Upload trainer_state.json with huggingface_hub
Browse files- trainer_state.json +3605 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,3605 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 2.4995098039215686,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 5099,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.004901960784313725,
|
| 13 |
+
"grad_norm": 4.5476274490356445,
|
| 14 |
+
"learning_rate": 7.84313725490196e-06,
|
| 15 |
+
"loss": 0.6369,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.00980392156862745,
|
| 20 |
+
"grad_norm": 2.52302885055542,
|
| 21 |
+
"learning_rate": 1.568627450980392e-05,
|
| 22 |
+
"loss": 0.484,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.014705882352941176,
|
| 27 |
+
"grad_norm": 3.1443543434143066,
|
| 28 |
+
"learning_rate": 2.3529411764705884e-05,
|
| 29 |
+
"loss": 0.3252,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0196078431372549,
|
| 34 |
+
"grad_norm": 2.1440389156341553,
|
| 35 |
+
"learning_rate": 3.137254901960784e-05,
|
| 36 |
+
"loss": 0.2779,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.024509803921568627,
|
| 41 |
+
"grad_norm": 1.7569645643234253,
|
| 42 |
+
"learning_rate": 3.9215686274509805e-05,
|
| 43 |
+
"loss": 0.2387,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.029411764705882353,
|
| 48 |
+
"grad_norm": 1.7137173414230347,
|
| 49 |
+
"learning_rate": 4.705882352941177e-05,
|
| 50 |
+
"loss": 0.2146,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03431372549019608,
|
| 55 |
+
"grad_norm": 1.0686582326889038,
|
| 56 |
+
"learning_rate": 5.490196078431373e-05,
|
| 57 |
+
"loss": 0.1638,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.0392156862745098,
|
| 62 |
+
"grad_norm": 1.834192156791687,
|
| 63 |
+
"learning_rate": 6.274509803921569e-05,
|
| 64 |
+
"loss": 0.1594,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.04411764705882353,
|
| 69 |
+
"grad_norm": 2.2320666313171387,
|
| 70 |
+
"learning_rate": 7.058823529411765e-05,
|
| 71 |
+
"loss": 0.156,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.049019607843137254,
|
| 76 |
+
"grad_norm": 1.2987866401672363,
|
| 77 |
+
"learning_rate": 7.843137254901961e-05,
|
| 78 |
+
"loss": 0.1447,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.05392156862745098,
|
| 83 |
+
"grad_norm": 1.5711545944213867,
|
| 84 |
+
"learning_rate": 8.627450980392158e-05,
|
| 85 |
+
"loss": 0.1449,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.058823529411764705,
|
| 90 |
+
"grad_norm": 0.8892576098442078,
|
| 91 |
+
"learning_rate": 9.411764705882353e-05,
|
| 92 |
+
"loss": 0.1513,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.06372549019607843,
|
| 97 |
+
"grad_norm": 1.1401337385177612,
|
| 98 |
+
"learning_rate": 0.00010196078431372549,
|
| 99 |
+
"loss": 0.1128,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.06862745098039216,
|
| 104 |
+
"grad_norm": 1.375543475151062,
|
| 105 |
+
"learning_rate": 0.00010980392156862746,
|
| 106 |
+
"loss": 0.1186,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.07352941176470588,
|
| 111 |
+
"grad_norm": 1.6076676845550537,
|
| 112 |
+
"learning_rate": 0.00011764705882352942,
|
| 113 |
+
"loss": 0.1396,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0784313725490196,
|
| 118 |
+
"grad_norm": 1.2637161016464233,
|
| 119 |
+
"learning_rate": 0.00012549019607843137,
|
| 120 |
+
"loss": 0.1187,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.08333333333333333,
|
| 125 |
+
"grad_norm": 0.5812987685203552,
|
| 126 |
+
"learning_rate": 0.00013333333333333334,
|
| 127 |
+
"loss": 0.1155,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.08823529411764706,
|
| 132 |
+
"grad_norm": 0.7302483916282654,
|
| 133 |
+
"learning_rate": 0.0001411764705882353,
|
| 134 |
+
"loss": 0.1068,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.09313725490196079,
|
| 139 |
+
"grad_norm": 0.6003187894821167,
|
| 140 |
+
"learning_rate": 0.00014901960784313728,
|
| 141 |
+
"loss": 0.1084,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.09803921568627451,
|
| 146 |
+
"grad_norm": 1.3157514333724976,
|
| 147 |
+
"learning_rate": 0.00015686274509803922,
|
| 148 |
+
"loss": 0.1127,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.10294117647058823,
|
| 153 |
+
"grad_norm": 0.8480639457702637,
|
| 154 |
+
"learning_rate": 0.0001647058823529412,
|
| 155 |
+
"loss": 0.1025,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.10784313725490197,
|
| 160 |
+
"grad_norm": 1.0640238523483276,
|
| 161 |
+
"learning_rate": 0.00017254901960784316,
|
| 162 |
+
"loss": 0.1214,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.11274509803921569,
|
| 167 |
+
"grad_norm": 0.7853420972824097,
|
| 168 |
+
"learning_rate": 0.0001803921568627451,
|
| 169 |
+
"loss": 0.1103,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.11764705882352941,
|
| 174 |
+
"grad_norm": 0.760675847530365,
|
| 175 |
+
"learning_rate": 0.00018823529411764707,
|
| 176 |
+
"loss": 0.1071,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.12254901960784313,
|
| 181 |
+
"grad_norm": 1.1404098272323608,
|
| 182 |
+
"learning_rate": 0.000196078431372549,
|
| 183 |
+
"loss": 0.1004,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.12745098039215685,
|
| 188 |
+
"grad_norm": 0.6359620690345764,
|
| 189 |
+
"learning_rate": 0.0001999994742235753,
|
| 190 |
+
"loss": 0.1065,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.1323529411764706,
|
| 195 |
+
"grad_norm": 0.7933241724967957,
|
| 196 |
+
"learning_rate": 0.00019999526804535039,
|
| 197 |
+
"loss": 0.0943,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.13725490196078433,
|
| 202 |
+
"grad_norm": 1.599077582359314,
|
| 203 |
+
"learning_rate": 0.00019998685586582082,
|
| 204 |
+
"loss": 0.1304,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.14215686274509803,
|
| 209 |
+
"grad_norm": 0.8844221234321594,
|
| 210 |
+
"learning_rate": 0.00019997423803881975,
|
| 211 |
+
"loss": 0.0917,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.14705882352941177,
|
| 216 |
+
"grad_norm": 1.2456647157669067,
|
| 217 |
+
"learning_rate": 0.00019995741509507825,
|
| 218 |
+
"loss": 0.111,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.15196078431372548,
|
| 223 |
+
"grad_norm": 0.6590626239776611,
|
| 224 |
+
"learning_rate": 0.00019993638774220307,
|
| 225 |
+
"loss": 0.1022,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.1568627450980392,
|
| 230 |
+
"grad_norm": 0.7061448693275452,
|
| 231 |
+
"learning_rate": 0.00019991115686464675,
|
| 232 |
+
"loss": 0.0938,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.16176470588235295,
|
| 237 |
+
"grad_norm": 1.0512727499008179,
|
| 238 |
+
"learning_rate": 0.00019988172352367056,
|
| 239 |
+
"loss": 0.1059,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.16666666666666666,
|
| 244 |
+
"grad_norm": 0.6884363889694214,
|
| 245 |
+
"learning_rate": 0.00019984808895729978,
|
| 246 |
+
"loss": 0.0801,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.1715686274509804,
|
| 251 |
+
"grad_norm": 0.8961064219474792,
|
| 252 |
+
"learning_rate": 0.00019981025458027169,
|
| 253 |
+
"loss": 0.0872,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.17647058823529413,
|
| 258 |
+
"grad_norm": 0.7410668730735779,
|
| 259 |
+
"learning_rate": 0.00019976822198397595,
|
| 260 |
+
"loss": 0.0935,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.18137254901960784,
|
| 265 |
+
"grad_norm": 0.8089532256126404,
|
| 266 |
+
"learning_rate": 0.00019972199293638777,
|
| 267 |
+
"loss": 0.0806,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.18627450980392157,
|
| 272 |
+
"grad_norm": 0.6644020676612854,
|
| 273 |
+
"learning_rate": 0.00019967156938199355,
|
| 274 |
+
"loss": 0.0885,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.19117647058823528,
|
| 279 |
+
"grad_norm": 0.8422799110412598,
|
| 280 |
+
"learning_rate": 0.00019961695344170895,
|
| 281 |
+
"loss": 0.0952,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.19607843137254902,
|
| 286 |
+
"grad_norm": 0.8162615299224854,
|
| 287 |
+
"learning_rate": 0.00019955814741278986,
|
| 288 |
+
"loss": 0.0802,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.20098039215686275,
|
| 293 |
+
"grad_norm": 0.7302709221839905,
|
| 294 |
+
"learning_rate": 0.0001994951537687357,
|
| 295 |
+
"loss": 0.0884,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.20588235294117646,
|
| 300 |
+
"grad_norm": 0.7032344937324524,
|
| 301 |
+
"learning_rate": 0.00019942797515918527,
|
| 302 |
+
"loss": 0.0896,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.2107843137254902,
|
| 307 |
+
"grad_norm": 0.8042428493499756,
|
| 308 |
+
"learning_rate": 0.00019935661440980554,
|
| 309 |
+
"loss": 0.0811,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.21568627450980393,
|
| 314 |
+
"grad_norm": 0.6656658053398132,
|
| 315 |
+
"learning_rate": 0.00019928107452217255,
|
| 316 |
+
"loss": 0.0856,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.22058823529411764,
|
| 321 |
+
"grad_norm": 1.2202825546264648,
|
| 322 |
+
"learning_rate": 0.00019920135867364534,
|
| 323 |
+
"loss": 0.0895,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.22549019607843138,
|
| 328 |
+
"grad_norm": 0.8210168480873108,
|
| 329 |
+
"learning_rate": 0.00019911747021723216,
|
| 330 |
+
"loss": 0.0807,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.23039215686274508,
|
| 335 |
+
"grad_norm": 0.7217456102371216,
|
| 336 |
+
"learning_rate": 0.0001990294126814496,
|
| 337 |
+
"loss": 0.0814,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.23529411764705882,
|
| 342 |
+
"grad_norm": 0.743126392364502,
|
| 343 |
+
"learning_rate": 0.00019893718977017402,
|
| 344 |
+
"loss": 0.0887,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.24019607843137256,
|
| 349 |
+
"grad_norm": 0.7190248370170593,
|
| 350 |
+
"learning_rate": 0.00019884080536248578,
|
| 351 |
+
"loss": 0.0859,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.24509803921568626,
|
| 356 |
+
"grad_norm": 0.5253967046737671,
|
| 357 |
+
"learning_rate": 0.00019874026351250623,
|
| 358 |
+
"loss": 0.0678,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.25,
|
| 363 |
+
"grad_norm": 0.8271141052246094,
|
| 364 |
+
"learning_rate": 0.00019863556844922696,
|
| 365 |
+
"loss": 0.0762,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.2549019607843137,
|
| 370 |
+
"grad_norm": 0.7656545639038086,
|
| 371 |
+
"learning_rate": 0.0001985267245763321,
|
| 372 |
+
"loss": 0.0724,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.25980392156862747,
|
| 377 |
+
"grad_norm": 0.6673869490623474,
|
| 378 |
+
"learning_rate": 0.00019841373647201297,
|
| 379 |
+
"loss": 0.0817,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.2647058823529412,
|
| 384 |
+
"grad_norm": 0.880395770072937,
|
| 385 |
+
"learning_rate": 0.00019829660888877565,
|
| 386 |
+
"loss": 0.0897,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.2696078431372549,
|
| 391 |
+
"grad_norm": 0.7278539538383484,
|
| 392 |
+
"learning_rate": 0.00019817534675324093,
|
| 393 |
+
"loss": 0.0808,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.27450980392156865,
|
| 398 |
+
"grad_norm": 0.5380986928939819,
|
| 399 |
+
"learning_rate": 0.00019804995516593712,
|
| 400 |
+
"loss": 0.077,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.27941176470588236,
|
| 405 |
+
"grad_norm": 0.9306485652923584,
|
| 406 |
+
"learning_rate": 0.00019792043940108564,
|
| 407 |
+
"loss": 0.0883,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.28431372549019607,
|
| 412 |
+
"grad_norm": 0.9304268956184387,
|
| 413 |
+
"learning_rate": 0.00019778680490637902,
|
| 414 |
+
"loss": 0.0899,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.28921568627450983,
|
| 419 |
+
"grad_norm": 0.7331838607788086,
|
| 420 |
+
"learning_rate": 0.00019764905730275184,
|
| 421 |
+
"loss": 0.0709,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.29411764705882354,
|
| 426 |
+
"grad_norm": 0.47580260038375854,
|
| 427 |
+
"learning_rate": 0.00019750720238414425,
|
| 428 |
+
"loss": 0.0857,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.29901960784313725,
|
| 433 |
+
"grad_norm": 0.5752102732658386,
|
| 434 |
+
"learning_rate": 0.0001973612461172583,
|
| 435 |
+
"loss": 0.0838,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.30392156862745096,
|
| 440 |
+
"grad_norm": 0.4644894599914551,
|
| 441 |
+
"learning_rate": 0.00019721119464130707,
|
| 442 |
+
"loss": 0.0851,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.3088235294117647,
|
| 447 |
+
"grad_norm": 0.7036497592926025,
|
| 448 |
+
"learning_rate": 0.00019705705426775616,
|
| 449 |
+
"loss": 0.0741,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.3137254901960784,
|
| 454 |
+
"grad_norm": 0.499897301197052,
|
| 455 |
+
"learning_rate": 0.0001968988314800585,
|
| 456 |
+
"loss": 0.0718,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.31862745098039214,
|
| 461 |
+
"grad_norm": 0.7211794853210449,
|
| 462 |
+
"learning_rate": 0.0001967365329333816,
|
| 463 |
+
"loss": 0.0798,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.3235294117647059,
|
| 468 |
+
"grad_norm": 0.6176502108573914,
|
| 469 |
+
"learning_rate": 0.0001965701654543274,
|
| 470 |
+
"loss": 0.0695,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.3284313725490196,
|
| 475 |
+
"grad_norm": 0.7395395636558533,
|
| 476 |
+
"learning_rate": 0.0001963997360406454,
|
| 477 |
+
"loss": 0.0581,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.3333333333333333,
|
| 482 |
+
"grad_norm": 0.5304160714149475,
|
| 483 |
+
"learning_rate": 0.00019622525186093818,
|
| 484 |
+
"loss": 0.0826,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.3382352941176471,
|
| 489 |
+
"grad_norm": 0.46235349774360657,
|
| 490 |
+
"learning_rate": 0.0001960467202543599,
|
| 491 |
+
"loss": 0.056,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.3431372549019608,
|
| 496 |
+
"grad_norm": 0.5242049098014832,
|
| 497 |
+
"learning_rate": 0.00019586414873030758,
|
| 498 |
+
"loss": 0.0728,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.3480392156862745,
|
| 503 |
+
"grad_norm": 0.552486777305603,
|
| 504 |
+
"learning_rate": 0.00019567754496810534,
|
| 505 |
+
"loss": 0.0806,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.35294117647058826,
|
| 510 |
+
"grad_norm": 0.5002785325050354,
|
| 511 |
+
"learning_rate": 0.0001954869168166812,
|
| 512 |
+
"loss": 0.0643,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.35784313725490197,
|
| 517 |
+
"grad_norm": 0.47353097796440125,
|
| 518 |
+
"learning_rate": 0.00019529227229423717,
|
| 519 |
+
"loss": 0.0838,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.3627450980392157,
|
| 524 |
+
"grad_norm": 0.4200286865234375,
|
| 525 |
+
"learning_rate": 0.00019509361958791174,
|
| 526 |
+
"loss": 0.0776,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.36764705882352944,
|
| 531 |
+
"grad_norm": 0.6603316068649292,
|
| 532 |
+
"learning_rate": 0.00019489096705343578,
|
| 533 |
+
"loss": 0.0705,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.37254901960784315,
|
| 538 |
+
"grad_norm": 0.37562692165374756,
|
| 539 |
+
"learning_rate": 0.0001946843232147809,
|
| 540 |
+
"loss": 0.072,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.37745098039215685,
|
| 545 |
+
"grad_norm": 0.6199838519096375,
|
| 546 |
+
"learning_rate": 0.0001944736967638009,
|
| 547 |
+
"loss": 0.0649,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.38235294117647056,
|
| 552 |
+
"grad_norm": 0.7614375948905945,
|
| 553 |
+
"learning_rate": 0.0001942590965598663,
|
| 554 |
+
"loss": 0.0735,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.3872549019607843,
|
| 559 |
+
"grad_norm": 0.671489953994751,
|
| 560 |
+
"learning_rate": 0.00019404053162949155,
|
| 561 |
+
"loss": 0.065,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.39215686274509803,
|
| 566 |
+
"grad_norm": 0.5170246362686157,
|
| 567 |
+
"learning_rate": 0.0001938180111659556,
|
| 568 |
+
"loss": 0.078,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.39705882352941174,
|
| 573 |
+
"grad_norm": 0.5392031073570251,
|
| 574 |
+
"learning_rate": 0.00019359154452891483,
|
| 575 |
+
"loss": 0.063,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.4019607843137255,
|
| 580 |
+
"grad_norm": 0.6858069896697998,
|
| 581 |
+
"learning_rate": 0.00019336114124400978,
|
| 582 |
+
"loss": 0.0783,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.4068627450980392,
|
| 587 |
+
"grad_norm": 0.7257099151611328,
|
| 588 |
+
"learning_rate": 0.0001931268110024642,
|
| 589 |
+
"loss": 0.0798,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.4117647058823529,
|
| 594 |
+
"grad_norm": 0.7296270132064819,
|
| 595 |
+
"learning_rate": 0.00019288856366067746,
|
| 596 |
+
"loss": 0.0619,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.4166666666666667,
|
| 601 |
+
"grad_norm": 0.6048017740249634,
|
| 602 |
+
"learning_rate": 0.0001926464092398101,
|
| 603 |
+
"loss": 0.0634,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.4215686274509804,
|
| 608 |
+
"grad_norm": 0.3223126232624054,
|
| 609 |
+
"learning_rate": 0.00019240035792536216,
|
| 610 |
+
"loss": 0.0755,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.4264705882352941,
|
| 615 |
+
"grad_norm": 0.45046138763427734,
|
| 616 |
+
"learning_rate": 0.0001921504200667449,
|
| 617 |
+
"loss": 0.0661,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.43137254901960786,
|
| 622 |
+
"grad_norm": 0.609027624130249,
|
| 623 |
+
"learning_rate": 0.00019189660617684537,
|
| 624 |
+
"loss": 0.0711,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.4362745098039216,
|
| 629 |
+
"grad_norm": 0.4166688323020935,
|
| 630 |
+
"learning_rate": 0.00019163892693158425,
|
| 631 |
+
"loss": 0.0644,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.4411764705882353,
|
| 636 |
+
"grad_norm": 0.6250641345977783,
|
| 637 |
+
"learning_rate": 0.00019137739316946685,
|
| 638 |
+
"loss": 0.0674,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.44607843137254904,
|
| 643 |
+
"grad_norm": 0.6781248450279236,
|
| 644 |
+
"learning_rate": 0.00019111201589112718,
|
| 645 |
+
"loss": 0.0657,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.45098039215686275,
|
| 650 |
+
"grad_norm": 0.9098891615867615,
|
| 651 |
+
"learning_rate": 0.00019084280625886516,
|
| 652 |
+
"loss": 0.0765,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.45588235294117646,
|
| 657 |
+
"grad_norm": 0.5926252603530884,
|
| 658 |
+
"learning_rate": 0.00019056977559617731,
|
| 659 |
+
"loss": 0.0896,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.46078431372549017,
|
| 664 |
+
"grad_norm": 0.6467915773391724,
|
| 665 |
+
"learning_rate": 0.0001902929353872803,
|
| 666 |
+
"loss": 0.0595,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.46568627450980393,
|
| 671 |
+
"grad_norm": 0.4950433671474457,
|
| 672 |
+
"learning_rate": 0.0001900122972766279,
|
| 673 |
+
"loss": 0.0651,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.47058823529411764,
|
| 678 |
+
"grad_norm": 0.6317784190177917,
|
| 679 |
+
"learning_rate": 0.0001897278730684213,
|
| 680 |
+
"loss": 0.08,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.47549019607843135,
|
| 685 |
+
"grad_norm": 0.47558578848838806,
|
| 686 |
+
"learning_rate": 0.0001894396747261125,
|
| 687 |
+
"loss": 0.0622,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.4803921568627451,
|
| 692 |
+
"grad_norm": 0.5610472559928894,
|
| 693 |
+
"learning_rate": 0.0001891477143719012,
|
| 694 |
+
"loss": 0.0667,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.4852941176470588,
|
| 699 |
+
"grad_norm": 0.7227151989936829,
|
| 700 |
+
"learning_rate": 0.00018885200428622474,
|
| 701 |
+
"loss": 0.0648,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.49019607843137253,
|
| 706 |
+
"grad_norm": 0.49453797936439514,
|
| 707 |
+
"learning_rate": 0.0001885525569072418,
|
| 708 |
+
"loss": 0.0663,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.4950980392156863,
|
| 713 |
+
"grad_norm": 0.4297734200954437,
|
| 714 |
+
"learning_rate": 0.000188249384830309,
|
| 715 |
+
"loss": 0.0779,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.5,
|
| 720 |
+
"grad_norm": 0.39416739344596863,
|
| 721 |
+
"learning_rate": 0.00018794250080745136,
|
| 722 |
+
"loss": 0.0577,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.5049019607843137,
|
| 727 |
+
"grad_norm": 0.6955050230026245,
|
| 728 |
+
"learning_rate": 0.0001876319177468256,
|
| 729 |
+
"loss": 0.0579,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.5098039215686274,
|
| 734 |
+
"grad_norm": 0.5533928871154785,
|
| 735 |
+
"learning_rate": 0.00018731764871217753,
|
| 736 |
+
"loss": 0.0583,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.5147058823529411,
|
| 741 |
+
"grad_norm": 0.4718644618988037,
|
| 742 |
+
"learning_rate": 0.00018699970692229233,
|
| 743 |
+
"loss": 0.0609,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.5196078431372549,
|
| 748 |
+
"grad_norm": 0.39921796321868896,
|
| 749 |
+
"learning_rate": 0.00018667810575043864,
|
| 750 |
+
"loss": 0.0612,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.5245098039215687,
|
| 755 |
+
"grad_norm": 0.34913963079452515,
|
| 756 |
+
"learning_rate": 0.0001863528587238061,
|
| 757 |
+
"loss": 0.0522,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.5294117647058824,
|
| 762 |
+
"grad_norm": 0.5829554796218872,
|
| 763 |
+
"learning_rate": 0.00018602397952293618,
|
| 764 |
+
"loss": 0.0651,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.5343137254901961,
|
| 769 |
+
"grad_norm": 0.7142338156700134,
|
| 770 |
+
"learning_rate": 0.00018569148198114695,
|
| 771 |
+
"loss": 0.0643,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.5392156862745098,
|
| 776 |
+
"grad_norm": 0.24581728875637054,
|
| 777 |
+
"learning_rate": 0.00018535538008395124,
|
| 778 |
+
"loss": 0.0537,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.5441176470588235,
|
| 783 |
+
"grad_norm": 0.41139382123947144,
|
| 784 |
+
"learning_rate": 0.0001850156879684681,
|
| 785 |
+
"loss": 0.0631,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.5490196078431373,
|
| 790 |
+
"grad_norm": 0.4532317519187927,
|
| 791 |
+
"learning_rate": 0.00018467241992282843,
|
| 792 |
+
"loss": 0.0573,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.553921568627451,
|
| 797 |
+
"grad_norm": 0.45865532755851746,
|
| 798 |
+
"learning_rate": 0.00018432559038557397,
|
| 799 |
+
"loss": 0.053,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.5588235294117647,
|
| 804 |
+
"grad_norm": 0.3976840078830719,
|
| 805 |
+
"learning_rate": 0.00018397521394504995,
|
| 806 |
+
"loss": 0.0529,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.5637254901960784,
|
| 811 |
+
"grad_norm": 0.47105035185813904,
|
| 812 |
+
"learning_rate": 0.00018362130533879133,
|
| 813 |
+
"loss": 0.0671,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.5686274509803921,
|
| 818 |
+
"grad_norm": 0.5433268547058105,
|
| 819 |
+
"learning_rate": 0.00018326387945290313,
|
| 820 |
+
"loss": 0.0529,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.5735294117647058,
|
| 825 |
+
"grad_norm": 0.6220820546150208,
|
| 826 |
+
"learning_rate": 0.00018290295132143415,
|
| 827 |
+
"loss": 0.0697,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.5784313725490197,
|
| 832 |
+
"grad_norm": 0.38451075553894043,
|
| 833 |
+
"learning_rate": 0.00018253853612574473,
|
| 834 |
+
"loss": 0.0621,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.5833333333333334,
|
| 839 |
+
"grad_norm": 0.49342110753059387,
|
| 840 |
+
"learning_rate": 0.00018217064919386807,
|
| 841 |
+
"loss": 0.0603,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.5882352941176471,
|
| 846 |
+
"grad_norm": 0.5240128636360168,
|
| 847 |
+
"learning_rate": 0.00018179930599986554,
|
| 848 |
+
"loss": 0.0614,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.5931372549019608,
|
| 853 |
+
"grad_norm": 0.5015797019004822,
|
| 854 |
+
"learning_rate": 0.0001814245221631758,
|
| 855 |
+
"loss": 0.0651,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.5980392156862745,
|
| 860 |
+
"grad_norm": 0.7029892802238464,
|
| 861 |
+
"learning_rate": 0.0001810463134479579,
|
| 862 |
+
"loss": 0.0598,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.6029411764705882,
|
| 867 |
+
"grad_norm": 0.3569225072860718,
|
| 868 |
+
"learning_rate": 0.00018066469576242806,
|
| 869 |
+
"loss": 0.0479,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.6078431372549019,
|
| 874 |
+
"grad_norm": 0.4940333068370819,
|
| 875 |
+
"learning_rate": 0.00018027968515819072,
|
| 876 |
+
"loss": 0.055,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.6127450980392157,
|
| 881 |
+
"grad_norm": 0.5233299732208252,
|
| 882 |
+
"learning_rate": 0.00017989129782956323,
|
| 883 |
+
"loss": 0.0555,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.6176470588235294,
|
| 888 |
+
"grad_norm": 0.35107848048210144,
|
| 889 |
+
"learning_rate": 0.00017949955011289465,
|
| 890 |
+
"loss": 0.0472,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.6225490196078431,
|
| 895 |
+
"grad_norm": 0.42003870010375977,
|
| 896 |
+
"learning_rate": 0.00017910445848587885,
|
| 897 |
+
"loss": 0.0454,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.6274509803921569,
|
| 902 |
+
"grad_norm": 0.24233393371105194,
|
| 903 |
+
"learning_rate": 0.00017870603956686117,
|
| 904 |
+
"loss": 0.0631,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.6323529411764706,
|
| 909 |
+
"grad_norm": 0.5557372570037842,
|
| 910 |
+
"learning_rate": 0.0001783043101141395,
|
| 911 |
+
"loss": 0.0628,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.6372549019607843,
|
| 916 |
+
"grad_norm": 0.33980950713157654,
|
| 917 |
+
"learning_rate": 0.00017789928702525952,
|
| 918 |
+
"loss": 0.0591,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.6421568627450981,
|
| 923 |
+
"grad_norm": 0.2716699242591858,
|
| 924 |
+
"learning_rate": 0.00017749098733630368,
|
| 925 |
+
"loss": 0.0584,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.6470588235294118,
|
| 930 |
+
"grad_norm": 0.42181700468063354,
|
| 931 |
+
"learning_rate": 0.00017707942822117495,
|
| 932 |
+
"loss": 0.0572,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.6519607843137255,
|
| 937 |
+
"grad_norm": 0.46250826120376587,
|
| 938 |
+
"learning_rate": 0.00017666462699087422,
|
| 939 |
+
"loss": 0.0614,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.6568627450980392,
|
| 944 |
+
"grad_norm": 0.7147281169891357,
|
| 945 |
+
"learning_rate": 0.00017624660109277223,
|
| 946 |
+
"loss": 0.0666,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.6617647058823529,
|
| 951 |
+
"grad_norm": 0.6105577945709229,
|
| 952 |
+
"learning_rate": 0.00017582536810987576,
|
| 953 |
+
"loss": 0.0508,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.6666666666666666,
|
| 958 |
+
"grad_norm": 0.5564696788787842,
|
| 959 |
+
"learning_rate": 0.00017540094576008796,
|
| 960 |
+
"loss": 0.0581,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.6715686274509803,
|
| 965 |
+
"grad_norm": 0.4955359399318695,
|
| 966 |
+
"learning_rate": 0.00017497335189546308,
|
| 967 |
+
"loss": 0.0569,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.6764705882352942,
|
| 972 |
+
"grad_norm": 0.40812528133392334,
|
| 973 |
+
"learning_rate": 0.0001745426045014558,
|
| 974 |
+
"loss": 0.065,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.6813725490196079,
|
| 979 |
+
"grad_norm": 0.31670013070106506,
|
| 980 |
+
"learning_rate": 0.00017410872169616447,
|
| 981 |
+
"loss": 0.0632,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.6862745098039216,
|
| 986 |
+
"grad_norm": 0.5076479911804199,
|
| 987 |
+
"learning_rate": 0.00017367172172956906,
|
| 988 |
+
"loss": 0.0558,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.6911764705882353,
|
| 993 |
+
"grad_norm": 0.5511890053749084,
|
| 994 |
+
"learning_rate": 0.0001732316229827637,
|
| 995 |
+
"loss": 0.0669,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.696078431372549,
|
| 1000 |
+
"grad_norm": 0.32073989510536194,
|
| 1001 |
+
"learning_rate": 0.00017278844396718336,
|
| 1002 |
+
"loss": 0.0543,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.7009803921568627,
|
| 1007 |
+
"grad_norm": 0.5955519080162048,
|
| 1008 |
+
"learning_rate": 0.00017234220332382528,
|
| 1009 |
+
"loss": 0.0594,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.7058823529411765,
|
| 1014 |
+
"grad_norm": 0.5735410451889038,
|
| 1015 |
+
"learning_rate": 0.00017189291982246493,
|
| 1016 |
+
"loss": 0.0498,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.7107843137254902,
|
| 1021 |
+
"grad_norm": 0.4336249530315399,
|
| 1022 |
+
"learning_rate": 0.0001714406123608665,
|
| 1023 |
+
"loss": 0.0577,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.7156862745098039,
|
| 1028 |
+
"grad_norm": 0.2760525047779083,
|
| 1029 |
+
"learning_rate": 0.00017098529996398796,
|
| 1030 |
+
"loss": 0.05,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.7205882352941176,
|
| 1035 |
+
"grad_norm": 0.466795951128006,
|
| 1036 |
+
"learning_rate": 0.00017052700178318088,
|
| 1037 |
+
"loss": 0.0435,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.7254901960784313,
|
| 1042 |
+
"grad_norm": 0.2780659794807434,
|
| 1043 |
+
"learning_rate": 0.00017006573709538492,
|
| 1044 |
+
"loss": 0.0516,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.7303921568627451,
|
| 1049 |
+
"grad_norm": 0.31002551317214966,
|
| 1050 |
+
"learning_rate": 0.00016960152530231696,
|
| 1051 |
+
"loss": 0.0494,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.7352941176470589,
|
| 1056 |
+
"grad_norm": 0.4070112407207489,
|
| 1057 |
+
"learning_rate": 0.00016913438592965497,
|
| 1058 |
+
"loss": 0.0594,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7401960784313726,
|
| 1063 |
+
"grad_norm": 0.5858240127563477,
|
| 1064 |
+
"learning_rate": 0.00016866433862621692,
|
| 1065 |
+
"loss": 0.0421,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.7450980392156863,
|
| 1070 |
+
"grad_norm": 0.2468300759792328,
|
| 1071 |
+
"learning_rate": 0.00016819140316313397,
|
| 1072 |
+
"loss": 0.0499,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.75,
|
| 1077 |
+
"grad_norm": 0.4881417155265808,
|
| 1078 |
+
"learning_rate": 0.00016771559943301926,
|
| 1079 |
+
"loss": 0.0557,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.7549019607843137,
|
| 1084 |
+
"grad_norm": 0.428586483001709,
|
| 1085 |
+
"learning_rate": 0.00016723694744913087,
|
| 1086 |
+
"loss": 0.0547,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.7598039215686274,
|
| 1091 |
+
"grad_norm": 0.4626677334308624,
|
| 1092 |
+
"learning_rate": 0.0001667554673445302,
|
| 1093 |
+
"loss": 0.0577,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.7647058823529411,
|
| 1098 |
+
"grad_norm": 0.44555196166038513,
|
| 1099 |
+
"learning_rate": 0.000166271179371235,
|
| 1100 |
+
"loss": 0.0498,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.7696078431372549,
|
| 1105 |
+
"grad_norm": 0.4426194131374359,
|
| 1106 |
+
"learning_rate": 0.0001657841038993677,
|
| 1107 |
+
"loss": 0.0491,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.7745098039215687,
|
| 1112 |
+
"grad_norm": 0.3198953866958618,
|
| 1113 |
+
"learning_rate": 0.00016529426141629843,
|
| 1114 |
+
"loss": 0.0472,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.7794117647058824,
|
| 1119 |
+
"grad_norm": 0.2759292423725128,
|
| 1120 |
+
"learning_rate": 0.0001648016725257834,
|
| 1121 |
+
"loss": 0.0508,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.7843137254901961,
|
| 1126 |
+
"grad_norm": 0.33216890692710876,
|
| 1127 |
+
"learning_rate": 0.00016430635794709817,
|
| 1128 |
+
"loss": 0.0516,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.7892156862745098,
|
| 1133 |
+
"grad_norm": 0.40735071897506714,
|
| 1134 |
+
"learning_rate": 0.0001638083385141662,
|
| 1135 |
+
"loss": 0.0549,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.7941176470588235,
|
| 1140 |
+
"grad_norm": 0.3739156424999237,
|
| 1141 |
+
"learning_rate": 0.0001633076351746827,
|
| 1142 |
+
"loss": 0.0543,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.7990196078431373,
|
| 1147 |
+
"grad_norm": 0.2865970730781555,
|
| 1148 |
+
"learning_rate": 0.0001628042689892331,
|
| 1149 |
+
"loss": 0.0557,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.803921568627451,
|
| 1154 |
+
"grad_norm": 0.6009504795074463,
|
| 1155 |
+
"learning_rate": 0.00016229826113040767,
|
| 1156 |
+
"loss": 0.0481,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.8088235294117647,
|
| 1161 |
+
"grad_norm": 0.2973253130912781,
|
| 1162 |
+
"learning_rate": 0.00016178963288191072,
|
| 1163 |
+
"loss": 0.0465,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.8137254901960784,
|
| 1168 |
+
"grad_norm": 0.36205539107322693,
|
| 1169 |
+
"learning_rate": 0.00016127840563766527,
|
| 1170 |
+
"loss": 0.0676,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.8186274509803921,
|
| 1175 |
+
"grad_norm": 0.41221827268600464,
|
| 1176 |
+
"learning_rate": 0.0001607646009009135,
|
| 1177 |
+
"loss": 0.0544,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.8235294117647058,
|
| 1182 |
+
"grad_norm": 0.7448700666427612,
|
| 1183 |
+
"learning_rate": 0.00016024824028331195,
|
| 1184 |
+
"loss": 0.0544,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.8284313725490197,
|
| 1189 |
+
"grad_norm": 0.5625671744346619,
|
| 1190 |
+
"learning_rate": 0.0001597293455040227,
|
| 1191 |
+
"loss": 0.0659,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.8333333333333334,
|
| 1196 |
+
"grad_norm": 0.5039217472076416,
|
| 1197 |
+
"learning_rate": 0.00015920793838879966,
|
| 1198 |
+
"loss": 0.0522,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.8382352941176471,
|
| 1203 |
+
"grad_norm": 0.6583822965621948,
|
| 1204 |
+
"learning_rate": 0.00015868404086907077,
|
| 1205 |
+
"loss": 0.0473,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.8431372549019608,
|
| 1210 |
+
"grad_norm": 0.4799802899360657,
|
| 1211 |
+
"learning_rate": 0.00015815767498101522,
|
| 1212 |
+
"loss": 0.0502,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.8480392156862745,
|
| 1217 |
+
"grad_norm": 0.38270649313926697,
|
| 1218 |
+
"learning_rate": 0.00015762886286463683,
|
| 1219 |
+
"loss": 0.0488,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.8529411764705882,
|
| 1224 |
+
"grad_norm": 0.5844431519508362,
|
| 1225 |
+
"learning_rate": 0.0001570976267628326,
|
| 1226 |
+
"loss": 0.0506,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.8578431372549019,
|
| 1231 |
+
"grad_norm": 0.3698306977748871,
|
| 1232 |
+
"learning_rate": 0.00015656398902045727,
|
| 1233 |
+
"loss": 0.0474,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.8627450980392157,
|
| 1238 |
+
"grad_norm": 0.4788702428340912,
|
| 1239 |
+
"learning_rate": 0.00015602797208338337,
|
| 1240 |
+
"loss": 0.0452,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.8676470588235294,
|
| 1245 |
+
"grad_norm": 0.47195127606391907,
|
| 1246 |
+
"learning_rate": 0.00015548959849755715,
|
| 1247 |
+
"loss": 0.0497,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.8725490196078431,
|
| 1252 |
+
"grad_norm": 0.3890342712402344,
|
| 1253 |
+
"learning_rate": 0.00015494889090805018,
|
| 1254 |
+
"loss": 0.0466,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.8774509803921569,
|
| 1259 |
+
"grad_norm": 0.24695785343647003,
|
| 1260 |
+
"learning_rate": 0.00015440587205810692,
|
| 1261 |
+
"loss": 0.0525,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.8823529411764706,
|
| 1266 |
+
"grad_norm": 0.41040509939193726,
|
| 1267 |
+
"learning_rate": 0.00015386056478818814,
|
| 1268 |
+
"loss": 0.0556,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.8872549019607843,
|
| 1273 |
+
"grad_norm": 0.5128910541534424,
|
| 1274 |
+
"learning_rate": 0.00015331299203501,
|
| 1275 |
+
"loss": 0.0485,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.8921568627450981,
|
| 1280 |
+
"grad_norm": 0.3270062804222107,
|
| 1281 |
+
"learning_rate": 0.0001527631768305796,
|
| 1282 |
+
"loss": 0.0423,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.8970588235294118,
|
| 1287 |
+
"grad_norm": 0.43017369508743286,
|
| 1288 |
+
"learning_rate": 0.00015221114230122584,
|
| 1289 |
+
"loss": 0.0461,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.9019607843137255,
|
| 1294 |
+
"grad_norm": 0.501888632774353,
|
| 1295 |
+
"learning_rate": 0.00015165691166662705,
|
| 1296 |
+
"loss": 0.0472,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.9068627450980392,
|
| 1301 |
+
"grad_norm": 0.286864310503006,
|
| 1302 |
+
"learning_rate": 0.00015110050823883406,
|
| 1303 |
+
"loss": 0.0418,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.9117647058823529,
|
| 1308 |
+
"grad_norm": 0.5109400153160095,
|
| 1309 |
+
"learning_rate": 0.00015054195542128968,
|
| 1310 |
+
"loss": 0.0426,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.9166666666666666,
|
| 1315 |
+
"grad_norm": 0.32924002408981323,
|
| 1316 |
+
"learning_rate": 0.00014998127670784448,
|
| 1317 |
+
"loss": 0.0389,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.9215686274509803,
|
| 1322 |
+
"grad_norm": 0.21681684255599976,
|
| 1323 |
+
"learning_rate": 0.0001494184956817684,
|
| 1324 |
+
"loss": 0.0487,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.9264705882352942,
|
| 1329 |
+
"grad_norm": 0.20006486773490906,
|
| 1330 |
+
"learning_rate": 0.00014885363601475888,
|
| 1331 |
+
"loss": 0.0521,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.9313725490196079,
|
| 1336 |
+
"grad_norm": 0.6207095980644226,
|
| 1337 |
+
"learning_rate": 0.00014828672146594511,
|
| 1338 |
+
"loss": 0.0542,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.9362745098039216,
|
| 1343 |
+
"grad_norm": 0.5559191703796387,
|
| 1344 |
+
"learning_rate": 0.00014771777588088884,
|
| 1345 |
+
"loss": 0.0446,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.9411764705882353,
|
| 1350 |
+
"grad_norm": 0.3139231503009796,
|
| 1351 |
+
"learning_rate": 0.00014714682319058112,
|
| 1352 |
+
"loss": 0.0403,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.946078431372549,
|
| 1357 |
+
"grad_norm": 0.3132689893245697,
|
| 1358 |
+
"learning_rate": 0.00014657388741043606,
|
| 1359 |
+
"loss": 0.0398,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.9509803921568627,
|
| 1364 |
+
"grad_norm": 0.2745281457901001,
|
| 1365 |
+
"learning_rate": 0.00014599899263928028,
|
| 1366 |
+
"loss": 0.0358,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.9558823529411765,
|
| 1371 |
+
"grad_norm": 0.2483411729335785,
|
| 1372 |
+
"learning_rate": 0.00014542216305833968,
|
| 1373 |
+
"loss": 0.0506,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.9607843137254902,
|
| 1378 |
+
"grad_norm": 0.3583033084869385,
|
| 1379 |
+
"learning_rate": 0.000144843422930222,
|
| 1380 |
+
"loss": 0.0442,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.9656862745098039,
|
| 1385 |
+
"grad_norm": 0.2526845932006836,
|
| 1386 |
+
"learning_rate": 0.00014426279659789651,
|
| 1387 |
+
"loss": 0.0458,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.9705882352941176,
|
| 1392 |
+
"grad_norm": 0.3186612129211426,
|
| 1393 |
+
"learning_rate": 0.00014368030848367,
|
| 1394 |
+
"loss": 0.052,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.9754901960784313,
|
| 1399 |
+
"grad_norm": 0.3088064193725586,
|
| 1400 |
+
"learning_rate": 0.00014309598308815945,
|
| 1401 |
+
"loss": 0.0453,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.9803921568627451,
|
| 1406 |
+
"grad_norm": 0.44008663296699524,
|
| 1407 |
+
"learning_rate": 0.00014250984498926167,
|
| 1408 |
+
"loss": 0.0449,
|
| 1409 |
+
"step": 2000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.9852941176470589,
|
| 1413 |
+
"grad_norm": 0.5169370174407959,
|
| 1414 |
+
"learning_rate": 0.0001419219188411194,
|
| 1415 |
+
"loss": 0.0411,
|
| 1416 |
+
"step": 2010
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.9901960784313726,
|
| 1420 |
+
"grad_norm": 0.32191041111946106,
|
| 1421 |
+
"learning_rate": 0.0001413322293730842,
|
| 1422 |
+
"loss": 0.0406,
|
| 1423 |
+
"step": 2020
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.9950980392156863,
|
| 1427 |
+
"grad_norm": 0.300047904253006,
|
| 1428 |
+
"learning_rate": 0.00014074080138867654,
|
| 1429 |
+
"loss": 0.0432,
|
| 1430 |
+
"step": 2030
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.0,
|
| 1434 |
+
"grad_norm": 0.6571453809738159,
|
| 1435 |
+
"learning_rate": 0.00014014765976454231,
|
| 1436 |
+
"loss": 0.0421,
|
| 1437 |
+
"step": 2040
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 1.0049019607843137,
|
| 1441 |
+
"grad_norm": 0.43792489171028137,
|
| 1442 |
+
"learning_rate": 0.00013955282944940652,
|
| 1443 |
+
"loss": 0.0389,
|
| 1444 |
+
"step": 2050
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 1.0098039215686274,
|
| 1448 |
+
"grad_norm": 0.40551769733428955,
|
| 1449 |
+
"learning_rate": 0.0001389563354630239,
|
| 1450 |
+
"loss": 0.0418,
|
| 1451 |
+
"step": 2060
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 1.0147058823529411,
|
| 1455 |
+
"grad_norm": 0.3299430012702942,
|
| 1456 |
+
"learning_rate": 0.0001383582028951265,
|
| 1457 |
+
"loss": 0.0465,
|
| 1458 |
+
"step": 2070
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 1.0196078431372548,
|
| 1462 |
+
"grad_norm": 0.2800670266151428,
|
| 1463 |
+
"learning_rate": 0.00013775845690436848,
|
| 1464 |
+
"loss": 0.0443,
|
| 1465 |
+
"step": 2080
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.0245098039215685,
|
| 1469 |
+
"grad_norm": 0.49185529351234436,
|
| 1470 |
+
"learning_rate": 0.00013715712271726772,
|
| 1471 |
+
"loss": 0.0415,
|
| 1472 |
+
"step": 2090
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.0294117647058822,
|
| 1476 |
+
"grad_norm": 0.6726065278053284,
|
| 1477 |
+
"learning_rate": 0.0001365542256271448,
|
| 1478 |
+
"loss": 0.038,
|
| 1479 |
+
"step": 2100
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 1.0343137254901962,
|
| 1483 |
+
"grad_norm": 0.5443005561828613,
|
| 1484 |
+
"learning_rate": 0.00013594979099305928,
|
| 1485 |
+
"loss": 0.0407,
|
| 1486 |
+
"step": 2110
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 1.0392156862745099,
|
| 1490 |
+
"grad_norm": 0.3882359564304352,
|
| 1491 |
+
"learning_rate": 0.00013534384423874272,
|
| 1492 |
+
"loss": 0.0479,
|
| 1493 |
+
"step": 2120
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 1.0441176470588236,
|
| 1497 |
+
"grad_norm": 0.3460078239440918,
|
| 1498 |
+
"learning_rate": 0.00013473641085152957,
|
| 1499 |
+
"loss": 0.0472,
|
| 1500 |
+
"step": 2130
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 1.0490196078431373,
|
| 1504 |
+
"grad_norm": 0.3264780342578888,
|
| 1505 |
+
"learning_rate": 0.00013412751638128503,
|
| 1506 |
+
"loss": 0.0374,
|
| 1507 |
+
"step": 2140
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.053921568627451,
|
| 1511 |
+
"grad_norm": 0.28663188219070435,
|
| 1512 |
+
"learning_rate": 0.0001335171864393304,
|
| 1513 |
+
"loss": 0.0386,
|
| 1514 |
+
"step": 2150
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.0588235294117647,
|
| 1518 |
+
"grad_norm": 0.22994215786457062,
|
| 1519 |
+
"learning_rate": 0.00013290544669736576,
|
| 1520 |
+
"loss": 0.0492,
|
| 1521 |
+
"step": 2160
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.0637254901960784,
|
| 1525 |
+
"grad_norm": 0.2018628716468811,
|
| 1526 |
+
"learning_rate": 0.0001322923228863902,
|
| 1527 |
+
"loss": 0.0367,
|
| 1528 |
+
"step": 2170
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.0686274509803921,
|
| 1532 |
+
"grad_norm": 0.20426690578460693,
|
| 1533 |
+
"learning_rate": 0.0001316778407956196,
|
| 1534 |
+
"loss": 0.0322,
|
| 1535 |
+
"step": 2180
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 1.0735294117647058,
|
| 1539 |
+
"grad_norm": 0.32902026176452637,
|
| 1540 |
+
"learning_rate": 0.00013106202627140163,
|
| 1541 |
+
"loss": 0.0321,
|
| 1542 |
+
"step": 2190
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 1.0784313725490196,
|
| 1546 |
+
"grad_norm": 0.3210899233818054,
|
| 1547 |
+
"learning_rate": 0.00013044490521612904,
|
| 1548 |
+
"loss": 0.0405,
|
| 1549 |
+
"step": 2200
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.0833333333333333,
|
| 1553 |
+
"grad_norm": 0.23476989567279816,
|
| 1554 |
+
"learning_rate": 0.00012982650358714967,
|
| 1555 |
+
"loss": 0.0416,
|
| 1556 |
+
"step": 2210
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 1.088235294117647,
|
| 1560 |
+
"grad_norm": 0.3046343922615051,
|
| 1561 |
+
"learning_rate": 0.000129206847395675,
|
| 1562 |
+
"loss": 0.0394,
|
| 1563 |
+
"step": 2220
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 1.093137254901961,
|
| 1567 |
+
"grad_norm": 0.27315208315849304,
|
| 1568 |
+
"learning_rate": 0.0001285859627056858,
|
| 1569 |
+
"loss": 0.0439,
|
| 1570 |
+
"step": 2230
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 1.0980392156862746,
|
| 1574 |
+
"grad_norm": 0.34966346621513367,
|
| 1575 |
+
"learning_rate": 0.00012796387563283605,
|
| 1576 |
+
"loss": 0.0387,
|
| 1577 |
+
"step": 2240
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.1029411764705883,
|
| 1581 |
+
"grad_norm": 0.24096551537513733,
|
| 1582 |
+
"learning_rate": 0.00012734061234335434,
|
| 1583 |
+
"loss": 0.0412,
|
| 1584 |
+
"step": 2250
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 1.107843137254902,
|
| 1588 |
+
"grad_norm": 0.26461055874824524,
|
| 1589 |
+
"learning_rate": 0.00012671619905294326,
|
| 1590 |
+
"loss": 0.0494,
|
| 1591 |
+
"step": 2260
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.1127450980392157,
|
| 1595 |
+
"grad_norm": 0.24669981002807617,
|
| 1596 |
+
"learning_rate": 0.0001260906620256767,
|
| 1597 |
+
"loss": 0.0396,
|
| 1598 |
+
"step": 2270
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 1.1176470588235294,
|
| 1602 |
+
"grad_norm": 0.37667712569236755,
|
| 1603 |
+
"learning_rate": 0.00012546402757289532,
|
| 1604 |
+
"loss": 0.0426,
|
| 1605 |
+
"step": 2280
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 1.1225490196078431,
|
| 1609 |
+
"grad_norm": 0.3720461130142212,
|
| 1610 |
+
"learning_rate": 0.00012483632205209953,
|
| 1611 |
+
"loss": 0.042,
|
| 1612 |
+
"step": 2290
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 1.1274509803921569,
|
| 1616 |
+
"grad_norm": 0.3975540101528168,
|
| 1617 |
+
"learning_rate": 0.0001242075718658411,
|
| 1618 |
+
"loss": 0.0464,
|
| 1619 |
+
"step": 2300
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.1323529411764706,
|
| 1623 |
+
"grad_norm": 0.34910279512405396,
|
| 1624 |
+
"learning_rate": 0.00012357780346061256,
|
| 1625 |
+
"loss": 0.0412,
|
| 1626 |
+
"step": 2310
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 1.1372549019607843,
|
| 1630 |
+
"grad_norm": 0.34874585270881653,
|
| 1631 |
+
"learning_rate": 0.00012294704332573462,
|
| 1632 |
+
"loss": 0.0458,
|
| 1633 |
+
"step": 2320
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.142156862745098,
|
| 1637 |
+
"grad_norm": 0.24936801195144653,
|
| 1638 |
+
"learning_rate": 0.0001223153179922423,
|
| 1639 |
+
"loss": 0.0437,
|
| 1640 |
+
"step": 2330
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 1.1470588235294117,
|
| 1644 |
+
"grad_norm": 0.39299267530441284,
|
| 1645 |
+
"learning_rate": 0.00012168265403176864,
|
| 1646 |
+
"loss": 0.0419,
|
| 1647 |
+
"step": 2340
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 1.1519607843137254,
|
| 1651 |
+
"grad_norm": 0.44494786858558655,
|
| 1652 |
+
"learning_rate": 0.0001210490780554274,
|
| 1653 |
+
"loss": 0.0391,
|
| 1654 |
+
"step": 2350
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 1.156862745098039,
|
| 1658 |
+
"grad_norm": 0.26526331901550293,
|
| 1659 |
+
"learning_rate": 0.00012041461671269337,
|
| 1660 |
+
"loss": 0.0338,
|
| 1661 |
+
"step": 2360
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.161764705882353,
|
| 1665 |
+
"grad_norm": 0.4189550280570984,
|
| 1666 |
+
"learning_rate": 0.00011977929669028174,
|
| 1667 |
+
"loss": 0.0441,
|
| 1668 |
+
"step": 2370
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 1.1666666666666667,
|
| 1672 |
+
"grad_norm": 0.2872006595134735,
|
| 1673 |
+
"learning_rate": 0.00011914314471102545,
|
| 1674 |
+
"loss": 0.0427,
|
| 1675 |
+
"step": 2380
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.1715686274509804,
|
| 1679 |
+
"grad_norm": 0.2823413908481598,
|
| 1680 |
+
"learning_rate": 0.0001185061875327512,
|
| 1681 |
+
"loss": 0.0443,
|
| 1682 |
+
"step": 2390
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 1.1764705882352942,
|
| 1686 |
+
"grad_norm": 0.26944294571876526,
|
| 1687 |
+
"learning_rate": 0.00011786845194715403,
|
| 1688 |
+
"loss": 0.0387,
|
| 1689 |
+
"step": 2400
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 1.1813725490196079,
|
| 1693 |
+
"grad_norm": 0.3672547936439514,
|
| 1694 |
+
"learning_rate": 0.00011722996477867026,
|
| 1695 |
+
"loss": 0.0397,
|
| 1696 |
+
"step": 2410
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 1.1862745098039216,
|
| 1700 |
+
"grad_norm": 0.2237931340932846,
|
| 1701 |
+
"learning_rate": 0.00011659075288334938,
|
| 1702 |
+
"loss": 0.0444,
|
| 1703 |
+
"step": 2420
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.1911764705882353,
|
| 1707 |
+
"grad_norm": 0.2790263295173645,
|
| 1708 |
+
"learning_rate": 0.00011595084314772429,
|
| 1709 |
+
"loss": 0.0358,
|
| 1710 |
+
"step": 2430
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 1.196078431372549,
|
| 1714 |
+
"grad_norm": 0.2825845777988434,
|
| 1715 |
+
"learning_rate": 0.00011531026248768048,
|
| 1716 |
+
"loss": 0.0368,
|
| 1717 |
+
"step": 2440
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 1.2009803921568627,
|
| 1721 |
+
"grad_norm": 0.27560582756996155,
|
| 1722 |
+
"learning_rate": 0.00011466903784732381,
|
| 1723 |
+
"loss": 0.0474,
|
| 1724 |
+
"step": 2450
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 1.2058823529411764,
|
| 1728 |
+
"grad_norm": 0.4237360656261444,
|
| 1729 |
+
"learning_rate": 0.00011402719619784734,
|
| 1730 |
+
"loss": 0.0375,
|
| 1731 |
+
"step": 2460
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 1.2107843137254901,
|
| 1735 |
+
"grad_norm": 0.3769036531448364,
|
| 1736 |
+
"learning_rate": 0.00011338476453639666,
|
| 1737 |
+
"loss": 0.0308,
|
| 1738 |
+
"step": 2470
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 1.215686274509804,
|
| 1742 |
+
"grad_norm": 0.32824084162712097,
|
| 1743 |
+
"learning_rate": 0.00011274176988493454,
|
| 1744 |
+
"loss": 0.0386,
|
| 1745 |
+
"step": 2480
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.2205882352941178,
|
| 1749 |
+
"grad_norm": 0.30048561096191406,
|
| 1750 |
+
"learning_rate": 0.0001120982392891042,
|
| 1751 |
+
"loss": 0.0391,
|
| 1752 |
+
"step": 2490
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 1.2254901960784315,
|
| 1756 |
+
"grad_norm": 0.30690962076187134,
|
| 1757 |
+
"learning_rate": 0.00011145419981709169,
|
| 1758 |
+
"loss": 0.0443,
|
| 1759 |
+
"step": 2500
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 1.2303921568627452,
|
| 1763 |
+
"grad_norm": 0.4496728479862213,
|
| 1764 |
+
"learning_rate": 0.00011080967855848755,
|
| 1765 |
+
"loss": 0.0447,
|
| 1766 |
+
"step": 2510
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 1.2352941176470589,
|
| 1770 |
+
"grad_norm": 0.25722137093544006,
|
| 1771 |
+
"learning_rate": 0.00011016470262314707,
|
| 1772 |
+
"loss": 0.0333,
|
| 1773 |
+
"step": 2520
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 1.2401960784313726,
|
| 1777 |
+
"grad_norm": 0.32415884733200073,
|
| 1778 |
+
"learning_rate": 0.00010951929914005033,
|
| 1779 |
+
"loss": 0.0375,
|
| 1780 |
+
"step": 2530
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 1.2450980392156863,
|
| 1784 |
+
"grad_norm": 0.33738696575164795,
|
| 1785 |
+
"learning_rate": 0.00010887349525616075,
|
| 1786 |
+
"loss": 0.0408,
|
| 1787 |
+
"step": 2540
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.25,
|
| 1791 |
+
"grad_norm": 0.3490372896194458,
|
| 1792 |
+
"learning_rate": 0.00010822731813528354,
|
| 1793 |
+
"loss": 0.0337,
|
| 1794 |
+
"step": 2550
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 1.2549019607843137,
|
| 1798 |
+
"grad_norm": 0.42863723635673523,
|
| 1799 |
+
"learning_rate": 0.00010758079495692294,
|
| 1800 |
+
"loss": 0.0442,
|
| 1801 |
+
"step": 2560
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 1.2598039215686274,
|
| 1805 |
+
"grad_norm": 0.29293495416641235,
|
| 1806 |
+
"learning_rate": 0.00010693395291513908,
|
| 1807 |
+
"loss": 0.0408,
|
| 1808 |
+
"step": 2570
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 1.2647058823529411,
|
| 1812 |
+
"grad_norm": 0.27024197578430176,
|
| 1813 |
+
"learning_rate": 0.00010628681921740414,
|
| 1814 |
+
"loss": 0.0377,
|
| 1815 |
+
"step": 2580
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 1.2696078431372548,
|
| 1819 |
+
"grad_norm": 0.35880324244499207,
|
| 1820 |
+
"learning_rate": 0.00010563942108345785,
|
| 1821 |
+
"loss": 0.0364,
|
| 1822 |
+
"step": 2590
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 1.2745098039215685,
|
| 1826 |
+
"grad_norm": 0.3253026306629181,
|
| 1827 |
+
"learning_rate": 0.0001049917857441628,
|
| 1828 |
+
"loss": 0.0374,
|
| 1829 |
+
"step": 2600
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 1.2794117647058822,
|
| 1833 |
+
"grad_norm": 0.33250367641448975,
|
| 1834 |
+
"learning_rate": 0.00010434394044035878,
|
| 1835 |
+
"loss": 0.0384,
|
| 1836 |
+
"step": 2610
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 1.284313725490196,
|
| 1840 |
+
"grad_norm": 0.34700486063957214,
|
| 1841 |
+
"learning_rate": 0.00010369591242171719,
|
| 1842 |
+
"loss": 0.0369,
|
| 1843 |
+
"step": 2620
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 1.2892156862745099,
|
| 1847 |
+
"grad_norm": 0.3287598788738251,
|
| 1848 |
+
"learning_rate": 0.00010304772894559475,
|
| 1849 |
+
"loss": 0.0419,
|
| 1850 |
+
"step": 2630
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 1.2941176470588236,
|
| 1854 |
+
"grad_norm": 0.3363092243671417,
|
| 1855 |
+
"learning_rate": 0.00010239941727588707,
|
| 1856 |
+
"loss": 0.0419,
|
| 1857 |
+
"step": 2640
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 1.2990196078431373,
|
| 1861 |
+
"grad_norm": 0.32619622349739075,
|
| 1862 |
+
"learning_rate": 0.0001017510046818817,
|
| 1863 |
+
"loss": 0.0353,
|
| 1864 |
+
"step": 2650
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 1.303921568627451,
|
| 1868 |
+
"grad_norm": 0.28630563616752625,
|
| 1869 |
+
"learning_rate": 0.00010110251843711149,
|
| 1870 |
+
"loss": 0.0317,
|
| 1871 |
+
"step": 2660
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 1.3088235294117647,
|
| 1875 |
+
"grad_norm": 0.470198392868042,
|
| 1876 |
+
"learning_rate": 0.00010045398581820702,
|
| 1877 |
+
"loss": 0.0397,
|
| 1878 |
+
"step": 2670
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 1.3137254901960784,
|
| 1882 |
+
"grad_norm": 0.34590962529182434,
|
| 1883 |
+
"learning_rate": 9.98054341037495e-05,
|
| 1884 |
+
"loss": 0.0336,
|
| 1885 |
+
"step": 2680
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 1.3186274509803921,
|
| 1889 |
+
"grad_norm": 0.18714579939842224,
|
| 1890 |
+
"learning_rate": 9.91568905731234e-05,
|
| 1891 |
+
"loss": 0.0268,
|
| 1892 |
+
"step": 2690
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 1.3235294117647058,
|
| 1896 |
+
"grad_norm": 0.3055776357650757,
|
| 1897 |
+
"learning_rate": 9.850838250536885e-05,
|
| 1898 |
+
"loss": 0.0384,
|
| 1899 |
+
"step": 2700
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 1.3284313725490196,
|
| 1903 |
+
"grad_norm": 0.33092889189720154,
|
| 1904 |
+
"learning_rate": 9.785993717803445e-05,
|
| 1905 |
+
"loss": 0.0323,
|
| 1906 |
+
"step": 2710
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 1.3333333333333333,
|
| 1910 |
+
"grad_norm": 0.26138371229171753,
|
| 1911 |
+
"learning_rate": 9.721158186602979e-05,
|
| 1912 |
+
"loss": 0.0391,
|
| 1913 |
+
"step": 2720
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 1.3382352941176472,
|
| 1917 |
+
"grad_norm": 0.2703782320022583,
|
| 1918 |
+
"learning_rate": 9.656334384047812e-05,
|
| 1919 |
+
"loss": 0.0268,
|
| 1920 |
+
"step": 2730
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 1.343137254901961,
|
| 1924 |
+
"grad_norm": 0.24206914007663727,
|
| 1925 |
+
"learning_rate": 9.591525036756952e-05,
|
| 1926 |
+
"loss": 0.032,
|
| 1927 |
+
"step": 2740
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 1.3480392156862746,
|
| 1931 |
+
"grad_norm": 0.3793281018733978,
|
| 1932 |
+
"learning_rate": 9.526732870741386e-05,
|
| 1933 |
+
"loss": 0.0399,
|
| 1934 |
+
"step": 2750
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 1.3529411764705883,
|
| 1938 |
+
"grad_norm": 0.3143618404865265,
|
| 1939 |
+
"learning_rate": 9.46196061128942e-05,
|
| 1940 |
+
"loss": 0.0365,
|
| 1941 |
+
"step": 2760
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 1.357843137254902,
|
| 1945 |
+
"grad_norm": 0.23540951311588287,
|
| 1946 |
+
"learning_rate": 9.397210982852053e-05,
|
| 1947 |
+
"loss": 0.0328,
|
| 1948 |
+
"step": 2770
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 1.3627450980392157,
|
| 1952 |
+
"grad_norm": 0.2023368775844574,
|
| 1953 |
+
"learning_rate": 9.332486708928373e-05,
|
| 1954 |
+
"loss": 0.0316,
|
| 1955 |
+
"step": 2780
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 1.3676470588235294,
|
| 1959 |
+
"grad_norm": 0.26689231395721436,
|
| 1960 |
+
"learning_rate": 9.267790511951015e-05,
|
| 1961 |
+
"loss": 0.0326,
|
| 1962 |
+
"step": 2790
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 1.3725490196078431,
|
| 1966 |
+
"grad_norm": 0.2792396545410156,
|
| 1967 |
+
"learning_rate": 9.203125113171631e-05,
|
| 1968 |
+
"loss": 0.0336,
|
| 1969 |
+
"step": 2800
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 1.3774509803921569,
|
| 1973 |
+
"grad_norm": 0.21045692265033722,
|
| 1974 |
+
"learning_rate": 9.13849323254645e-05,
|
| 1975 |
+
"loss": 0.0296,
|
| 1976 |
+
"step": 2810
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 1.3823529411764706,
|
| 1980 |
+
"grad_norm": 0.26224854588508606,
|
| 1981 |
+
"learning_rate": 9.073897588621853e-05,
|
| 1982 |
+
"loss": 0.0311,
|
| 1983 |
+
"step": 2820
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 1.3872549019607843,
|
| 1987 |
+
"grad_norm": 0.30219170451164246,
|
| 1988 |
+
"learning_rate": 9.009340898420029e-05,
|
| 1989 |
+
"loss": 0.0379,
|
| 1990 |
+
"step": 2830
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 1.392156862745098,
|
| 1994 |
+
"grad_norm": 0.19660678505897522,
|
| 1995 |
+
"learning_rate": 8.944825877324708e-05,
|
| 1996 |
+
"loss": 0.035,
|
| 1997 |
+
"step": 2840
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.3970588235294117,
|
| 2001 |
+
"grad_norm": 0.2348472774028778,
|
| 2002 |
+
"learning_rate": 8.880355238966923e-05,
|
| 2003 |
+
"loss": 0.0366,
|
| 2004 |
+
"step": 2850
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 1.4019607843137254,
|
| 2008 |
+
"grad_norm": 0.44347622990608215,
|
| 2009 |
+
"learning_rate": 8.815931695110885e-05,
|
| 2010 |
+
"loss": 0.0333,
|
| 2011 |
+
"step": 2860
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 1.406862745098039,
|
| 2015 |
+
"grad_norm": 0.34309887886047363,
|
| 2016 |
+
"learning_rate": 8.751557955539915e-05,
|
| 2017 |
+
"loss": 0.0394,
|
| 2018 |
+
"step": 2870
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 1.4117647058823528,
|
| 2022 |
+
"grad_norm": 0.3023855984210968,
|
| 2023 |
+
"learning_rate": 8.687236727942465e-05,
|
| 2024 |
+
"loss": 0.0308,
|
| 2025 |
+
"step": 2880
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 1.4166666666666667,
|
| 2029 |
+
"grad_norm": 0.25127673149108887,
|
| 2030 |
+
"learning_rate": 8.622970717798227e-05,
|
| 2031 |
+
"loss": 0.0384,
|
| 2032 |
+
"step": 2890
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 1.4215686274509804,
|
| 2036 |
+
"grad_norm": 0.17014305293560028,
|
| 2037 |
+
"learning_rate": 8.558762628264345e-05,
|
| 2038 |
+
"loss": 0.0331,
|
| 2039 |
+
"step": 2900
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 1.4264705882352942,
|
| 2043 |
+
"grad_norm": 0.32725852727890015,
|
| 2044 |
+
"learning_rate": 8.494615160061694e-05,
|
| 2045 |
+
"loss": 0.0326,
|
| 2046 |
+
"step": 2910
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 1.4313725490196079,
|
| 2050 |
+
"grad_norm": 0.2895604968070984,
|
| 2051 |
+
"learning_rate": 8.430531011361298e-05,
|
| 2052 |
+
"loss": 0.0319,
|
| 2053 |
+
"step": 2920
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 1.4362745098039216,
|
| 2057 |
+
"grad_norm": 0.3882890045642853,
|
| 2058 |
+
"learning_rate": 8.366512877670842e-05,
|
| 2059 |
+
"loss": 0.0331,
|
| 2060 |
+
"step": 2930
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 1.4411764705882353,
|
| 2064 |
+
"grad_norm": 0.27492624521255493,
|
| 2065 |
+
"learning_rate": 8.302563451721282e-05,
|
| 2066 |
+
"loss": 0.0384,
|
| 2067 |
+
"step": 2940
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 1.446078431372549,
|
| 2071 |
+
"grad_norm": 0.23316094279289246,
|
| 2072 |
+
"learning_rate": 8.238685423353588e-05,
|
| 2073 |
+
"loss": 0.0426,
|
| 2074 |
+
"step": 2950
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 1.4509803921568627,
|
| 2078 |
+
"grad_norm": 0.395353227853775,
|
| 2079 |
+
"learning_rate": 8.174881479405607e-05,
|
| 2080 |
+
"loss": 0.0347,
|
| 2081 |
+
"step": 2960
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 1.4558823529411764,
|
| 2085 |
+
"grad_norm": 0.5179559588432312,
|
| 2086 |
+
"learning_rate": 8.111154303599049e-05,
|
| 2087 |
+
"loss": 0.0371,
|
| 2088 |
+
"step": 2970
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 1.4607843137254901,
|
| 2092 |
+
"grad_norm": 0.22947251796722412,
|
| 2093 |
+
"learning_rate": 8.047506576426596e-05,
|
| 2094 |
+
"loss": 0.03,
|
| 2095 |
+
"step": 2980
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 1.465686274509804,
|
| 2099 |
+
"grad_norm": 0.3831785023212433,
|
| 2100 |
+
"learning_rate": 7.983940975039166e-05,
|
| 2101 |
+
"loss": 0.0346,
|
| 2102 |
+
"step": 2990
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 1.4705882352941178,
|
| 2106 |
+
"grad_norm": 0.34753093123435974,
|
| 2107 |
+
"learning_rate": 7.920460173133304e-05,
|
| 2108 |
+
"loss": 0.0464,
|
| 2109 |
+
"step": 3000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 1.4754901960784315,
|
| 2113 |
+
"grad_norm": 0.24026577174663544,
|
| 2114 |
+
"learning_rate": 7.85706684083871e-05,
|
| 2115 |
+
"loss": 0.0334,
|
| 2116 |
+
"step": 3010
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 1.4803921568627452,
|
| 2120 |
+
"grad_norm": 0.32264992594718933,
|
| 2121 |
+
"learning_rate": 7.793763644605947e-05,
|
| 2122 |
+
"loss": 0.0329,
|
| 2123 |
+
"step": 3020
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 1.4852941176470589,
|
| 2127 |
+
"grad_norm": 0.25292444229125977,
|
| 2128 |
+
"learning_rate": 7.730553247094266e-05,
|
| 2129 |
+
"loss": 0.0306,
|
| 2130 |
+
"step": 3030
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 1.4901960784313726,
|
| 2134 |
+
"grad_norm": 0.21395047008991241,
|
| 2135 |
+
"learning_rate": 7.667438307059627e-05,
|
| 2136 |
+
"loss": 0.0346,
|
| 2137 |
+
"step": 3040
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 1.4950980392156863,
|
| 2141 |
+
"grad_norm": 0.29967001080513,
|
| 2142 |
+
"learning_rate": 7.604421479242846e-05,
|
| 2143 |
+
"loss": 0.0403,
|
| 2144 |
+
"step": 3050
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 1.5,
|
| 2148 |
+
"grad_norm": 0.2828430235385895,
|
| 2149 |
+
"learning_rate": 7.541505414257959e-05,
|
| 2150 |
+
"loss": 0.0327,
|
| 2151 |
+
"step": 3060
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 1.5049019607843137,
|
| 2155 |
+
"grad_norm": 0.43027809262275696,
|
| 2156 |
+
"learning_rate": 7.478692758480698e-05,
|
| 2157 |
+
"loss": 0.0331,
|
| 2158 |
+
"step": 3070
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 1.5098039215686274,
|
| 2162 |
+
"grad_norm": 0.34473538398742676,
|
| 2163 |
+
"learning_rate": 7.415986153937202e-05,
|
| 2164 |
+
"loss": 0.0444,
|
| 2165 |
+
"step": 3080
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 1.5147058823529411,
|
| 2169 |
+
"grad_norm": 0.3717981278896332,
|
| 2170 |
+
"learning_rate": 7.353388238192892e-05,
|
| 2171 |
+
"loss": 0.042,
|
| 2172 |
+
"step": 3090
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 1.5196078431372548,
|
| 2176 |
+
"grad_norm": 0.15127846598625183,
|
| 2177 |
+
"learning_rate": 7.29090164424151e-05,
|
| 2178 |
+
"loss": 0.0296,
|
| 2179 |
+
"step": 3100
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 1.5245098039215685,
|
| 2183 |
+
"grad_norm": 0.33397403359413147,
|
| 2184 |
+
"learning_rate": 7.22852900039438e-05,
|
| 2185 |
+
"loss": 0.0314,
|
| 2186 |
+
"step": 3110
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 1.5294117647058822,
|
| 2190 |
+
"grad_norm": 0.25943371653556824,
|
| 2191 |
+
"learning_rate": 7.166272930169861e-05,
|
| 2192 |
+
"loss": 0.0342,
|
| 2193 |
+
"step": 3120
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 1.534313725490196,
|
| 2197 |
+
"grad_norm": 0.3120077848434448,
|
| 2198 |
+
"learning_rate": 7.104136052182992e-05,
|
| 2199 |
+
"loss": 0.0317,
|
| 2200 |
+
"step": 3130
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 1.5392156862745097,
|
| 2204 |
+
"grad_norm": 0.24377594888210297,
|
| 2205 |
+
"learning_rate": 7.042120980035346e-05,
|
| 2206 |
+
"loss": 0.0284,
|
| 2207 |
+
"step": 3140
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 1.5441176470588234,
|
| 2211 |
+
"grad_norm": 0.19070957601070404,
|
| 2212 |
+
"learning_rate": 6.980230322205099e-05,
|
| 2213 |
+
"loss": 0.0343,
|
| 2214 |
+
"step": 3150
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 1.5490196078431373,
|
| 2218 |
+
"grad_norm": 0.22296807169914246,
|
| 2219 |
+
"learning_rate": 6.918466681937308e-05,
|
| 2220 |
+
"loss": 0.0299,
|
| 2221 |
+
"step": 3160
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 1.553921568627451,
|
| 2225 |
+
"grad_norm": 0.2279416173696518,
|
| 2226 |
+
"learning_rate": 6.856832657134424e-05,
|
| 2227 |
+
"loss": 0.0333,
|
| 2228 |
+
"step": 3170
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 1.5588235294117647,
|
| 2232 |
+
"grad_norm": 0.2242182493209839,
|
| 2233 |
+
"learning_rate": 6.795330840247006e-05,
|
| 2234 |
+
"loss": 0.0331,
|
| 2235 |
+
"step": 3180
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 1.5637254901960784,
|
| 2239 |
+
"grad_norm": 0.2774062752723694,
|
| 2240 |
+
"learning_rate": 6.733963818164686e-05,
|
| 2241 |
+
"loss": 0.0266,
|
| 2242 |
+
"step": 3190
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 1.5686274509803921,
|
| 2246 |
+
"grad_norm": 0.37312522530555725,
|
| 2247 |
+
"learning_rate": 6.672734172107354e-05,
|
| 2248 |
+
"loss": 0.0376,
|
| 2249 |
+
"step": 3200
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 1.5735294117647058,
|
| 2253 |
+
"grad_norm": 0.23322941362857819,
|
| 2254 |
+
"learning_rate": 6.611644477516595e-05,
|
| 2255 |
+
"loss": 0.0282,
|
| 2256 |
+
"step": 3210
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 1.5784313725490198,
|
| 2260 |
+
"grad_norm": 0.32735109329223633,
|
| 2261 |
+
"learning_rate": 6.550697303947345e-05,
|
| 2262 |
+
"loss": 0.0294,
|
| 2263 |
+
"step": 3220
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 1.5833333333333335,
|
| 2267 |
+
"grad_norm": 0.21853038668632507,
|
| 2268 |
+
"learning_rate": 6.489895214959828e-05,
|
| 2269 |
+
"loss": 0.0259,
|
| 2270 |
+
"step": 3230
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 1.5882352941176472,
|
| 2274 |
+
"grad_norm": 0.3016158640384674,
|
| 2275 |
+
"learning_rate": 6.429240768011719e-05,
|
| 2276 |
+
"loss": 0.028,
|
| 2277 |
+
"step": 3240
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 1.593137254901961,
|
| 2281 |
+
"grad_norm": 0.20449745655059814,
|
| 2282 |
+
"learning_rate": 6.368736514350568e-05,
|
| 2283 |
+
"loss": 0.0303,
|
| 2284 |
+
"step": 3250
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 1.5980392156862746,
|
| 2288 |
+
"grad_norm": 0.2439008206129074,
|
| 2289 |
+
"learning_rate": 6.308384998906506e-05,
|
| 2290 |
+
"loss": 0.027,
|
| 2291 |
+
"step": 3260
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 1.6029411764705883,
|
| 2295 |
+
"grad_norm": 0.28377825021743774,
|
| 2296 |
+
"learning_rate": 6.248188760185173e-05,
|
| 2297 |
+
"loss": 0.0302,
|
| 2298 |
+
"step": 3270
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 1.607843137254902,
|
| 2302 |
+
"grad_norm": 0.32138168811798096,
|
| 2303 |
+
"learning_rate": 6.188150330160971e-05,
|
| 2304 |
+
"loss": 0.0255,
|
| 2305 |
+
"step": 3280
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 1.6127450980392157,
|
| 2309 |
+
"grad_norm": 0.23931661248207092,
|
| 2310 |
+
"learning_rate": 6.128272234170547e-05,
|
| 2311 |
+
"loss": 0.0284,
|
| 2312 |
+
"step": 3290
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 1.6176470588235294,
|
| 2316 |
+
"grad_norm": 0.26550066471099854,
|
| 2317 |
+
"learning_rate": 6.068556990806579e-05,
|
| 2318 |
+
"loss": 0.039,
|
| 2319 |
+
"step": 3300
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 1.6225490196078431,
|
| 2323 |
+
"grad_norm": 0.20831480622291565,
|
| 2324 |
+
"learning_rate": 6.0090071118118355e-05,
|
| 2325 |
+
"loss": 0.0248,
|
| 2326 |
+
"step": 3310
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 1.6274509803921569,
|
| 2330 |
+
"grad_norm": 0.26879703998565674,
|
| 2331 |
+
"learning_rate": 5.949625101973527e-05,
|
| 2332 |
+
"loss": 0.0303,
|
| 2333 |
+
"step": 3320
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 1.6323529411764706,
|
| 2337 |
+
"grad_norm": 0.34190261363983154,
|
| 2338 |
+
"learning_rate": 5.890413459017958e-05,
|
| 2339 |
+
"loss": 0.0296,
|
| 2340 |
+
"step": 3330
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 1.6372549019607843,
|
| 2344 |
+
"grad_norm": 0.3451369106769562,
|
| 2345 |
+
"learning_rate": 5.8313746735054544e-05,
|
| 2346 |
+
"loss": 0.0274,
|
| 2347 |
+
"step": 3340
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 1.642156862745098,
|
| 2351 |
+
"grad_norm": 0.2692447304725647,
|
| 2352 |
+
"learning_rate": 5.77251122872561e-05,
|
| 2353 |
+
"loss": 0.0303,
|
| 2354 |
+
"step": 3350
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 1.6470588235294117,
|
| 2358 |
+
"grad_norm": 0.28057631850242615,
|
| 2359 |
+
"learning_rate": 5.713825600592841e-05,
|
| 2360 |
+
"loss": 0.0335,
|
| 2361 |
+
"step": 3360
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 1.6519607843137254,
|
| 2365 |
+
"grad_norm": 0.20118731260299683,
|
| 2366 |
+
"learning_rate": 5.6553202575422385e-05,
|
| 2367 |
+
"loss": 0.0339,
|
| 2368 |
+
"step": 3370
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 1.656862745098039,
|
| 2372 |
+
"grad_norm": 0.27384528517723083,
|
| 2373 |
+
"learning_rate": 5.596997660425746e-05,
|
| 2374 |
+
"loss": 0.0296,
|
| 2375 |
+
"step": 3380
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 1.6617647058823528,
|
| 2379 |
+
"grad_norm": 0.27839264273643494,
|
| 2380 |
+
"learning_rate": 5.538860262408632e-05,
|
| 2381 |
+
"loss": 0.0306,
|
| 2382 |
+
"step": 3390
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 1.6666666666666665,
|
| 2386 |
+
"grad_norm": 0.26187360286712646,
|
| 2387 |
+
"learning_rate": 5.480910508866333e-05,
|
| 2388 |
+
"loss": 0.0327,
|
| 2389 |
+
"step": 3400
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 1.6715686274509802,
|
| 2393 |
+
"grad_norm": 0.16635389626026154,
|
| 2394 |
+
"learning_rate": 5.423150837281585e-05,
|
| 2395 |
+
"loss": 0.0268,
|
| 2396 |
+
"step": 3410
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 1.6764705882352942,
|
| 2400 |
+
"grad_norm": 0.3123128414154053,
|
| 2401 |
+
"learning_rate": 5.365583677141883e-05,
|
| 2402 |
+
"loss": 0.0345,
|
| 2403 |
+
"step": 3420
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 1.6813725490196079,
|
| 2407 |
+
"grad_norm": 0.30305176973342896,
|
| 2408 |
+
"learning_rate": 5.308211449837315e-05,
|
| 2409 |
+
"loss": 0.0264,
|
| 2410 |
+
"step": 3430
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 1.6862745098039216,
|
| 2414 |
+
"grad_norm": 0.19436487555503845,
|
| 2415 |
+
"learning_rate": 5.2510365685587026e-05,
|
| 2416 |
+
"loss": 0.0318,
|
| 2417 |
+
"step": 3440
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 1.6911764705882353,
|
| 2421 |
+
"grad_norm": 0.2048874795436859,
|
| 2422 |
+
"learning_rate": 5.1940614381961004e-05,
|
| 2423 |
+
"loss": 0.0296,
|
| 2424 |
+
"step": 3450
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 1.696078431372549,
|
| 2428 |
+
"grad_norm": 0.32843217253685,
|
| 2429 |
+
"learning_rate": 5.137288455237627e-05,
|
| 2430 |
+
"loss": 0.0288,
|
| 2431 |
+
"step": 3460
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 1.7009803921568627,
|
| 2435 |
+
"grad_norm": 0.26734060049057007,
|
| 2436 |
+
"learning_rate": 5.080720007668689e-05,
|
| 2437 |
+
"loss": 0.0309,
|
| 2438 |
+
"step": 3470
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 1.7058823529411766,
|
| 2442 |
+
"grad_norm": 0.2050999402999878,
|
| 2443 |
+
"learning_rate": 5.0243584748715235e-05,
|
| 2444 |
+
"loss": 0.0283,
|
| 2445 |
+
"step": 3480
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 1.7107843137254903,
|
| 2449 |
+
"grad_norm": 0.32800912857055664,
|
| 2450 |
+
"learning_rate": 4.968206227525111e-05,
|
| 2451 |
+
"loss": 0.0356,
|
| 2452 |
+
"step": 3490
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 1.715686274509804,
|
| 2456 |
+
"grad_norm": 0.2212320864200592,
|
| 2457 |
+
"learning_rate": 4.912265627505468e-05,
|
| 2458 |
+
"loss": 0.0278,
|
| 2459 |
+
"step": 3500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 1.7205882352941178,
|
| 2463 |
+
"grad_norm": 0.22088485956192017,
|
| 2464 |
+
"learning_rate": 4.856539027786305e-05,
|
| 2465 |
+
"loss": 0.0315,
|
| 2466 |
+
"step": 3510
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 1.7254901960784315,
|
| 2470 |
+
"grad_norm": 0.2626785635948181,
|
| 2471 |
+
"learning_rate": 4.8010287723400494e-05,
|
| 2472 |
+
"loss": 0.0395,
|
| 2473 |
+
"step": 3520
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 1.7303921568627452,
|
| 2477 |
+
"grad_norm": 0.343022882938385,
|
| 2478 |
+
"learning_rate": 4.745737196039259e-05,
|
| 2479 |
+
"loss": 0.0235,
|
| 2480 |
+
"step": 3530
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 1.7352941176470589,
|
| 2484 |
+
"grad_norm": 0.4321844279766083,
|
| 2485 |
+
"learning_rate": 4.6906666245583965e-05,
|
| 2486 |
+
"loss": 0.0393,
|
| 2487 |
+
"step": 3540
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 1.7401960784313726,
|
| 2491 |
+
"grad_norm": 0.34605053067207336,
|
| 2492 |
+
"learning_rate": 4.6358193742760305e-05,
|
| 2493 |
+
"loss": 0.0214,
|
| 2494 |
+
"step": 3550
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 1.7450980392156863,
|
| 2498 |
+
"grad_norm": 0.2320283055305481,
|
| 2499 |
+
"learning_rate": 4.5811977521773906e-05,
|
| 2500 |
+
"loss": 0.0331,
|
| 2501 |
+
"step": 3560
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 1.75,
|
| 2505 |
+
"grad_norm": 0.21285474300384521,
|
| 2506 |
+
"learning_rate": 4.526804055757328e-05,
|
| 2507 |
+
"loss": 0.0333,
|
| 2508 |
+
"step": 3570
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 1.7549019607843137,
|
| 2512 |
+
"grad_norm": 0.185288667678833,
|
| 2513 |
+
"learning_rate": 4.472640572923687e-05,
|
| 2514 |
+
"loss": 0.0269,
|
| 2515 |
+
"step": 3580
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 1.7598039215686274,
|
| 2519 |
+
"grad_norm": 0.2527756989002228,
|
| 2520 |
+
"learning_rate": 4.4187095819010674e-05,
|
| 2521 |
+
"loss": 0.0296,
|
| 2522 |
+
"step": 3590
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 1.7647058823529411,
|
| 2526 |
+
"grad_norm": 0.1697828769683838,
|
| 2527 |
+
"learning_rate": 4.365013351135001e-05,
|
| 2528 |
+
"loss": 0.0272,
|
| 2529 |
+
"step": 3600
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 1.7696078431372548,
|
| 2533 |
+
"grad_norm": 0.215946227312088,
|
| 2534 |
+
"learning_rate": 4.311554139196522e-05,
|
| 2535 |
+
"loss": 0.0262,
|
| 2536 |
+
"step": 3610
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 1.7745098039215685,
|
| 2540 |
+
"grad_norm": 0.41971561312675476,
|
| 2541 |
+
"learning_rate": 4.258334194687188e-05,
|
| 2542 |
+
"loss": 0.0282,
|
| 2543 |
+
"step": 3620
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 1.7794117647058822,
|
| 2547 |
+
"grad_norm": 0.18381306529045105,
|
| 2548 |
+
"learning_rate": 4.205355756144489e-05,
|
| 2549 |
+
"loss": 0.0293,
|
| 2550 |
+
"step": 3630
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 1.784313725490196,
|
| 2554 |
+
"grad_norm": 0.24697428941726685,
|
| 2555 |
+
"learning_rate": 4.152621051947682e-05,
|
| 2556 |
+
"loss": 0.0205,
|
| 2557 |
+
"step": 3640
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 1.7892156862745097,
|
| 2561 |
+
"grad_norm": 0.23640835285186768,
|
| 2562 |
+
"learning_rate": 4.1001323002240754e-05,
|
| 2563 |
+
"loss": 0.0304,
|
| 2564 |
+
"step": 3650
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 1.7941176470588234,
|
| 2568 |
+
"grad_norm": 0.21955536305904388,
|
| 2569 |
+
"learning_rate": 4.047891708755724e-05,
|
| 2570 |
+
"loss": 0.0281,
|
| 2571 |
+
"step": 3660
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 1.7990196078431373,
|
| 2575 |
+
"grad_norm": 0.34594038128852844,
|
| 2576 |
+
"learning_rate": 3.995901474886568e-05,
|
| 2577 |
+
"loss": 0.0344,
|
| 2578 |
+
"step": 3670
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 1.803921568627451,
|
| 2582 |
+
"grad_norm": 0.4010615646839142,
|
| 2583 |
+
"learning_rate": 3.944163785429992e-05,
|
| 2584 |
+
"loss": 0.0241,
|
| 2585 |
+
"step": 3680
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 1.8088235294117647,
|
| 2589 |
+
"grad_norm": 0.24570715427398682,
|
| 2590 |
+
"learning_rate": 3.8926808165768715e-05,
|
| 2591 |
+
"loss": 0.0269,
|
| 2592 |
+
"step": 3690
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 1.8137254901960784,
|
| 2596 |
+
"grad_norm": 0.2714114785194397,
|
| 2597 |
+
"learning_rate": 3.841454733804016e-05,
|
| 2598 |
+
"loss": 0.0287,
|
| 2599 |
+
"step": 3700
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 1.8186274509803921,
|
| 2603 |
+
"grad_norm": 0.29776889085769653,
|
| 2604 |
+
"learning_rate": 3.790487691783099e-05,
|
| 2605 |
+
"loss": 0.0325,
|
| 2606 |
+
"step": 3710
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 1.8235294117647058,
|
| 2610 |
+
"grad_norm": 0.20955297350883484,
|
| 2611 |
+
"learning_rate": 3.739781834290006e-05,
|
| 2612 |
+
"loss": 0.0319,
|
| 2613 |
+
"step": 3720
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 1.8284313725490198,
|
| 2617 |
+
"grad_norm": 0.2985910475254059,
|
| 2618 |
+
"learning_rate": 3.689339294114692e-05,
|
| 2619 |
+
"loss": 0.0244,
|
| 2620 |
+
"step": 3730
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 1.8333333333333335,
|
| 2624 |
+
"grad_norm": 0.33746784925460815,
|
| 2625 |
+
"learning_rate": 3.639162192971457e-05,
|
| 2626 |
+
"loss": 0.0272,
|
| 2627 |
+
"step": 3740
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 1.8382352941176472,
|
| 2631 |
+
"grad_norm": 0.2537771761417389,
|
| 2632 |
+
"learning_rate": 3.5892526414096925e-05,
|
| 2633 |
+
"loss": 0.0317,
|
| 2634 |
+
"step": 3750
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 1.843137254901961,
|
| 2638 |
+
"grad_norm": 0.30214208364486694,
|
| 2639 |
+
"learning_rate": 3.53961273872513e-05,
|
| 2640 |
+
"loss": 0.0207,
|
| 2641 |
+
"step": 3760
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 1.8480392156862746,
|
| 2645 |
+
"grad_norm": 0.3083361089229584,
|
| 2646 |
+
"learning_rate": 3.490244572871524e-05,
|
| 2647 |
+
"loss": 0.0216,
|
| 2648 |
+
"step": 3770
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 1.8529411764705883,
|
| 2652 |
+
"grad_norm": 0.24969086050987244,
|
| 2653 |
+
"learning_rate": 3.44115022037284e-05,
|
| 2654 |
+
"loss": 0.0254,
|
| 2655 |
+
"step": 3780
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 1.857843137254902,
|
| 2659 |
+
"grad_norm": 0.2084352970123291,
|
| 2660 |
+
"learning_rate": 3.3923317462358905e-05,
|
| 2661 |
+
"loss": 0.0305,
|
| 2662 |
+
"step": 3790
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 1.8627450980392157,
|
| 2666 |
+
"grad_norm": 0.21671414375305176,
|
| 2667 |
+
"learning_rate": 3.3437912038635056e-05,
|
| 2668 |
+
"loss": 0.0303,
|
| 2669 |
+
"step": 3800
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 1.8676470588235294,
|
| 2673 |
+
"grad_norm": 0.2956879436969757,
|
| 2674 |
+
"learning_rate": 3.295530634968147e-05,
|
| 2675 |
+
"loss": 0.0298,
|
| 2676 |
+
"step": 3810
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 1.8725490196078431,
|
| 2680 |
+
"grad_norm": 0.29368528723716736,
|
| 2681 |
+
"learning_rate": 3.24755206948602e-05,
|
| 2682 |
+
"loss": 0.0261,
|
| 2683 |
+
"step": 3820
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 1.8774509803921569,
|
| 2687 |
+
"grad_norm": 0.13201627135276794,
|
| 2688 |
+
"learning_rate": 3.199857525491714e-05,
|
| 2689 |
+
"loss": 0.0217,
|
| 2690 |
+
"step": 3830
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 1.8823529411764706,
|
| 2694 |
+
"grad_norm": 0.21656860411167145,
|
| 2695 |
+
"learning_rate": 3.1524490091133e-05,
|
| 2696 |
+
"loss": 0.0288,
|
| 2697 |
+
"step": 3840
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 1.8872549019607843,
|
| 2701 |
+
"grad_norm": 0.20571519434452057,
|
| 2702 |
+
"learning_rate": 3.105328514447957e-05,
|
| 2703 |
+
"loss": 0.0254,
|
| 2704 |
+
"step": 3850
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 1.892156862745098,
|
| 2708 |
+
"grad_norm": 0.2538784444332123,
|
| 2709 |
+
"learning_rate": 3.0584980234780916e-05,
|
| 2710 |
+
"loss": 0.0325,
|
| 2711 |
+
"step": 3860
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 1.8970588235294117,
|
| 2715 |
+
"grad_norm": 0.282520592212677,
|
| 2716 |
+
"learning_rate": 3.0119595059879678e-05,
|
| 2717 |
+
"loss": 0.0292,
|
| 2718 |
+
"step": 3870
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 1.9019607843137254,
|
| 2722 |
+
"grad_norm": 0.2727642357349396,
|
| 2723 |
+
"learning_rate": 2.965714919480872e-05,
|
| 2724 |
+
"loss": 0.0264,
|
| 2725 |
+
"step": 3880
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 1.906862745098039,
|
| 2729 |
+
"grad_norm": 0.1925729215145111,
|
| 2730 |
+
"learning_rate": 2.9197662090967625e-05,
|
| 2731 |
+
"loss": 0.0282,
|
| 2732 |
+
"step": 3890
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 1.9117647058823528,
|
| 2736 |
+
"grad_norm": 0.23374304175376892,
|
| 2737 |
+
"learning_rate": 2.8741153075304438e-05,
|
| 2738 |
+
"loss": 0.0266,
|
| 2739 |
+
"step": 3900
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 1.9166666666666665,
|
| 2743 |
+
"grad_norm": 0.24565072357654572,
|
| 2744 |
+
"learning_rate": 2.828764134950297e-05,
|
| 2745 |
+
"loss": 0.0198,
|
| 2746 |
+
"step": 3910
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 1.9215686274509802,
|
| 2750 |
+
"grad_norm": 0.1271701604127884,
|
| 2751 |
+
"learning_rate": 2.7837145989174974e-05,
|
| 2752 |
+
"loss": 0.0206,
|
| 2753 |
+
"step": 3920
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 1.9264705882352942,
|
| 2757 |
+
"grad_norm": 0.25117677450180054,
|
| 2758 |
+
"learning_rate": 2.7389685943057852e-05,
|
| 2759 |
+
"loss": 0.0249,
|
| 2760 |
+
"step": 3930
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 1.9313725490196079,
|
| 2764 |
+
"grad_norm": 0.33383405208587646,
|
| 2765 |
+
"learning_rate": 2.6945280032217535e-05,
|
| 2766 |
+
"loss": 0.0298,
|
| 2767 |
+
"step": 3940
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 1.9362745098039216,
|
| 2771 |
+
"grad_norm": 0.22882990539073944,
|
| 2772 |
+
"learning_rate": 2.6503946949256974e-05,
|
| 2773 |
+
"loss": 0.0273,
|
| 2774 |
+
"step": 3950
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 1.9411764705882353,
|
| 2778 |
+
"grad_norm": 0.2781018316745758,
|
| 2779 |
+
"learning_rate": 2.6065705257529848e-05,
|
| 2780 |
+
"loss": 0.0345,
|
| 2781 |
+
"step": 3960
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 1.946078431372549,
|
| 2785 |
+
"grad_norm": 0.23163281381130219,
|
| 2786 |
+
"learning_rate": 2.5630573390359624e-05,
|
| 2787 |
+
"loss": 0.024,
|
| 2788 |
+
"step": 3970
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 1.9509803921568627,
|
| 2792 |
+
"grad_norm": 0.30135810375213623,
|
| 2793 |
+
"learning_rate": 2.5198569650264403e-05,
|
| 2794 |
+
"loss": 0.0245,
|
| 2795 |
+
"step": 3980
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 1.9558823529411766,
|
| 2799 |
+
"grad_norm": 0.2169611006975174,
|
| 2800 |
+
"learning_rate": 2.4769712208186967e-05,
|
| 2801 |
+
"loss": 0.0217,
|
| 2802 |
+
"step": 3990
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 1.9607843137254903,
|
| 2806 |
+
"grad_norm": 0.2509106397628784,
|
| 2807 |
+
"learning_rate": 2.4344019102730542e-05,
|
| 2808 |
+
"loss": 0.0225,
|
| 2809 |
+
"step": 4000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 1.965686274509804,
|
| 2813 |
+
"grad_norm": 0.3170604407787323,
|
| 2814 |
+
"learning_rate": 2.3921508239399913e-05,
|
| 2815 |
+
"loss": 0.027,
|
| 2816 |
+
"step": 4010
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 1.9705882352941178,
|
| 2820 |
+
"grad_norm": 0.1460844874382019,
|
| 2821 |
+
"learning_rate": 2.350219738984849e-05,
|
| 2822 |
+
"loss": 0.021,
|
| 2823 |
+
"step": 4020
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 1.9754901960784315,
|
| 2827 |
+
"grad_norm": 0.27143651247024536,
|
| 2828 |
+
"learning_rate": 2.3086104191130643e-05,
|
| 2829 |
+
"loss": 0.0262,
|
| 2830 |
+
"step": 4030
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 1.9803921568627452,
|
| 2834 |
+
"grad_norm": 0.2494950294494629,
|
| 2835 |
+
"learning_rate": 2.2673246144959935e-05,
|
| 2836 |
+
"loss": 0.0249,
|
| 2837 |
+
"step": 4040
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 1.9852941176470589,
|
| 2841 |
+
"grad_norm": 0.26637178659439087,
|
| 2842 |
+
"learning_rate": 2.226364061697287e-05,
|
| 2843 |
+
"loss": 0.0325,
|
| 2844 |
+
"step": 4050
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 1.9901960784313726,
|
| 2848 |
+
"grad_norm": 0.2871919274330139,
|
| 2849 |
+
"learning_rate": 2.185730483599856e-05,
|
| 2850 |
+
"loss": 0.0286,
|
| 2851 |
+
"step": 4060
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 1.9950980392156863,
|
| 2855 |
+
"grad_norm": 0.2098357230424881,
|
| 2856 |
+
"learning_rate": 2.1454255893334064e-05,
|
| 2857 |
+
"loss": 0.0332,
|
| 2858 |
+
"step": 4070
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 2.0,
|
| 2862 |
+
"grad_norm": 0.2278250902891159,
|
| 2863 |
+
"learning_rate": 2.10545107420253e-05,
|
| 2864 |
+
"loss": 0.0234,
|
| 2865 |
+
"step": 4080
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 2.0049019607843137,
|
| 2869 |
+
"grad_norm": 0.266368567943573,
|
| 2870 |
+
"learning_rate": 2.0658086196154236e-05,
|
| 2871 |
+
"loss": 0.0226,
|
| 2872 |
+
"step": 4090
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 2.0098039215686274,
|
| 2876 |
+
"grad_norm": 0.10192721337080002,
|
| 2877 |
+
"learning_rate": 2.026499893013144e-05,
|
| 2878 |
+
"loss": 0.0217,
|
| 2879 |
+
"step": 4100
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 2.014705882352941,
|
| 2883 |
+
"grad_norm": 0.2021161913871765,
|
| 2884 |
+
"learning_rate": 1.9875265477994875e-05,
|
| 2885 |
+
"loss": 0.0219,
|
| 2886 |
+
"step": 4110
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 2.019607843137255,
|
| 2890 |
+
"grad_norm": 0.18436592817306519,
|
| 2891 |
+
"learning_rate": 1.9488902232714267e-05,
|
| 2892 |
+
"loss": 0.0216,
|
| 2893 |
+
"step": 4120
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 2.0245098039215685,
|
| 2897 |
+
"grad_norm": 0.18808206915855408,
|
| 2898 |
+
"learning_rate": 1.9105925445501794e-05,
|
| 2899 |
+
"loss": 0.0184,
|
| 2900 |
+
"step": 4130
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 2.0294117647058822,
|
| 2904 |
+
"grad_norm": 0.2091018408536911,
|
| 2905 |
+
"learning_rate": 1.87263512251284e-05,
|
| 2906 |
+
"loss": 0.0237,
|
| 2907 |
+
"step": 4140
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 2.034313725490196,
|
| 2911 |
+
"grad_norm": 0.29868388175964355,
|
| 2912 |
+
"learning_rate": 1.8350195537246184e-05,
|
| 2913 |
+
"loss": 0.0251,
|
| 2914 |
+
"step": 4150
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 2.0392156862745097,
|
| 2918 |
+
"grad_norm": 0.15926848351955414,
|
| 2919 |
+
"learning_rate": 1.797747420371699e-05,
|
| 2920 |
+
"loss": 0.0214,
|
| 2921 |
+
"step": 4160
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 2.0441176470588234,
|
| 2925 |
+
"grad_norm": 0.24016976356506348,
|
| 2926 |
+
"learning_rate": 1.7608202901946826e-05,
|
| 2927 |
+
"loss": 0.0206,
|
| 2928 |
+
"step": 4170
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 2.049019607843137,
|
| 2932 |
+
"grad_norm": 0.18072175979614258,
|
| 2933 |
+
"learning_rate": 1.7242397164226452e-05,
|
| 2934 |
+
"loss": 0.0192,
|
| 2935 |
+
"step": 4180
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 2.053921568627451,
|
| 2939 |
+
"grad_norm": 0.21760503947734833,
|
| 2940 |
+
"learning_rate": 1.6880072377078026e-05,
|
| 2941 |
+
"loss": 0.0237,
|
| 2942 |
+
"step": 4190
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 2.0588235294117645,
|
| 2946 |
+
"grad_norm": 0.28834259510040283,
|
| 2947 |
+
"learning_rate": 1.6521243780607974e-05,
|
| 2948 |
+
"loss": 0.0185,
|
| 2949 |
+
"step": 4200
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 2.063725490196078,
|
| 2953 |
+
"grad_norm": 0.14096632599830627,
|
| 2954 |
+
"learning_rate": 1.616592646786599e-05,
|
| 2955 |
+
"loss": 0.0184,
|
| 2956 |
+
"step": 4210
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 2.0686274509803924,
|
| 2960 |
+
"grad_norm": 0.15555402636528015,
|
| 2961 |
+
"learning_rate": 1.5814135384210026e-05,
|
| 2962 |
+
"loss": 0.02,
|
| 2963 |
+
"step": 4220
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 2.073529411764706,
|
| 2967 |
+
"grad_norm": 0.09052202850580215,
|
| 2968 |
+
"learning_rate": 1.5465885326677897e-05,
|
| 2969 |
+
"loss": 0.019,
|
| 2970 |
+
"step": 4230
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 2.0784313725490198,
|
| 2974 |
+
"grad_norm": 0.1283917874097824,
|
| 2975 |
+
"learning_rate": 1.512119094336466e-05,
|
| 2976 |
+
"loss": 0.0194,
|
| 2977 |
+
"step": 4240
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 2.0833333333333335,
|
| 2981 |
+
"grad_norm": 0.268288791179657,
|
| 2982 |
+
"learning_rate": 1.4780066732806663e-05,
|
| 2983 |
+
"loss": 0.024,
|
| 2984 |
+
"step": 4250
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 2.088235294117647,
|
| 2988 |
+
"grad_norm": 0.1498635858297348,
|
| 2989 |
+
"learning_rate": 1.4442527043371622e-05,
|
| 2990 |
+
"loss": 0.0226,
|
| 2991 |
+
"step": 4260
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 2.093137254901961,
|
| 2995 |
+
"grad_norm": 0.31013911962509155,
|
| 2996 |
+
"learning_rate": 1.4108586072655062e-05,
|
| 2997 |
+
"loss": 0.0198,
|
| 2998 |
+
"step": 4270
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 2.0980392156862746,
|
| 3002 |
+
"grad_norm": 0.19611553847789764,
|
| 3003 |
+
"learning_rate": 1.377825786688326e-05,
|
| 3004 |
+
"loss": 0.029,
|
| 3005 |
+
"step": 4280
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 2.1029411764705883,
|
| 3009 |
+
"grad_norm": 0.15280510485172272,
|
| 3010 |
+
"learning_rate": 1.3451556320322344e-05,
|
| 3011 |
+
"loss": 0.0305,
|
| 3012 |
+
"step": 4290
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 2.107843137254902,
|
| 3016 |
+
"grad_norm": 0.1233508512377739,
|
| 3017 |
+
"learning_rate": 1.3128495174693833e-05,
|
| 3018 |
+
"loss": 0.0214,
|
| 3019 |
+
"step": 4300
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 2.1127450980392157,
|
| 3023 |
+
"grad_norm": 0.23667648434638977,
|
| 3024 |
+
"learning_rate": 1.280908801859676e-05,
|
| 3025 |
+
"loss": 0.017,
|
| 3026 |
+
"step": 4310
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 2.1176470588235294,
|
| 3030 |
+
"grad_norm": 0.2285485714673996,
|
| 3031 |
+
"learning_rate": 1.2493348286936013e-05,
|
| 3032 |
+
"loss": 0.019,
|
| 3033 |
+
"step": 4320
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 2.122549019607843,
|
| 3037 |
+
"grad_norm": 0.17249974608421326,
|
| 3038 |
+
"learning_rate": 1.2181289260357265e-05,
|
| 3039 |
+
"loss": 0.0233,
|
| 3040 |
+
"step": 4330
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 2.127450980392157,
|
| 3044 |
+
"grad_norm": 0.18264269828796387,
|
| 3045 |
+
"learning_rate": 1.1872924064688328e-05,
|
| 3046 |
+
"loss": 0.0176,
|
| 3047 |
+
"step": 4340
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 2.1323529411764706,
|
| 3051 |
+
"grad_norm": 0.23498280346393585,
|
| 3052 |
+
"learning_rate": 1.1568265670387125e-05,
|
| 3053 |
+
"loss": 0.0216,
|
| 3054 |
+
"step": 4350
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 2.1372549019607843,
|
| 3058 |
+
"grad_norm": 0.20697841048240662,
|
| 3059 |
+
"learning_rate": 1.12673268919961e-05,
|
| 3060 |
+
"loss": 0.0221,
|
| 3061 |
+
"step": 4360
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 2.142156862745098,
|
| 3065 |
+
"grad_norm": 0.20883601903915405,
|
| 3066 |
+
"learning_rate": 1.0970120387603122e-05,
|
| 3067 |
+
"loss": 0.0211,
|
| 3068 |
+
"step": 4370
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 2.1470588235294117,
|
| 3072 |
+
"grad_norm": 0.2525753676891327,
|
| 3073 |
+
"learning_rate": 1.0676658658309225e-05,
|
| 3074 |
+
"loss": 0.0182,
|
| 3075 |
+
"step": 4380
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 2.1519607843137254,
|
| 3079 |
+
"grad_norm": 0.20986422896385193,
|
| 3080 |
+
"learning_rate": 1.0386954047702646e-05,
|
| 3081 |
+
"loss": 0.0222,
|
| 3082 |
+
"step": 4390
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 2.156862745098039,
|
| 3086 |
+
"grad_norm": 0.18922965228557587,
|
| 3087 |
+
"learning_rate": 1.010101874133973e-05,
|
| 3088 |
+
"loss": 0.0215,
|
| 3089 |
+
"step": 4400
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 2.161764705882353,
|
| 3093 |
+
"grad_norm": 0.23508426547050476,
|
| 3094 |
+
"learning_rate": 9.81886476623226e-06,
|
| 3095 |
+
"loss": 0.0158,
|
| 3096 |
+
"step": 4410
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 2.1666666666666665,
|
| 3100 |
+
"grad_norm": 0.17150916159152985,
|
| 3101 |
+
"learning_rate": 9.540503990341743e-06,
|
| 3102 |
+
"loss": 0.0204,
|
| 3103 |
+
"step": 4420
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 2.1715686274509802,
|
| 3107 |
+
"grad_norm": 0.12821370363235474,
|
| 3108 |
+
"learning_rate": 9.265948122080048e-06,
|
| 3109 |
+
"loss": 0.0162,
|
| 3110 |
+
"step": 4430
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 2.176470588235294,
|
| 3114 |
+
"grad_norm": 0.14045512676239014,
|
| 3115 |
+
"learning_rate": 8.995208709817071e-06,
|
| 3116 |
+
"loss": 0.0231,
|
| 3117 |
+
"step": 4440
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 2.1813725490196076,
|
| 3121 |
+
"grad_norm": 0.24774880707263947,
|
| 3122 |
+
"learning_rate": 8.728297141394858e-06,
|
| 3123 |
+
"loss": 0.0205,
|
| 3124 |
+
"step": 4450
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 2.186274509803922,
|
| 3128 |
+
"grad_norm": 0.20708513259887695,
|
| 3129 |
+
"learning_rate": 8.465224643648728e-06,
|
| 3130 |
+
"loss": 0.0217,
|
| 3131 |
+
"step": 4460
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 2.1911764705882355,
|
| 3135 |
+
"grad_norm": 0.1988290697336197,
|
| 3136 |
+
"learning_rate": 8.206002281934977e-06,
|
| 3137 |
+
"loss": 0.0181,
|
| 3138 |
+
"step": 4470
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 2.196078431372549,
|
| 3142 |
+
"grad_norm": 0.24248534440994263,
|
| 3143 |
+
"learning_rate": 7.950640959665457e-06,
|
| 3144 |
+
"loss": 0.014,
|
| 3145 |
+
"step": 4480
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 2.200980392156863,
|
| 3149 |
+
"grad_norm": 0.199833944439888,
|
| 3150 |
+
"learning_rate": 7.69915141784896e-06,
|
| 3151 |
+
"loss": 0.0177,
|
| 3152 |
+
"step": 4490
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 2.2058823529411766,
|
| 3156 |
+
"grad_norm": 0.18880146741867065,
|
| 3157 |
+
"learning_rate": 7.451544234639473e-06,
|
| 3158 |
+
"loss": 0.0309,
|
| 3159 |
+
"step": 4500
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 2.2107843137254903,
|
| 3163 |
+
"grad_norm": 0.19817057251930237,
|
| 3164 |
+
"learning_rate": 7.207829824891199e-06,
|
| 3165 |
+
"loss": 0.0135,
|
| 3166 |
+
"step": 4510
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 2.215686274509804,
|
| 3170 |
+
"grad_norm": 0.22702986001968384,
|
| 3171 |
+
"learning_rate": 6.968018439720414e-06,
|
| 3172 |
+
"loss": 0.0242,
|
| 3173 |
+
"step": 4520
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 2.2205882352941178,
|
| 3177 |
+
"grad_norm": 0.23108519613742828,
|
| 3178 |
+
"learning_rate": 6.732120166074441e-06,
|
| 3179 |
+
"loss": 0.0266,
|
| 3180 |
+
"step": 4530
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 2.2254901960784315,
|
| 3184 |
+
"grad_norm": 0.2683473229408264,
|
| 3185 |
+
"learning_rate": 6.500144926307295e-06,
|
| 3186 |
+
"loss": 0.0138,
|
| 3187 |
+
"step": 4540
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 2.230392156862745,
|
| 3191 |
+
"grad_norm": 0.3906085193157196,
|
| 3192 |
+
"learning_rate": 6.272102477762254e-06,
|
| 3193 |
+
"loss": 0.0257,
|
| 3194 |
+
"step": 4550
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 2.235294117647059,
|
| 3198 |
+
"grad_norm": 0.16685569286346436,
|
| 3199 |
+
"learning_rate": 6.048002412361598e-06,
|
| 3200 |
+
"loss": 0.0197,
|
| 3201 |
+
"step": 4560
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 2.2401960784313726,
|
| 3205 |
+
"grad_norm": 0.19442564249038696,
|
| 3206 |
+
"learning_rate": 5.827854156203017e-06,
|
| 3207 |
+
"loss": 0.0183,
|
| 3208 |
+
"step": 4570
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 2.2450980392156863,
|
| 3212 |
+
"grad_norm": 0.22825686633586884,
|
| 3213 |
+
"learning_rate": 5.611666969163243e-06,
|
| 3214 |
+
"loss": 0.0192,
|
| 3215 |
+
"step": 4580
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 2.25,
|
| 3219 |
+
"grad_norm": 0.13165044784545898,
|
| 3220 |
+
"learning_rate": 5.399449944508439e-06,
|
| 3221 |
+
"loss": 0.0181,
|
| 3222 |
+
"step": 4590
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 2.2549019607843137,
|
| 3226 |
+
"grad_norm": 0.32770487666130066,
|
| 3227 |
+
"learning_rate": 5.1912120085118365e-06,
|
| 3228 |
+
"loss": 0.0194,
|
| 3229 |
+
"step": 4600
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 2.2598039215686274,
|
| 3233 |
+
"grad_norm": 0.20789536833763123,
|
| 3234 |
+
"learning_rate": 4.986961920078204e-06,
|
| 3235 |
+
"loss": 0.0274,
|
| 3236 |
+
"step": 4610
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 2.264705882352941,
|
| 3240 |
+
"grad_norm": 0.3107619881629944,
|
| 3241 |
+
"learning_rate": 4.786708270375462e-06,
|
| 3242 |
+
"loss": 0.0244,
|
| 3243 |
+
"step": 4620
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 2.269607843137255,
|
| 3247 |
+
"grad_norm": 0.11234085261821747,
|
| 3248 |
+
"learning_rate": 4.590459482473286e-06,
|
| 3249 |
+
"loss": 0.0125,
|
| 3250 |
+
"step": 4630
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 2.2745098039215685,
|
| 3254 |
+
"grad_norm": 0.2838551700115204,
|
| 3255 |
+
"learning_rate": 4.398223810988866e-06,
|
| 3256 |
+
"loss": 0.0209,
|
| 3257 |
+
"step": 4640
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 2.2794117647058822,
|
| 3261 |
+
"grad_norm": 0.2697765529155731,
|
| 3262 |
+
"learning_rate": 4.2100093417396845e-06,
|
| 3263 |
+
"loss": 0.0245,
|
| 3264 |
+
"step": 4650
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 2.284313725490196,
|
| 3268 |
+
"grad_norm": 0.1846655011177063,
|
| 3269 |
+
"learning_rate": 4.0258239914033765e-06,
|
| 3270 |
+
"loss": 0.0271,
|
| 3271 |
+
"step": 4660
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 2.2892156862745097,
|
| 3275 |
+
"grad_norm": 0.20711906254291534,
|
| 3276 |
+
"learning_rate": 3.8456755071847765e-06,
|
| 3277 |
+
"loss": 0.0262,
|
| 3278 |
+
"step": 4670
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 2.2941176470588234,
|
| 3282 |
+
"grad_norm": 0.23577173054218292,
|
| 3283 |
+
"learning_rate": 3.6695714664900293e-06,
|
| 3284 |
+
"loss": 0.0147,
|
| 3285 |
+
"step": 4680
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 2.299019607843137,
|
| 3289 |
+
"grad_norm": 0.21056267619132996,
|
| 3290 |
+
"learning_rate": 3.49751927660793e-06,
|
| 3291 |
+
"loss": 0.0242,
|
| 3292 |
+
"step": 4690
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 2.303921568627451,
|
| 3296 |
+
"grad_norm": 0.1903896927833557,
|
| 3297 |
+
"learning_rate": 3.329526174398223e-06,
|
| 3298 |
+
"loss": 0.0199,
|
| 3299 |
+
"step": 4700
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 2.3088235294117645,
|
| 3303 |
+
"grad_norm": 0.1740642637014389,
|
| 3304 |
+
"learning_rate": 3.165599225987381e-06,
|
| 3305 |
+
"loss": 0.0241,
|
| 3306 |
+
"step": 4710
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 2.313725490196078,
|
| 3310 |
+
"grad_norm": 0.23523923754692078,
|
| 3311 |
+
"learning_rate": 3.005745326471254e-06,
|
| 3312 |
+
"loss": 0.0196,
|
| 3313 |
+
"step": 4720
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 2.318627450980392,
|
| 3317 |
+
"grad_norm": 0.17898762226104736,
|
| 3318 |
+
"learning_rate": 2.849971199625112e-06,
|
| 3319 |
+
"loss": 0.0178,
|
| 3320 |
+
"step": 4730
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 2.323529411764706,
|
| 3324 |
+
"grad_norm": 0.18050484359264374,
|
| 3325 |
+
"learning_rate": 2.6982833976208043e-06,
|
| 3326 |
+
"loss": 0.0264,
|
| 3327 |
+
"step": 4740
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 2.3284313725490198,
|
| 3331 |
+
"grad_norm": 0.14259324967861176,
|
| 3332 |
+
"learning_rate": 2.5506883007511695e-06,
|
| 3333 |
+
"loss": 0.0163,
|
| 3334 |
+
"step": 4750
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 2.3333333333333335,
|
| 3338 |
+
"grad_norm": 0.25935426354408264,
|
| 3339 |
+
"learning_rate": 2.407192117161683e-06,
|
| 3340 |
+
"loss": 0.0215,
|
| 3341 |
+
"step": 4760
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 2.338235294117647,
|
| 3345 |
+
"grad_norm": 0.24692723155021667,
|
| 3346 |
+
"learning_rate": 2.2678008825893106e-06,
|
| 3347 |
+
"loss": 0.0185,
|
| 3348 |
+
"step": 4770
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 2.343137254901961,
|
| 3352 |
+
"grad_norm": 0.22854046523571014,
|
| 3353 |
+
"learning_rate": 2.1325204601086222e-06,
|
| 3354 |
+
"loss": 0.0235,
|
| 3355 |
+
"step": 4780
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 2.3480392156862746,
|
| 3359 |
+
"grad_norm": 0.31760168075561523,
|
| 3360 |
+
"learning_rate": 2.001356539885213e-06,
|
| 3361 |
+
"loss": 0.0279,
|
| 3362 |
+
"step": 4790
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 2.3529411764705883,
|
| 3366 |
+
"grad_norm": 0.19233286380767822,
|
| 3367 |
+
"learning_rate": 1.8743146389363474e-06,
|
| 3368 |
+
"loss": 0.0174,
|
| 3369 |
+
"step": 4800
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 2.357843137254902,
|
| 3373 |
+
"grad_norm": 0.28652098774909973,
|
| 3374 |
+
"learning_rate": 1.7514001008988923e-06,
|
| 3375 |
+
"loss": 0.0193,
|
| 3376 |
+
"step": 4810
|
| 3377 |
+
},
|
| 3378 |
+
{
|
| 3379 |
+
"epoch": 2.3627450980392157,
|
| 3380 |
+
"grad_norm": 0.16845420002937317,
|
| 3381 |
+
"learning_rate": 1.6326180958045502e-06,
|
| 3382 |
+
"loss": 0.0177,
|
| 3383 |
+
"step": 4820
|
| 3384 |
+
},
|
| 3385 |
+
{
|
| 3386 |
+
"epoch": 2.3676470588235294,
|
| 3387 |
+
"grad_norm": 0.12303224951028824,
|
| 3388 |
+
"learning_rate": 1.517973619862445e-06,
|
| 3389 |
+
"loss": 0.0205,
|
| 3390 |
+
"step": 4830
|
| 3391 |
+
},
|
| 3392 |
+
{
|
| 3393 |
+
"epoch": 2.372549019607843,
|
| 3394 |
+
"grad_norm": 0.1442280411720276,
|
| 3395 |
+
"learning_rate": 1.4074714952489132e-06,
|
| 3396 |
+
"loss": 0.0215,
|
| 3397 |
+
"step": 4840
|
| 3398 |
+
},
|
| 3399 |
+
{
|
| 3400 |
+
"epoch": 2.377450980392157,
|
| 3401 |
+
"grad_norm": 0.18173760175704956,
|
| 3402 |
+
"learning_rate": 1.3011163699046758e-06,
|
| 3403 |
+
"loss": 0.0166,
|
| 3404 |
+
"step": 4850
|
| 3405 |
+
},
|
| 3406 |
+
{
|
| 3407 |
+
"epoch": 2.3823529411764706,
|
| 3408 |
+
"grad_norm": 0.2543584406375885,
|
| 3409 |
+
"learning_rate": 1.1989127173393955e-06,
|
| 3410 |
+
"loss": 0.0195,
|
| 3411 |
+
"step": 4860
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 2.3872549019607843,
|
| 3415 |
+
"grad_norm": 0.14100182056427002,
|
| 3416 |
+
"learning_rate": 1.1008648364434493e-06,
|
| 3417 |
+
"loss": 0.0252,
|
| 3418 |
+
"step": 4870
|
| 3419 |
+
},
|
| 3420 |
+
{
|
| 3421 |
+
"epoch": 2.392156862745098,
|
| 3422 |
+
"grad_norm": 0.22381238639354706,
|
| 3423 |
+
"learning_rate": 1.0069768513071287e-06,
|
| 3424 |
+
"loss": 0.0184,
|
| 3425 |
+
"step": 4880
|
| 3426 |
+
},
|
| 3427 |
+
{
|
| 3428 |
+
"epoch": 2.3970588235294117,
|
| 3429 |
+
"grad_norm": 0.1522032916545868,
|
| 3430 |
+
"learning_rate": 9.172527110472007e-07,
|
| 3431 |
+
"loss": 0.0226,
|
| 3432 |
+
"step": 4890
|
| 3433 |
+
},
|
| 3434 |
+
{
|
| 3435 |
+
"epoch": 2.4019607843137254,
|
| 3436 |
+
"grad_norm": 0.151298388838768,
|
| 3437 |
+
"learning_rate": 8.316961896407293e-07,
|
| 3438 |
+
"loss": 0.0161,
|
| 3439 |
+
"step": 4900
|
| 3440 |
+
},
|
| 3441 |
+
{
|
| 3442 |
+
"epoch": 2.406862745098039,
|
| 3443 |
+
"grad_norm": 0.10205589234828949,
|
| 3444 |
+
"learning_rate": 7.503108857664476e-07,
|
| 3445 |
+
"loss": 0.0221,
|
| 3446 |
+
"step": 4910
|
| 3447 |
+
},
|
| 3448 |
+
{
|
| 3449 |
+
"epoch": 2.411764705882353,
|
| 3450 |
+
"grad_norm": 0.29145070910453796,
|
| 3451 |
+
"learning_rate": 6.731002226532557e-07,
|
| 3452 |
+
"loss": 0.0256,
|
| 3453 |
+
"step": 4920
|
| 3454 |
+
},
|
| 3455 |
+
{
|
| 3456 |
+
"epoch": 2.4166666666666665,
|
| 3457 |
+
"grad_norm": 0.11820173263549805,
|
| 3458 |
+
"learning_rate": 6.000674479363366e-07,
|
| 3459 |
+
"loss": 0.0192,
|
| 3460 |
+
"step": 4930
|
| 3461 |
+
},
|
| 3462 |
+
{
|
| 3463 |
+
"epoch": 2.4215686274509802,
|
| 3464 |
+
"grad_norm": 0.34024494886398315,
|
| 3465 |
+
"learning_rate": 5.312156335205098e-07,
|
| 3466 |
+
"loss": 0.0184,
|
| 3467 |
+
"step": 4940
|
| 3468 |
+
},
|
| 3469 |
+
{
|
| 3470 |
+
"epoch": 2.426470588235294,
|
| 3471 |
+
"grad_norm": 0.2437254935503006,
|
| 3472 |
+
"learning_rate": 4.665476754510234e-07,
|
| 3473 |
+
"loss": 0.0172,
|
| 3474 |
+
"step": 4950
|
| 3475 |
+
},
|
| 3476 |
+
{
|
| 3477 |
+
"epoch": 2.431372549019608,
|
| 3478 |
+
"grad_norm": 0.22836612164974213,
|
| 3479 |
+
"learning_rate": 4.0606629379175143e-07,
|
| 3480 |
+
"loss": 0.0236,
|
| 3481 |
+
"step": 4960
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 2.436274509803922,
|
| 3485 |
+
"grad_norm": 0.30892133712768555,
|
| 3486 |
+
"learning_rate": 3.497740325107746e-07,
|
| 3487 |
+
"loss": 0.021,
|
| 3488 |
+
"step": 4970
|
| 3489 |
+
},
|
| 3490 |
+
{
|
| 3491 |
+
"epoch": 2.4411764705882355,
|
| 3492 |
+
"grad_norm": 0.1466980129480362,
|
| 3493 |
+
"learning_rate": 2.9767325937338775e-07,
|
| 3494 |
+
"loss": 0.0158,
|
| 3495 |
+
"step": 4980
|
| 3496 |
+
},
|
| 3497 |
+
{
|
| 3498 |
+
"epoch": 2.446078431372549,
|
| 3499 |
+
"grad_norm": 0.17119024693965912,
|
| 3500 |
+
"learning_rate": 2.497661658424688e-07,
|
| 3501 |
+
"loss": 0.0211,
|
| 3502 |
+
"step": 4990
|
| 3503 |
+
},
|
| 3504 |
+
{
|
| 3505 |
+
"epoch": 2.450980392156863,
|
| 3506 |
+
"grad_norm": 0.24735775589942932,
|
| 3507 |
+
"learning_rate": 2.0605476698636328e-07,
|
| 3508 |
+
"loss": 0.0198,
|
| 3509 |
+
"step": 5000
|
| 3510 |
+
},
|
| 3511 |
+
{
|
| 3512 |
+
"epoch": 2.4558823529411766,
|
| 3513 |
+
"grad_norm": 0.22227314114570618,
|
| 3514 |
+
"learning_rate": 1.6654090139408551e-07,
|
| 3515 |
+
"loss": 0.0222,
|
| 3516 |
+
"step": 5010
|
| 3517 |
+
},
|
| 3518 |
+
{
|
| 3519 |
+
"epoch": 2.4607843137254903,
|
| 3520 |
+
"grad_norm": 0.2925474941730499,
|
| 3521 |
+
"learning_rate": 1.3122623109795839e-07,
|
| 3522 |
+
"loss": 0.0212,
|
| 3523 |
+
"step": 5020
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 2.465686274509804,
|
| 3527 |
+
"grad_norm": 0.11938751488924026,
|
| 3528 |
+
"learning_rate": 1.0011224150379139e-07,
|
| 3529 |
+
"loss": 0.0206,
|
| 3530 |
+
"step": 5030
|
| 3531 |
+
},
|
| 3532 |
+
{
|
| 3533 |
+
"epoch": 2.4705882352941178,
|
| 3534 |
+
"grad_norm": 0.18019632995128632,
|
| 3535 |
+
"learning_rate": 7.320024132829729e-08,
|
| 3536 |
+
"loss": 0.0175,
|
| 3537 |
+
"step": 5040
|
| 3538 |
+
},
|
| 3539 |
+
{
|
| 3540 |
+
"epoch": 2.4754901960784315,
|
| 3541 |
+
"grad_norm": 0.1969996839761734,
|
| 3542 |
+
"learning_rate": 5.049136254413611e-08,
|
| 3543 |
+
"loss": 0.0196,
|
| 3544 |
+
"step": 5050
|
| 3545 |
+
},
|
| 3546 |
+
{
|
| 3547 |
+
"epoch": 2.480392156862745,
|
| 3548 |
+
"grad_norm": 0.26969224214553833,
|
| 3549 |
+
"learning_rate": 3.1986560332242234e-08,
|
| 3550 |
+
"loss": 0.0159,
|
| 3551 |
+
"step": 5060
|
| 3552 |
+
},
|
| 3553 |
+
{
|
| 3554 |
+
"epoch": 2.485294117647059,
|
| 3555 |
+
"grad_norm": 0.24808961153030396,
|
| 3556 |
+
"learning_rate": 1.768661304166752e-08,
|
| 3557 |
+
"loss": 0.0235,
|
| 3558 |
+
"step": 5070
|
| 3559 |
+
},
|
| 3560 |
+
{
|
| 3561 |
+
"epoch": 2.4901960784313726,
|
| 3562 |
+
"grad_norm": 0.08070988208055496,
|
| 3563 |
+
"learning_rate": 7.592122156829806e-09,
|
| 3564 |
+
"loss": 0.0196,
|
| 3565 |
+
"step": 5080
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 2.4950980392156863,
|
| 3569 |
+
"grad_norm": 0.3003765642642975,
|
| 3570 |
+
"learning_rate": 1.7035122722663943e-09,
|
| 3571 |
+
"loss": 0.0214,
|
| 3572 |
+
"step": 5090
|
| 3573 |
+
},
|
| 3574 |
+
{
|
| 3575 |
+
"epoch": 2.4995098039215686,
|
| 3576 |
+
"step": 5099,
|
| 3577 |
+
"total_flos": 1.0010895665363712e+17,
|
| 3578 |
+
"train_loss": 0.049077389520344766,
|
| 3579 |
+
"train_runtime": 1743.2855,
|
| 3580 |
+
"train_samples_per_second": 46.799,
|
| 3581 |
+
"train_steps_per_second": 2.925
|
| 3582 |
+
}
|
| 3583 |
+
],
|
| 3584 |
+
"logging_steps": 10,
|
| 3585 |
+
"max_steps": 5099,
|
| 3586 |
+
"num_input_tokens_seen": 0,
|
| 3587 |
+
"num_train_epochs": 3,
|
| 3588 |
+
"save_steps": 500,
|
| 3589 |
+
"stateful_callbacks": {
|
| 3590 |
+
"TrainerControl": {
|
| 3591 |
+
"args": {
|
| 3592 |
+
"should_epoch_stop": false,
|
| 3593 |
+
"should_evaluate": false,
|
| 3594 |
+
"should_log": false,
|
| 3595 |
+
"should_save": true,
|
| 3596 |
+
"should_training_stop": true
|
| 3597 |
+
},
|
| 3598 |
+
"attributes": {}
|
| 3599 |
+
}
|
| 3600 |
+
},
|
| 3601 |
+
"total_flos": 1.0010895665363712e+17,
|
| 3602 |
+
"train_batch_size": 16,
|
| 3603 |
+
"trial_name": null,
|
| 3604 |
+
"trial_params": null
|
| 3605 |
+
}
|