Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +3675 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,3675 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 19.965384615384615,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5191,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.038461538461538464,
|
| 13 |
+
"grad_norm": 7.0167975425720215,
|
| 14 |
+
"learning_rate": 7.692307692307694e-06,
|
| 15 |
+
"loss": 1.1358,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.07692307692307693,
|
| 20 |
+
"grad_norm": 3.2799055576324463,
|
| 21 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 22 |
+
"loss": 0.8043,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.11538461538461539,
|
| 27 |
+
"grad_norm": 4.641620635986328,
|
| 28 |
+
"learning_rate": 2.307692307692308e-05,
|
| 29 |
+
"loss": 0.4223,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.15384615384615385,
|
| 34 |
+
"grad_norm": 0.8846292495727539,
|
| 35 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 36 |
+
"loss": 0.2722,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.19230769230769232,
|
| 41 |
+
"grad_norm": 0.9276573657989502,
|
| 42 |
+
"learning_rate": 3.846153846153846e-05,
|
| 43 |
+
"loss": 0.1947,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.23076923076923078,
|
| 48 |
+
"grad_norm": 1.1657756567001343,
|
| 49 |
+
"learning_rate": 4.615384615384616e-05,
|
| 50 |
+
"loss": 0.1486,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.2692307692307692,
|
| 55 |
+
"grad_norm": 0.9154715538024902,
|
| 56 |
+
"learning_rate": 5.384615384615385e-05,
|
| 57 |
+
"loss": 0.1246,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.3076923076923077,
|
| 62 |
+
"grad_norm": 1.1416549682617188,
|
| 63 |
+
"learning_rate": 6.153846153846155e-05,
|
| 64 |
+
"loss": 0.1014,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.34615384615384615,
|
| 69 |
+
"grad_norm": 0.8195381760597229,
|
| 70 |
+
"learning_rate": 6.923076923076924e-05,
|
| 71 |
+
"loss": 0.0926,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.38461538461538464,
|
| 76 |
+
"grad_norm": 0.45692142844200134,
|
| 77 |
+
"learning_rate": 7.692307692307693e-05,
|
| 78 |
+
"loss": 0.0815,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.4230769230769231,
|
| 83 |
+
"grad_norm": 0.4946421980857849,
|
| 84 |
+
"learning_rate": 8.461538461538461e-05,
|
| 85 |
+
"loss": 0.0832,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.46153846153846156,
|
| 90 |
+
"grad_norm": 0.6091125011444092,
|
| 91 |
+
"learning_rate": 9.230769230769232e-05,
|
| 92 |
+
"loss": 0.0721,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.5,
|
| 97 |
+
"grad_norm": 0.610546886920929,
|
| 98 |
+
"learning_rate": 0.0001,
|
| 99 |
+
"loss": 0.0632,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.5384615384615384,
|
| 104 |
+
"grad_norm": 0.5763479471206665,
|
| 105 |
+
"learning_rate": 0.0001076923076923077,
|
| 106 |
+
"loss": 0.0673,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.5769230769230769,
|
| 111 |
+
"grad_norm": 0.821746826171875,
|
| 112 |
+
"learning_rate": 0.00011538461538461538,
|
| 113 |
+
"loss": 0.0641,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.6153846153846154,
|
| 118 |
+
"grad_norm": 0.5488954186439514,
|
| 119 |
+
"learning_rate": 0.0001230769230769231,
|
| 120 |
+
"loss": 0.0625,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.6538461538461539,
|
| 125 |
+
"grad_norm": 0.6605976819992065,
|
| 126 |
+
"learning_rate": 0.00013076923076923077,
|
| 127 |
+
"loss": 0.0579,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.6923076923076923,
|
| 132 |
+
"grad_norm": 0.8009259104728699,
|
| 133 |
+
"learning_rate": 0.00013846153846153847,
|
| 134 |
+
"loss": 0.0523,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.7307692307692307,
|
| 139 |
+
"grad_norm": 0.4878791272640228,
|
| 140 |
+
"learning_rate": 0.00014615384615384615,
|
| 141 |
+
"loss": 0.0538,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.7692307692307693,
|
| 146 |
+
"grad_norm": 0.3475745618343353,
|
| 147 |
+
"learning_rate": 0.00015384615384615385,
|
| 148 |
+
"loss": 0.0519,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.8076923076923077,
|
| 153 |
+
"grad_norm": 0.454436719417572,
|
| 154 |
+
"learning_rate": 0.00016153846153846155,
|
| 155 |
+
"loss": 0.0514,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.8461538461538461,
|
| 160 |
+
"grad_norm": 0.7808321118354797,
|
| 161 |
+
"learning_rate": 0.00016923076923076923,
|
| 162 |
+
"loss": 0.0517,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.8846153846153846,
|
| 167 |
+
"grad_norm": 0.6214762926101685,
|
| 168 |
+
"learning_rate": 0.00017692307692307693,
|
| 169 |
+
"loss": 0.0524,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.9230769230769231,
|
| 174 |
+
"grad_norm": 0.47517019510269165,
|
| 175 |
+
"learning_rate": 0.00018461538461538463,
|
| 176 |
+
"loss": 0.0545,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.9615384615384616,
|
| 181 |
+
"grad_norm": 0.7772664427757263,
|
| 182 |
+
"learning_rate": 0.00019230769230769233,
|
| 183 |
+
"loss": 0.0491,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 1.0,
|
| 188 |
+
"grad_norm": 0.863855242729187,
|
| 189 |
+
"learning_rate": 0.0002,
|
| 190 |
+
"loss": 0.0482,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.0384615384615385,
|
| 195 |
+
"grad_norm": 0.34070315957069397,
|
| 196 |
+
"learning_rate": 0.00019999797045691374,
|
| 197 |
+
"loss": 0.0467,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 1.0769230769230769,
|
| 202 |
+
"grad_norm": 0.38879549503326416,
|
| 203 |
+
"learning_rate": 0.00019999188191003583,
|
| 204 |
+
"loss": 0.0392,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 1.1153846153846154,
|
| 209 |
+
"grad_norm": 0.4438105821609497,
|
| 210 |
+
"learning_rate": 0.00019998173460650565,
|
| 211 |
+
"loss": 0.0424,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.1538461538461537,
|
| 216 |
+
"grad_norm": 0.9575906991958618,
|
| 217 |
+
"learning_rate": 0.00019996752895821098,
|
| 218 |
+
"loss": 0.0503,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.1923076923076923,
|
| 223 |
+
"grad_norm": 0.5975697636604309,
|
| 224 |
+
"learning_rate": 0.00019994926554177134,
|
| 225 |
+
"loss": 0.0487,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 1.2307692307692308,
|
| 230 |
+
"grad_norm": 0.5060175061225891,
|
| 231 |
+
"learning_rate": 0.00019992694509851454,
|
| 232 |
+
"loss": 0.0376,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 1.2692307692307692,
|
| 237 |
+
"grad_norm": 0.35833704471588135,
|
| 238 |
+
"learning_rate": 0.0001999005685344466,
|
| 239 |
+
"loss": 0.0378,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 1.3076923076923077,
|
| 244 |
+
"grad_norm": 0.5617273449897766,
|
| 245 |
+
"learning_rate": 0.000199870136920215,
|
| 246 |
+
"loss": 0.0476,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 1.3461538461538463,
|
| 251 |
+
"grad_norm": 0.611136794090271,
|
| 252 |
+
"learning_rate": 0.00019983565149106513,
|
| 253 |
+
"loss": 0.0442,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.3846153846153846,
|
| 258 |
+
"grad_norm": 0.3568519949913025,
|
| 259 |
+
"learning_rate": 0.00019979711364679037,
|
| 260 |
+
"loss": 0.042,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.4230769230769231,
|
| 265 |
+
"grad_norm": 0.3810388445854187,
|
| 266 |
+
"learning_rate": 0.00019975452495167494,
|
| 267 |
+
"loss": 0.0375,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.4615384615384617,
|
| 272 |
+
"grad_norm": 0.5420308113098145,
|
| 273 |
+
"learning_rate": 0.00019970788713443073,
|
| 274 |
+
"loss": 0.0359,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.5,
|
| 279 |
+
"grad_norm": 0.5682479739189148,
|
| 280 |
+
"learning_rate": 0.00019965720208812692,
|
| 281 |
+
"loss": 0.0371,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.5384615384615383,
|
| 286 |
+
"grad_norm": 0.2949374318122864,
|
| 287 |
+
"learning_rate": 0.00019960247187011322,
|
| 288 |
+
"loss": 0.0368,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.5769230769230769,
|
| 293 |
+
"grad_norm": 0.44852954149246216,
|
| 294 |
+
"learning_rate": 0.0001995436987019363,
|
| 295 |
+
"loss": 0.0375,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.6153846153846154,
|
| 300 |
+
"grad_norm": 0.2495189756155014,
|
| 301 |
+
"learning_rate": 0.00019948088496924972,
|
| 302 |
+
"loss": 0.0363,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.6538461538461537,
|
| 307 |
+
"grad_norm": 0.4514489769935608,
|
| 308 |
+
"learning_rate": 0.00019941403322171707,
|
| 309 |
+
"loss": 0.037,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.6923076923076923,
|
| 314 |
+
"grad_norm": 0.26508140563964844,
|
| 315 |
+
"learning_rate": 0.00019934314617290832,
|
| 316 |
+
"loss": 0.0337,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.7307692307692308,
|
| 321 |
+
"grad_norm": 0.23346170783042908,
|
| 322 |
+
"learning_rate": 0.00019926822670018993,
|
| 323 |
+
"loss": 0.0344,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.7692307692307692,
|
| 328 |
+
"grad_norm": 0.3835722506046295,
|
| 329 |
+
"learning_rate": 0.00019918927784460781,
|
| 330 |
+
"loss": 0.0324,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.8076923076923077,
|
| 335 |
+
"grad_norm": 0.3047066330909729,
|
| 336 |
+
"learning_rate": 0.00019910630281076407,
|
| 337 |
+
"loss": 0.0305,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.8461538461538463,
|
| 342 |
+
"grad_norm": 0.22565406560897827,
|
| 343 |
+
"learning_rate": 0.0001990193049666868,
|
| 344 |
+
"loss": 0.033,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.8846153846153846,
|
| 349 |
+
"grad_norm": 0.39445042610168457,
|
| 350 |
+
"learning_rate": 0.00019892828784369345,
|
| 351 |
+
"loss": 0.0301,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.9230769230769231,
|
| 356 |
+
"grad_norm": 0.4797164797782898,
|
| 357 |
+
"learning_rate": 0.00019883325513624756,
|
| 358 |
+
"loss": 0.0343,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.9615384615384617,
|
| 363 |
+
"grad_norm": 0.5288494825363159,
|
| 364 |
+
"learning_rate": 0.00019873421070180857,
|
| 365 |
+
"loss": 0.0318,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 2.0,
|
| 370 |
+
"grad_norm": 0.3639772832393646,
|
| 371 |
+
"learning_rate": 0.00019863115856067536,
|
| 372 |
+
"loss": 0.0318,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 2.0384615384615383,
|
| 377 |
+
"grad_norm": 0.4338575303554535,
|
| 378 |
+
"learning_rate": 0.0001985241028958232,
|
| 379 |
+
"loss": 0.029,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.076923076923077,
|
| 384 |
+
"grad_norm": 0.25063231587409973,
|
| 385 |
+
"learning_rate": 0.00019841304805273382,
|
| 386 |
+
"loss": 0.0299,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 2.1153846153846154,
|
| 391 |
+
"grad_norm": 0.17708472907543182,
|
| 392 |
+
"learning_rate": 0.0001982979985392189,
|
| 393 |
+
"loss": 0.024,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 2.1538461538461537,
|
| 398 |
+
"grad_norm": 0.33794716000556946,
|
| 399 |
+
"learning_rate": 0.00019817895902523745,
|
| 400 |
+
"loss": 0.0285,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 2.1923076923076925,
|
| 405 |
+
"grad_norm": 0.22588586807250977,
|
| 406 |
+
"learning_rate": 0.00019805593434270584,
|
| 407 |
+
"loss": 0.0309,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 2.230769230769231,
|
| 412 |
+
"grad_norm": 0.27854955196380615,
|
| 413 |
+
"learning_rate": 0.00019792892948530195,
|
| 414 |
+
"loss": 0.0287,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 2.269230769230769,
|
| 419 |
+
"grad_norm": 0.24123087525367737,
|
| 420 |
+
"learning_rate": 0.0001977979496082624,
|
| 421 |
+
"loss": 0.0259,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.3076923076923075,
|
| 426 |
+
"grad_norm": 0.311997652053833,
|
| 427 |
+
"learning_rate": 0.0001976630000281733,
|
| 428 |
+
"loss": 0.0273,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.3461538461538463,
|
| 433 |
+
"grad_norm": 0.2107621133327484,
|
| 434 |
+
"learning_rate": 0.00019752408622275436,
|
| 435 |
+
"loss": 0.0267,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 2.3846153846153846,
|
| 440 |
+
"grad_norm": 0.49752968549728394,
|
| 441 |
+
"learning_rate": 0.0001973812138306366,
|
| 442 |
+
"loss": 0.0305,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 2.423076923076923,
|
| 447 |
+
"grad_norm": 0.2595117688179016,
|
| 448 |
+
"learning_rate": 0.00019723438865113367,
|
| 449 |
+
"loss": 0.0272,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 2.4615384615384617,
|
| 454 |
+
"grad_norm": 0.35266903042793274,
|
| 455 |
+
"learning_rate": 0.00019708361664400598,
|
| 456 |
+
"loss": 0.0289,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 2.5,
|
| 461 |
+
"grad_norm": 0.3101799190044403,
|
| 462 |
+
"learning_rate": 0.00019692890392921934,
|
| 463 |
+
"loss": 0.0304,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.5384615384615383,
|
| 468 |
+
"grad_norm": 0.3763001561164856,
|
| 469 |
+
"learning_rate": 0.00019677025678669607,
|
| 470 |
+
"loss": 0.0261,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.5769230769230766,
|
| 475 |
+
"grad_norm": 0.17876401543617249,
|
| 476 |
+
"learning_rate": 0.00019660768165606046,
|
| 477 |
+
"loss": 0.0251,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 2.6153846153846154,
|
| 482 |
+
"grad_norm": 0.2988975942134857,
|
| 483 |
+
"learning_rate": 0.00019644118513637715,
|
| 484 |
+
"loss": 0.0265,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 2.6538461538461537,
|
| 489 |
+
"grad_norm": 0.32994213700294495,
|
| 490 |
+
"learning_rate": 0.00019627077398588335,
|
| 491 |
+
"loss": 0.0247,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 2.6923076923076925,
|
| 496 |
+
"grad_norm": 0.25379645824432373,
|
| 497 |
+
"learning_rate": 0.00019609645512171442,
|
| 498 |
+
"loss": 0.0241,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 2.730769230769231,
|
| 503 |
+
"grad_norm": 0.3905366361141205,
|
| 504 |
+
"learning_rate": 0.0001959182356196234,
|
| 505 |
+
"loss": 0.0249,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.769230769230769,
|
| 510 |
+
"grad_norm": 0.22633816301822662,
|
| 511 |
+
"learning_rate": 0.00019573612271369336,
|
| 512 |
+
"loss": 0.0267,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 2.8076923076923075,
|
| 517 |
+
"grad_norm": 0.2320065051317215,
|
| 518 |
+
"learning_rate": 0.00019555012379604417,
|
| 519 |
+
"loss": 0.0255,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 2.8461538461538463,
|
| 524 |
+
"grad_norm": 0.21357214450836182,
|
| 525 |
+
"learning_rate": 0.0001953602464165321,
|
| 526 |
+
"loss": 0.0289,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 2.8846153846153846,
|
| 531 |
+
"grad_norm": 0.23385067284107208,
|
| 532 |
+
"learning_rate": 0.00019516649828244363,
|
| 533 |
+
"loss": 0.023,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.9230769230769234,
|
| 538 |
+
"grad_norm": 0.2579933702945709,
|
| 539 |
+
"learning_rate": 0.0001949688872581825,
|
| 540 |
+
"loss": 0.0258,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 2.9615384615384617,
|
| 545 |
+
"grad_norm": 0.3598341643810272,
|
| 546 |
+
"learning_rate": 0.0001947674213649504,
|
| 547 |
+
"loss": 0.0282,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 3.0,
|
| 552 |
+
"grad_norm": 0.2527090907096863,
|
| 553 |
+
"learning_rate": 0.00019456210878042164,
|
| 554 |
+
"loss": 0.0277,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 3.0384615384615383,
|
| 559 |
+
"grad_norm": 0.3657386898994446,
|
| 560 |
+
"learning_rate": 0.00019435295783841092,
|
| 561 |
+
"loss": 0.0246,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 3.076923076923077,
|
| 566 |
+
"grad_norm": 0.35666748881340027,
|
| 567 |
+
"learning_rate": 0.0001941399770285352,
|
| 568 |
+
"loss": 0.024,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 3.1153846153846154,
|
| 573 |
+
"grad_norm": 0.19721713662147522,
|
| 574 |
+
"learning_rate": 0.00019392317499586907,
|
| 575 |
+
"loss": 0.0204,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 3.1538461538461537,
|
| 580 |
+
"grad_norm": 0.23313289880752563,
|
| 581 |
+
"learning_rate": 0.00019370256054059386,
|
| 582 |
+
"loss": 0.0217,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 3.1923076923076925,
|
| 587 |
+
"grad_norm": 0.2258790284395218,
|
| 588 |
+
"learning_rate": 0.00019347814261764046,
|
| 589 |
+
"loss": 0.0233,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.230769230769231,
|
| 594 |
+
"grad_norm": 0.21273957192897797,
|
| 595 |
+
"learning_rate": 0.00019324993033632568,
|
| 596 |
+
"loss": 0.0215,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 3.269230769230769,
|
| 601 |
+
"grad_norm": 0.2613265812397003,
|
| 602 |
+
"learning_rate": 0.00019301793295998273,
|
| 603 |
+
"loss": 0.0226,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 3.3076923076923075,
|
| 608 |
+
"grad_norm": 0.35958781838417053,
|
| 609 |
+
"learning_rate": 0.00019278215990558504,
|
| 610 |
+
"loss": 0.0258,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 3.3461538461538463,
|
| 615 |
+
"grad_norm": 0.27915194630622864,
|
| 616 |
+
"learning_rate": 0.000192542620743364,
|
| 617 |
+
"loss": 0.0255,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 3.3846153846153846,
|
| 622 |
+
"grad_norm": 0.23008808493614197,
|
| 623 |
+
"learning_rate": 0.00019229932519642068,
|
| 624 |
+
"loss": 0.0259,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 3.423076923076923,
|
| 629 |
+
"grad_norm": 0.3263101279735565,
|
| 630 |
+
"learning_rate": 0.00019205228314033093,
|
| 631 |
+
"loss": 0.0247,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.4615384615384617,
|
| 636 |
+
"grad_norm": 0.27318865060806274,
|
| 637 |
+
"learning_rate": 0.00019180150460274478,
|
| 638 |
+
"loss": 0.0246,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 3.5,
|
| 643 |
+
"grad_norm": 0.38367944955825806,
|
| 644 |
+
"learning_rate": 0.00019154699976297907,
|
| 645 |
+
"loss": 0.0218,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 3.5384615384615383,
|
| 650 |
+
"grad_norm": 0.24464184045791626,
|
| 651 |
+
"learning_rate": 0.00019128877895160465,
|
| 652 |
+
"loss": 0.0205,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 3.5769230769230766,
|
| 657 |
+
"grad_norm": 0.2480219155550003,
|
| 658 |
+
"learning_rate": 0.00019102685265002666,
|
| 659 |
+
"loss": 0.0248,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 3.6153846153846154,
|
| 664 |
+
"grad_norm": 0.3331535756587982,
|
| 665 |
+
"learning_rate": 0.0001907612314900595,
|
| 666 |
+
"loss": 0.0232,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 3.6538461538461537,
|
| 671 |
+
"grad_norm": 0.22589264810085297,
|
| 672 |
+
"learning_rate": 0.00019049192625349487,
|
| 673 |
+
"loss": 0.0208,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.6923076923076925,
|
| 678 |
+
"grad_norm": 0.29576578736305237,
|
| 679 |
+
"learning_rate": 0.00019021894787166443,
|
| 680 |
+
"loss": 0.0239,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 3.730769230769231,
|
| 685 |
+
"grad_norm": 0.249545156955719,
|
| 686 |
+
"learning_rate": 0.00018994230742499593,
|
| 687 |
+
"loss": 0.0225,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 3.769230769230769,
|
| 692 |
+
"grad_norm": 0.2573404312133789,
|
| 693 |
+
"learning_rate": 0.00018966201614256347,
|
| 694 |
+
"loss": 0.0209,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 3.8076923076923075,
|
| 699 |
+
"grad_norm": 0.32673683762550354,
|
| 700 |
+
"learning_rate": 0.00018937808540163173,
|
| 701 |
+
"loss": 0.023,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 3.8461538461538463,
|
| 706 |
+
"grad_norm": 0.29287075996398926,
|
| 707 |
+
"learning_rate": 0.0001890905267271942,
|
| 708 |
+
"loss": 0.0231,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 3.8846153846153846,
|
| 713 |
+
"grad_norm": 0.2039879858493805,
|
| 714 |
+
"learning_rate": 0.0001887993517915052,
|
| 715 |
+
"loss": 0.0184,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 3.9230769230769234,
|
| 720 |
+
"grad_norm": 0.24603568017482758,
|
| 721 |
+
"learning_rate": 0.00018850457241360635,
|
| 722 |
+
"loss": 0.0262,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 3.9615384615384617,
|
| 727 |
+
"grad_norm": 0.2239179164171219,
|
| 728 |
+
"learning_rate": 0.00018820620055884658,
|
| 729 |
+
"loss": 0.0218,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 4.0,
|
| 734 |
+
"grad_norm": 0.17946171760559082,
|
| 735 |
+
"learning_rate": 0.0001879042483383966,
|
| 736 |
+
"loss": 0.0181,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 4.038461538461538,
|
| 741 |
+
"grad_norm": 0.11465075612068176,
|
| 742 |
+
"learning_rate": 0.00018759872800875726,
|
| 743 |
+
"loss": 0.0172,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 4.076923076923077,
|
| 748 |
+
"grad_norm": 0.17597812414169312,
|
| 749 |
+
"learning_rate": 0.000187289651971262,
|
| 750 |
+
"loss": 0.0169,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 4.115384615384615,
|
| 755 |
+
"grad_norm": 0.19563104212284088,
|
| 756 |
+
"learning_rate": 0.00018697703277157347,
|
| 757 |
+
"loss": 0.0198,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.153846153846154,
|
| 762 |
+
"grad_norm": 0.18617744743824005,
|
| 763 |
+
"learning_rate": 0.00018666088309917447,
|
| 764 |
+
"loss": 0.0218,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 4.1923076923076925,
|
| 769 |
+
"grad_norm": 0.23704974353313446,
|
| 770 |
+
"learning_rate": 0.00018634121578685255,
|
| 771 |
+
"loss": 0.0186,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 4.230769230769231,
|
| 776 |
+
"grad_norm": 0.32059594988822937,
|
| 777 |
+
"learning_rate": 0.00018601804381017945,
|
| 778 |
+
"loss": 0.0211,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 4.269230769230769,
|
| 783 |
+
"grad_norm": 0.28677231073379517,
|
| 784 |
+
"learning_rate": 0.00018569138028698415,
|
| 785 |
+
"loss": 0.0159,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 4.3076923076923075,
|
| 790 |
+
"grad_norm": 0.24357284605503082,
|
| 791 |
+
"learning_rate": 0.00018536123847682056,
|
| 792 |
+
"loss": 0.0189,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 4.346153846153846,
|
| 797 |
+
"grad_norm": 0.1930132806301117,
|
| 798 |
+
"learning_rate": 0.00018502763178042923,
|
| 799 |
+
"loss": 0.0202,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.384615384615385,
|
| 804 |
+
"grad_norm": 0.1683141440153122,
|
| 805 |
+
"learning_rate": 0.00018469057373919347,
|
| 806 |
+
"loss": 0.0202,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 4.423076923076923,
|
| 811 |
+
"grad_norm": 0.1439371407032013,
|
| 812 |
+
"learning_rate": 0.00018435007803458965,
|
| 813 |
+
"loss": 0.0186,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 4.461538461538462,
|
| 818 |
+
"grad_norm": 0.23720888793468475,
|
| 819 |
+
"learning_rate": 0.00018400615848763175,
|
| 820 |
+
"loss": 0.0209,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 4.5,
|
| 825 |
+
"grad_norm": 0.15303802490234375,
|
| 826 |
+
"learning_rate": 0.00018365882905831065,
|
| 827 |
+
"loss": 0.0232,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 4.538461538461538,
|
| 832 |
+
"grad_norm": 0.140608549118042,
|
| 833 |
+
"learning_rate": 0.0001833081038450271,
|
| 834 |
+
"loss": 0.0192,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 4.576923076923077,
|
| 839 |
+
"grad_norm": 0.31198394298553467,
|
| 840 |
+
"learning_rate": 0.00018295399708401975,
|
| 841 |
+
"loss": 0.0168,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.615384615384615,
|
| 846 |
+
"grad_norm": 0.1934264749288559,
|
| 847 |
+
"learning_rate": 0.00018259652314878721,
|
| 848 |
+
"loss": 0.0207,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 4.653846153846154,
|
| 853 |
+
"grad_norm": 0.33610162138938904,
|
| 854 |
+
"learning_rate": 0.00018223569654950454,
|
| 855 |
+
"loss": 0.0196,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 4.6923076923076925,
|
| 860 |
+
"grad_norm": 0.2503054738044739,
|
| 861 |
+
"learning_rate": 0.00018187153193243436,
|
| 862 |
+
"loss": 0.0201,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 4.730769230769231,
|
| 867 |
+
"grad_norm": 0.20254291594028473,
|
| 868 |
+
"learning_rate": 0.0001815040440793322,
|
| 869 |
+
"loss": 0.0199,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 4.769230769230769,
|
| 874 |
+
"grad_norm": 0.24005241692066193,
|
| 875 |
+
"learning_rate": 0.0001811332479068468,
|
| 876 |
+
"loss": 0.0214,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 4.8076923076923075,
|
| 881 |
+
"grad_norm": 0.21922606229782104,
|
| 882 |
+
"learning_rate": 0.00018075915846591428,
|
| 883 |
+
"loss": 0.0204,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 4.846153846153846,
|
| 888 |
+
"grad_norm": 0.2370215505361557,
|
| 889 |
+
"learning_rate": 0.00018038179094114742,
|
| 890 |
+
"loss": 0.022,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 4.884615384615385,
|
| 895 |
+
"grad_norm": 0.2668112814426422,
|
| 896 |
+
"learning_rate": 0.0001800011606502192,
|
| 897 |
+
"loss": 0.0188,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 4.923076923076923,
|
| 902 |
+
"grad_norm": 0.15615129470825195,
|
| 903 |
+
"learning_rate": 0.00017961728304324116,
|
| 904 |
+
"loss": 0.0182,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 4.961538461538462,
|
| 909 |
+
"grad_norm": 0.23041196167469025,
|
| 910 |
+
"learning_rate": 0.00017923017370213615,
|
| 911 |
+
"loss": 0.0188,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 5.0,
|
| 916 |
+
"grad_norm": 0.21733583509922028,
|
| 917 |
+
"learning_rate": 0.0001788398483400059,
|
| 918 |
+
"loss": 0.0198,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 5.038461538461538,
|
| 923 |
+
"grad_norm": 0.2194267213344574,
|
| 924 |
+
"learning_rate": 0.00017844632280049328,
|
| 925 |
+
"loss": 0.0181,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 5.076923076923077,
|
| 930 |
+
"grad_norm": 0.14587156474590302,
|
| 931 |
+
"learning_rate": 0.00017804961305713896,
|
| 932 |
+
"loss": 0.0151,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 5.115384615384615,
|
| 937 |
+
"grad_norm": 0.16885510087013245,
|
| 938 |
+
"learning_rate": 0.00017764973521273332,
|
| 939 |
+
"loss": 0.0215,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 5.153846153846154,
|
| 944 |
+
"grad_norm": 0.19568288326263428,
|
| 945 |
+
"learning_rate": 0.00017724670549866268,
|
| 946 |
+
"loss": 0.0191,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 5.1923076923076925,
|
| 951 |
+
"grad_norm": 0.15629301965236664,
|
| 952 |
+
"learning_rate": 0.00017684054027425035,
|
| 953 |
+
"loss": 0.0174,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 5.230769230769231,
|
| 958 |
+
"grad_norm": 0.17440061271190643,
|
| 959 |
+
"learning_rate": 0.00017643125602609288,
|
| 960 |
+
"loss": 0.0145,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 5.269230769230769,
|
| 965 |
+
"grad_norm": 0.21910502016544342,
|
| 966 |
+
"learning_rate": 0.0001760188693673905,
|
| 967 |
+
"loss": 0.0181,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.3076923076923075,
|
| 972 |
+
"grad_norm": 0.21728090941905975,
|
| 973 |
+
"learning_rate": 0.00017560339703727315,
|
| 974 |
+
"loss": 0.0137,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 5.346153846153846,
|
| 979 |
+
"grad_norm": 0.15279999375343323,
|
| 980 |
+
"learning_rate": 0.00017518485590012068,
|
| 981 |
+
"loss": 0.0152,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 5.384615384615385,
|
| 986 |
+
"grad_norm": 0.33734527230262756,
|
| 987 |
+
"learning_rate": 0.00017476326294487852,
|
| 988 |
+
"loss": 0.0169,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 5.423076923076923,
|
| 993 |
+
"grad_norm": 0.15465325117111206,
|
| 994 |
+
"learning_rate": 0.00017433863528436804,
|
| 995 |
+
"loss": 0.0159,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 5.461538461538462,
|
| 1000 |
+
"grad_norm": 0.1761886179447174,
|
| 1001 |
+
"learning_rate": 0.00017391099015459186,
|
| 1002 |
+
"loss": 0.0179,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 5.5,
|
| 1007 |
+
"grad_norm": 0.1088884100317955,
|
| 1008 |
+
"learning_rate": 0.00017348034491403432,
|
| 1009 |
+
"loss": 0.0163,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.538461538461538,
|
| 1014 |
+
"grad_norm": 0.3224456012248993,
|
| 1015 |
+
"learning_rate": 0.00017304671704295685,
|
| 1016 |
+
"loss": 0.016,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 5.576923076923077,
|
| 1021 |
+
"grad_norm": 0.2751074433326721,
|
| 1022 |
+
"learning_rate": 0.0001726101241426884,
|
| 1023 |
+
"loss": 0.0152,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 5.615384615384615,
|
| 1028 |
+
"grad_norm": 0.20135682821273804,
|
| 1029 |
+
"learning_rate": 0.00017217058393491098,
|
| 1030 |
+
"loss": 0.0213,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 5.653846153846154,
|
| 1035 |
+
"grad_norm": 0.13225609064102173,
|
| 1036 |
+
"learning_rate": 0.00017172811426094045,
|
| 1037 |
+
"loss": 0.0151,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 5.6923076923076925,
|
| 1042 |
+
"grad_norm": 0.24074223637580872,
|
| 1043 |
+
"learning_rate": 0.0001712827330810021,
|
| 1044 |
+
"loss": 0.0181,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 5.730769230769231,
|
| 1049 |
+
"grad_norm": 0.2650367319583893,
|
| 1050 |
+
"learning_rate": 0.00017083445847350186,
|
| 1051 |
+
"loss": 0.0181,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 5.769230769230769,
|
| 1056 |
+
"grad_norm": 0.17957966029644012,
|
| 1057 |
+
"learning_rate": 0.00017038330863429236,
|
| 1058 |
+
"loss": 0.0153,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 5.8076923076923075,
|
| 1063 |
+
"grad_norm": 0.1726105660200119,
|
| 1064 |
+
"learning_rate": 0.00016992930187593425,
|
| 1065 |
+
"loss": 0.0178,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 5.846153846153846,
|
| 1070 |
+
"grad_norm": 0.29331403970718384,
|
| 1071 |
+
"learning_rate": 0.00016947245662695317,
|
| 1072 |
+
"loss": 0.0164,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 5.884615384615385,
|
| 1077 |
+
"grad_norm": 0.1539272964000702,
|
| 1078 |
+
"learning_rate": 0.0001690127914310914,
|
| 1079 |
+
"loss": 0.0185,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 5.923076923076923,
|
| 1084 |
+
"grad_norm": 0.2472430318593979,
|
| 1085 |
+
"learning_rate": 0.00016855032494655537,
|
| 1086 |
+
"loss": 0.0163,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 5.961538461538462,
|
| 1091 |
+
"grad_norm": 0.23108868300914764,
|
| 1092 |
+
"learning_rate": 0.00016808507594525822,
|
| 1093 |
+
"loss": 0.0163,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 6.0,
|
| 1098 |
+
"grad_norm": 0.1876610368490219,
|
| 1099 |
+
"learning_rate": 0.00016761706331205783,
|
| 1100 |
+
"loss": 0.0183,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 6.038461538461538,
|
| 1105 |
+
"grad_norm": 0.2372998148202896,
|
| 1106 |
+
"learning_rate": 0.00016714630604399021,
|
| 1107 |
+
"loss": 0.0184,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 6.076923076923077,
|
| 1112 |
+
"grad_norm": 0.23431254923343658,
|
| 1113 |
+
"learning_rate": 0.00016667282324949863,
|
| 1114 |
+
"loss": 0.0177,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 6.115384615384615,
|
| 1119 |
+
"grad_norm": 0.18878187239170074,
|
| 1120 |
+
"learning_rate": 0.00016619663414765764,
|
| 1121 |
+
"loss": 0.0176,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 6.153846153846154,
|
| 1126 |
+
"grad_norm": 0.17434252798557281,
|
| 1127 |
+
"learning_rate": 0.00016571775806739335,
|
| 1128 |
+
"loss": 0.0182,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 6.1923076923076925,
|
| 1133 |
+
"grad_norm": 0.21172362565994263,
|
| 1134 |
+
"learning_rate": 0.00016523621444669836,
|
| 1135 |
+
"loss": 0.0186,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.230769230769231,
|
| 1140 |
+
"grad_norm": 0.3032718300819397,
|
| 1141 |
+
"learning_rate": 0.00016475202283184332,
|
| 1142 |
+
"loss": 0.0158,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 6.269230769230769,
|
| 1147 |
+
"grad_norm": 0.2207326591014862,
|
| 1148 |
+
"learning_rate": 0.00016426520287658307,
|
| 1149 |
+
"loss": 0.0148,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 6.3076923076923075,
|
| 1154 |
+
"grad_norm": 0.1728544682264328,
|
| 1155 |
+
"learning_rate": 0.00016377577434135908,
|
| 1156 |
+
"loss": 0.018,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 6.346153846153846,
|
| 1161 |
+
"grad_norm": 0.2233678251504898,
|
| 1162 |
+
"learning_rate": 0.00016328375709249738,
|
| 1163 |
+
"loss": 0.0143,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 6.384615384615385,
|
| 1168 |
+
"grad_norm": 0.156491219997406,
|
| 1169 |
+
"learning_rate": 0.00016278917110140205,
|
| 1170 |
+
"loss": 0.015,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 6.423076923076923,
|
| 1175 |
+
"grad_norm": 0.17857709527015686,
|
| 1176 |
+
"learning_rate": 0.00016229203644374475,
|
| 1177 |
+
"loss": 0.0164,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.461538461538462,
|
| 1182 |
+
"grad_norm": 0.1415674239397049,
|
| 1183 |
+
"learning_rate": 0.00016179237329864952,
|
| 1184 |
+
"loss": 0.0157,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 6.5,
|
| 1189 |
+
"grad_norm": 0.1668516844511032,
|
| 1190 |
+
"learning_rate": 0.00016129020194787403,
|
| 1191 |
+
"loss": 0.0168,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 6.538461538461538,
|
| 1196 |
+
"grad_norm": 0.1995563507080078,
|
| 1197 |
+
"learning_rate": 0.00016078554277498618,
|
| 1198 |
+
"loss": 0.018,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 6.576923076923077,
|
| 1203 |
+
"grad_norm": 0.18094874918460846,
|
| 1204 |
+
"learning_rate": 0.00016027841626453665,
|
| 1205 |
+
"loss": 0.0198,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 6.615384615384615,
|
| 1210 |
+
"grad_norm": 0.17089825868606567,
|
| 1211 |
+
"learning_rate": 0.00015976884300122746,
|
| 1212 |
+
"loss": 0.0158,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 6.653846153846154,
|
| 1217 |
+
"grad_norm": 0.17652678489685059,
|
| 1218 |
+
"learning_rate": 0.00015925684366907653,
|
| 1219 |
+
"loss": 0.0187,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 6.6923076923076925,
|
| 1224 |
+
"grad_norm": 0.17994622886180878,
|
| 1225 |
+
"learning_rate": 0.00015874243905057798,
|
| 1226 |
+
"loss": 0.0156,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 6.730769230769231,
|
| 1231 |
+
"grad_norm": 0.19167155027389526,
|
| 1232 |
+
"learning_rate": 0.0001582256500258585,
|
| 1233 |
+
"loss": 0.0198,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 6.769230769230769,
|
| 1238 |
+
"grad_norm": 0.18259945511817932,
|
| 1239 |
+
"learning_rate": 0.00015770649757182993,
|
| 1240 |
+
"loss": 0.0164,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 6.8076923076923075,
|
| 1245 |
+
"grad_norm": 0.22917133569717407,
|
| 1246 |
+
"learning_rate": 0.0001571850027613378,
|
| 1247 |
+
"loss": 0.0169,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 6.846153846153846,
|
| 1252 |
+
"grad_norm": 0.20906826853752136,
|
| 1253 |
+
"learning_rate": 0.00015666118676230576,
|
| 1254 |
+
"loss": 0.0179,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 6.884615384615385,
|
| 1259 |
+
"grad_norm": 0.3887175917625427,
|
| 1260 |
+
"learning_rate": 0.00015613507083687668,
|
| 1261 |
+
"loss": 0.0178,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 6.923076923076923,
|
| 1266 |
+
"grad_norm": 0.1430666148662567,
|
| 1267 |
+
"learning_rate": 0.00015560667634054932,
|
| 1268 |
+
"loss": 0.0169,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 6.961538461538462,
|
| 1273 |
+
"grad_norm": 0.34076541662216187,
|
| 1274 |
+
"learning_rate": 0.00015507602472131162,
|
| 1275 |
+
"loss": 0.0176,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 7.0,
|
| 1280 |
+
"grad_norm": 0.248886838555336,
|
| 1281 |
+
"learning_rate": 0.00015454313751877004,
|
| 1282 |
+
"loss": 0.0178,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 7.038461538461538,
|
| 1287 |
+
"grad_norm": 0.21486853063106537,
|
| 1288 |
+
"learning_rate": 0.0001540080363632754,
|
| 1289 |
+
"loss": 0.0162,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 7.076923076923077,
|
| 1294 |
+
"grad_norm": 0.19121983647346497,
|
| 1295 |
+
"learning_rate": 0.00015347074297504464,
|
| 1296 |
+
"loss": 0.0147,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 7.115384615384615,
|
| 1301 |
+
"grad_norm": 0.20204658806324005,
|
| 1302 |
+
"learning_rate": 0.00015293127916327944,
|
| 1303 |
+
"loss": 0.0167,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 7.153846153846154,
|
| 1308 |
+
"grad_norm": 0.17696718871593475,
|
| 1309 |
+
"learning_rate": 0.0001523896668252808,
|
| 1310 |
+
"loss": 0.017,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 7.1923076923076925,
|
| 1315 |
+
"grad_norm": 0.19161172211170197,
|
| 1316 |
+
"learning_rate": 0.0001518459279455602,
|
| 1317 |
+
"loss": 0.0143,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 7.230769230769231,
|
| 1322 |
+
"grad_norm": 0.2208317518234253,
|
| 1323 |
+
"learning_rate": 0.0001513000845949473,
|
| 1324 |
+
"loss": 0.016,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 7.269230769230769,
|
| 1329 |
+
"grad_norm": 0.16817378997802734,
|
| 1330 |
+
"learning_rate": 0.00015075215892969418,
|
| 1331 |
+
"loss": 0.0177,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 7.3076923076923075,
|
| 1336 |
+
"grad_norm": 0.21887525916099548,
|
| 1337 |
+
"learning_rate": 0.0001502021731905756,
|
| 1338 |
+
"loss": 0.0142,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 7.346153846153846,
|
| 1343 |
+
"grad_norm": 0.14425648748874664,
|
| 1344 |
+
"learning_rate": 0.00014965014970198676,
|
| 1345 |
+
"loss": 0.015,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.384615384615385,
|
| 1350 |
+
"grad_norm": 0.2206214964389801,
|
| 1351 |
+
"learning_rate": 0.0001490961108710367,
|
| 1352 |
+
"loss": 0.0158,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 7.423076923076923,
|
| 1357 |
+
"grad_norm": 0.13397306203842163,
|
| 1358 |
+
"learning_rate": 0.00014854007918663903,
|
| 1359 |
+
"loss": 0.0172,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 7.461538461538462,
|
| 1364 |
+
"grad_norm": 0.2873019278049469,
|
| 1365 |
+
"learning_rate": 0.00014798207721859898,
|
| 1366 |
+
"loss": 0.0167,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 7.5,
|
| 1371 |
+
"grad_norm": 0.20510359108448029,
|
| 1372 |
+
"learning_rate": 0.00014742212761669724,
|
| 1373 |
+
"loss": 0.0149,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 7.538461538461538,
|
| 1378 |
+
"grad_norm": 0.2240099161863327,
|
| 1379 |
+
"learning_rate": 0.00014686025310977065,
|
| 1380 |
+
"loss": 0.0149,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 7.576923076923077,
|
| 1385 |
+
"grad_norm": 0.2232583910226822,
|
| 1386 |
+
"learning_rate": 0.0001462964765047897,
|
| 1387 |
+
"loss": 0.0148,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 7.615384615384615,
|
| 1392 |
+
"grad_norm": 0.16749520599842072,
|
| 1393 |
+
"learning_rate": 0.00014573082068593254,
|
| 1394 |
+
"loss": 0.0165,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 7.653846153846154,
|
| 1399 |
+
"grad_norm": 0.18627752363681793,
|
| 1400 |
+
"learning_rate": 0.00014516330861365638,
|
| 1401 |
+
"loss": 0.0152,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 7.6923076923076925,
|
| 1406 |
+
"grad_norm": 0.2765505313873291,
|
| 1407 |
+
"learning_rate": 0.00014459396332376517,
|
| 1408 |
+
"loss": 0.0154,
|
| 1409 |
+
"step": 2000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 7.730769230769231,
|
| 1413 |
+
"grad_norm": 0.2526029050350189,
|
| 1414 |
+
"learning_rate": 0.00014402280792647493,
|
| 1415 |
+
"loss": 0.0149,
|
| 1416 |
+
"step": 2010
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 7.769230769230769,
|
| 1420 |
+
"grad_norm": 0.12884902954101562,
|
| 1421 |
+
"learning_rate": 0.00014344986560547537,
|
| 1422 |
+
"loss": 0.0164,
|
| 1423 |
+
"step": 2020
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 7.8076923076923075,
|
| 1427 |
+
"grad_norm": 0.10823468118906021,
|
| 1428 |
+
"learning_rate": 0.00014287515961698902,
|
| 1429 |
+
"loss": 0.0149,
|
| 1430 |
+
"step": 2030
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 7.846153846153846,
|
| 1434 |
+
"grad_norm": 0.15259294211864471,
|
| 1435 |
+
"learning_rate": 0.0001422987132888272,
|
| 1436 |
+
"loss": 0.0143,
|
| 1437 |
+
"step": 2040
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 7.884615384615385,
|
| 1441 |
+
"grad_norm": 0.13543514907360077,
|
| 1442 |
+
"learning_rate": 0.00014172055001944312,
|
| 1443 |
+
"loss": 0.0134,
|
| 1444 |
+
"step": 2050
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 7.923076923076923,
|
| 1448 |
+
"grad_norm": 0.2171320766210556,
|
| 1449 |
+
"learning_rate": 0.0001411406932769821,
|
| 1450 |
+
"loss": 0.0156,
|
| 1451 |
+
"step": 2060
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 7.961538461538462,
|
| 1455 |
+
"grad_norm": 0.09700020402669907,
|
| 1456 |
+
"learning_rate": 0.000140559166598329,
|
| 1457 |
+
"loss": 0.0141,
|
| 1458 |
+
"step": 2070
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 8.0,
|
| 1462 |
+
"grad_norm": 0.19601549208164215,
|
| 1463 |
+
"learning_rate": 0.00013997599358815275,
|
| 1464 |
+
"loss": 0.0147,
|
| 1465 |
+
"step": 2080
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 8.038461538461538,
|
| 1469 |
+
"grad_norm": 0.18070325255393982,
|
| 1470 |
+
"learning_rate": 0.0001393911979179485,
|
| 1471 |
+
"loss": 0.0159,
|
| 1472 |
+
"step": 2090
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 8.076923076923077,
|
| 1476 |
+
"grad_norm": 0.17102424800395966,
|
| 1477 |
+
"learning_rate": 0.0001388048033250763,
|
| 1478 |
+
"loss": 0.0171,
|
| 1479 |
+
"step": 2100
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 8.115384615384615,
|
| 1483 |
+
"grad_norm": 0.15269245207309723,
|
| 1484 |
+
"learning_rate": 0.00013821683361179809,
|
| 1485 |
+
"loss": 0.0136,
|
| 1486 |
+
"step": 2110
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 8.153846153846153,
|
| 1490 |
+
"grad_norm": 0.14295409619808197,
|
| 1491 |
+
"learning_rate": 0.00013762731264431116,
|
| 1492 |
+
"loss": 0.0116,
|
| 1493 |
+
"step": 2120
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 8.192307692307692,
|
| 1497 |
+
"grad_norm": 0.13391715288162231,
|
| 1498 |
+
"learning_rate": 0.00013703626435177956,
|
| 1499 |
+
"loss": 0.017,
|
| 1500 |
+
"step": 2130
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 8.23076923076923,
|
| 1504 |
+
"grad_norm": 0.18101289868354797,
|
| 1505 |
+
"learning_rate": 0.00013644371272536284,
|
| 1506 |
+
"loss": 0.0145,
|
| 1507 |
+
"step": 2140
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 8.26923076923077,
|
| 1511 |
+
"grad_norm": 0.13576363027095795,
|
| 1512 |
+
"learning_rate": 0.00013584968181724213,
|
| 1513 |
+
"loss": 0.0139,
|
| 1514 |
+
"step": 2150
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.307692307692308,
|
| 1518 |
+
"grad_norm": 0.14105717837810516,
|
| 1519 |
+
"learning_rate": 0.00013525419573964386,
|
| 1520 |
+
"loss": 0.0132,
|
| 1521 |
+
"step": 2160
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 8.346153846153847,
|
| 1525 |
+
"grad_norm": 0.09896686673164368,
|
| 1526 |
+
"learning_rate": 0.00013465727866386108,
|
| 1527 |
+
"loss": 0.0107,
|
| 1528 |
+
"step": 2170
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 8.384615384615385,
|
| 1532 |
+
"grad_norm": 0.27290794253349304,
|
| 1533 |
+
"learning_rate": 0.0001340589548192723,
|
| 1534 |
+
"loss": 0.0143,
|
| 1535 |
+
"step": 2180
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 8.423076923076923,
|
| 1539 |
+
"grad_norm": 0.1378784477710724,
|
| 1540 |
+
"learning_rate": 0.00013345924849235793,
|
| 1541 |
+
"loss": 0.0136,
|
| 1542 |
+
"step": 2190
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 8.461538461538462,
|
| 1546 |
+
"grad_norm": 0.18884290754795074,
|
| 1547 |
+
"learning_rate": 0.00013285818402571458,
|
| 1548 |
+
"loss": 0.0153,
|
| 1549 |
+
"step": 2200
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 8.5,
|
| 1553 |
+
"grad_norm": 0.13120310008525848,
|
| 1554 |
+
"learning_rate": 0.00013225578581706687,
|
| 1555 |
+
"loss": 0.0145,
|
| 1556 |
+
"step": 2210
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 8.538461538461538,
|
| 1560 |
+
"grad_norm": 0.1788390427827835,
|
| 1561 |
+
"learning_rate": 0.00013165207831827724,
|
| 1562 |
+
"loss": 0.0152,
|
| 1563 |
+
"step": 2220
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 8.576923076923077,
|
| 1567 |
+
"grad_norm": 0.16243167221546173,
|
| 1568 |
+
"learning_rate": 0.00013104708603435325,
|
| 1569 |
+
"loss": 0.0161,
|
| 1570 |
+
"step": 2230
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 8.615384615384615,
|
| 1574 |
+
"grad_norm": 0.24895057082176208,
|
| 1575 |
+
"learning_rate": 0.00013044083352245308,
|
| 1576 |
+
"loss": 0.0145,
|
| 1577 |
+
"step": 2240
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 8.653846153846153,
|
| 1581 |
+
"grad_norm": 0.15210068225860596,
|
| 1582 |
+
"learning_rate": 0.0001298333453908886,
|
| 1583 |
+
"loss": 0.0135,
|
| 1584 |
+
"step": 2250
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 8.692307692307692,
|
| 1588 |
+
"grad_norm": 0.19090086221694946,
|
| 1589 |
+
"learning_rate": 0.00012922464629812653,
|
| 1590 |
+
"loss": 0.0167,
|
| 1591 |
+
"step": 2260
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 8.73076923076923,
|
| 1595 |
+
"grad_norm": 0.2009580284357071,
|
| 1596 |
+
"learning_rate": 0.0001286147609517876,
|
| 1597 |
+
"loss": 0.0134,
|
| 1598 |
+
"step": 2270
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 8.76923076923077,
|
| 1602 |
+
"grad_norm": 0.20718801021575928,
|
| 1603 |
+
"learning_rate": 0.00012800371410764356,
|
| 1604 |
+
"loss": 0.0122,
|
| 1605 |
+
"step": 2280
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 8.807692307692308,
|
| 1609 |
+
"grad_norm": 0.18385674059391022,
|
| 1610 |
+
"learning_rate": 0.00012739153056861243,
|
| 1611 |
+
"loss": 0.0145,
|
| 1612 |
+
"step": 2290
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 8.846153846153847,
|
| 1616 |
+
"grad_norm": 0.19775976240634918,
|
| 1617 |
+
"learning_rate": 0.00012677823518375146,
|
| 1618 |
+
"loss": 0.0143,
|
| 1619 |
+
"step": 2300
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 8.884615384615385,
|
| 1623 |
+
"grad_norm": 0.14536786079406738,
|
| 1624 |
+
"learning_rate": 0.00012616385284724898,
|
| 1625 |
+
"loss": 0.0136,
|
| 1626 |
+
"step": 2310
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 8.923076923076923,
|
| 1630 |
+
"grad_norm": 0.1978665441274643,
|
| 1631 |
+
"learning_rate": 0.00012554840849741332,
|
| 1632 |
+
"loss": 0.0144,
|
| 1633 |
+
"step": 2320
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 8.961538461538462,
|
| 1637 |
+
"grad_norm": 0.16081364452838898,
|
| 1638 |
+
"learning_rate": 0.00012493192711566104,
|
| 1639 |
+
"loss": 0.0146,
|
| 1640 |
+
"step": 2330
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 9.0,
|
| 1644 |
+
"grad_norm": 0.18740719556808472,
|
| 1645 |
+
"learning_rate": 0.00012431443372550266,
|
| 1646 |
+
"loss": 0.017,
|
| 1647 |
+
"step": 2340
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 9.038461538461538,
|
| 1651 |
+
"grad_norm": 0.1770515739917755,
|
| 1652 |
+
"learning_rate": 0.00012369595339152702,
|
| 1653 |
+
"loss": 0.0154,
|
| 1654 |
+
"step": 2350
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 9.076923076923077,
|
| 1658 |
+
"grad_norm": 0.12580080330371857,
|
| 1659 |
+
"learning_rate": 0.0001230765112183838,
|
| 1660 |
+
"loss": 0.0149,
|
| 1661 |
+
"step": 2360
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 9.115384615384615,
|
| 1665 |
+
"grad_norm": 0.19087755680084229,
|
| 1666 |
+
"learning_rate": 0.00012245613234976462,
|
| 1667 |
+
"loss": 0.0145,
|
| 1668 |
+
"step": 2370
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 9.153846153846153,
|
| 1672 |
+
"grad_norm": 0.10227272659540176,
|
| 1673 |
+
"learning_rate": 0.00012183484196738233,
|
| 1674 |
+
"loss": 0.0126,
|
| 1675 |
+
"step": 2380
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 9.192307692307692,
|
| 1679 |
+
"grad_norm": 0.17450454831123352,
|
| 1680 |
+
"learning_rate": 0.00012121266528994895,
|
| 1681 |
+
"loss": 0.0145,
|
| 1682 |
+
"step": 2390
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 9.23076923076923,
|
| 1686 |
+
"grad_norm": 0.15834400057792664,
|
| 1687 |
+
"learning_rate": 0.00012058962757215195,
|
| 1688 |
+
"loss": 0.0122,
|
| 1689 |
+
"step": 2400
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 9.26923076923077,
|
| 1693 |
+
"grad_norm": 0.13027557730674744,
|
| 1694 |
+
"learning_rate": 0.00011996575410362919,
|
| 1695 |
+
"loss": 0.0137,
|
| 1696 |
+
"step": 2410
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 9.307692307692308,
|
| 1700 |
+
"grad_norm": 0.18278013169765472,
|
| 1701 |
+
"learning_rate": 0.00011934107020794239,
|
| 1702 |
+
"loss": 0.0133,
|
| 1703 |
+
"step": 2420
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 9.346153846153847,
|
| 1707 |
+
"grad_norm": 0.15627382695674896,
|
| 1708 |
+
"learning_rate": 0.00011871560124154912,
|
| 1709 |
+
"loss": 0.0158,
|
| 1710 |
+
"step": 2430
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 9.384615384615385,
|
| 1714 |
+
"grad_norm": 0.16056114435195923,
|
| 1715 |
+
"learning_rate": 0.00011808937259277377,
|
| 1716 |
+
"loss": 0.0131,
|
| 1717 |
+
"step": 2440
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 9.423076923076923,
|
| 1721 |
+
"grad_norm": 0.11140304803848267,
|
| 1722 |
+
"learning_rate": 0.00011746240968077682,
|
| 1723 |
+
"loss": 0.0142,
|
| 1724 |
+
"step": 2450
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 9.461538461538462,
|
| 1728 |
+
"grad_norm": 0.17418035864830017,
|
| 1729 |
+
"learning_rate": 0.0001168347379545231,
|
| 1730 |
+
"loss": 0.0123,
|
| 1731 |
+
"step": 2460
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 9.5,
|
| 1735 |
+
"grad_norm": 0.0985954999923706,
|
| 1736 |
+
"learning_rate": 0.0001162063828917489,
|
| 1737 |
+
"loss": 0.0123,
|
| 1738 |
+
"step": 2470
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 9.538461538461538,
|
| 1742 |
+
"grad_norm": 0.22951987385749817,
|
| 1743 |
+
"learning_rate": 0.00011557736999792765,
|
| 1744 |
+
"loss": 0.0127,
|
| 1745 |
+
"step": 2480
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 9.576923076923077,
|
| 1749 |
+
"grad_norm": 0.13590703904628754,
|
| 1750 |
+
"learning_rate": 0.00011494772480523482,
|
| 1751 |
+
"loss": 0.0135,
|
| 1752 |
+
"step": 2490
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 9.615384615384615,
|
| 1756 |
+
"grad_norm": 0.1775355190038681,
|
| 1757 |
+
"learning_rate": 0.00011431747287151125,
|
| 1758 |
+
"loss": 0.013,
|
| 1759 |
+
"step": 2500
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 9.653846153846153,
|
| 1763 |
+
"grad_norm": 0.22847086191177368,
|
| 1764 |
+
"learning_rate": 0.00011368663977922613,
|
| 1765 |
+
"loss": 0.0136,
|
| 1766 |
+
"step": 2510
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 9.692307692307692,
|
| 1770 |
+
"grad_norm": 0.1462956815958023,
|
| 1771 |
+
"learning_rate": 0.00011305525113443822,
|
| 1772 |
+
"loss": 0.0141,
|
| 1773 |
+
"step": 2520
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 9.73076923076923,
|
| 1777 |
+
"grad_norm": 0.1806713193655014,
|
| 1778 |
+
"learning_rate": 0.00011242333256575677,
|
| 1779 |
+
"loss": 0.0118,
|
| 1780 |
+
"step": 2530
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 9.76923076923077,
|
| 1784 |
+
"grad_norm": 0.14892129600048065,
|
| 1785 |
+
"learning_rate": 0.00011179090972330093,
|
| 1786 |
+
"loss": 0.012,
|
| 1787 |
+
"step": 2540
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 9.807692307692308,
|
| 1791 |
+
"grad_norm": 0.1719180941581726,
|
| 1792 |
+
"learning_rate": 0.0001111580082776589,
|
| 1793 |
+
"loss": 0.0134,
|
| 1794 |
+
"step": 2550
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 9.846153846153847,
|
| 1798 |
+
"grad_norm": 0.1946127712726593,
|
| 1799 |
+
"learning_rate": 0.00011052465391884575,
|
| 1800 |
+
"loss": 0.0146,
|
| 1801 |
+
"step": 2560
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 9.884615384615385,
|
| 1805 |
+
"grad_norm": 0.16166430711746216,
|
| 1806 |
+
"learning_rate": 0.00010989087235526068,
|
| 1807 |
+
"loss": 0.0145,
|
| 1808 |
+
"step": 2570
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 9.923076923076923,
|
| 1812 |
+
"grad_norm": 0.18934275209903717,
|
| 1813 |
+
"learning_rate": 0.00010925668931264346,
|
| 1814 |
+
"loss": 0.0109,
|
| 1815 |
+
"step": 2580
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 9.961538461538462,
|
| 1819 |
+
"grad_norm": 0.14358612895011902,
|
| 1820 |
+
"learning_rate": 0.00010862213053303037,
|
| 1821 |
+
"loss": 0.0099,
|
| 1822 |
+
"step": 2590
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 10.0,
|
| 1826 |
+
"grad_norm": 0.2569468319416046,
|
| 1827 |
+
"learning_rate": 0.00010798722177370897,
|
| 1828 |
+
"loss": 0.0147,
|
| 1829 |
+
"step": 2600
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 10.038461538461538,
|
| 1833 |
+
"grad_norm": 0.21339739859104156,
|
| 1834 |
+
"learning_rate": 0.00010735198880617303,
|
| 1835 |
+
"loss": 0.0137,
|
| 1836 |
+
"step": 2610
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 10.076923076923077,
|
| 1840 |
+
"grad_norm": 0.12013400346040726,
|
| 1841 |
+
"learning_rate": 0.00010671645741507603,
|
| 1842 |
+
"loss": 0.011,
|
| 1843 |
+
"step": 2620
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 10.115384615384615,
|
| 1847 |
+
"grad_norm": 0.18901623785495758,
|
| 1848 |
+
"learning_rate": 0.00010608065339718483,
|
| 1849 |
+
"loss": 0.0109,
|
| 1850 |
+
"step": 2630
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 10.153846153846153,
|
| 1854 |
+
"grad_norm": 0.10211600363254547,
|
| 1855 |
+
"learning_rate": 0.00010544460256033237,
|
| 1856 |
+
"loss": 0.0132,
|
| 1857 |
+
"step": 2640
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 10.192307692307692,
|
| 1861 |
+
"grad_norm": 0.17099827527999878,
|
| 1862 |
+
"learning_rate": 0.00010480833072237022,
|
| 1863 |
+
"loss": 0.0139,
|
| 1864 |
+
"step": 2650
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 10.23076923076923,
|
| 1868 |
+
"grad_norm": 0.15852560102939606,
|
| 1869 |
+
"learning_rate": 0.00010417186371012063,
|
| 1870 |
+
"loss": 0.0116,
|
| 1871 |
+
"step": 2660
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 10.26923076923077,
|
| 1875 |
+
"grad_norm": 0.17198602855205536,
|
| 1876 |
+
"learning_rate": 0.000103535227358328,
|
| 1877 |
+
"loss": 0.0114,
|
| 1878 |
+
"step": 2670
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 10.307692307692308,
|
| 1882 |
+
"grad_norm": 0.21135368943214417,
|
| 1883 |
+
"learning_rate": 0.00010289844750861053,
|
| 1884 |
+
"loss": 0.0126,
|
| 1885 |
+
"step": 2680
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 10.346153846153847,
|
| 1889 |
+
"grad_norm": 0.13126599788665771,
|
| 1890 |
+
"learning_rate": 0.000102261550008411,
|
| 1891 |
+
"loss": 0.013,
|
| 1892 |
+
"step": 2690
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 10.384615384615385,
|
| 1896 |
+
"grad_norm": 0.16771045327186584,
|
| 1897 |
+
"learning_rate": 0.00010162456070994781,
|
| 1898 |
+
"loss": 0.0111,
|
| 1899 |
+
"step": 2700
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 10.423076923076923,
|
| 1903 |
+
"grad_norm": 0.1977088302373886,
|
| 1904 |
+
"learning_rate": 0.00010098750546916547,
|
| 1905 |
+
"loss": 0.0123,
|
| 1906 |
+
"step": 2710
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 10.461538461538462,
|
| 1910 |
+
"grad_norm": 0.12443752586841583,
|
| 1911 |
+
"learning_rate": 0.00010035041014468518,
|
| 1912 |
+
"loss": 0.0114,
|
| 1913 |
+
"step": 2720
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 10.5,
|
| 1917 |
+
"grad_norm": 0.2196878045797348,
|
| 1918 |
+
"learning_rate": 9.971330059675518e-05,
|
| 1919 |
+
"loss": 0.0151,
|
| 1920 |
+
"step": 2730
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 10.538461538461538,
|
| 1924 |
+
"grad_norm": 0.1408672332763672,
|
| 1925 |
+
"learning_rate": 9.907620268620098e-05,
|
| 1926 |
+
"loss": 0.011,
|
| 1927 |
+
"step": 2740
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 10.576923076923077,
|
| 1931 |
+
"grad_norm": 0.14648661017417908,
|
| 1932 |
+
"learning_rate": 9.84391422733758e-05,
|
| 1933 |
+
"loss": 0.0101,
|
| 1934 |
+
"step": 2750
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 10.615384615384615,
|
| 1938 |
+
"grad_norm": 0.14232785999774933,
|
| 1939 |
+
"learning_rate": 9.78021452171108e-05,
|
| 1940 |
+
"loss": 0.0104,
|
| 1941 |
+
"step": 2760
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 10.653846153846153,
|
| 1945 |
+
"grad_norm": 0.16160699725151062,
|
| 1946 |
+
"learning_rate": 9.716523737366537e-05,
|
| 1947 |
+
"loss": 0.0127,
|
| 1948 |
+
"step": 2770
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 10.692307692307692,
|
| 1952 |
+
"grad_norm": 0.20870442688465118,
|
| 1953 |
+
"learning_rate": 9.652844459567777e-05,
|
| 1954 |
+
"loss": 0.0103,
|
| 1955 |
+
"step": 2780
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 10.73076923076923,
|
| 1959 |
+
"grad_norm": 0.15344356000423431,
|
| 1960 |
+
"learning_rate": 9.589179273111551e-05,
|
| 1961 |
+
"loss": 0.0101,
|
| 1962 |
+
"step": 2790
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 10.76923076923077,
|
| 1966 |
+
"grad_norm": 0.11750971525907516,
|
| 1967 |
+
"learning_rate": 9.525530762222645e-05,
|
| 1968 |
+
"loss": 0.0111,
|
| 1969 |
+
"step": 2800
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 10.807692307692308,
|
| 1973 |
+
"grad_norm": 0.117418073117733,
|
| 1974 |
+
"learning_rate": 9.461901510448962e-05,
|
| 1975 |
+
"loss": 0.0128,
|
| 1976 |
+
"step": 2810
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 10.846153846153847,
|
| 1980 |
+
"grad_norm": 0.317682683467865,
|
| 1981 |
+
"learning_rate": 9.398294100556668e-05,
|
| 1982 |
+
"loss": 0.0124,
|
| 1983 |
+
"step": 2820
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 10.884615384615385,
|
| 1987 |
+
"grad_norm": 0.14449702203273773,
|
| 1988 |
+
"learning_rate": 9.334711114425335e-05,
|
| 1989 |
+
"loss": 0.0117,
|
| 1990 |
+
"step": 2830
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 10.923076923076923,
|
| 1994 |
+
"grad_norm": 0.09939802438020706,
|
| 1995 |
+
"learning_rate": 9.27115513294316e-05,
|
| 1996 |
+
"loss": 0.012,
|
| 1997 |
+
"step": 2840
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 10.961538461538462,
|
| 2001 |
+
"grad_norm": 0.16178478300571442,
|
| 2002 |
+
"learning_rate": 9.207628735902205e-05,
|
| 2003 |
+
"loss": 0.0138,
|
| 2004 |
+
"step": 2850
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 11.0,
|
| 2008 |
+
"grad_norm": 0.21062442660331726,
|
| 2009 |
+
"learning_rate": 9.144134501893671e-05,
|
| 2010 |
+
"loss": 0.0122,
|
| 2011 |
+
"step": 2860
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 11.038461538461538,
|
| 2015 |
+
"grad_norm": 0.17146876454353333,
|
| 2016 |
+
"learning_rate": 9.080675008203221e-05,
|
| 2017 |
+
"loss": 0.0148,
|
| 2018 |
+
"step": 2870
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 11.076923076923077,
|
| 2022 |
+
"grad_norm": 0.13412779569625854,
|
| 2023 |
+
"learning_rate": 9.017252830706394e-05,
|
| 2024 |
+
"loss": 0.0099,
|
| 2025 |
+
"step": 2880
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 11.115384615384615,
|
| 2029 |
+
"grad_norm": 0.14295507967472076,
|
| 2030 |
+
"learning_rate": 8.953870543764026e-05,
|
| 2031 |
+
"loss": 0.0116,
|
| 2032 |
+
"step": 2890
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 11.153846153846153,
|
| 2036 |
+
"grad_norm": 0.13120034337043762,
|
| 2037 |
+
"learning_rate": 8.890530720117767e-05,
|
| 2038 |
+
"loss": 0.0127,
|
| 2039 |
+
"step": 2900
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 11.192307692307692,
|
| 2043 |
+
"grad_norm": 0.18492452800273895,
|
| 2044 |
+
"learning_rate": 8.827235930785636e-05,
|
| 2045 |
+
"loss": 0.0106,
|
| 2046 |
+
"step": 2910
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 11.23076923076923,
|
| 2050 |
+
"grad_norm": 0.15923990309238434,
|
| 2051 |
+
"learning_rate": 8.763988744957672e-05,
|
| 2052 |
+
"loss": 0.0134,
|
| 2053 |
+
"step": 2920
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 11.26923076923077,
|
| 2057 |
+
"grad_norm": 0.10775059461593628,
|
| 2058 |
+
"learning_rate": 8.700791729891651e-05,
|
| 2059 |
+
"loss": 0.0141,
|
| 2060 |
+
"step": 2930
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 11.307692307692308,
|
| 2064 |
+
"grad_norm": 0.1554538756608963,
|
| 2065 |
+
"learning_rate": 8.637647450808879e-05,
|
| 2066 |
+
"loss": 0.0103,
|
| 2067 |
+
"step": 2940
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 11.346153846153847,
|
| 2071 |
+
"grad_norm": 0.16043995320796967,
|
| 2072 |
+
"learning_rate": 8.574558470790053e-05,
|
| 2073 |
+
"loss": 0.0147,
|
| 2074 |
+
"step": 2950
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 11.384615384615385,
|
| 2078 |
+
"grad_norm": 0.1335715800523758,
|
| 2079 |
+
"learning_rate": 8.511527350671233e-05,
|
| 2080 |
+
"loss": 0.0115,
|
| 2081 |
+
"step": 2960
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 11.423076923076923,
|
| 2085 |
+
"grad_norm": 0.13701431453227997,
|
| 2086 |
+
"learning_rate": 8.448556648939907e-05,
|
| 2087 |
+
"loss": 0.0108,
|
| 2088 |
+
"step": 2970
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 11.461538461538462,
|
| 2092 |
+
"grad_norm": 0.09677750617265701,
|
| 2093 |
+
"learning_rate": 8.385648921631116e-05,
|
| 2094 |
+
"loss": 0.0095,
|
| 2095 |
+
"step": 2980
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 11.5,
|
| 2099 |
+
"grad_norm": 0.08720875531435013,
|
| 2100 |
+
"learning_rate": 8.322806722223725e-05,
|
| 2101 |
+
"loss": 0.0086,
|
| 2102 |
+
"step": 2990
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 11.538461538461538,
|
| 2106 |
+
"grad_norm": 0.10705883800983429,
|
| 2107 |
+
"learning_rate": 8.260032601536758e-05,
|
| 2108 |
+
"loss": 0.0103,
|
| 2109 |
+
"step": 3000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 11.576923076923077,
|
| 2113 |
+
"grad_norm": 0.1649639904499054,
|
| 2114 |
+
"learning_rate": 8.19732910762587e-05,
|
| 2115 |
+
"loss": 0.0093,
|
| 2116 |
+
"step": 3010
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 11.615384615384615,
|
| 2120 |
+
"grad_norm": 0.0938439667224884,
|
| 2121 |
+
"learning_rate": 8.134698785679909e-05,
|
| 2122 |
+
"loss": 0.0087,
|
| 2123 |
+
"step": 3020
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 11.653846153846153,
|
| 2127 |
+
"grad_norm": 0.15994666516780853,
|
| 2128 |
+
"learning_rate": 8.072144177917615e-05,
|
| 2129 |
+
"loss": 0.0097,
|
| 2130 |
+
"step": 3030
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 11.692307692307692,
|
| 2134 |
+
"grad_norm": 0.16346988081932068,
|
| 2135 |
+
"learning_rate": 8.009667823484424e-05,
|
| 2136 |
+
"loss": 0.0087,
|
| 2137 |
+
"step": 3040
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 11.73076923076923,
|
| 2141 |
+
"grad_norm": 0.12906785309314728,
|
| 2142 |
+
"learning_rate": 7.947272258349396e-05,
|
| 2143 |
+
"loss": 0.0113,
|
| 2144 |
+
"step": 3050
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 11.76923076923077,
|
| 2148 |
+
"grad_norm": 0.1393740475177765,
|
| 2149 |
+
"learning_rate": 7.884960015202289e-05,
|
| 2150 |
+
"loss": 0.0078,
|
| 2151 |
+
"step": 3060
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 11.807692307692308,
|
| 2155 |
+
"grad_norm": 0.11777480691671371,
|
| 2156 |
+
"learning_rate": 7.822733623350749e-05,
|
| 2157 |
+
"loss": 0.0108,
|
| 2158 |
+
"step": 3070
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 11.846153846153847,
|
| 2162 |
+
"grad_norm": 0.11602358520030975,
|
| 2163 |
+
"learning_rate": 7.760595608617646e-05,
|
| 2164 |
+
"loss": 0.0078,
|
| 2165 |
+
"step": 3080
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 11.884615384615385,
|
| 2169 |
+
"grad_norm": 0.09884156286716461,
|
| 2170 |
+
"learning_rate": 7.698548493238537e-05,
|
| 2171 |
+
"loss": 0.0118,
|
| 2172 |
+
"step": 3090
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 11.923076923076923,
|
| 2176 |
+
"grad_norm": 0.16078071296215057,
|
| 2177 |
+
"learning_rate": 7.636594795759307e-05,
|
| 2178 |
+
"loss": 0.0108,
|
| 2179 |
+
"step": 3100
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 11.961538461538462,
|
| 2183 |
+
"grad_norm": 0.07723239809274673,
|
| 2184 |
+
"learning_rate": 7.574737030933921e-05,
|
| 2185 |
+
"loss": 0.0095,
|
| 2186 |
+
"step": 3110
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 12.0,
|
| 2190 |
+
"grad_norm": 0.1335863620042801,
|
| 2191 |
+
"learning_rate": 7.512977709622362e-05,
|
| 2192 |
+
"loss": 0.0108,
|
| 2193 |
+
"step": 3120
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 12.038461538461538,
|
| 2197 |
+
"grad_norm": 0.20832855999469757,
|
| 2198 |
+
"learning_rate": 7.451319338688705e-05,
|
| 2199 |
+
"loss": 0.0107,
|
| 2200 |
+
"step": 3130
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 12.076923076923077,
|
| 2204 |
+
"grad_norm": 0.0902862548828125,
|
| 2205 |
+
"learning_rate": 7.389764420899348e-05,
|
| 2206 |
+
"loss": 0.0116,
|
| 2207 |
+
"step": 3140
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 12.115384615384615,
|
| 2211 |
+
"grad_norm": 0.0983407124876976,
|
| 2212 |
+
"learning_rate": 7.328315454821455e-05,
|
| 2213 |
+
"loss": 0.0117,
|
| 2214 |
+
"step": 3150
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 12.153846153846153,
|
| 2218 |
+
"grad_norm": 0.10898351669311523,
|
| 2219 |
+
"learning_rate": 7.266974934721506e-05,
|
| 2220 |
+
"loss": 0.0087,
|
| 2221 |
+
"step": 3160
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 12.192307692307692,
|
| 2225 |
+
"grad_norm": 0.12344610691070557,
|
| 2226 |
+
"learning_rate": 7.205745350464081e-05,
|
| 2227 |
+
"loss": 0.0076,
|
| 2228 |
+
"step": 3170
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 12.23076923076923,
|
| 2232 |
+
"grad_norm": 0.1027444526553154,
|
| 2233 |
+
"learning_rate": 7.144629187410756e-05,
|
| 2234 |
+
"loss": 0.0089,
|
| 2235 |
+
"step": 3180
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 12.26923076923077,
|
| 2239 |
+
"grad_norm": 0.13259388506412506,
|
| 2240 |
+
"learning_rate": 7.083628926319259e-05,
|
| 2241 |
+
"loss": 0.0093,
|
| 2242 |
+
"step": 3190
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 12.307692307692308,
|
| 2246 |
+
"grad_norm": 0.15261848270893097,
|
| 2247 |
+
"learning_rate": 7.022747043242753e-05,
|
| 2248 |
+
"loss": 0.0088,
|
| 2249 |
+
"step": 3200
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 12.346153846153847,
|
| 2253 |
+
"grad_norm": 0.18796484172344208,
|
| 2254 |
+
"learning_rate": 6.961986009429342e-05,
|
| 2255 |
+
"loss": 0.0094,
|
| 2256 |
+
"step": 3210
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 12.384615384615385,
|
| 2260 |
+
"grad_norm": 0.12565913796424866,
|
| 2261 |
+
"learning_rate": 6.901348291221737e-05,
|
| 2262 |
+
"loss": 0.0092,
|
| 2263 |
+
"step": 3220
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 12.423076923076923,
|
| 2267 |
+
"grad_norm": 0.11127506196498871,
|
| 2268 |
+
"learning_rate": 6.840836349957179e-05,
|
| 2269 |
+
"loss": 0.0078,
|
| 2270 |
+
"step": 3230
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 12.461538461538462,
|
| 2274 |
+
"grad_norm": 0.17799825966358185,
|
| 2275 |
+
"learning_rate": 6.780452641867508e-05,
|
| 2276 |
+
"loss": 0.0098,
|
| 2277 |
+
"step": 3240
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 12.5,
|
| 2281 |
+
"grad_norm": 0.18342117965221405,
|
| 2282 |
+
"learning_rate": 6.720199617979468e-05,
|
| 2283 |
+
"loss": 0.0083,
|
| 2284 |
+
"step": 3250
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 12.538461538461538,
|
| 2288 |
+
"grad_norm": 0.14820097386837006,
|
| 2289 |
+
"learning_rate": 6.660079724015226e-05,
|
| 2290 |
+
"loss": 0.0087,
|
| 2291 |
+
"step": 3260
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 12.576923076923077,
|
| 2295 |
+
"grad_norm": 0.12141295522451401,
|
| 2296 |
+
"learning_rate": 6.600095400293078e-05,
|
| 2297 |
+
"loss": 0.0108,
|
| 2298 |
+
"step": 3270
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 12.615384615384615,
|
| 2302 |
+
"grad_norm": 0.10008122026920319,
|
| 2303 |
+
"learning_rate": 6.540249081628416e-05,
|
| 2304 |
+
"loss": 0.0097,
|
| 2305 |
+
"step": 3280
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 12.653846153846153,
|
| 2309 |
+
"grad_norm": 0.14060907065868378,
|
| 2310 |
+
"learning_rate": 6.480543197234886e-05,
|
| 2311 |
+
"loss": 0.0111,
|
| 2312 |
+
"step": 3290
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 12.692307692307692,
|
| 2316 |
+
"grad_norm": 0.16219834983348846,
|
| 2317 |
+
"learning_rate": 6.420980170625788e-05,
|
| 2318 |
+
"loss": 0.0113,
|
| 2319 |
+
"step": 3300
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 12.73076923076923,
|
| 2323 |
+
"grad_norm": 0.1932661533355713,
|
| 2324 |
+
"learning_rate": 6.361562419515693e-05,
|
| 2325 |
+
"loss": 0.0093,
|
| 2326 |
+
"step": 3310
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 12.76923076923077,
|
| 2330 |
+
"grad_norm": 0.14726077020168304,
|
| 2331 |
+
"learning_rate": 6.302292355722325e-05,
|
| 2332 |
+
"loss": 0.0099,
|
| 2333 |
+
"step": 3320
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 12.807692307692308,
|
| 2337 |
+
"grad_norm": 0.1557319462299347,
|
| 2338 |
+
"learning_rate": 6.243172385068646e-05,
|
| 2339 |
+
"loss": 0.0087,
|
| 2340 |
+
"step": 3330
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 12.846153846153847,
|
| 2344 |
+
"grad_norm": 0.21162110567092896,
|
| 2345 |
+
"learning_rate": 6.184204907285215e-05,
|
| 2346 |
+
"loss": 0.0098,
|
| 2347 |
+
"step": 3340
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 12.884615384615385,
|
| 2351 |
+
"grad_norm": 0.11167220026254654,
|
| 2352 |
+
"learning_rate": 6.125392315912761e-05,
|
| 2353 |
+
"loss": 0.0089,
|
| 2354 |
+
"step": 3350
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 12.923076923076923,
|
| 2358 |
+
"grad_norm": 0.10831273347139359,
|
| 2359 |
+
"learning_rate": 6.066736998205054e-05,
|
| 2360 |
+
"loss": 0.0088,
|
| 2361 |
+
"step": 3360
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 12.961538461538462,
|
| 2365 |
+
"grad_norm": 0.07814571261405945,
|
| 2366 |
+
"learning_rate": 6.00824133503198e-05,
|
| 2367 |
+
"loss": 0.0092,
|
| 2368 |
+
"step": 3370
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 13.0,
|
| 2372 |
+
"grad_norm": 0.11000088602304459,
|
| 2373 |
+
"learning_rate": 5.94990770078292e-05,
|
| 2374 |
+
"loss": 0.0096,
|
| 2375 |
+
"step": 3380
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 13.038461538461538,
|
| 2379 |
+
"grad_norm": 0.1632835417985916,
|
| 2380 |
+
"learning_rate": 5.8917384632703534e-05,
|
| 2381 |
+
"loss": 0.0068,
|
| 2382 |
+
"step": 3390
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 13.076923076923077,
|
| 2386 |
+
"grad_norm": 0.12230722606182098,
|
| 2387 |
+
"learning_rate": 5.8337359836337526e-05,
|
| 2388 |
+
"loss": 0.0092,
|
| 2389 |
+
"step": 3400
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 13.115384615384615,
|
| 2393 |
+
"grad_norm": 0.12406644970178604,
|
| 2394 |
+
"learning_rate": 5.775902616243751e-05,
|
| 2395 |
+
"loss": 0.0095,
|
| 2396 |
+
"step": 3410
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 13.153846153846153,
|
| 2400 |
+
"grad_norm": 0.10183268785476685,
|
| 2401 |
+
"learning_rate": 5.7182407086065706e-05,
|
| 2402 |
+
"loss": 0.0074,
|
| 2403 |
+
"step": 3420
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 13.192307692307692,
|
| 2407 |
+
"grad_norm": 0.11906790733337402,
|
| 2408 |
+
"learning_rate": 5.660752601268731e-05,
|
| 2409 |
+
"loss": 0.0064,
|
| 2410 |
+
"step": 3430
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 13.23076923076923,
|
| 2414 |
+
"grad_norm": 0.13318245112895966,
|
| 2415 |
+
"learning_rate": 5.603440627722043e-05,
|
| 2416 |
+
"loss": 0.0097,
|
| 2417 |
+
"step": 3440
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 13.26923076923077,
|
| 2421 |
+
"grad_norm": 0.10035967826843262,
|
| 2422 |
+
"learning_rate": 5.546307114308902e-05,
|
| 2423 |
+
"loss": 0.0105,
|
| 2424 |
+
"step": 3450
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 13.307692307692308,
|
| 2428 |
+
"grad_norm": 0.11527876555919647,
|
| 2429 |
+
"learning_rate": 5.4893543801278514e-05,
|
| 2430 |
+
"loss": 0.0077,
|
| 2431 |
+
"step": 3460
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 13.346153846153847,
|
| 2435 |
+
"grad_norm": 0.18029938638210297,
|
| 2436 |
+
"learning_rate": 5.4325847369394486e-05,
|
| 2437 |
+
"loss": 0.0099,
|
| 2438 |
+
"step": 3470
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 13.384615384615385,
|
| 2442 |
+
"grad_norm": 0.13629525899887085,
|
| 2443 |
+
"learning_rate": 5.376000489072437e-05,
|
| 2444 |
+
"loss": 0.0082,
|
| 2445 |
+
"step": 3480
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 13.423076923076923,
|
| 2449 |
+
"grad_norm": 0.0793733075261116,
|
| 2450 |
+
"learning_rate": 5.3196039333301864e-05,
|
| 2451 |
+
"loss": 0.0072,
|
| 2452 |
+
"step": 3490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 13.461538461538462,
|
| 2456 |
+
"grad_norm": 0.14016038179397583,
|
| 2457 |
+
"learning_rate": 5.263397358897499e-05,
|
| 2458 |
+
"loss": 0.0101,
|
| 2459 |
+
"step": 3500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 13.5,
|
| 2463 |
+
"grad_norm": 0.13849851489067078,
|
| 2464 |
+
"learning_rate": 5.207383047247665e-05,
|
| 2465 |
+
"loss": 0.0086,
|
| 2466 |
+
"step": 3510
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 13.538461538461538,
|
| 2470 |
+
"grad_norm": 0.10788647085428238,
|
| 2471 |
+
"learning_rate": 5.1515632720498666e-05,
|
| 2472 |
+
"loss": 0.0075,
|
| 2473 |
+
"step": 3520
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 13.576923076923077,
|
| 2477 |
+
"grad_norm": 0.14721888303756714,
|
| 2478 |
+
"learning_rate": 5.095940299076874e-05,
|
| 2479 |
+
"loss": 0.0111,
|
| 2480 |
+
"step": 3530
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 13.615384615384615,
|
| 2484 |
+
"grad_norm": 0.11311782151460648,
|
| 2485 |
+
"learning_rate": 5.0405163861130925e-05,
|
| 2486 |
+
"loss": 0.0083,
|
| 2487 |
+
"step": 3540
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 13.653846153846153,
|
| 2491 |
+
"grad_norm": 0.18341895937919617,
|
| 2492 |
+
"learning_rate": 4.985293782862913e-05,
|
| 2493 |
+
"loss": 0.0074,
|
| 2494 |
+
"step": 3550
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 13.692307692307692,
|
| 2498 |
+
"grad_norm": 0.10047091543674469,
|
| 2499 |
+
"learning_rate": 4.930274730859391e-05,
|
| 2500 |
+
"loss": 0.0076,
|
| 2501 |
+
"step": 3560
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 13.73076923076923,
|
| 2505 |
+
"grad_norm": 0.10870840400457382,
|
| 2506 |
+
"learning_rate": 4.875461463373251e-05,
|
| 2507 |
+
"loss": 0.0078,
|
| 2508 |
+
"step": 3570
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 13.76923076923077,
|
| 2512 |
+
"grad_norm": 0.12459917366504669,
|
| 2513 |
+
"learning_rate": 4.820856205322257e-05,
|
| 2514 |
+
"loss": 0.0093,
|
| 2515 |
+
"step": 3580
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 13.807692307692308,
|
| 2519 |
+
"grad_norm": 0.07429447025060654,
|
| 2520 |
+
"learning_rate": 4.766461173180888e-05,
|
| 2521 |
+
"loss": 0.0086,
|
| 2522 |
+
"step": 3590
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 13.846153846153847,
|
| 2526 |
+
"grad_norm": 0.08076323568820953,
|
| 2527 |
+
"learning_rate": 4.7122785748903755e-05,
|
| 2528 |
+
"loss": 0.0079,
|
| 2529 |
+
"step": 3600
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 13.884615384615385,
|
| 2533 |
+
"grad_norm": 0.09467586874961853,
|
| 2534 |
+
"learning_rate": 4.6583106097690744e-05,
|
| 2535 |
+
"loss": 0.0069,
|
| 2536 |
+
"step": 3610
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 13.923076923076923,
|
| 2540 |
+
"grad_norm": 0.10673276335000992,
|
| 2541 |
+
"learning_rate": 4.604559468423188e-05,
|
| 2542 |
+
"loss": 0.0084,
|
| 2543 |
+
"step": 3620
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 13.961538461538462,
|
| 2547 |
+
"grad_norm": 0.12460212409496307,
|
| 2548 |
+
"learning_rate": 4.551027332657868e-05,
|
| 2549 |
+
"loss": 0.0067,
|
| 2550 |
+
"step": 3630
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 14.0,
|
| 2554 |
+
"grad_norm": 0.13634030520915985,
|
| 2555 |
+
"learning_rate": 4.4977163753886366e-05,
|
| 2556 |
+
"loss": 0.008,
|
| 2557 |
+
"step": 3640
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 14.038461538461538,
|
| 2561 |
+
"grad_norm": 0.10553108155727386,
|
| 2562 |
+
"learning_rate": 4.4446287605531826e-05,
|
| 2563 |
+
"loss": 0.0064,
|
| 2564 |
+
"step": 3650
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 14.076923076923077,
|
| 2568 |
+
"grad_norm": 0.11521629989147186,
|
| 2569 |
+
"learning_rate": 4.391766643023542e-05,
|
| 2570 |
+
"loss": 0.0084,
|
| 2571 |
+
"step": 3660
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 14.115384615384615,
|
| 2575 |
+
"grad_norm": 0.10882071405649185,
|
| 2576 |
+
"learning_rate": 4.3391321685186184e-05,
|
| 2577 |
+
"loss": 0.0078,
|
| 2578 |
+
"step": 3670
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 14.153846153846153,
|
| 2582 |
+
"grad_norm": 0.061356205493211746,
|
| 2583 |
+
"learning_rate": 4.2867274735170914e-05,
|
| 2584 |
+
"loss": 0.0066,
|
| 2585 |
+
"step": 3680
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 14.192307692307692,
|
| 2589 |
+
"grad_norm": 0.12024400383234024,
|
| 2590 |
+
"learning_rate": 4.234554685170683e-05,
|
| 2591 |
+
"loss": 0.0062,
|
| 2592 |
+
"step": 3690
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 14.23076923076923,
|
| 2596 |
+
"grad_norm": 0.1351236253976822,
|
| 2597 |
+
"learning_rate": 4.182615921217835e-05,
|
| 2598 |
+
"loss": 0.0084,
|
| 2599 |
+
"step": 3700
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 14.26923076923077,
|
| 2603 |
+
"grad_norm": 0.06800480931997299,
|
| 2604 |
+
"learning_rate": 4.130913289897732e-05,
|
| 2605 |
+
"loss": 0.0056,
|
| 2606 |
+
"step": 3710
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 14.307692307692308,
|
| 2610 |
+
"grad_norm": 0.11439716070890427,
|
| 2611 |
+
"learning_rate": 4.0794488898647365e-05,
|
| 2612 |
+
"loss": 0.0083,
|
| 2613 |
+
"step": 3720
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 14.346153846153847,
|
| 2617 |
+
"grad_norm": 0.07181525975465775,
|
| 2618 |
+
"learning_rate": 4.0282248101031884e-05,
|
| 2619 |
+
"loss": 0.007,
|
| 2620 |
+
"step": 3730
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 14.384615384615385,
|
| 2624 |
+
"grad_norm": 0.08257249742746353,
|
| 2625 |
+
"learning_rate": 3.977243129842627e-05,
|
| 2626 |
+
"loss": 0.0067,
|
| 2627 |
+
"step": 3740
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 14.423076923076923,
|
| 2631 |
+
"grad_norm": 0.06434876471757889,
|
| 2632 |
+
"learning_rate": 3.9265059184733934e-05,
|
| 2633 |
+
"loss": 0.0074,
|
| 2634 |
+
"step": 3750
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 14.461538461538462,
|
| 2638 |
+
"grad_norm": 0.11237592250108719,
|
| 2639 |
+
"learning_rate": 3.8760152354626075e-05,
|
| 2640 |
+
"loss": 0.0065,
|
| 2641 |
+
"step": 3760
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 14.5,
|
| 2645 |
+
"grad_norm": 0.1452643871307373,
|
| 2646 |
+
"learning_rate": 3.825773130270609e-05,
|
| 2647 |
+
"loss": 0.0081,
|
| 2648 |
+
"step": 3770
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 14.538461538461538,
|
| 2652 |
+
"grad_norm": 0.08552118390798569,
|
| 2653 |
+
"learning_rate": 3.7757816422677406e-05,
|
| 2654 |
+
"loss": 0.0069,
|
| 2655 |
+
"step": 3780
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 14.576923076923077,
|
| 2659 |
+
"grad_norm": 0.12589380145072937,
|
| 2660 |
+
"learning_rate": 3.7260428006515835e-05,
|
| 2661 |
+
"loss": 0.0056,
|
| 2662 |
+
"step": 3790
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 14.615384615384615,
|
| 2666 |
+
"grad_norm": 0.0928504690527916,
|
| 2667 |
+
"learning_rate": 3.6765586243645735e-05,
|
| 2668 |
+
"loss": 0.0082,
|
| 2669 |
+
"step": 3800
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 14.653846153846153,
|
| 2673 |
+
"grad_norm": 0.11787436157464981,
|
| 2674 |
+
"learning_rate": 3.6273311220120696e-05,
|
| 2675 |
+
"loss": 0.0082,
|
| 2676 |
+
"step": 3810
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 14.692307692307692,
|
| 2680 |
+
"grad_norm": 0.1571059226989746,
|
| 2681 |
+
"learning_rate": 3.578362291780818e-05,
|
| 2682 |
+
"loss": 0.0065,
|
| 2683 |
+
"step": 3820
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 14.73076923076923,
|
| 2687 |
+
"grad_norm": 0.09746450930833817,
|
| 2688 |
+
"learning_rate": 3.529654121357827e-05,
|
| 2689 |
+
"loss": 0.0073,
|
| 2690 |
+
"step": 3830
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 14.76923076923077,
|
| 2694 |
+
"grad_norm": 0.10719246417284012,
|
| 2695 |
+
"learning_rate": 3.481208587849712e-05,
|
| 2696 |
+
"loss": 0.0075,
|
| 2697 |
+
"step": 3840
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 14.807692307692308,
|
| 2701 |
+
"grad_norm": 0.09190470725297928,
|
| 2702 |
+
"learning_rate": 3.4330276577024255e-05,
|
| 2703 |
+
"loss": 0.0062,
|
| 2704 |
+
"step": 3850
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 14.846153846153847,
|
| 2708 |
+
"grad_norm": 0.1003689393401146,
|
| 2709 |
+
"learning_rate": 3.3851132866214394e-05,
|
| 2710 |
+
"loss": 0.0063,
|
| 2711 |
+
"step": 3860
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 14.884615384615385,
|
| 2715 |
+
"grad_norm": 0.10899902880191803,
|
| 2716 |
+
"learning_rate": 3.337467419492368e-05,
|
| 2717 |
+
"loss": 0.0069,
|
| 2718 |
+
"step": 3870
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 14.923076923076923,
|
| 2722 |
+
"grad_norm": 0.10312393307685852,
|
| 2723 |
+
"learning_rate": 3.290091990302011e-05,
|
| 2724 |
+
"loss": 0.0065,
|
| 2725 |
+
"step": 3880
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 14.961538461538462,
|
| 2729 |
+
"grad_norm": 0.10359740257263184,
|
| 2730 |
+
"learning_rate": 3.2429889220598655e-05,
|
| 2731 |
+
"loss": 0.0062,
|
| 2732 |
+
"step": 3890
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 15.0,
|
| 2736 |
+
"grad_norm": 0.07049980014562607,
|
| 2737 |
+
"learning_rate": 3.19616012672006e-05,
|
| 2738 |
+
"loss": 0.0064,
|
| 2739 |
+
"step": 3900
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 15.038461538461538,
|
| 2743 |
+
"grad_norm": 0.06978161633014679,
|
| 2744 |
+
"learning_rate": 3.1496075051037574e-05,
|
| 2745 |
+
"loss": 0.0065,
|
| 2746 |
+
"step": 3910
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 15.076923076923077,
|
| 2750 |
+
"grad_norm": 0.16202770173549652,
|
| 2751 |
+
"learning_rate": 3.1033329468219775e-05,
|
| 2752 |
+
"loss": 0.01,
|
| 2753 |
+
"step": 3920
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 15.115384615384615,
|
| 2757 |
+
"grad_norm": 0.09229037910699844,
|
| 2758 |
+
"learning_rate": 3.057338330198919e-05,
|
| 2759 |
+
"loss": 0.0068,
|
| 2760 |
+
"step": 3930
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 15.153846153846153,
|
| 2764 |
+
"grad_norm": 0.12501412630081177,
|
| 2765 |
+
"learning_rate": 3.0116255221957047e-05,
|
| 2766 |
+
"loss": 0.0061,
|
| 2767 |
+
"step": 3940
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 15.192307692307692,
|
| 2771 |
+
"grad_norm": 0.07364345341920853,
|
| 2772 |
+
"learning_rate": 2.9661963783346058e-05,
|
| 2773 |
+
"loss": 0.0063,
|
| 2774 |
+
"step": 3950
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 15.23076923076923,
|
| 2778 |
+
"grad_norm": 0.129119873046875,
|
| 2779 |
+
"learning_rate": 2.92105274262372e-05,
|
| 2780 |
+
"loss": 0.007,
|
| 2781 |
+
"step": 3960
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 15.26923076923077,
|
| 2785 |
+
"grad_norm": 0.09468358010053635,
|
| 2786 |
+
"learning_rate": 2.876196447482117e-05,
|
| 2787 |
+
"loss": 0.0065,
|
| 2788 |
+
"step": 3970
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 15.307692307692308,
|
| 2792 |
+
"grad_norm": 0.11235329508781433,
|
| 2793 |
+
"learning_rate": 2.8316293136654725e-05,
|
| 2794 |
+
"loss": 0.0087,
|
| 2795 |
+
"step": 3980
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 15.346153846153847,
|
| 2799 |
+
"grad_norm": 0.1262669563293457,
|
| 2800 |
+
"learning_rate": 2.7873531501921545e-05,
|
| 2801 |
+
"loss": 0.0106,
|
| 2802 |
+
"step": 3990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 15.384615384615385,
|
| 2806 |
+
"grad_norm": 0.10660561174154282,
|
| 2807 |
+
"learning_rate": 2.7433697542697946e-05,
|
| 2808 |
+
"loss": 0.0099,
|
| 2809 |
+
"step": 4000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 15.423076923076923,
|
| 2813 |
+
"grad_norm": 0.07230859249830246,
|
| 2814 |
+
"learning_rate": 2.6996809112223277e-05,
|
| 2815 |
+
"loss": 0.0075,
|
| 2816 |
+
"step": 4010
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 15.461538461538462,
|
| 2820 |
+
"grad_norm": 0.15842293202877045,
|
| 2821 |
+
"learning_rate": 2.6562883944175456e-05,
|
| 2822 |
+
"loss": 0.008,
|
| 2823 |
+
"step": 4020
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 15.5,
|
| 2827 |
+
"grad_norm": 0.0934629887342453,
|
| 2828 |
+
"learning_rate": 2.6131939651950966e-05,
|
| 2829 |
+
"loss": 0.0056,
|
| 2830 |
+
"step": 4030
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 15.538461538461538,
|
| 2834 |
+
"grad_norm": 0.131788432598114,
|
| 2835 |
+
"learning_rate": 2.5703993727950016e-05,
|
| 2836 |
+
"loss": 0.0071,
|
| 2837 |
+
"step": 4040
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 15.576923076923077,
|
| 2841 |
+
"grad_norm": 0.09212549030780792,
|
| 2842 |
+
"learning_rate": 2.5279063542866376e-05,
|
| 2843 |
+
"loss": 0.0055,
|
| 2844 |
+
"step": 4050
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 15.615384615384615,
|
| 2848 |
+
"grad_norm": 0.1430416852235794,
|
| 2849 |
+
"learning_rate": 2.4857166344982462e-05,
|
| 2850 |
+
"loss": 0.0065,
|
| 2851 |
+
"step": 4060
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 15.653846153846153,
|
| 2855 |
+
"grad_norm": 0.08664490282535553,
|
| 2856 |
+
"learning_rate": 2.44383192594691e-05,
|
| 2857 |
+
"loss": 0.0059,
|
| 2858 |
+
"step": 4070
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 15.692307692307692,
|
| 2862 |
+
"grad_norm": 0.08127816766500473,
|
| 2863 |
+
"learning_rate": 2.4022539287690416e-05,
|
| 2864 |
+
"loss": 0.0066,
|
| 2865 |
+
"step": 4080
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 15.73076923076923,
|
| 2869 |
+
"grad_norm": 0.10413156449794769,
|
| 2870 |
+
"learning_rate": 2.360984330651378e-05,
|
| 2871 |
+
"loss": 0.0069,
|
| 2872 |
+
"step": 4090
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 15.76923076923077,
|
| 2876 |
+
"grad_norm": 0.08507948368787766,
|
| 2877 |
+
"learning_rate": 2.3200248067624654e-05,
|
| 2878 |
+
"loss": 0.0071,
|
| 2879 |
+
"step": 4100
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 15.807692307692308,
|
| 2883 |
+
"grad_norm": 0.0667034238576889,
|
| 2884 |
+
"learning_rate": 2.2793770196846743e-05,
|
| 2885 |
+
"loss": 0.0061,
|
| 2886 |
+
"step": 4110
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 15.846153846153847,
|
| 2890 |
+
"grad_norm": 0.09667922556400299,
|
| 2891 |
+
"learning_rate": 2.239042619346713e-05,
|
| 2892 |
+
"loss": 0.0075,
|
| 2893 |
+
"step": 4120
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 15.884615384615385,
|
| 2897 |
+
"grad_norm": 0.0950574055314064,
|
| 2898 |
+
"learning_rate": 2.1990232429566484e-05,
|
| 2899 |
+
"loss": 0.005,
|
| 2900 |
+
"step": 4130
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 15.923076923076923,
|
| 2904 |
+
"grad_norm": 0.09804387390613556,
|
| 2905 |
+
"learning_rate": 2.159320514935449e-05,
|
| 2906 |
+
"loss": 0.0073,
|
| 2907 |
+
"step": 4140
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 15.961538461538462,
|
| 2911 |
+
"grad_norm": 0.06155986711382866,
|
| 2912 |
+
"learning_rate": 2.119936046851062e-05,
|
| 2913 |
+
"loss": 0.0084,
|
| 2914 |
+
"step": 4150
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 16.0,
|
| 2918 |
+
"grad_norm": 0.12021949142217636,
|
| 2919 |
+
"learning_rate": 2.080871437352985e-05,
|
| 2920 |
+
"loss": 0.0056,
|
| 2921 |
+
"step": 4160
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 16.03846153846154,
|
| 2925 |
+
"grad_norm": 0.11103606224060059,
|
| 2926 |
+
"learning_rate": 2.042128272107382e-05,
|
| 2927 |
+
"loss": 0.0074,
|
| 2928 |
+
"step": 4170
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 16.076923076923077,
|
| 2932 |
+
"grad_norm": 0.07725587487220764,
|
| 2933 |
+
"learning_rate": 2.0037081237327125e-05,
|
| 2934 |
+
"loss": 0.0073,
|
| 2935 |
+
"step": 4180
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 16.115384615384617,
|
| 2939 |
+
"grad_norm": 0.081053227186203,
|
| 2940 |
+
"learning_rate": 1.965612551735908e-05,
|
| 2941 |
+
"loss": 0.0053,
|
| 2942 |
+
"step": 4190
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 16.153846153846153,
|
| 2946 |
+
"grad_norm": 0.09564225375652313,
|
| 2947 |
+
"learning_rate": 1.927843102449065e-05,
|
| 2948 |
+
"loss": 0.0059,
|
| 2949 |
+
"step": 4200
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 16.192307692307693,
|
| 2953 |
+
"grad_norm": 0.10796211659908295,
|
| 2954 |
+
"learning_rate": 1.890401308966676e-05,
|
| 2955 |
+
"loss": 0.0052,
|
| 2956 |
+
"step": 4210
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 16.23076923076923,
|
| 2960 |
+
"grad_norm": 0.05812790244817734,
|
| 2961 |
+
"learning_rate": 1.853288691083407e-05,
|
| 2962 |
+
"loss": 0.0054,
|
| 2963 |
+
"step": 4220
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 16.26923076923077,
|
| 2967 |
+
"grad_norm": 0.07313115149736404,
|
| 2968 |
+
"learning_rate": 1.816506755232391e-05,
|
| 2969 |
+
"loss": 0.006,
|
| 2970 |
+
"step": 4230
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 16.307692307692307,
|
| 2974 |
+
"grad_norm": 0.07596628367900848,
|
| 2975 |
+
"learning_rate": 1.7800569944241053e-05,
|
| 2976 |
+
"loss": 0.0054,
|
| 2977 |
+
"step": 4240
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 16.346153846153847,
|
| 2981 |
+
"grad_norm": 0.10533738136291504,
|
| 2982 |
+
"learning_rate": 1.74394088818575e-05,
|
| 2983 |
+
"loss": 0.007,
|
| 2984 |
+
"step": 4250
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 16.384615384615383,
|
| 2988 |
+
"grad_norm": 0.0967307761311531,
|
| 2989 |
+
"learning_rate": 1.708159902501201e-05,
|
| 2990 |
+
"loss": 0.0056,
|
| 2991 |
+
"step": 4260
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 16.423076923076923,
|
| 2995 |
+
"grad_norm": 0.06359889358282089,
|
| 2996 |
+
"learning_rate": 1.672715489751496e-05,
|
| 2997 |
+
"loss": 0.006,
|
| 2998 |
+
"step": 4270
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 16.46153846153846,
|
| 3002 |
+
"grad_norm": 0.08679856359958649,
|
| 3003 |
+
"learning_rate": 1.637609088655896e-05,
|
| 3004 |
+
"loss": 0.0053,
|
| 3005 |
+
"step": 4280
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 16.5,
|
| 3009 |
+
"grad_norm": 0.07200346887111664,
|
| 3010 |
+
"learning_rate": 1.6028421242134707e-05,
|
| 3011 |
+
"loss": 0.0051,
|
| 3012 |
+
"step": 4290
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 16.53846153846154,
|
| 3016 |
+
"grad_norm": 0.06287889182567596,
|
| 3017 |
+
"learning_rate": 1.568416007645269e-05,
|
| 3018 |
+
"loss": 0.0055,
|
| 3019 |
+
"step": 4300
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 16.576923076923077,
|
| 3023 |
+
"grad_norm": 0.08229623734951019,
|
| 3024 |
+
"learning_rate": 1.53433213633703e-05,
|
| 3025 |
+
"loss": 0.0062,
|
| 3026 |
+
"step": 4310
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 16.615384615384617,
|
| 3030 |
+
"grad_norm": 0.060530465096235275,
|
| 3031 |
+
"learning_rate": 1.5005918937824548e-05,
|
| 3032 |
+
"loss": 0.0067,
|
| 3033 |
+
"step": 4320
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 16.653846153846153,
|
| 3037 |
+
"grad_norm": 0.094852514564991,
|
| 3038 |
+
"learning_rate": 1.4671966495270672e-05,
|
| 3039 |
+
"loss": 0.0057,
|
| 3040 |
+
"step": 4330
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 16.692307692307693,
|
| 3044 |
+
"grad_norm": 0.09454119205474854,
|
| 3045 |
+
"learning_rate": 1.4341477591126096e-05,
|
| 3046 |
+
"loss": 0.0055,
|
| 3047 |
+
"step": 4340
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 16.73076923076923,
|
| 3051 |
+
"grad_norm": 0.06650886684656143,
|
| 3052 |
+
"learning_rate": 1.4014465640220243e-05,
|
| 3053 |
+
"loss": 0.0067,
|
| 3054 |
+
"step": 4350
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 16.76923076923077,
|
| 3058 |
+
"grad_norm": 0.0640147477388382,
|
| 3059 |
+
"learning_rate": 1.3690943916249943e-05,
|
| 3060 |
+
"loss": 0.0056,
|
| 3061 |
+
"step": 4360
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 16.807692307692307,
|
| 3065 |
+
"grad_norm": 0.0917372778058052,
|
| 3066 |
+
"learning_rate": 1.3370925551240798e-05,
|
| 3067 |
+
"loss": 0.006,
|
| 3068 |
+
"step": 4370
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 16.846153846153847,
|
| 3072 |
+
"grad_norm": 0.09550613164901733,
|
| 3073 |
+
"learning_rate": 1.3054423535014005e-05,
|
| 3074 |
+
"loss": 0.0063,
|
| 3075 |
+
"step": 4380
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 16.884615384615383,
|
| 3079 |
+
"grad_norm": 0.0998285710811615,
|
| 3080 |
+
"learning_rate": 1.2741450714659153e-05,
|
| 3081 |
+
"loss": 0.0055,
|
| 3082 |
+
"step": 4390
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 16.923076923076923,
|
| 3086 |
+
"grad_norm": 0.16476678848266602,
|
| 3087 |
+
"learning_rate": 1.2432019794012684e-05,
|
| 3088 |
+
"loss": 0.0069,
|
| 3089 |
+
"step": 4400
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 16.96153846153846,
|
| 3093 |
+
"grad_norm": 0.18762223422527313,
|
| 3094 |
+
"learning_rate": 1.2126143333142337e-05,
|
| 3095 |
+
"loss": 0.007,
|
| 3096 |
+
"step": 4410
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 17.0,
|
| 3100 |
+
"grad_norm": 0.07807065546512604,
|
| 3101 |
+
"learning_rate": 1.1823833747837233e-05,
|
| 3102 |
+
"loss": 0.006,
|
| 3103 |
+
"step": 4420
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 17.03846153846154,
|
| 3107 |
+
"grad_norm": 0.086389921605587,
|
| 3108 |
+
"learning_rate": 1.1525103309103934e-05,
|
| 3109 |
+
"loss": 0.0046,
|
| 3110 |
+
"step": 4430
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 17.076923076923077,
|
| 3114 |
+
"grad_norm": 0.09868927299976349,
|
| 3115 |
+
"learning_rate": 1.1229964142668425e-05,
|
| 3116 |
+
"loss": 0.0061,
|
| 3117 |
+
"step": 4440
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 17.115384615384617,
|
| 3121 |
+
"grad_norm": 0.06974314898252487,
|
| 3122 |
+
"learning_rate": 1.0938428228483733e-05,
|
| 3123 |
+
"loss": 0.0049,
|
| 3124 |
+
"step": 4450
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 17.153846153846153,
|
| 3128 |
+
"grad_norm": 0.06945253908634186,
|
| 3129 |
+
"learning_rate": 1.0650507400243859e-05,
|
| 3130 |
+
"loss": 0.0065,
|
| 3131 |
+
"step": 4460
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 17.192307692307693,
|
| 3135 |
+
"grad_norm": 0.08085139840841293,
|
| 3136 |
+
"learning_rate": 1.0366213344903342e-05,
|
| 3137 |
+
"loss": 0.005,
|
| 3138 |
+
"step": 4470
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 17.23076923076923,
|
| 3142 |
+
"grad_norm": 0.08081810176372528,
|
| 3143 |
+
"learning_rate": 1.00855576022029e-05,
|
| 3144 |
+
"loss": 0.0051,
|
| 3145 |
+
"step": 4480
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 17.26923076923077,
|
| 3149 |
+
"grad_norm": 0.04182474687695503,
|
| 3150 |
+
"learning_rate": 9.808551564200908e-06,
|
| 3151 |
+
"loss": 0.0056,
|
| 3152 |
+
"step": 4490
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 17.307692307692307,
|
| 3156 |
+
"grad_norm": 0.1365455687046051,
|
| 3157 |
+
"learning_rate": 9.535206474811198e-06,
|
| 3158 |
+
"loss": 0.0068,
|
| 3159 |
+
"step": 4500
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 17.346153846153847,
|
| 3163 |
+
"grad_norm": 0.10072525590658188,
|
| 3164 |
+
"learning_rate": 9.265533429346484e-06,
|
| 3165 |
+
"loss": 0.0058,
|
| 3166 |
+
"step": 4510
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 17.384615384615383,
|
| 3170 |
+
"grad_norm": 0.0997309759259224,
|
| 3171 |
+
"learning_rate": 8.999543374068098e-06,
|
| 3172 |
+
"loss": 0.0055,
|
| 3173 |
+
"step": 4520
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 17.423076923076923,
|
| 3177 |
+
"grad_norm": 0.05286664515733719,
|
| 3178 |
+
"learning_rate": 8.737247105741542e-06,
|
| 3179 |
+
"loss": 0.0036,
|
| 3180 |
+
"step": 4530
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 17.46153846153846,
|
| 3184 |
+
"grad_norm": 0.0948430672287941,
|
| 3185 |
+
"learning_rate": 8.478655271198389e-06,
|
| 3186 |
+
"loss": 0.0057,
|
| 3187 |
+
"step": 4540
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 17.5,
|
| 3191 |
+
"grad_norm": 0.043210502713918686,
|
| 3192 |
+
"learning_rate": 8.223778366904056e-06,
|
| 3193 |
+
"loss": 0.0058,
|
| 3194 |
+
"step": 4550
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 17.53846153846154,
|
| 3198 |
+
"grad_norm": 0.08293162286281586,
|
| 3199 |
+
"learning_rate": 7.972626738531708e-06,
|
| 3200 |
+
"loss": 0.0057,
|
| 3201 |
+
"step": 4560
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 17.576923076923077,
|
| 3205 |
+
"grad_norm": 0.049041640013456345,
|
| 3206 |
+
"learning_rate": 7.725210580542387e-06,
|
| 3207 |
+
"loss": 0.0049,
|
| 3208 |
+
"step": 4570
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 17.615384615384617,
|
| 3212 |
+
"grad_norm": 0.08410376310348511,
|
| 3213 |
+
"learning_rate": 7.48153993577112e-06,
|
| 3214 |
+
"loss": 0.005,
|
| 3215 |
+
"step": 4580
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 17.653846153846153,
|
| 3219 |
+
"grad_norm": 0.07442956417798996,
|
| 3220 |
+
"learning_rate": 7.241624695019366e-06,
|
| 3221 |
+
"loss": 0.0046,
|
| 3222 |
+
"step": 4590
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 17.692307692307693,
|
| 3226 |
+
"grad_norm": 0.04315830394625664,
|
| 3227 |
+
"learning_rate": 7.005474596653494e-06,
|
| 3228 |
+
"loss": 0.0063,
|
| 3229 |
+
"step": 4600
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 17.73076923076923,
|
| 3233 |
+
"grad_norm": 0.10330154001712799,
|
| 3234 |
+
"learning_rate": 6.773099226209501e-06,
|
| 3235 |
+
"loss": 0.0056,
|
| 3236 |
+
"step": 4610
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 17.76923076923077,
|
| 3240 |
+
"grad_norm": 0.0930248498916626,
|
| 3241 |
+
"learning_rate": 6.544508016003903e-06,
|
| 3242 |
+
"loss": 0.0058,
|
| 3243 |
+
"step": 4620
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 17.807692307692307,
|
| 3247 |
+
"grad_norm": 0.22107692062854767,
|
| 3248 |
+
"learning_rate": 6.319710244750898e-06,
|
| 3249 |
+
"loss": 0.0058,
|
| 3250 |
+
"step": 4630
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 17.846153846153847,
|
| 3254 |
+
"grad_norm": 0.07021880894899368,
|
| 3255 |
+
"learning_rate": 6.0987150371857534e-06,
|
| 3256 |
+
"loss": 0.005,
|
| 3257 |
+
"step": 4640
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 17.884615384615383,
|
| 3261 |
+
"grad_norm": 0.12697994709014893,
|
| 3262 |
+
"learning_rate": 5.88153136369437e-06,
|
| 3263 |
+
"loss": 0.0053,
|
| 3264 |
+
"step": 4650
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 17.923076923076923,
|
| 3268 |
+
"grad_norm": 0.0982799381017685,
|
| 3269 |
+
"learning_rate": 5.668168039949206e-06,
|
| 3270 |
+
"loss": 0.006,
|
| 3271 |
+
"step": 4660
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 17.96153846153846,
|
| 3275 |
+
"grad_norm": 0.07556256651878357,
|
| 3276 |
+
"learning_rate": 5.458633726551443e-06,
|
| 3277 |
+
"loss": 0.0052,
|
| 3278 |
+
"step": 4670
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 18.0,
|
| 3282 |
+
"grad_norm": 0.0795573741197586,
|
| 3283 |
+
"learning_rate": 5.252936928679431e-06,
|
| 3284 |
+
"loss": 0.0057,
|
| 3285 |
+
"step": 4680
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 18.03846153846154,
|
| 3289 |
+
"grad_norm": 0.0670771524310112,
|
| 3290 |
+
"learning_rate": 5.051085995743421e-06,
|
| 3291 |
+
"loss": 0.0076,
|
| 3292 |
+
"step": 4690
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 18.076923076923077,
|
| 3296 |
+
"grad_norm": 0.11251858621835709,
|
| 3297 |
+
"learning_rate": 4.853089121046739e-06,
|
| 3298 |
+
"loss": 0.0065,
|
| 3299 |
+
"step": 4700
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 18.115384615384617,
|
| 3303 |
+
"grad_norm": 0.10217051953077316,
|
| 3304 |
+
"learning_rate": 4.6589543414531364e-06,
|
| 3305 |
+
"loss": 0.0043,
|
| 3306 |
+
"step": 4710
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 18.153846153846153,
|
| 3310 |
+
"grad_norm": 0.050674889236688614,
|
| 3311 |
+
"learning_rate": 4.468689537060633e-06,
|
| 3312 |
+
"loss": 0.0049,
|
| 3313 |
+
"step": 4720
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 18.192307692307693,
|
| 3317 |
+
"grad_norm": 0.053426776081323624,
|
| 3318 |
+
"learning_rate": 4.282302430881568e-06,
|
| 3319 |
+
"loss": 0.0059,
|
| 3320 |
+
"step": 4730
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 18.23076923076923,
|
| 3324 |
+
"grad_norm": 0.09936069697141647,
|
| 3325 |
+
"learning_rate": 4.099800588529212e-06,
|
| 3326 |
+
"loss": 0.0057,
|
| 3327 |
+
"step": 4740
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 18.26923076923077,
|
| 3331 |
+
"grad_norm": 0.1067737266421318,
|
| 3332 |
+
"learning_rate": 3.921191417910608e-06,
|
| 3333 |
+
"loss": 0.0048,
|
| 3334 |
+
"step": 4750
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 18.307692307692307,
|
| 3338 |
+
"grad_norm": 0.058368340134620667,
|
| 3339 |
+
"learning_rate": 3.746482168925902e-06,
|
| 3340 |
+
"loss": 0.0068,
|
| 3341 |
+
"step": 4760
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 18.346153846153847,
|
| 3345 |
+
"grad_norm": 0.10528615117073059,
|
| 3346 |
+
"learning_rate": 3.5756799331740588e-06,
|
| 3347 |
+
"loss": 0.0064,
|
| 3348 |
+
"step": 4770
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 18.384615384615383,
|
| 3352 |
+
"grad_norm": 0.0884661078453064,
|
| 3353 |
+
"learning_rate": 3.4087916436650104e-06,
|
| 3354 |
+
"loss": 0.0066,
|
| 3355 |
+
"step": 4780
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 18.423076923076923,
|
| 3359 |
+
"grad_norm": 0.03825301676988602,
|
| 3360 |
+
"learning_rate": 3.2458240745382617e-06,
|
| 3361 |
+
"loss": 0.005,
|
| 3362 |
+
"step": 4790
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 18.46153846153846,
|
| 3366 |
+
"grad_norm": 0.08059193193912506,
|
| 3367 |
+
"learning_rate": 3.0867838407878436e-06,
|
| 3368 |
+
"loss": 0.0064,
|
| 3369 |
+
"step": 4800
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 18.5,
|
| 3373 |
+
"grad_norm": 0.0991053432226181,
|
| 3374 |
+
"learning_rate": 2.931677397993915e-06,
|
| 3375 |
+
"loss": 0.0048,
|
| 3376 |
+
"step": 4810
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 18.53846153846154,
|
| 3380 |
+
"grad_norm": 0.05811507999897003,
|
| 3381 |
+
"learning_rate": 2.7805110420606407e-06,
|
| 3382 |
+
"loss": 0.0073,
|
| 3383 |
+
"step": 4820
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 18.576923076923077,
|
| 3387 |
+
"grad_norm": 0.10655184090137482,
|
| 3388 |
+
"learning_rate": 2.633290908960684e-06,
|
| 3389 |
+
"loss": 0.004,
|
| 3390 |
+
"step": 4830
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 18.615384615384617,
|
| 3394 |
+
"grad_norm": 0.07305694371461868,
|
| 3395 |
+
"learning_rate": 2.490022974486095e-06,
|
| 3396 |
+
"loss": 0.0053,
|
| 3397 |
+
"step": 4840
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 18.653846153846153,
|
| 3401 |
+
"grad_norm": 0.06774792075157166,
|
| 3402 |
+
"learning_rate": 2.350713054005793e-06,
|
| 3403 |
+
"loss": 0.0049,
|
| 3404 |
+
"step": 4850
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 18.692307692307693,
|
| 3408 |
+
"grad_norm": 0.049805670976638794,
|
| 3409 |
+
"learning_rate": 2.2153668022295015e-06,
|
| 3410 |
+
"loss": 0.0044,
|
| 3411 |
+
"step": 4860
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 18.73076923076923,
|
| 3415 |
+
"grad_norm": 0.048138655722141266,
|
| 3416 |
+
"learning_rate": 2.0839897129782183e-06,
|
| 3417 |
+
"loss": 0.0055,
|
| 3418 |
+
"step": 4870
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 18.76923076923077,
|
| 3422 |
+
"grad_norm": 0.090663842856884,
|
| 3423 |
+
"learning_rate": 1.956587118961195e-06,
|
| 3424 |
+
"loss": 0.0047,
|
| 3425 |
+
"step": 4880
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 18.807692307692307,
|
| 3429 |
+
"grad_norm": 0.11241704970598221,
|
| 3430 |
+
"learning_rate": 1.833164191559511e-06,
|
| 3431 |
+
"loss": 0.0044,
|
| 3432 |
+
"step": 4890
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 18.846153846153847,
|
| 3436 |
+
"grad_norm": 0.04880325496196747,
|
| 3437 |
+
"learning_rate": 1.713725940616151e-06,
|
| 3438 |
+
"loss": 0.0038,
|
| 3439 |
+
"step": 4900
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 18.884615384615383,
|
| 3443 |
+
"grad_norm": 0.04214165359735489,
|
| 3444 |
+
"learning_rate": 1.5982772142326462e-06,
|
| 3445 |
+
"loss": 0.0043,
|
| 3446 |
+
"step": 4910
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 18.923076923076923,
|
| 3450 |
+
"grad_norm": 0.07064809650182724,
|
| 3451 |
+
"learning_rate": 1.4868226985722877e-06,
|
| 3452 |
+
"loss": 0.0056,
|
| 3453 |
+
"step": 4920
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 18.96153846153846,
|
| 3457 |
+
"grad_norm": 0.06411836296319962,
|
| 3458 |
+
"learning_rate": 1.3793669176698887e-06,
|
| 3459 |
+
"loss": 0.0056,
|
| 3460 |
+
"step": 4930
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 19.0,
|
| 3464 |
+
"grad_norm": 0.10901802033185959,
|
| 3465 |
+
"learning_rate": 1.2759142332482099e-06,
|
| 3466 |
+
"loss": 0.0057,
|
| 3467 |
+
"step": 4940
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 19.03846153846154,
|
| 3471 |
+
"grad_norm": 0.05795470252633095,
|
| 3472 |
+
"learning_rate": 1.1764688445408677e-06,
|
| 3473 |
+
"loss": 0.0049,
|
| 3474 |
+
"step": 4950
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 19.076923076923077,
|
| 3478 |
+
"grad_norm": 0.03703946992754936,
|
| 3479 |
+
"learning_rate": 1.0810347881218708e-06,
|
| 3480 |
+
"loss": 0.0059,
|
| 3481 |
+
"step": 4960
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 19.115384615384617,
|
| 3485 |
+
"grad_norm": 0.05625122785568237,
|
| 3486 |
+
"learning_rate": 9.896159377418057e-07,
|
| 3487 |
+
"loss": 0.0065,
|
| 3488 |
+
"step": 4970
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 19.153846153846153,
|
| 3492 |
+
"grad_norm": 0.08023211359977722,
|
| 3493 |
+
"learning_rate": 9.022160041705974e-07,
|
| 3494 |
+
"loss": 0.0053,
|
| 3495 |
+
"step": 4980
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 19.192307692307693,
|
| 3499 |
+
"grad_norm": 0.06640934944152832,
|
| 3500 |
+
"learning_rate": 8.188385350468508e-07,
|
| 3501 |
+
"loss": 0.0046,
|
| 3502 |
+
"step": 4990
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 19.23076923076923,
|
| 3506 |
+
"grad_norm": 0.10493505746126175,
|
| 3507 |
+
"learning_rate": 7.394869147339112e-07,
|
| 3508 |
+
"loss": 0.0061,
|
| 3509 |
+
"step": 5000
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 19.26923076923077,
|
| 3513 |
+
"grad_norm": 0.09153269976377487,
|
| 3514 |
+
"learning_rate": 6.641643641823958e-07,
|
| 3515 |
+
"loss": 0.0067,
|
| 3516 |
+
"step": 5010
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 19.307692307692307,
|
| 3520 |
+
"grad_norm": 0.06915758550167084,
|
| 3521 |
+
"learning_rate": 5.928739407995543e-07,
|
| 3522 |
+
"loss": 0.0052,
|
| 3523 |
+
"step": 5020
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 19.346153846153847,
|
| 3527 |
+
"grad_norm": 0.08392194658517838,
|
| 3528 |
+
"learning_rate": 5.256185383251011e-07,
|
| 3529 |
+
"loss": 0.0061,
|
| 3530 |
+
"step": 5030
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 19.384615384615383,
|
| 3534 |
+
"grad_norm": 0.04674766585230827,
|
| 3535 |
+
"learning_rate": 4.6240088671378746e-07,
|
| 3536 |
+
"loss": 0.0046,
|
| 3537 |
+
"step": 5040
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 19.423076923076923,
|
| 3541 |
+
"grad_norm": 0.052531130611896515,
|
| 3542 |
+
"learning_rate": 4.0322355202455644e-07,
|
| 3543 |
+
"loss": 0.006,
|
| 3544 |
+
"step": 5050
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 19.46153846153846,
|
| 3548 |
+
"grad_norm": 0.062260791659355164,
|
| 3549 |
+
"learning_rate": 3.4808893631642636e-07,
|
| 3550 |
+
"loss": 0.0043,
|
| 3551 |
+
"step": 5060
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 19.5,
|
| 3555 |
+
"grad_norm": 0.06215400993824005,
|
| 3556 |
+
"learning_rate": 2.969992775509578e-07,
|
| 3557 |
+
"loss": 0.0045,
|
| 3558 |
+
"step": 5070
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 19.53846153846154,
|
| 3562 |
+
"grad_norm": 0.07574478536844254,
|
| 3563 |
+
"learning_rate": 2.4995664950141493e-07,
|
| 3564 |
+
"loss": 0.0038,
|
| 3565 |
+
"step": 5080
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 19.576923076923077,
|
| 3569 |
+
"grad_norm": 0.05041692778468132,
|
| 3570 |
+
"learning_rate": 2.0696296166862194e-07,
|
| 3571 |
+
"loss": 0.0038,
|
| 3572 |
+
"step": 5090
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 19.615384615384617,
|
| 3576 |
+
"grad_norm": 0.10670372098684311,
|
| 3577 |
+
"learning_rate": 1.6801995920341375e-07,
|
| 3578 |
+
"loss": 0.0049,
|
| 3579 |
+
"step": 5100
|
| 3580 |
+
},
|
| 3581 |
+
{
|
| 3582 |
+
"epoch": 19.653846153846153,
|
| 3583 |
+
"grad_norm": 0.05979029834270477,
|
| 3584 |
+
"learning_rate": 1.3312922283581496e-07,
|
| 3585 |
+
"loss": 0.0048,
|
| 3586 |
+
"step": 5110
|
| 3587 |
+
},
|
| 3588 |
+
{
|
| 3589 |
+
"epoch": 19.692307692307693,
|
| 3590 |
+
"grad_norm": 0.05246563255786896,
|
| 3591 |
+
"learning_rate": 1.0229216881088022e-07,
|
| 3592 |
+
"loss": 0.0047,
|
| 3593 |
+
"step": 5120
|
| 3594 |
+
},
|
| 3595 |
+
{
|
| 3596 |
+
"epoch": 19.73076923076923,
|
| 3597 |
+
"grad_norm": 0.08008202165365219,
|
| 3598 |
+
"learning_rate": 7.551004883120661e-08,
|
| 3599 |
+
"loss": 0.0045,
|
| 3600 |
+
"step": 5130
|
| 3601 |
+
},
|
| 3602 |
+
{
|
| 3603 |
+
"epoch": 19.76923076923077,
|
| 3604 |
+
"grad_norm": 0.04513508453965187,
|
| 3605 |
+
"learning_rate": 5.278395000611891e-08,
|
| 3606 |
+
"loss": 0.0047,
|
| 3607 |
+
"step": 5140
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 19.807692307692307,
|
| 3611 |
+
"grad_norm": 0.07471077144145966,
|
| 3612 |
+
"learning_rate": 3.4114794807571514e-08,
|
| 3613 |
+
"loss": 0.0056,
|
| 3614 |
+
"step": 5150
|
| 3615 |
+
},
|
| 3616 |
+
{
|
| 3617 |
+
"epoch": 19.846153846153847,
|
| 3618 |
+
"grad_norm": 0.07386625558137894,
|
| 3619 |
+
"learning_rate": 1.9503341032633958e-08,
|
| 3620 |
+
"loss": 0.0052,
|
| 3621 |
+
"step": 5160
|
| 3622 |
+
},
|
| 3623 |
+
{
|
| 3624 |
+
"epoch": 19.884615384615383,
|
| 3625 |
+
"grad_norm": 0.05373087525367737,
|
| 3626 |
+
"learning_rate": 8.950181772826582e-09,
|
| 3627 |
+
"loss": 0.0062,
|
| 3628 |
+
"step": 5170
|
| 3629 |
+
},
|
| 3630 |
+
{
|
| 3631 |
+
"epoch": 19.923076923076923,
|
| 3632 |
+
"grad_norm": 0.0932120531797409,
|
| 3633 |
+
"learning_rate": 2.455745389962072e-09,
|
| 3634 |
+
"loss": 0.0047,
|
| 3635 |
+
"step": 5180
|
| 3636 |
+
},
|
| 3637 |
+
{
|
| 3638 |
+
"epoch": 19.96153846153846,
|
| 3639 |
+
"grad_norm": 0.08532081544399261,
|
| 3640 |
+
"learning_rate": 2.0295498825984027e-11,
|
| 3641 |
+
"loss": 0.0037,
|
| 3642 |
+
"step": 5190
|
| 3643 |
+
},
|
| 3644 |
+
{
|
| 3645 |
+
"epoch": 19.965384615384615,
|
| 3646 |
+
"step": 5191,
|
| 3647 |
+
"total_flos": 7.394345918457876e+17,
|
| 3648 |
+
"train_loss": 0.02140558246369566,
|
| 3649 |
+
"train_runtime": 6034.0474,
|
| 3650 |
+
"train_samples_per_second": 55.058,
|
| 3651 |
+
"train_steps_per_second": 0.86
|
| 3652 |
+
}
|
| 3653 |
+
],
|
| 3654 |
+
"logging_steps": 10,
|
| 3655 |
+
"max_steps": 5191,
|
| 3656 |
+
"num_input_tokens_seen": 0,
|
| 3657 |
+
"num_train_epochs": 20,
|
| 3658 |
+
"save_steps": 10000,
|
| 3659 |
+
"stateful_callbacks": {
|
| 3660 |
+
"TrainerControl": {
|
| 3661 |
+
"args": {
|
| 3662 |
+
"should_epoch_stop": false,
|
| 3663 |
+
"should_evaluate": false,
|
| 3664 |
+
"should_log": false,
|
| 3665 |
+
"should_save": true,
|
| 3666 |
+
"should_training_stop": true
|
| 3667 |
+
},
|
| 3668 |
+
"attributes": {}
|
| 3669 |
+
}
|
| 3670 |
+
},
|
| 3671 |
+
"total_flos": 7.394345918457876e+17,
|
| 3672 |
+
"train_batch_size": 64,
|
| 3673 |
+
"trial_name": null,
|
| 3674 |
+
"trial_params": null
|
| 3675 |
+
}
|