning_rate": 4.434086711810747e-06,
|
| 3249 |
+
"loss": 0.1581,
|
| 3250 |
+
"step": 26550
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 2.34,
|
| 3254 |
+
"learning_rate": 4.404772374168206e-06,
|
| 3255 |
+
"loss": 0.1969,
|
| 3256 |
+
"step": 26600
|
| 3257 |
+
},
|
| 3258 |
+
{
|
| 3259 |
+
"epoch": 2.34,
|
| 3260 |
+
"learning_rate": 4.375458036525665e-06,
|
| 3261 |
+
"loss": 0.168,
|
| 3262 |
+
"step": 26650
|
| 3263 |
+
},
|
| 3264 |
+
{
|
| 3265 |
+
"epoch": 2.35,
|
| 3266 |
+
"learning_rate": 4.346143698883124e-06,
|
| 3267 |
+
"loss": 0.1771,
|
| 3268 |
+
"step": 26700
|
| 3269 |
+
},
|
| 3270 |
+
{
|
| 3271 |
+
"epoch": 2.35,
|
| 3272 |
+
"learning_rate": 4.316829361240583e-06,
|
| 3273 |
+
"loss": 0.1569,
|
| 3274 |
+
"step": 26750
|
| 3275 |
+
},
|
| 3276 |
+
{
|
| 3277 |
+
"epoch": 2.36,
|
| 3278 |
+
"learning_rate": 4.2875150235980425e-06,
|
| 3279 |
+
"loss": 0.1493,
|
| 3280 |
+
"step": 26800
|
| 3281 |
+
},
|
| 3282 |
+
{
|
| 3283 |
+
"epoch": 2.36,
|
| 3284 |
+
"learning_rate": 4.2582006859555014e-06,
|
| 3285 |
+
"loss": 0.1958,
|
| 3286 |
+
"step": 26850
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 2.37,
|
| 3290 |
+
"learning_rate": 4.22888634831296e-06,
|
| 3291 |
+
"loss": 0.1786,
|
| 3292 |
+
"step": 26900
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 2.37,
|
| 3296 |
+
"learning_rate": 4.199572010670419e-06,
|
| 3297 |
+
"loss": 0.1958,
|
| 3298 |
+
"step": 26950
|
| 3299 |
+
},
|
| 3300 |
+
{
|
| 3301 |
+
"epoch": 2.37,
|
| 3302 |
+
"learning_rate": 4.170257673027878e-06,
|
| 3303 |
+
"loss": 0.1769,
|
| 3304 |
+
"step": 27000
|
| 3305 |
+
},
|
| 3306 |
+
{
|
| 3307 |
+
"epoch": 2.38,
|
| 3308 |
+
"learning_rate": 4.140943335385337e-06,
|
| 3309 |
+
"loss": 0.1805,
|
| 3310 |
+
"step": 27050
|
| 3311 |
+
},
|
| 3312 |
+
{
|
| 3313 |
+
"epoch": 2.38,
|
| 3314 |
+
"learning_rate": 4.111628997742797e-06,
|
| 3315 |
+
"loss": 0.1883,
|
| 3316 |
+
"step": 27100
|
| 3317 |
+
},
|
| 3318 |
+
{
|
| 3319 |
+
"epoch": 2.39,
|
| 3320 |
+
"learning_rate": 4.082314660100256e-06,
|
| 3321 |
+
"loss": 0.1765,
|
| 3322 |
+
"step": 27150
|
| 3323 |
+
},
|
| 3324 |
+
{
|
| 3325 |
+
"epoch": 2.39,
|
| 3326 |
+
"learning_rate": 4.053000322457714e-06,
|
| 3327 |
+
"loss": 0.1831,
|
| 3328 |
+
"step": 27200
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 2.4,
|
| 3332 |
+
"learning_rate": 4.023685984815174e-06,
|
| 3333 |
+
"loss": 0.1805,
|
| 3334 |
+
"step": 27250
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 2.4,
|
| 3338 |
+
"learning_rate": 3.9943716471726325e-06,
|
| 3339 |
+
"loss": 0.1805,
|
| 3340 |
+
"step": 27300
|
| 3341 |
+
},
|
| 3342 |
+
{
|
| 3343 |
+
"epoch": 2.41,
|
| 3344 |
+
"learning_rate": 3.965057309530091e-06,
|
| 3345 |
+
"loss": 0.1683,
|
| 3346 |
+
"step": 27350
|
| 3347 |
+
},
|
| 3348 |
+
{
|
| 3349 |
+
"epoch": 2.41,
|
| 3350 |
+
"learning_rate": 3.93574297188755e-06,
|
| 3351 |
+
"loss": 0.1859,
|
| 3352 |
+
"step": 27400
|
| 3353 |
+
},
|
| 3354 |
+
{
|
| 3355 |
+
"epoch": 2.41,
|
| 3356 |
+
"learning_rate": 3.90642863424501e-06,
|
| 3357 |
+
"loss": 0.1763,
|
| 3358 |
+
"step": 27450
|
| 3359 |
+
},
|
| 3360 |
+
{
|
| 3361 |
+
"epoch": 2.42,
|
| 3362 |
+
"learning_rate": 3.877114296602469e-06,
|
| 3363 |
+
"loss": 0.1791,
|
| 3364 |
+
"step": 27500
|
| 3365 |
+
},
|
| 3366 |
+
{
|
| 3367 |
+
"epoch": 2.42,
|
| 3368 |
+
"learning_rate": 3.847799958959927e-06,
|
| 3369 |
+
"loss": 0.1715,
|
| 3370 |
+
"step": 27550
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 2.43,
|
| 3374 |
+
"learning_rate": 3.818485621317387e-06,
|
| 3375 |
+
"loss": 0.1885,
|
| 3376 |
+
"step": 27600
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 2.43,
|
| 3380 |
+
"learning_rate": 3.7891712836748457e-06,
|
| 3381 |
+
"loss": 0.151,
|
| 3382 |
+
"step": 27650
|
| 3383 |
+
},
|
| 3384 |
+
{
|
| 3385 |
+
"epoch": 2.44,
|
| 3386 |
+
"learning_rate": 3.7598569460323047e-06,
|
| 3387 |
+
"loss": 0.1733,
|
| 3388 |
+
"step": 27700
|
| 3389 |
+
},
|
| 3390 |
+
{
|
| 3391 |
+
"epoch": 2.44,
|
| 3392 |
+
"learning_rate": 3.730542608389764e-06,
|
| 3393 |
+
"loss": 0.1892,
|
| 3394 |
+
"step": 27750
|
| 3395 |
+
},
|
| 3396 |
+
{
|
| 3397 |
+
"epoch": 2.44,
|
| 3398 |
+
"learning_rate": 3.701228270747223e-06,
|
| 3399 |
+
"loss": 0.1922,
|
| 3400 |
+
"step": 27800
|
| 3401 |
+
},
|
| 3402 |
+
{
|
| 3403 |
+
"epoch": 2.45,
|
| 3404 |
+
"learning_rate": 3.6719139331046814e-06,
|
| 3405 |
+
"loss": 0.1585,
|
| 3406 |
+
"step": 27850
|
| 3407 |
+
},
|
| 3408 |
+
{
|
| 3409 |
+
"epoch": 2.45,
|
| 3410 |
+
"learning_rate": 3.6425995954621407e-06,
|
| 3411 |
+
"loss": 0.1685,
|
| 3412 |
+
"step": 27900
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 2.46,
|
| 3416 |
+
"learning_rate": 3.6132852578195996e-06,
|
| 3417 |
+
"loss": 0.1873,
|
| 3418 |
+
"step": 27950
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 2.46,
|
| 3422 |
+
"learning_rate": 3.583970920177059e-06,
|
| 3423 |
+
"loss": 0.1941,
|
| 3424 |
+
"step": 28000
|
| 3425 |
+
},
|
| 3426 |
+
{
|
| 3427 |
+
"epoch": 2.47,
|
| 3428 |
+
"learning_rate": 3.554656582534518e-06,
|
| 3429 |
+
"loss": 0.1764,
|
| 3430 |
+
"step": 28050
|
| 3431 |
+
},
|
| 3432 |
+
{
|
| 3433 |
+
"epoch": 2.47,
|
| 3434 |
+
"learning_rate": 3.5253422448919772e-06,
|
| 3435 |
+
"loss": 0.182,
|
| 3436 |
+
"step": 28100
|
| 3437 |
+
},
|
| 3438 |
+
{
|
| 3439 |
+
"epoch": 2.48,
|
| 3440 |
+
"learning_rate": 3.4960279072494357e-06,
|
| 3441 |
+
"loss": 0.2042,
|
| 3442 |
+
"step": 28150
|
| 3443 |
+
},
|
| 3444 |
+
{
|
| 3445 |
+
"epoch": 2.48,
|
| 3446 |
+
"learning_rate": 3.466713569606895e-06,
|
| 3447 |
+
"loss": 0.1863,
|
| 3448 |
+
"step": 28200
|
| 3449 |
+
},
|
| 3450 |
+
{
|
| 3451 |
+
"epoch": 2.48,
|
| 3452 |
+
"learning_rate": 3.437399231964354e-06,
|
| 3453 |
+
"loss": 0.172,
|
| 3454 |
+
"step": 28250
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 2.49,
|
| 3458 |
+
"learning_rate": 3.4080848943218133e-06,
|
| 3459 |
+
"loss": 0.1928,
|
| 3460 |
+
"step": 28300
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 2.49,
|
| 3464 |
+
"learning_rate": 3.3787705566792722e-06,
|
| 3465 |
+
"loss": 0.1892,
|
| 3466 |
+
"step": 28350
|
| 3467 |
+
},
|
| 3468 |
+
{
|
| 3469 |
+
"epoch": 2.5,
|
| 3470 |
+
"learning_rate": 3.3494562190367316e-06,
|
| 3471 |
+
"loss": 0.1878,
|
| 3472 |
+
"step": 28400
|
| 3473 |
+
},
|
| 3474 |
+
{
|
| 3475 |
+
"epoch": 2.5,
|
| 3476 |
+
"learning_rate": 3.32014188139419e-06,
|
| 3477 |
+
"loss": 0.1673,
|
| 3478 |
+
"step": 28450
|
| 3479 |
+
},
|
| 3480 |
+
{
|
| 3481 |
+
"epoch": 2.51,
|
| 3482 |
+
"learning_rate": 3.290827543751649e-06,
|
| 3483 |
+
"loss": 0.1936,
|
| 3484 |
+
"step": 28500
|
| 3485 |
+
},
|
| 3486 |
+
{
|
| 3487 |
+
"epoch": 2.51,
|
| 3488 |
+
"learning_rate": 3.2615132061091083e-06,
|
| 3489 |
+
"loss": 0.1826,
|
| 3490 |
+
"step": 28550
|
| 3491 |
+
},
|
| 3492 |
+
{
|
| 3493 |
+
"epoch": 2.52,
|
| 3494 |
+
"learning_rate": 3.232198868466567e-06,
|
| 3495 |
+
"loss": 0.1773,
|
| 3496 |
+
"step": 28600
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 2.52,
|
| 3500 |
+
"learning_rate": 3.2028845308240265e-06,
|
| 3501 |
+
"loss": 0.1682,
|
| 3502 |
+
"step": 28650
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 2.52,
|
| 3506 |
+
"learning_rate": 3.1735701931814855e-06,
|
| 3507 |
+
"loss": 0.1768,
|
| 3508 |
+
"step": 28700
|
| 3509 |
+
},
|
| 3510 |
+
{
|
| 3511 |
+
"epoch": 2.53,
|
| 3512 |
+
"learning_rate": 3.144255855538944e-06,
|
| 3513 |
+
"loss": 0.1816,
|
| 3514 |
+
"step": 28750
|
| 3515 |
+
},
|
| 3516 |
+
{
|
| 3517 |
+
"epoch": 2.53,
|
| 3518 |
+
"learning_rate": 3.1149415178964033e-06,
|
| 3519 |
+
"loss": 0.1692,
|
| 3520 |
+
"step": 28800
|
| 3521 |
+
},
|
| 3522 |
+
{
|
| 3523 |
+
"epoch": 2.54,
|
| 3524 |
+
"learning_rate": 3.085627180253862e-06,
|
| 3525 |
+
"loss": 0.1745,
|
| 3526 |
+
"step": 28850
|
| 3527 |
+
},
|
| 3528 |
+
{
|
| 3529 |
+
"epoch": 2.54,
|
| 3530 |
+
"learning_rate": 3.0563128426113215e-06,
|
| 3531 |
+
"loss": 0.1848,
|
| 3532 |
+
"step": 28900
|
| 3533 |
+
},
|
| 3534 |
+
{
|
| 3535 |
+
"epoch": 2.55,
|
| 3536 |
+
"learning_rate": 3.0269985049687804e-06,
|
| 3537 |
+
"loss": 0.1407,
|
| 3538 |
+
"step": 28950
|
| 3539 |
+
},
|
| 3540 |
+
{
|
| 3541 |
+
"epoch": 2.55,
|
| 3542 |
+
"learning_rate": 2.9976841673262398e-06,
|
| 3543 |
+
"loss": 0.1686,
|
| 3544 |
+
"step": 29000
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 2.55,
|
| 3548 |
+
"learning_rate": 2.9683698296836987e-06,
|
| 3549 |
+
"loss": 0.1956,
|
| 3550 |
+
"step": 29050
|
| 3551 |
+
},
|
| 3552 |
+
{
|
| 3553 |
+
"epoch": 2.56,
|
| 3554 |
+
"learning_rate": 2.939055492041157e-06,
|
| 3555 |
+
"loss": 0.1682,
|
| 3556 |
+
"step": 29100
|
| 3557 |
+
},
|
| 3558 |
+
{
|
| 3559 |
+
"epoch": 2.56,
|
| 3560 |
+
"learning_rate": 2.9097411543986165e-06,
|
| 3561 |
+
"loss": 0.1623,
|
| 3562 |
+
"step": 29150
|
| 3563 |
+
},
|
| 3564 |
+
{
|
| 3565 |
+
"epoch": 2.57,
|
| 3566 |
+
"learning_rate": 2.8804268167560754e-06,
|
| 3567 |
+
"loss": 0.1719,
|
| 3568 |
+
"step": 29200
|
| 3569 |
+
},
|
| 3570 |
+
{
|
| 3571 |
+
"epoch": 2.57,
|
| 3572 |
+
"learning_rate": 2.8511124791135348e-06,
|
| 3573 |
+
"loss": 0.1943,
|
| 3574 |
+
"step": 29250
|
| 3575 |
+
},
|
| 3576 |
+
{
|
| 3577 |
+
"epoch": 2.58,
|
| 3578 |
+
"learning_rate": 2.8217981414709937e-06,
|
| 3579 |
+
"loss": 0.181,
|
| 3580 |
+
"step": 29300
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"epoch": 2.58,
|
| 3584 |
+
"learning_rate": 2.792483803828453e-06,
|
| 3585 |
+
"loss": 0.192,
|
| 3586 |
+
"step": 29350
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 2.59,
|
| 3590 |
+
"learning_rate": 2.7631694661859115e-06,
|
| 3591 |
+
"loss": 0.1913,
|
| 3592 |
+
"step": 29400
|
| 3593 |
+
},
|
| 3594 |
+
{
|
| 3595 |
+
"epoch": 2.59,
|
| 3596 |
+
"learning_rate": 2.733855128543371e-06,
|
| 3597 |
+
"loss": 0.1489,
|
| 3598 |
+
"step": 29450
|
| 3599 |
+
},
|
| 3600 |
+
{
|
| 3601 |
+
"epoch": 2.59,
|
| 3602 |
+
"learning_rate": 2.7045407909008298e-06,
|
| 3603 |
+
"loss": 0.1676,
|
| 3604 |
+
"step": 29500
|
| 3605 |
+
},
|
| 3606 |
+
{
|
| 3607 |
+
"epoch": 2.6,
|
| 3608 |
+
"learning_rate": 2.675226453258289e-06,
|
| 3609 |
+
"loss": 0.1583,
|
| 3610 |
+
"step": 29550
|
| 3611 |
+
},
|
| 3612 |
+
{
|
| 3613 |
+
"epoch": 2.6,
|
| 3614 |
+
"learning_rate": 2.645912115615748e-06,
|
| 3615 |
+
"loss": 0.1836,
|
| 3616 |
+
"step": 29600
|
| 3617 |
+
},
|
| 3618 |
+
{
|
| 3619 |
+
"epoch": 2.61,
|
| 3620 |
+
"learning_rate": 2.616597777973207e-06,
|
| 3621 |
+
"loss": 0.1613,
|
| 3622 |
+
"step": 29650
|
| 3623 |
+
},
|
| 3624 |
+
{
|
| 3625 |
+
"epoch": 2.61,
|
| 3626 |
+
"learning_rate": 2.587283440330666e-06,
|
| 3627 |
+
"loss": 0.175,
|
| 3628 |
+
"step": 29700
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 2.62,
|
| 3632 |
+
"learning_rate": 2.5579691026881247e-06,
|
| 3633 |
+
"loss": 0.1924,
|
| 3634 |
+
"step": 29750
|
| 3635 |
+
},
|
| 3636 |
+
{
|
| 3637 |
+
"epoch": 2.62,
|
| 3638 |
+
"learning_rate": 2.528654765045584e-06,
|
| 3639 |
+
"loss": 0.1799,
|
| 3640 |
+
"step": 29800
|
| 3641 |
+
},
|
| 3642 |
+
{
|
| 3643 |
+
"epoch": 2.63,
|
| 3644 |
+
"learning_rate": 2.499340427403043e-06,
|
| 3645 |
+
"loss": 0.1701,
|
| 3646 |
+
"step": 29850
|
| 3647 |
+
},
|
| 3648 |
+
{
|
| 3649 |
+
"epoch": 2.63,
|
| 3650 |
+
"learning_rate": 2.4700260897605023e-06,
|
| 3651 |
+
"loss": 0.1517,
|
| 3652 |
+
"step": 29900
|
| 3653 |
+
},
|
| 3654 |
+
{
|
| 3655 |
+
"epoch": 2.63,
|
| 3656 |
+
"learning_rate": 2.440711752117961e-06,
|
| 3657 |
+
"loss": 0.1427,
|
| 3658 |
+
"step": 29950
|
| 3659 |
+
},
|
| 3660 |
+
{
|
| 3661 |
+
"epoch": 2.64,
|
| 3662 |
+
"learning_rate": 2.41139741447542e-06,
|
| 3663 |
+
"loss": 0.1752,
|
| 3664 |
+
"step": 30000
|
| 3665 |
+
},
|
| 3666 |
+
{
|
| 3667 |
+
"epoch": 2.64,
|
| 3668 |
+
"eval_accuracy": 0.8895127380657927,
|
| 3669 |
+
"eval_combined_score": 0.8696125410209191,
|
| 3670 |
+
"eval_f1": 0.8497123439760456,
|
| 3671 |
+
"eval_loss": 0.3174465298652649,
|
| 3672 |
+
"eval_runtime": 51.9237,
|
| 3673 |
+
"eval_samples_per_second": 778.643,
|
| 3674 |
+
"eval_steps_per_second": 48.668,
|
| 3675 |
+
"step": 30000
|
| 3676 |
+
},
|
| 3677 |
+
{
|
| 3678 |
+
"epoch": 2.64,
|
| 3679 |
+
"learning_rate": 2.382083076832879e-06,
|
| 3680 |
+
"loss": 0.1827,
|
| 3681 |
+
"step": 30050
|
| 3682 |
+
},
|
| 3683 |
+
{
|
| 3684 |
+
"epoch": 2.65,
|
| 3685 |
+
"learning_rate": 2.352768739190338e-06,
|
| 3686 |
+
"loss": 0.1958,
|
| 3687 |
+
"step": 30100
|
| 3688 |
+
},
|
| 3689 |
+
{
|
| 3690 |
+
"epoch": 2.65,
|
| 3691 |
+
"learning_rate": 2.3234544015477973e-06,
|
| 3692 |
+
"loss": 0.1718,
|
| 3693 |
+
"step": 30150
|
| 3694 |
+
},
|
| 3695 |
+
{
|
| 3696 |
+
"epoch": 2.66,
|
| 3697 |
+
"learning_rate": 2.2941400639052562e-06,
|
| 3698 |
+
"loss": 0.1857,
|
| 3699 |
+
"step": 30200
|
| 3700 |
+
},
|
| 3701 |
+
{
|
| 3702 |
+
"epoch": 2.66,
|
| 3703 |
+
"learning_rate": 2.264825726262715e-06,
|
| 3704 |
+
"loss": 0.1652,
|
| 3705 |
+
"step": 30250
|
| 3706 |
+
},
|
| 3707 |
+
{
|
| 3708 |
+
"epoch": 2.66,
|
| 3709 |
+
"learning_rate": 2.2355113886201745e-06,
|
| 3710 |
+
"loss": 0.1764,
|
| 3711 |
+
"step": 30300
|
| 3712 |
+
},
|
| 3713 |
+
{
|
| 3714 |
+
"epoch": 2.67,
|
| 3715 |
+
"learning_rate": 2.2061970509776334e-06,
|
| 3716 |
+
"loss": 0.1637,
|
| 3717 |
+
"step": 30350
|
| 3718 |
+
},
|
| 3719 |
+
{
|
| 3720 |
+
"epoch": 2.67,
|
| 3721 |
+
"learning_rate": 2.1768827133350923e-06,
|
| 3722 |
+
"loss": 0.1635,
|
| 3723 |
+
"step": 30400
|
| 3724 |
+
},
|
| 3725 |
+
{
|
| 3726 |
+
"epoch": 2.68,
|
| 3727 |
+
"learning_rate": 2.1475683756925512e-06,
|
| 3728 |
+
"loss": 0.169,
|
| 3729 |
+
"step": 30450
|
| 3730 |
+
},
|
| 3731 |
+
{
|
| 3732 |
+
"epoch": 2.68,
|
| 3733 |
+
"learning_rate": 2.1182540380500106e-06,
|
| 3734 |
+
"loss": 0.1804,
|
| 3735 |
+
"step": 30500
|
| 3736 |
+
},
|
| 3737 |
+
{
|
| 3738 |
+
"epoch": 2.69,
|
| 3739 |
+
"learning_rate": 2.0889397004074695e-06,
|
| 3740 |
+
"loss": 0.1613,
|
| 3741 |
+
"step": 30550
|
| 3742 |
+
},
|
| 3743 |
+
{
|
| 3744 |
+
"epoch": 2.69,
|
| 3745 |
+
"learning_rate": 2.0596253627649284e-06,
|
| 3746 |
+
"loss": 0.1771,
|
| 3747 |
+
"step": 30600
|
| 3748 |
+
},
|
| 3749 |
+
{
|
| 3750 |
+
"epoch": 2.7,
|
| 3751 |
+
"learning_rate": 2.0303110251223877e-06,
|
| 3752 |
+
"loss": 0.1969,
|
| 3753 |
+
"step": 30650
|
| 3754 |
+
},
|
| 3755 |
+
{
|
| 3756 |
+
"epoch": 2.7,
|
| 3757 |
+
"learning_rate": 2.0009966874798466e-06,
|
| 3758 |
+
"loss": 0.1701,
|
| 3759 |
+
"step": 30700
|
| 3760 |
+
},
|
| 3761 |
+
{
|
| 3762 |
+
"epoch": 2.7,
|
| 3763 |
+
"learning_rate": 1.9716823498373055e-06,
|
| 3764 |
+
"loss": 0.1634,
|
| 3765 |
+
"step": 30750
|
| 3766 |
+
},
|
| 3767 |
+
{
|
| 3768 |
+
"epoch": 2.71,
|
| 3769 |
+
"learning_rate": 1.9423680121947645e-06,
|
| 3770 |
+
"loss": 0.1758,
|
| 3771 |
+
"step": 30800
|
| 3772 |
+
},
|
| 3773 |
+
{
|
| 3774 |
+
"epoch": 2.71,
|
| 3775 |
+
"learning_rate": 1.9130536745522234e-06,
|
| 3776 |
+
"loss": 0.1659,
|
| 3777 |
+
"step": 30850
|
| 3778 |
+
},
|
| 3779 |
+
{
|
| 3780 |
+
"epoch": 2.72,
|
| 3781 |
+
"learning_rate": 1.8837393369096827e-06,
|
| 3782 |
+
"loss": 0.1865,
|
| 3783 |
+
"step": 30900
|
| 3784 |
+
},
|
| 3785 |
+
{
|
| 3786 |
+
"epoch": 2.72,
|
| 3787 |
+
"learning_rate": 1.8544249992671418e-06,
|
| 3788 |
+
"loss": 0.2038,
|
| 3789 |
+
"step": 30950
|
| 3790 |
+
},
|
| 3791 |
+
{
|
| 3792 |
+
"epoch": 2.73,
|
| 3793 |
+
"learning_rate": 1.825110661624601e-06,
|
| 3794 |
+
"loss": 0.1647,
|
| 3795 |
+
"step": 31000
|
| 3796 |
+
},
|
| 3797 |
+
{
|
| 3798 |
+
"epoch": 2.73,
|
| 3799 |
+
"learning_rate": 1.7957963239820597e-06,
|
| 3800 |
+
"loss": 0.1783,
|
| 3801 |
+
"step": 31050
|
| 3802 |
+
},
|
| 3803 |
+
{
|
| 3804 |
+
"epoch": 2.74,
|
| 3805 |
+
"learning_rate": 1.7664819863395188e-06,
|
| 3806 |
+
"loss": 0.167,
|
| 3807 |
+
"step": 31100
|
| 3808 |
+
},
|
| 3809 |
+
{
|
| 3810 |
+
"epoch": 2.74,
|
| 3811 |
+
"learning_rate": 1.737167648696978e-06,
|
| 3812 |
+
"loss": 0.1748,
|
| 3813 |
+
"step": 31150
|
| 3814 |
+
},
|
| 3815 |
+
{
|
| 3816 |
+
"epoch": 2.74,
|
| 3817 |
+
"learning_rate": 1.7078533110544368e-06,
|
| 3818 |
+
"loss": 0.1682,
|
| 3819 |
+
"step": 31200
|
| 3820 |
+
},
|
| 3821 |
+
{
|
| 3822 |
+
"epoch": 2.75,
|
| 3823 |
+
"learning_rate": 1.678538973411896e-06,
|
| 3824 |
+
"loss": 0.1964,
|
| 3825 |
+
"step": 31250
|
| 3826 |
+
},
|
| 3827 |
+
{
|
| 3828 |
+
"epoch": 2.75,
|
| 3829 |
+
"learning_rate": 1.649224635769355e-06,
|
| 3830 |
+
"loss": 0.1587,
|
| 3831 |
+
"step": 31300
|
| 3832 |
+
},
|
| 3833 |
+
{
|
| 3834 |
+
"epoch": 2.76,
|
| 3835 |
+
"learning_rate": 1.619910298126814e-06,
|
| 3836 |
+
"loss": 0.1753,
|
| 3837 |
+
"step": 31350
|
| 3838 |
+
},
|
| 3839 |
+
{
|
| 3840 |
+
"epoch": 2.76,
|
| 3841 |
+
"learning_rate": 1.590595960484273e-06,
|
| 3842 |
+
"loss": 0.1782,
|
| 3843 |
+
"step": 31400
|
| 3844 |
+
},
|
| 3845 |
+
{
|
| 3846 |
+
"epoch": 2.77,
|
| 3847 |
+
"learning_rate": 1.561281622841732e-06,
|
| 3848 |
+
"loss": 0.1825,
|
| 3849 |
+
"step": 31450
|
| 3850 |
+
},
|
| 3851 |
+
{
|
| 3852 |
+
"epoch": 2.77,
|
| 3853 |
+
"learning_rate": 1.531967285199191e-06,
|
| 3854 |
+
"loss": 0.2297,
|
| 3855 |
+
"step": 31500
|
| 3856 |
+
},
|
| 3857 |
+
{
|
| 3858 |
+
"epoch": 2.77,
|
| 3859 |
+
"learning_rate": 1.50265294755665e-06,
|
| 3860 |
+
"loss": 0.1535,
|
| 3861 |
+
"step": 31550
|
| 3862 |
+
},
|
| 3863 |
+
{
|
| 3864 |
+
"epoch": 2.78,
|
| 3865 |
+
"learning_rate": 1.4733386099141092e-06,
|
| 3866 |
+
"loss": 0.182,
|
| 3867 |
+
"step": 31600
|
| 3868 |
+
},
|
| 3869 |
+
{
|
| 3870 |
+
"epoch": 2.78,
|
| 3871 |
+
"learning_rate": 1.444024272271568e-06,
|
| 3872 |
+
"loss": 0.1816,
|
| 3873 |
+
"step": 31650
|
| 3874 |
+
},
|
| 3875 |
+
{
|
| 3876 |
+
"epoch": 2.79,
|
| 3877 |
+
"learning_rate": 1.4147099346290272e-06,
|
| 3878 |
+
"loss": 0.1621,
|
| 3879 |
+
"step": 31700
|
| 3880 |
+
},
|
| 3881 |
+
{
|
| 3882 |
+
"epoch": 2.79,
|
| 3883 |
+
"learning_rate": 1.3853955969864864e-06,
|
| 3884 |
+
"loss": 0.1623,
|
| 3885 |
+
"step": 31750
|
| 3886 |
+
},
|
| 3887 |
+
{
|
| 3888 |
+
"epoch": 2.8,
|
| 3889 |
+
"learning_rate": 1.356081259343945e-06,
|
| 3890 |
+
"loss": 0.1721,
|
| 3891 |
+
"step": 31800
|
| 3892 |
+
},
|
| 3893 |
+
{
|
| 3894 |
+
"epoch": 2.8,
|
| 3895 |
+
"learning_rate": 1.3267669217014042e-06,
|
| 3896 |
+
"loss": 0.1665,
|
| 3897 |
+
"step": 31850
|
| 3898 |
+
},
|
| 3899 |
+
{
|
| 3900 |
+
"epoch": 2.81,
|
| 3901 |
+
"learning_rate": 1.2974525840588633e-06,
|
| 3902 |
+
"loss": 0.1588,
|
| 3903 |
+
"step": 31900
|
| 3904 |
+
},
|
| 3905 |
+
{
|
| 3906 |
+
"epoch": 2.81,
|
| 3907 |
+
"learning_rate": 1.2681382464163222e-06,
|
| 3908 |
+
"loss": 0.1764,
|
| 3909 |
+
"step": 31950
|
| 3910 |
+
},
|
| 3911 |
+
{
|
| 3912 |
+
"epoch": 2.81,
|
| 3913 |
+
"learning_rate": 1.2388239087737813e-06,
|
| 3914 |
+
"loss": 0.1601,
|
| 3915 |
+
"step": 32000
|
| 3916 |
+
},
|
| 3917 |
+
{
|
| 3918 |
+
"epoch": 2.82,
|
| 3919 |
+
"learning_rate": 1.2095095711312405e-06,
|
| 3920 |
+
"loss": 0.186,
|
| 3921 |
+
"step": 32050
|
| 3922 |
+
},
|
| 3923 |
+
{
|
| 3924 |
+
"epoch": 2.82,
|
| 3925 |
+
"learning_rate": 1.1801952334886994e-06,
|
| 3926 |
+
"loss": 0.1764,
|
| 3927 |
+
"step": 32100
|
| 3928 |
+
},
|
| 3929 |
+
{
|
| 3930 |
+
"epoch": 2.83,
|
| 3931 |
+
"learning_rate": 1.1508808958461585e-06,
|
| 3932 |
+
"loss": 0.1659,
|
| 3933 |
+
"step": 32150
|
| 3934 |
+
},
|
| 3935 |
+
{
|
| 3936 |
+
"epoch": 2.83,
|
| 3937 |
+
"learning_rate": 1.1215665582036176e-06,
|
| 3938 |
+
"loss": 0.1906,
|
| 3939 |
+
"step": 32200
|
| 3940 |
+
},
|
| 3941 |
+
{
|
| 3942 |
+
"epoch": 2.84,
|
| 3943 |
+
"learning_rate": 1.0922522205610765e-06,
|
| 3944 |
+
"loss": 0.189,
|
| 3945 |
+
"step": 32250
|
| 3946 |
+
},
|
| 3947 |
+
{
|
| 3948 |
+
"epoch": 2.84,
|
| 3949 |
+
"learning_rate": 1.0629378829185355e-06,
|
| 3950 |
+
"loss": 0.1501,
|
| 3951 |
+
"step": 32300
|
| 3952 |
+
},
|
| 3953 |
+
{
|
| 3954 |
+
"epoch": 2.84,
|
| 3955 |
+
"learning_rate": 1.0336235452759946e-06,
|
| 3956 |
+
"loss": 0.1963,
|
| 3957 |
+
"step": 32350
|
| 3958 |
+
},
|
| 3959 |
+
{
|
| 3960 |
+
"epoch": 2.85,
|
| 3961 |
+
"learning_rate": 1.0043092076334537e-06,
|
| 3962 |
+
"loss": 0.1888,
|
| 3963 |
+
"step": 32400
|
| 3964 |
+
},
|
| 3965 |
+
{
|
| 3966 |
+
"epoch": 2.85,
|
| 3967 |
+
"learning_rate": 9.749948699909126e-07,
|
| 3968 |
+
"loss": 0.1784,
|
| 3969 |
+
"step": 32450
|
| 3970 |
+
},
|
| 3971 |
+
{
|
| 3972 |
+
"epoch": 2.86,
|
| 3973 |
+
"learning_rate": 9.456805323483717e-07,
|
| 3974 |
+
"loss": 0.181,
|
| 3975 |
+
"step": 32500
|
| 3976 |
+
},
|
| 3977 |
+
{
|
| 3978 |
+
"epoch": 2.86,
|
| 3979 |
+
"learning_rate": 9.163661947058307e-07,
|
| 3980 |
+
"loss": 0.1643,
|
| 3981 |
+
"step": 32550
|
| 3982 |
+
},
|
| 3983 |
+
{
|
| 3984 |
+
"epoch": 2.87,
|
| 3985 |
+
"learning_rate": 8.870518570632897e-07,
|
| 3986 |
+
"loss": 0.1599,
|
| 3987 |
+
"step": 32600
|
| 3988 |
+
},
|
| 3989 |
+
{
|
| 3990 |
+
"epoch": 2.87,
|
| 3991 |
+
"learning_rate": 8.577375194207488e-07,
|
| 3992 |
+
"loss": 0.1685,
|
| 3993 |
+
"step": 32650
|
| 3994 |
+
},
|
| 3995 |
+
{
|
| 3996 |
+
"epoch": 2.88,
|
| 3997 |
+
"learning_rate": 8.284231817782078e-07,
|
| 3998 |
+
"loss": 0.1667,
|
| 3999 |
+
"step": 32700
|
| 4000 |
+
},
|
| 4001 |
+
{
|
| 4002 |
+
"epoch": 2.88,
|
| 4003 |
+
"learning_rate": 7.991088441356669e-07,
|
| 4004 |
+
"loss": 0.2037,
|
| 4005 |
+
"step": 32750
|
| 4006 |
+
},
|
| 4007 |
+
{
|
| 4008 |
+
"epoch": 2.88,
|
| 4009 |
+
"learning_rate": 7.697945064931259e-07,
|
| 4010 |
+
"loss": 0.1674,
|
| 4011 |
+
"step": 32800
|
| 4012 |
+
},
|
| 4013 |
+
{
|
| 4014 |
+
"epoch": 2.89,
|
| 4015 |
+
"learning_rate": 7.404801688505849e-07,
|
| 4016 |
+
"loss": 0.1995,
|
| 4017 |
+
"step": 32850
|
| 4018 |
+
},
|
| 4019 |
+
{
|
| 4020 |
+
"epoch": 2.89,
|
| 4021 |
+
"learning_rate": 7.11165831208044e-07,
|
| 4022 |
+
"loss": 0.1593,
|
| 4023 |
+
"step": 32900
|
| 4024 |
+
},
|
| 4025 |
+
{
|
| 4026 |
+
"epoch": 2.9,
|
| 4027 |
+
"learning_rate": 6.818514935655029e-07,
|
| 4028 |
+
"loss": 0.1797,
|
| 4029 |
+
"step": 32950
|
| 4030 |
+
},
|
| 4031 |
+
{
|
| 4032 |
+
"epoch": 2.9,
|
| 4033 |
+
"learning_rate": 6.525371559229619e-07,
|
| 4034 |
+
"loss": 0.1789,
|
| 4035 |
+
"step": 33000
|
| 4036 |
+
},
|
| 4037 |
+
{
|
| 4038 |
+
"epoch": 2.91,
|
| 4039 |
+
"learning_rate": 6.23222818280421e-07,
|
| 4040 |
+
"loss": 0.1835,
|
| 4041 |
+
"step": 33050
|
| 4042 |
+
},
|
| 4043 |
+
{
|
| 4044 |
+
"epoch": 2.91,
|
| 4045 |
+
"learning_rate": 5.939084806378801e-07,
|
| 4046 |
+
"loss": 0.1651,
|
| 4047 |
+
"step": 33100
|
| 4048 |
+
},
|
| 4049 |
+
{
|
| 4050 |
+
"epoch": 2.92,
|
| 4051 |
+
"learning_rate": 5.645941429953391e-07,
|
| 4052 |
+
"loss": 0.1973,
|
| 4053 |
+
"step": 33150
|
| 4054 |
+
},
|
| 4055 |
+
{
|
| 4056 |
+
"epoch": 2.92,
|
| 4057 |
+
"learning_rate": 5.352798053527981e-07,
|
| 4058 |
+
"loss": 0.164,
|
| 4059 |
+
"step": 33200
|
| 4060 |
+
},
|
| 4061 |
+
{
|
| 4062 |
+
"epoch": 2.92,
|
| 4063 |
+
"learning_rate": 5.059654677102571e-07,
|
| 4064 |
+
"loss": 0.1716,
|
| 4065 |
+
"step": 33250
|
| 4066 |
+
},
|
| 4067 |
+
{
|
| 4068 |
+
"epoch": 2.93,
|
| 4069 |
+
"learning_rate": 4.7665113006771615e-07,
|
| 4070 |
+
"loss": 0.1556,
|
| 4071 |
+
"step": 33300
|
| 4072 |
+
},
|
| 4073 |
+
{
|
| 4074 |
+
"epoch": 2.93,
|
| 4075 |
+
"learning_rate": 4.473367924251752e-07,
|
| 4076 |
+
"loss": 0.1659,
|
| 4077 |
+
"step": 33350
|
| 4078 |
+
},
|
| 4079 |
+
{
|
| 4080 |
+
"epoch": 2.94,
|
| 4081 |
+
"learning_rate": 4.180224547826342e-07,
|
| 4082 |
+
"loss": 0.1732,
|
| 4083 |
+
"step": 33400
|
| 4084 |
+
},
|
| 4085 |
+
{
|
| 4086 |
+
"epoch": 2.94,
|
| 4087 |
+
"learning_rate": 3.8870811714009326e-07,
|
| 4088 |
+
"loss": 0.1693,
|
| 4089 |
+
"step": 33450
|
| 4090 |
+
},
|
| 4091 |
+
{
|
| 4092 |
+
"epoch": 2.95,
|
| 4093 |
+
"learning_rate": 3.5939377949755233e-07,
|
| 4094 |
+
"loss": 0.1827,
|
| 4095 |
+
"step": 33500
|
| 4096 |
+
},
|
| 4097 |
+
{
|
| 4098 |
+
"epoch": 2.95,
|
| 4099 |
+
"learning_rate": 3.300794418550113e-07,
|
| 4100 |
+
"loss": 0.1695,
|
| 4101 |
+
"step": 33550
|
| 4102 |
+
},
|
| 4103 |
+
{
|
| 4104 |
+
"epoch": 2.95,
|
| 4105 |
+
"learning_rate": 3.007651042124703e-07,
|
| 4106 |
+
"loss": 0.1522,
|
| 4107 |
+
"step": 33600
|
| 4108 |
+
},
|
| 4109 |
+
{
|
| 4110 |
+
"epoch": 2.96,
|
| 4111 |
+
"learning_rate": 2.714507665699294e-07,
|
| 4112 |
+
"loss": 0.2081,
|
| 4113 |
+
"step": 33650
|
| 4114 |
+
},
|
| 4115 |
+
{
|
| 4116 |
+
"epoch": 2.96,
|
| 4117 |
+
"learning_rate": 2.421364289273884e-07,
|
| 4118 |
+
"loss": 0.1725,
|
| 4119 |
+
"step": 33700
|
| 4120 |
+
},
|
| 4121 |
+
{
|
| 4122 |
+
"epoch": 2.97,
|
| 4123 |
+
"learning_rate": 2.1282209128484745e-07,
|
| 4124 |
+
"loss": 0.1617,
|
| 4125 |
+
"step": 33750
|
| 4126 |
+
},
|
| 4127 |
+
{
|
| 4128 |
+
"epoch": 2.97,
|
| 4129 |
+
"learning_rate": 1.8350775364230647e-07,
|
| 4130 |
+
"loss": 0.1989,
|
| 4131 |
+
"step": 33800
|
| 4132 |
+
},
|
| 4133 |
+
{
|
| 4134 |
+
"epoch": 2.98,
|
| 4135 |
+
"learning_rate": 1.5419341599976551e-07,
|
| 4136 |
+
"loss": 0.1559,
|
| 4137 |
+
"step": 33850
|
| 4138 |
+
},
|
| 4139 |
+
{
|
| 4140 |
+
"epoch": 2.98,
|
| 4141 |
+
"learning_rate": 1.2487907835722453e-07,
|
| 4142 |
+
"loss": 0.1603,
|
| 4143 |
+
"step": 33900
|
| 4144 |
+
},
|
| 4145 |
+
{
|
| 4146 |
+
"epoch": 2.99,
|
| 4147 |
+
"learning_rate": 9.556474071468355e-08,
|
| 4148 |
+
"loss": 0.1928,
|
| 4149 |
+
"step": 33950
|
| 4150 |
+
},
|
| 4151 |
+
{
|
| 4152 |
+
"epoch": 2.99,
|
| 4153 |
+
"learning_rate": 6.62504030721426e-08,
|
| 4154 |
+
"loss": 0.1633,
|
| 4155 |
+
"step": 34000
|
| 4156 |
+
},
|
| 4157 |
+
{
|
| 4158 |
+
"epoch": 2.99,
|
| 4159 |
+
"learning_rate": 3.693606542960162e-08,
|
| 4160 |
+
"loss": 0.1599,
|
| 4161 |
+
"step": 34050
|
| 4162 |
+
},
|
| 4163 |
+
{
|
| 4164 |
+
"epoch": 3.0,
|
| 4165 |
+
"learning_rate": 7.621727787060652e-09,
|
| 4166 |
+
"loss": 0.1707,
|
| 4167 |
+
"step": 34100
|
| 4168 |
+
},
|
| 4169 |
+
{
|
| 4170 |
+
"epoch": 3.0,
|
| 4171 |
+
"step": 34113,
|
| 4172 |
+
"total_flos": 4.99485813607428e+16,
|
| 4173 |
+
"train_loss": 0.25306039583650514,
|
| 4174 |
+
"train_runtime": 4825.8215,
|
| 4175 |
+
"train_samples_per_second": 226.187,
|
| 4176 |
+
"train_steps_per_second": 7.069
|
| 4177 |
+
}
|
| 4178 |
+
],
|
| 4179 |
+
"logging_steps": 50,
|
| 4180 |
+
"max_steps": 34113,
|
| 4181 |
+
"num_train_epochs": 3,
|
| 4182 |
+
"save_steps": 5000,
|
| 4183 |
+
"total_flos": 4.99485813607428e+16,
|
| 4184 |
+
"trial_name": null,
|
| 4185 |
+
"trial_params": null
|
| 4186 |
+
}
|