Upload multi_teacher_fantasy_alchemy_without_cot/checkpoint-5400/trainer_state.json with huggingface_hub
Browse files
multi_teacher_fantasy_alchemy_without_cot/checkpoint-5400/trainer_state.json
ADDED
|
@@ -0,0 +1,3814 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 4.437140509449466,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5400,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.008216926869350863,
|
| 14 |
+
"grad_norm": 20.6521664868754,
|
| 15 |
+
"learning_rate": 9.240246406570842e-07,
|
| 16 |
+
"loss": 8.1344,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.016433853738701727,
|
| 21 |
+
"grad_norm": 18.75711536193637,
|
| 22 |
+
"learning_rate": 1.9507186858316226e-06,
|
| 23 |
+
"loss": 7.9653,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.024650780608052588,
|
| 28 |
+
"grad_norm": 11.644294667992769,
|
| 29 |
+
"learning_rate": 2.9774127310061602e-06,
|
| 30 |
+
"loss": 7.2337,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.03286770747740345,
|
| 35 |
+
"grad_norm": 7.186299732523691,
|
| 36 |
+
"learning_rate": 4.004106776180698e-06,
|
| 37 |
+
"loss": 6.2436,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.041084634346754315,
|
| 42 |
+
"grad_norm": 6.028680766018775,
|
| 43 |
+
"learning_rate": 5.030800821355237e-06,
|
| 44 |
+
"loss": 5.4582,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.049301561216105176,
|
| 49 |
+
"grad_norm": 8.01146787060327,
|
| 50 |
+
"learning_rate": 6.057494866529775e-06,
|
| 51 |
+
"loss": 4.6383,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.05751848808545604,
|
| 56 |
+
"grad_norm": 6.474137612808116,
|
| 57 |
+
"learning_rate": 7.084188911704312e-06,
|
| 58 |
+
"loss": 3.41,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.0657354149548069,
|
| 63 |
+
"grad_norm": 2.3678549728521503,
|
| 64 |
+
"learning_rate": 8.11088295687885e-06,
|
| 65 |
+
"loss": 2.3809,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.07395234182415776,
|
| 70 |
+
"grad_norm": 1.7789962021111834,
|
| 71 |
+
"learning_rate": 9.137577002053388e-06,
|
| 72 |
+
"loss": 2.0251,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.08216926869350863,
|
| 77 |
+
"grad_norm": 2.0296841787136595,
|
| 78 |
+
"learning_rate": 1.0164271047227925e-05,
|
| 79 |
+
"loss": 1.9094,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.09038619556285948,
|
| 84 |
+
"grad_norm": 4.5379547550775925,
|
| 85 |
+
"learning_rate": 1.1190965092402464e-05,
|
| 86 |
+
"loss": 1.8227,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.09860312243221035,
|
| 91 |
+
"grad_norm": 4.133214822477088,
|
| 92 |
+
"learning_rate": 1.2217659137577003e-05,
|
| 93 |
+
"loss": 1.7539,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.10682004930156122,
|
| 98 |
+
"grad_norm": 12.868168725927204,
|
| 99 |
+
"learning_rate": 1.324435318275154e-05,
|
| 100 |
+
"loss": 1.7031,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.11503697617091208,
|
| 105 |
+
"grad_norm": 6.450060429921103,
|
| 106 |
+
"learning_rate": 1.427104722792608e-05,
|
| 107 |
+
"loss": 1.6525,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.12325390304026294,
|
| 112 |
+
"grad_norm": 8.671152101355725,
|
| 113 |
+
"learning_rate": 1.5297741273100617e-05,
|
| 114 |
+
"loss": 1.6144,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.1314708299096138,
|
| 119 |
+
"grad_norm": 7.494958448355011,
|
| 120 |
+
"learning_rate": 1.6324435318275156e-05,
|
| 121 |
+
"loss": 1.575,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.13968775677896467,
|
| 126 |
+
"grad_norm": 7.749660614586779,
|
| 127 |
+
"learning_rate": 1.735112936344969e-05,
|
| 128 |
+
"loss": 1.5372,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.14790468364831552,
|
| 133 |
+
"grad_norm": 8.182648325800557,
|
| 134 |
+
"learning_rate": 1.837782340862423e-05,
|
| 135 |
+
"loss": 1.4918,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.1561216105176664,
|
| 140 |
+
"grad_norm": 5.782818345230798,
|
| 141 |
+
"learning_rate": 1.940451745379877e-05,
|
| 142 |
+
"loss": 1.4442,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.16433853738701726,
|
| 147 |
+
"grad_norm": 4.377767445583269,
|
| 148 |
+
"learning_rate": 2.0431211498973308e-05,
|
| 149 |
+
"loss": 1.4135,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.1725554642563681,
|
| 154 |
+
"grad_norm": 6.917683584431286,
|
| 155 |
+
"learning_rate": 2.1457905544147843e-05,
|
| 156 |
+
"loss": 1.3559,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.18077239112571897,
|
| 161 |
+
"grad_norm": 5.4201085598769065,
|
| 162 |
+
"learning_rate": 2.2484599589322382e-05,
|
| 163 |
+
"loss": 1.3334,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.18898931799506985,
|
| 168 |
+
"grad_norm": 1.7147042968760517,
|
| 169 |
+
"learning_rate": 2.351129363449692e-05,
|
| 170 |
+
"loss": 1.2822,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.1972062448644207,
|
| 175 |
+
"grad_norm": 2.7781476506938207,
|
| 176 |
+
"learning_rate": 2.453798767967146e-05,
|
| 177 |
+
"loss": 1.2538,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.20542317173377156,
|
| 182 |
+
"grad_norm": 3.1382479212042522,
|
| 183 |
+
"learning_rate": 2.5564681724845996e-05,
|
| 184 |
+
"loss": 1.2093,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.21364009860312244,
|
| 189 |
+
"grad_norm": 3.84263025478836,
|
| 190 |
+
"learning_rate": 2.6591375770020538e-05,
|
| 191 |
+
"loss": 1.1789,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.2218570254724733,
|
| 196 |
+
"grad_norm": 4.56729430742156,
|
| 197 |
+
"learning_rate": 2.7618069815195074e-05,
|
| 198 |
+
"loss": 1.1537,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.23007395234182415,
|
| 203 |
+
"grad_norm": 3.396912773118433,
|
| 204 |
+
"learning_rate": 2.8644763860369613e-05,
|
| 205 |
+
"loss": 1.1055,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.23829087921117503,
|
| 210 |
+
"grad_norm": 2.9099056698892802,
|
| 211 |
+
"learning_rate": 2.9671457905544148e-05,
|
| 212 |
+
"loss": 1.0825,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.2465078060805259,
|
| 217 |
+
"grad_norm": 5.08766339955455,
|
| 218 |
+
"learning_rate": 3.069815195071869e-05,
|
| 219 |
+
"loss": 1.0362,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.25472473294987674,
|
| 224 |
+
"grad_norm": 3.071437319717274,
|
| 225 |
+
"learning_rate": 3.172484599589323e-05,
|
| 226 |
+
"loss": 1.0277,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.2629416598192276,
|
| 231 |
+
"grad_norm": 2.783117166830873,
|
| 232 |
+
"learning_rate": 3.2751540041067765e-05,
|
| 233 |
+
"loss": 0.9931,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.27115858668857845,
|
| 238 |
+
"grad_norm": 3.2281961993022934,
|
| 239 |
+
"learning_rate": 3.37782340862423e-05,
|
| 240 |
+
"loss": 0.9558,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.27937551355792933,
|
| 245 |
+
"grad_norm": 3.5461115086836696,
|
| 246 |
+
"learning_rate": 3.480492813141684e-05,
|
| 247 |
+
"loss": 0.9276,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2875924404272802,
|
| 252 |
+
"grad_norm": 3.23867135269811,
|
| 253 |
+
"learning_rate": 3.583162217659138e-05,
|
| 254 |
+
"loss": 0.9029,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.29580936729663104,
|
| 259 |
+
"grad_norm": 2.9653265132671325,
|
| 260 |
+
"learning_rate": 3.6858316221765914e-05,
|
| 261 |
+
"loss": 0.8759,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.3040262941659819,
|
| 266 |
+
"grad_norm": 2.6616367806431422,
|
| 267 |
+
"learning_rate": 3.788501026694045e-05,
|
| 268 |
+
"loss": 0.8583,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.3122432210353328,
|
| 273 |
+
"grad_norm": 4.802167335262461,
|
| 274 |
+
"learning_rate": 3.891170431211499e-05,
|
| 275 |
+
"loss": 0.8257,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.32046014790468363,
|
| 280 |
+
"grad_norm": 3.5139799826616427,
|
| 281 |
+
"learning_rate": 3.993839835728953e-05,
|
| 282 |
+
"loss": 0.8065,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.3286770747740345,
|
| 287 |
+
"grad_norm": 3.5566795680934153,
|
| 288 |
+
"learning_rate": 4.096509240246407e-05,
|
| 289 |
+
"loss": 0.7766,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.3368940016433854,
|
| 294 |
+
"grad_norm": 5.016388912688849,
|
| 295 |
+
"learning_rate": 4.1991786447638605e-05,
|
| 296 |
+
"loss": 0.7521,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.3451109285127362,
|
| 301 |
+
"grad_norm": 4.100894602879236,
|
| 302 |
+
"learning_rate": 4.301848049281315e-05,
|
| 303 |
+
"loss": 0.7388,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.3533278553820871,
|
| 308 |
+
"grad_norm": 3.3274885019517004,
|
| 309 |
+
"learning_rate": 4.404517453798768e-05,
|
| 310 |
+
"loss": 0.7158,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.36154478225143794,
|
| 315 |
+
"grad_norm": 3.664275856908218,
|
| 316 |
+
"learning_rate": 4.507186858316222e-05,
|
| 317 |
+
"loss": 0.691,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.3697617091207888,
|
| 322 |
+
"grad_norm": 3.1992131800713386,
|
| 323 |
+
"learning_rate": 4.6098562628336754e-05,
|
| 324 |
+
"loss": 0.6696,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.3779786359901397,
|
| 329 |
+
"grad_norm": 3.673357196076266,
|
| 330 |
+
"learning_rate": 4.7125256673511296e-05,
|
| 331 |
+
"loss": 0.648,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.3861955628594905,
|
| 336 |
+
"grad_norm": 4.2513870810342365,
|
| 337 |
+
"learning_rate": 4.815195071868583e-05,
|
| 338 |
+
"loss": 0.6228,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.3944124897288414,
|
| 343 |
+
"grad_norm": 2.745076649884309,
|
| 344 |
+
"learning_rate": 4.9178644763860374e-05,
|
| 345 |
+
"loss": 0.6025,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.4026294165981923,
|
| 350 |
+
"grad_norm": 3.764105619282273,
|
| 351 |
+
"learning_rate": 5.020533880903491e-05,
|
| 352 |
+
"loss": 0.5847,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.4108463434675431,
|
| 357 |
+
"grad_norm": 2.9431726886958955,
|
| 358 |
+
"learning_rate": 5.1232032854209445e-05,
|
| 359 |
+
"loss": 0.5653,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.419063270336894,
|
| 364 |
+
"grad_norm": 3.7883418470145864,
|
| 365 |
+
"learning_rate": 5.225872689938398e-05,
|
| 366 |
+
"loss": 0.5481,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.4272801972062449,
|
| 371 |
+
"grad_norm": 3.2826984002457023,
|
| 372 |
+
"learning_rate": 5.328542094455853e-05,
|
| 373 |
+
"loss": 0.5267,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.4354971240755957,
|
| 378 |
+
"grad_norm": 3.7751689108988358,
|
| 379 |
+
"learning_rate": 5.4312114989733065e-05,
|
| 380 |
+
"loss": 0.5136,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.4437140509449466,
|
| 385 |
+
"grad_norm": 3.1468929057036275,
|
| 386 |
+
"learning_rate": 5.53388090349076e-05,
|
| 387 |
+
"loss": 0.4923,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.4519309778142975,
|
| 392 |
+
"grad_norm": 3.332657724668719,
|
| 393 |
+
"learning_rate": 5.6365503080082136e-05,
|
| 394 |
+
"loss": 0.4814,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.4601479046836483,
|
| 399 |
+
"grad_norm": 3.4721014951966556,
|
| 400 |
+
"learning_rate": 5.739219712525667e-05,
|
| 401 |
+
"loss": 0.463,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.4683648315529992,
|
| 406 |
+
"grad_norm": 2.883971781790991,
|
| 407 |
+
"learning_rate": 5.8418891170431214e-05,
|
| 408 |
+
"loss": 0.4407,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.47658175842235007,
|
| 413 |
+
"grad_norm": 3.858236605679481,
|
| 414 |
+
"learning_rate": 5.944558521560575e-05,
|
| 415 |
+
"loss": 0.4301,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.4847986852917009,
|
| 420 |
+
"grad_norm": 2.7193833088424952,
|
| 421 |
+
"learning_rate": 6.04722792607803e-05,
|
| 422 |
+
"loss": 0.4184,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.4930156121610518,
|
| 427 |
+
"grad_norm": 3.4810813188988776,
|
| 428 |
+
"learning_rate": 6.149897330595483e-05,
|
| 429 |
+
"loss": 0.4063,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.5012325390304027,
|
| 434 |
+
"grad_norm": 2.56285006880157,
|
| 435 |
+
"learning_rate": 6.252566735112937e-05,
|
| 436 |
+
"loss": 0.3899,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.5094494658997535,
|
| 441 |
+
"grad_norm": 2.71953277054627,
|
| 442 |
+
"learning_rate": 6.35523613963039e-05,
|
| 443 |
+
"loss": 0.3733,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.5176663927691043,
|
| 448 |
+
"grad_norm": 2.5203449713953696,
|
| 449 |
+
"learning_rate": 6.457905544147844e-05,
|
| 450 |
+
"loss": 0.3516,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.5258833196384552,
|
| 455 |
+
"grad_norm": 3.0318245975311515,
|
| 456 |
+
"learning_rate": 6.560574948665298e-05,
|
| 457 |
+
"loss": 0.3485,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.5341002465078061,
|
| 462 |
+
"grad_norm": 2.5920740384939553,
|
| 463 |
+
"learning_rate": 6.663244353182751e-05,
|
| 464 |
+
"loss": 0.3376,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.5423171733771569,
|
| 469 |
+
"grad_norm": 2.2652774245755842,
|
| 470 |
+
"learning_rate": 6.765913757700205e-05,
|
| 471 |
+
"loss": 0.3252,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.5505341002465078,
|
| 476 |
+
"grad_norm": 3.067932138213513,
|
| 477 |
+
"learning_rate": 6.86858316221766e-05,
|
| 478 |
+
"loss": 0.3102,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.5587510271158587,
|
| 483 |
+
"grad_norm": 2.334836395372571,
|
| 484 |
+
"learning_rate": 6.971252566735113e-05,
|
| 485 |
+
"loss": 0.3075,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.5669679539852095,
|
| 490 |
+
"grad_norm": 2.00073050652139,
|
| 491 |
+
"learning_rate": 7.073921971252567e-05,
|
| 492 |
+
"loss": 0.2944,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.5751848808545604,
|
| 497 |
+
"grad_norm": 2.4192633170485887,
|
| 498 |
+
"learning_rate": 7.17659137577002e-05,
|
| 499 |
+
"loss": 0.2859,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.5834018077239113,
|
| 504 |
+
"grad_norm": 2.191465248513388,
|
| 505 |
+
"learning_rate": 7.279260780287475e-05,
|
| 506 |
+
"loss": 0.2808,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.5916187345932621,
|
| 511 |
+
"grad_norm": 2.1743578076452295,
|
| 512 |
+
"learning_rate": 7.381930184804929e-05,
|
| 513 |
+
"loss": 0.2649,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.599835661462613,
|
| 518 |
+
"grad_norm": 2.3765522408619666,
|
| 519 |
+
"learning_rate": 7.484599589322382e-05,
|
| 520 |
+
"loss": 0.2619,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.6080525883319639,
|
| 525 |
+
"grad_norm": 1.9037287843676427,
|
| 526 |
+
"learning_rate": 7.587268993839836e-05,
|
| 527 |
+
"loss": 0.2537,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.6162695152013147,
|
| 532 |
+
"grad_norm": 2.0158389895405175,
|
| 533 |
+
"learning_rate": 7.689938398357291e-05,
|
| 534 |
+
"loss": 0.247,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.6244864420706656,
|
| 539 |
+
"grad_norm": 2.342705818216856,
|
| 540 |
+
"learning_rate": 7.792607802874744e-05,
|
| 541 |
+
"loss": 0.2431,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.6327033689400164,
|
| 546 |
+
"grad_norm": 2.3111970572134433,
|
| 547 |
+
"learning_rate": 7.895277207392198e-05,
|
| 548 |
+
"loss": 0.2359,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.6409202958093673,
|
| 553 |
+
"grad_norm": 2.021365837854459,
|
| 554 |
+
"learning_rate": 7.997946611909651e-05,
|
| 555 |
+
"loss": 0.2306,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.6491372226787182,
|
| 560 |
+
"grad_norm": 1.6933294146500717,
|
| 561 |
+
"learning_rate": 8.100616016427105e-05,
|
| 562 |
+
"loss": 0.2205,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.657354149548069,
|
| 567 |
+
"grad_norm": 1.9476560876825286,
|
| 568 |
+
"learning_rate": 8.203285420944559e-05,
|
| 569 |
+
"loss": 0.2191,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.6655710764174199,
|
| 574 |
+
"grad_norm": 1.3694311196851179,
|
| 575 |
+
"learning_rate": 8.305954825462012e-05,
|
| 576 |
+
"loss": 0.2134,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.6737880032867708,
|
| 581 |
+
"grad_norm": 1.6387129939415674,
|
| 582 |
+
"learning_rate": 8.408624229979466e-05,
|
| 583 |
+
"loss": 0.2091,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.6820049301561216,
|
| 588 |
+
"grad_norm": 1.8243619220976712,
|
| 589 |
+
"learning_rate": 8.51129363449692e-05,
|
| 590 |
+
"loss": 0.2037,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.6902218570254725,
|
| 595 |
+
"grad_norm": 1.878468903383571,
|
| 596 |
+
"learning_rate": 8.613963039014374e-05,
|
| 597 |
+
"loss": 0.201,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.6984387838948234,
|
| 602 |
+
"grad_norm": 1.5456037490386814,
|
| 603 |
+
"learning_rate": 8.716632443531828e-05,
|
| 604 |
+
"loss": 0.1917,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.7066557107641742,
|
| 609 |
+
"grad_norm": 1.4587668759474102,
|
| 610 |
+
"learning_rate": 8.819301848049281e-05,
|
| 611 |
+
"loss": 0.187,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.714872637633525,
|
| 616 |
+
"grad_norm": 1.3297873752100995,
|
| 617 |
+
"learning_rate": 8.921971252566735e-05,
|
| 618 |
+
"loss": 0.1871,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.7230895645028759,
|
| 623 |
+
"grad_norm": 1.4624135144902266,
|
| 624 |
+
"learning_rate": 9.02464065708419e-05,
|
| 625 |
+
"loss": 0.1836,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.7313064913722268,
|
| 630 |
+
"grad_norm": 1.6852040534292057,
|
| 631 |
+
"learning_rate": 9.127310061601643e-05,
|
| 632 |
+
"loss": 0.1785,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.7395234182415776,
|
| 637 |
+
"grad_norm": 1.2757673968230296,
|
| 638 |
+
"learning_rate": 9.229979466119097e-05,
|
| 639 |
+
"loss": 0.1697,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.7477403451109285,
|
| 644 |
+
"grad_norm": 1.2025916192441726,
|
| 645 |
+
"learning_rate": 9.332648870636552e-05,
|
| 646 |
+
"loss": 0.1702,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.7559572719802794,
|
| 651 |
+
"grad_norm": 1.1424432696511961,
|
| 652 |
+
"learning_rate": 9.435318275154005e-05,
|
| 653 |
+
"loss": 0.1641,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.7641741988496302,
|
| 658 |
+
"grad_norm": 1.319419967704341,
|
| 659 |
+
"learning_rate": 9.537987679671459e-05,
|
| 660 |
+
"loss": 0.165,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.772391125718981,
|
| 665 |
+
"grad_norm": 1.328833007141129,
|
| 666 |
+
"learning_rate": 9.640657084188912e-05,
|
| 667 |
+
"loss": 0.1585,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.780608052588332,
|
| 672 |
+
"grad_norm": 1.257239962817816,
|
| 673 |
+
"learning_rate": 9.743326488706366e-05,
|
| 674 |
+
"loss": 0.1579,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.7888249794576828,
|
| 679 |
+
"grad_norm": 1.3160289143517214,
|
| 680 |
+
"learning_rate": 9.84599589322382e-05,
|
| 681 |
+
"loss": 0.1523,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.7970419063270336,
|
| 686 |
+
"grad_norm": 1.257470705040321,
|
| 687 |
+
"learning_rate": 9.948665297741273e-05,
|
| 688 |
+
"loss": 0.1485,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.8052588331963846,
|
| 693 |
+
"grad_norm": 1.3654898318565305,
|
| 694 |
+
"learning_rate": 9.999991965235563e-05,
|
| 695 |
+
"loss": 0.1448,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.8134757600657354,
|
| 700 |
+
"grad_norm": 1.0497934285629105,
|
| 701 |
+
"learning_rate": 9.999927687274992e-05,
|
| 702 |
+
"loss": 0.1407,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.8216926869350862,
|
| 707 |
+
"grad_norm": 1.2404666893886855,
|
| 708 |
+
"learning_rate": 9.999799132180184e-05,
|
| 709 |
+
"loss": 0.1406,
|
| 710 |
+
"step": 1000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.8299096138044372,
|
| 714 |
+
"grad_norm": 1.1134285871781415,
|
| 715 |
+
"learning_rate": 9.999606301603793e-05,
|
| 716 |
+
"loss": 0.1352,
|
| 717 |
+
"step": 1010
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.838126540673788,
|
| 721 |
+
"grad_norm": 1.0341905175457653,
|
| 722 |
+
"learning_rate": 9.999349198024775e-05,
|
| 723 |
+
"loss": 0.1306,
|
| 724 |
+
"step": 1020
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.8463434675431388,
|
| 728 |
+
"grad_norm": 1.1978341840529492,
|
| 729 |
+
"learning_rate": 9.999027824748352e-05,
|
| 730 |
+
"loss": 0.1308,
|
| 731 |
+
"step": 1030
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.8545603944124898,
|
| 735 |
+
"grad_norm": 0.9476848708735395,
|
| 736 |
+
"learning_rate": 9.998642185905977e-05,
|
| 737 |
+
"loss": 0.1275,
|
| 738 |
+
"step": 1040
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.8627773212818406,
|
| 742 |
+
"grad_norm": 1.151501305833549,
|
| 743 |
+
"learning_rate": 9.998192286455273e-05,
|
| 744 |
+
"loss": 0.1237,
|
| 745 |
+
"step": 1050
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.8709942481511914,
|
| 749 |
+
"grad_norm": 1.058908290806141,
|
| 750 |
+
"learning_rate": 9.997678132179969e-05,
|
| 751 |
+
"loss": 0.122,
|
| 752 |
+
"step": 1060
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.8792111750205424,
|
| 756 |
+
"grad_norm": 0.9773537576111707,
|
| 757 |
+
"learning_rate": 9.997099729689839e-05,
|
| 758 |
+
"loss": 0.1187,
|
| 759 |
+
"step": 1070
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.8874281018898932,
|
| 763 |
+
"grad_norm": 0.9239568401977911,
|
| 764 |
+
"learning_rate": 9.996457086420598e-05,
|
| 765 |
+
"loss": 0.1178,
|
| 766 |
+
"step": 1080
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.895645028759244,
|
| 770 |
+
"grad_norm": 0.8018713550987714,
|
| 771 |
+
"learning_rate": 9.995750210633822e-05,
|
| 772 |
+
"loss": 0.1156,
|
| 773 |
+
"step": 1090
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.903861955628595,
|
| 777 |
+
"grad_norm": 0.910903730189005,
|
| 778 |
+
"learning_rate": 9.994979111416829e-05,
|
| 779 |
+
"loss": 0.1117,
|
| 780 |
+
"step": 1100
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.9120788824979458,
|
| 784 |
+
"grad_norm": 0.8641295674337285,
|
| 785 |
+
"learning_rate": 9.994143798682573e-05,
|
| 786 |
+
"loss": 0.11,
|
| 787 |
+
"step": 1110
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.9202958093672966,
|
| 791 |
+
"grad_norm": 0.8312917811297335,
|
| 792 |
+
"learning_rate": 9.993244283169513e-05,
|
| 793 |
+
"loss": 0.1072,
|
| 794 |
+
"step": 1120
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.9285127362366475,
|
| 798 |
+
"grad_norm": 0.932166905539136,
|
| 799 |
+
"learning_rate": 9.99228057644147e-05,
|
| 800 |
+
"loss": 0.1043,
|
| 801 |
+
"step": 1130
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.9367296631059984,
|
| 805 |
+
"grad_norm": 0.7442839862957729,
|
| 806 |
+
"learning_rate": 9.991252690887485e-05,
|
| 807 |
+
"loss": 0.1016,
|
| 808 |
+
"step": 1140
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.9449465899753492,
|
| 812 |
+
"grad_norm": 0.9433805635704259,
|
| 813 |
+
"learning_rate": 9.990160639721657e-05,
|
| 814 |
+
"loss": 0.0995,
|
| 815 |
+
"step": 1150
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.9531635168447001,
|
| 819 |
+
"grad_norm": 0.6781275157765514,
|
| 820 |
+
"learning_rate": 9.989004436982974e-05,
|
| 821 |
+
"loss": 0.0955,
|
| 822 |
+
"step": 1160
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.961380443714051,
|
| 826 |
+
"grad_norm": 0.8437772770709154,
|
| 827 |
+
"learning_rate": 9.987784097535127e-05,
|
| 828 |
+
"loss": 0.0942,
|
| 829 |
+
"step": 1170
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.9695973705834018,
|
| 833 |
+
"grad_norm": 0.8251028566653184,
|
| 834 |
+
"learning_rate": 9.986499637066331e-05,
|
| 835 |
+
"loss": 0.093,
|
| 836 |
+
"step": 1180
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.9778142974527527,
|
| 840 |
+
"grad_norm": 0.7770190924053046,
|
| 841 |
+
"learning_rate": 9.985151072089112e-05,
|
| 842 |
+
"loss": 0.0906,
|
| 843 |
+
"step": 1190
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.9860312243221035,
|
| 847 |
+
"grad_norm": 0.8013950177509783,
|
| 848 |
+
"learning_rate": 9.983738419940096e-05,
|
| 849 |
+
"loss": 0.0891,
|
| 850 |
+
"step": 1200
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.9942481511914544,
|
| 854 |
+
"grad_norm": 0.8098380947976941,
|
| 855 |
+
"learning_rate": 9.982261698779795e-05,
|
| 856 |
+
"loss": 0.0864,
|
| 857 |
+
"step": 1210
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 1.0024650780608053,
|
| 861 |
+
"grad_norm": 0.7476525561924587,
|
| 862 |
+
"learning_rate": 9.980720927592366e-05,
|
| 863 |
+
"loss": 0.0841,
|
| 864 |
+
"step": 1220
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 1.0106820049301561,
|
| 868 |
+
"grad_norm": 0.6812631846413963,
|
| 869 |
+
"learning_rate": 9.979116126185364e-05,
|
| 870 |
+
"loss": 0.0771,
|
| 871 |
+
"step": 1230
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.018898931799507,
|
| 875 |
+
"grad_norm": 0.7035140667346095,
|
| 876 |
+
"learning_rate": 9.977447315189493e-05,
|
| 877 |
+
"loss": 0.0745,
|
| 878 |
+
"step": 1240
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 1.0271158586688578,
|
| 882 |
+
"grad_norm": 0.6884460379070533,
|
| 883 |
+
"learning_rate": 9.975714516058344e-05,
|
| 884 |
+
"loss": 0.0725,
|
| 885 |
+
"step": 1250
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 1.0353327855382086,
|
| 889 |
+
"grad_norm": 0.6413028251652191,
|
| 890 |
+
"learning_rate": 9.973917751068111e-05,
|
| 891 |
+
"loss": 0.0729,
|
| 892 |
+
"step": 1260
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 1.0435497124075597,
|
| 896 |
+
"grad_norm": 0.6746774697903619,
|
| 897 |
+
"learning_rate": 9.972057043317308e-05,
|
| 898 |
+
"loss": 0.0693,
|
| 899 |
+
"step": 1270
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 1.0517666392769105,
|
| 903 |
+
"grad_norm": 0.6126131445159512,
|
| 904 |
+
"learning_rate": 9.970132416726474e-05,
|
| 905 |
+
"loss": 0.067,
|
| 906 |
+
"step": 1280
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 1.0599835661462613,
|
| 910 |
+
"grad_norm": 0.7620795693209419,
|
| 911 |
+
"learning_rate": 9.968143896037864e-05,
|
| 912 |
+
"loss": 0.0669,
|
| 913 |
+
"step": 1290
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.0682004930156122,
|
| 917 |
+
"grad_norm": 0.7363014416948239,
|
| 918 |
+
"learning_rate": 9.966091506815128e-05,
|
| 919 |
+
"loss": 0.0672,
|
| 920 |
+
"step": 1300
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.076417419884963,
|
| 924 |
+
"grad_norm": 0.7081570706805334,
|
| 925 |
+
"learning_rate": 9.963975275442989e-05,
|
| 926 |
+
"loss": 0.0659,
|
| 927 |
+
"step": 1310
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 1.0846343467543138,
|
| 931 |
+
"grad_norm": 0.651054946068983,
|
| 932 |
+
"learning_rate": 9.961795229126896e-05,
|
| 933 |
+
"loss": 0.0641,
|
| 934 |
+
"step": 1320
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 1.0928512736236646,
|
| 938 |
+
"grad_norm": 0.6220299750434742,
|
| 939 |
+
"learning_rate": 9.959551395892681e-05,
|
| 940 |
+
"loss": 0.0617,
|
| 941 |
+
"step": 1330
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.1010682004930157,
|
| 945 |
+
"grad_norm": 0.683835515314829,
|
| 946 |
+
"learning_rate": 9.957243804586197e-05,
|
| 947 |
+
"loss": 0.0617,
|
| 948 |
+
"step": 1340
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.1092851273623665,
|
| 952 |
+
"grad_norm": 0.6807680658530081,
|
| 953 |
+
"learning_rate": 9.954872484872944e-05,
|
| 954 |
+
"loss": 0.0612,
|
| 955 |
+
"step": 1350
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.1175020542317173,
|
| 959 |
+
"grad_norm": 0.6514282591375157,
|
| 960 |
+
"learning_rate": 9.952437467237687e-05,
|
| 961 |
+
"loss": 0.06,
|
| 962 |
+
"step": 1360
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 1.1257189811010682,
|
| 966 |
+
"grad_norm": 0.6914058703103726,
|
| 967 |
+
"learning_rate": 9.949938782984072e-05,
|
| 968 |
+
"loss": 0.0596,
|
| 969 |
+
"step": 1370
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 1.133935907970419,
|
| 973 |
+
"grad_norm": 0.6568080555118521,
|
| 974 |
+
"learning_rate": 9.947376464234215e-05,
|
| 975 |
+
"loss": 0.0605,
|
| 976 |
+
"step": 1380
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 1.14215283483977,
|
| 980 |
+
"grad_norm": 0.6527114530677591,
|
| 981 |
+
"learning_rate": 9.944750543928297e-05,
|
| 982 |
+
"loss": 0.0581,
|
| 983 |
+
"step": 1390
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.1503697617091209,
|
| 987 |
+
"grad_norm": 0.6275835834331879,
|
| 988 |
+
"learning_rate": 9.942061055824127e-05,
|
| 989 |
+
"loss": 0.056,
|
| 990 |
+
"step": 1400
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 1.1585866885784717,
|
| 994 |
+
"grad_norm": 0.5795880116604848,
|
| 995 |
+
"learning_rate": 9.939308034496728e-05,
|
| 996 |
+
"loss": 0.0551,
|
| 997 |
+
"step": 1410
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.1668036154478225,
|
| 1001 |
+
"grad_norm": 0.5725267494550338,
|
| 1002 |
+
"learning_rate": 9.936491515337874e-05,
|
| 1003 |
+
"loss": 0.0543,
|
| 1004 |
+
"step": 1420
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 1.1750205423171733,
|
| 1008 |
+
"grad_norm": 0.541713230823975,
|
| 1009 |
+
"learning_rate": 9.933611534555644e-05,
|
| 1010 |
+
"loss": 0.0534,
|
| 1011 |
+
"step": 1430
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 1.1832374691865242,
|
| 1015 |
+
"grad_norm": 0.566900586795121,
|
| 1016 |
+
"learning_rate": 9.930668129173956e-05,
|
| 1017 |
+
"loss": 0.0534,
|
| 1018 |
+
"step": 1440
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 1.191454396055875,
|
| 1022 |
+
"grad_norm": 0.6604801022732167,
|
| 1023 |
+
"learning_rate": 9.927661337032093e-05,
|
| 1024 |
+
"loss": 0.0526,
|
| 1025 |
+
"step": 1450
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.199671322925226,
|
| 1029 |
+
"grad_norm": 0.6055038306197658,
|
| 1030 |
+
"learning_rate": 9.924591196784205e-05,
|
| 1031 |
+
"loss": 0.0526,
|
| 1032 |
+
"step": 1460
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 1.2078882497945769,
|
| 1036 |
+
"grad_norm": 0.5987814107613061,
|
| 1037 |
+
"learning_rate": 9.92145774789883e-05,
|
| 1038 |
+
"loss": 0.0513,
|
| 1039 |
+
"step": 1470
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.2161051766639277,
|
| 1043 |
+
"grad_norm": 0.6362437102114854,
|
| 1044 |
+
"learning_rate": 9.918261030658371e-05,
|
| 1045 |
+
"loss": 0.0515,
|
| 1046 |
+
"step": 1480
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 1.2243221035332785,
|
| 1050 |
+
"grad_norm": 0.6509204932067915,
|
| 1051 |
+
"learning_rate": 9.915001086158588e-05,
|
| 1052 |
+
"loss": 0.0513,
|
| 1053 |
+
"step": 1490
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 1.2325390304026294,
|
| 1057 |
+
"grad_norm": 0.5288885085851819,
|
| 1058 |
+
"learning_rate": 9.911677956308068e-05,
|
| 1059 |
+
"loss": 0.0506,
|
| 1060 |
+
"step": 1500
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 1.2407559572719804,
|
| 1064 |
+
"grad_norm": 0.5726585625989773,
|
| 1065 |
+
"learning_rate": 9.908291683827678e-05,
|
| 1066 |
+
"loss": 0.0486,
|
| 1067 |
+
"step": 1510
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.2489728841413312,
|
| 1071 |
+
"grad_norm": 0.589121204058476,
|
| 1072 |
+
"learning_rate": 9.904842312250024e-05,
|
| 1073 |
+
"loss": 0.0477,
|
| 1074 |
+
"step": 1520
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 1.257189811010682,
|
| 1078 |
+
"grad_norm": 0.6631593501137325,
|
| 1079 |
+
"learning_rate": 9.901329885918898e-05,
|
| 1080 |
+
"loss": 0.0483,
|
| 1081 |
+
"step": 1530
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.2654067378800329,
|
| 1085 |
+
"grad_norm": 0.5197654457910325,
|
| 1086 |
+
"learning_rate": 9.897754449988687e-05,
|
| 1087 |
+
"loss": 0.0457,
|
| 1088 |
+
"step": 1540
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 1.2736236647493837,
|
| 1092 |
+
"grad_norm": 0.5929482681421497,
|
| 1093 |
+
"learning_rate": 9.894116050423812e-05,
|
| 1094 |
+
"loss": 0.0453,
|
| 1095 |
+
"step": 1550
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 1.2818405916187345,
|
| 1099 |
+
"grad_norm": 0.4975008211438788,
|
| 1100 |
+
"learning_rate": 9.890414733998131e-05,
|
| 1101 |
+
"loss": 0.0435,
|
| 1102 |
+
"step": 1560
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 1.2900575184880854,
|
| 1106 |
+
"grad_norm": 0.4818366382204066,
|
| 1107 |
+
"learning_rate": 9.886650548294333e-05,
|
| 1108 |
+
"loss": 0.0439,
|
| 1109 |
+
"step": 1570
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.2982744453574364,
|
| 1113 |
+
"grad_norm": 0.5522264928264335,
|
| 1114 |
+
"learning_rate": 9.882823541703332e-05,
|
| 1115 |
+
"loss": 0.0431,
|
| 1116 |
+
"step": 1580
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 1.3064913722267872,
|
| 1120 |
+
"grad_norm": 0.542238275005933,
|
| 1121 |
+
"learning_rate": 9.878933763423644e-05,
|
| 1122 |
+
"loss": 0.0445,
|
| 1123 |
+
"step": 1590
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.314708299096138,
|
| 1127 |
+
"grad_norm": 0.4697141826718836,
|
| 1128 |
+
"learning_rate": 9.874981263460749e-05,
|
| 1129 |
+
"loss": 0.0418,
|
| 1130 |
+
"step": 1600
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 1.322925225965489,
|
| 1134 |
+
"grad_norm": 0.5054746485369349,
|
| 1135 |
+
"learning_rate": 9.87096609262646e-05,
|
| 1136 |
+
"loss": 0.0421,
|
| 1137 |
+
"step": 1610
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 1.3311421528348397,
|
| 1141 |
+
"grad_norm": 0.5395613302218427,
|
| 1142 |
+
"learning_rate": 9.866888302538255e-05,
|
| 1143 |
+
"loss": 0.0417,
|
| 1144 |
+
"step": 1620
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 1.3393590797041908,
|
| 1148 |
+
"grad_norm": 0.5157634368299763,
|
| 1149 |
+
"learning_rate": 9.862747945618629e-05,
|
| 1150 |
+
"loss": 0.0415,
|
| 1151 |
+
"step": 1630
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.3475760065735414,
|
| 1155 |
+
"grad_norm": 0.5519915955496593,
|
| 1156 |
+
"learning_rate": 9.858545075094402e-05,
|
| 1157 |
+
"loss": 0.0413,
|
| 1158 |
+
"step": 1640
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 1.3557929334428924,
|
| 1162 |
+
"grad_norm": 0.49777066445549617,
|
| 1163 |
+
"learning_rate": 9.854279744996054e-05,
|
| 1164 |
+
"loss": 0.0396,
|
| 1165 |
+
"step": 1650
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.3640098603122432,
|
| 1169 |
+
"grad_norm": 0.4940013153776274,
|
| 1170 |
+
"learning_rate": 9.849952010157014e-05,
|
| 1171 |
+
"loss": 0.0381,
|
| 1172 |
+
"step": 1660
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 1.372226787181594,
|
| 1176 |
+
"grad_norm": 0.5431277400396429,
|
| 1177 |
+
"learning_rate": 9.845561926212967e-05,
|
| 1178 |
+
"loss": 0.0386,
|
| 1179 |
+
"step": 1670
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 1.380443714050945,
|
| 1183 |
+
"grad_norm": 0.48822750198841763,
|
| 1184 |
+
"learning_rate": 9.841109549601133e-05,
|
| 1185 |
+
"loss": 0.0377,
|
| 1186 |
+
"step": 1680
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 1.3886606409202957,
|
| 1190 |
+
"grad_norm": 0.5171364389810015,
|
| 1191 |
+
"learning_rate": 9.83659493755954e-05,
|
| 1192 |
+
"loss": 0.0396,
|
| 1193 |
+
"step": 1690
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.3968775677896468,
|
| 1197 |
+
"grad_norm": 0.4649307579718927,
|
| 1198 |
+
"learning_rate": 9.832018148126291e-05,
|
| 1199 |
+
"loss": 0.0373,
|
| 1200 |
+
"step": 1700
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 1.4050944946589976,
|
| 1204 |
+
"grad_norm": 0.5377696639648694,
|
| 1205 |
+
"learning_rate": 9.827379240138824e-05,
|
| 1206 |
+
"loss": 0.0383,
|
| 1207 |
+
"step": 1710
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.4133114215283484,
|
| 1211 |
+
"grad_norm": 0.42020238886616096,
|
| 1212 |
+
"learning_rate": 9.822678273233139e-05,
|
| 1213 |
+
"loss": 0.0371,
|
| 1214 |
+
"step": 1720
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 1.4215283483976993,
|
| 1218 |
+
"grad_norm": 0.4583445207331194,
|
| 1219 |
+
"learning_rate": 9.817915307843046e-05,
|
| 1220 |
+
"loss": 0.036,
|
| 1221 |
+
"step": 1730
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 1.42974527526705,
|
| 1225 |
+
"grad_norm": 0.48217230605106,
|
| 1226 |
+
"learning_rate": 9.813090405199388e-05,
|
| 1227 |
+
"loss": 0.0355,
|
| 1228 |
+
"step": 1740
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 1.437962202136401,
|
| 1232 |
+
"grad_norm": 0.47263814058277104,
|
| 1233 |
+
"learning_rate": 9.80820362732924e-05,
|
| 1234 |
+
"loss": 0.0355,
|
| 1235 |
+
"step": 1750
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.4461791290057517,
|
| 1239 |
+
"grad_norm": 0.4649301210888705,
|
| 1240 |
+
"learning_rate": 9.803255037055128e-05,
|
| 1241 |
+
"loss": 0.034,
|
| 1242 |
+
"step": 1760
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 1.4543960558751028,
|
| 1246 |
+
"grad_norm": 0.42623498164530615,
|
| 1247 |
+
"learning_rate": 9.798244697994214e-05,
|
| 1248 |
+
"loss": 0.0339,
|
| 1249 |
+
"step": 1770
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.4626129827444536,
|
| 1253 |
+
"grad_norm": 0.4962026330356662,
|
| 1254 |
+
"learning_rate": 9.793172674557475e-05,
|
| 1255 |
+
"loss": 0.0338,
|
| 1256 |
+
"step": 1780
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.4708299096138044,
|
| 1260 |
+
"grad_norm": 0.46548952637723423,
|
| 1261 |
+
"learning_rate": 9.78803903194888e-05,
|
| 1262 |
+
"loss": 0.032,
|
| 1263 |
+
"step": 1790
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.4790468364831553,
|
| 1267 |
+
"grad_norm": 0.4102052300529563,
|
| 1268 |
+
"learning_rate": 9.782843836164553e-05,
|
| 1269 |
+
"loss": 0.0319,
|
| 1270 |
+
"step": 1800
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 1.487263763352506,
|
| 1274 |
+
"grad_norm": 0.42444897423048955,
|
| 1275 |
+
"learning_rate": 9.777587153991916e-05,
|
| 1276 |
+
"loss": 0.031,
|
| 1277 |
+
"step": 1810
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 1.4954806902218571,
|
| 1281 |
+
"grad_norm": 0.4484267423165579,
|
| 1282 |
+
"learning_rate": 9.77226905300884e-05,
|
| 1283 |
+
"loss": 0.032,
|
| 1284 |
+
"step": 1820
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 1.5036976170912077,
|
| 1288 |
+
"grad_norm": 0.4420656908365349,
|
| 1289 |
+
"learning_rate": 9.766889601582774e-05,
|
| 1290 |
+
"loss": 0.033,
|
| 1291 |
+
"step": 1830
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.5119145439605588,
|
| 1295 |
+
"grad_norm": 0.539138218093657,
|
| 1296 |
+
"learning_rate": 9.761448868869863e-05,
|
| 1297 |
+
"loss": 0.0317,
|
| 1298 |
+
"step": 1840
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.5201314708299096,
|
| 1302 |
+
"grad_norm": 0.5049031583303247,
|
| 1303 |
+
"learning_rate": 9.755946924814055e-05,
|
| 1304 |
+
"loss": 0.0307,
|
| 1305 |
+
"step": 1850
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 1.5283483976992605,
|
| 1309 |
+
"grad_norm": 0.44305034547894634,
|
| 1310 |
+
"learning_rate": 9.750383840146218e-05,
|
| 1311 |
+
"loss": 0.0311,
|
| 1312 |
+
"step": 1860
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 1.5365653245686115,
|
| 1316 |
+
"grad_norm": 0.4579452379069676,
|
| 1317 |
+
"learning_rate": 9.744759686383211e-05,
|
| 1318 |
+
"loss": 0.0325,
|
| 1319 |
+
"step": 1870
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 1.544782251437962,
|
| 1323 |
+
"grad_norm": 0.37927816066142095,
|
| 1324 |
+
"learning_rate": 9.739074535826978e-05,
|
| 1325 |
+
"loss": 0.0302,
|
| 1326 |
+
"step": 1880
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 1.5529991783073132,
|
| 1330 |
+
"grad_norm": 0.4459201122125563,
|
| 1331 |
+
"learning_rate": 9.733328461563612e-05,
|
| 1332 |
+
"loss": 0.0299,
|
| 1333 |
+
"step": 1890
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.561216105176664,
|
| 1337 |
+
"grad_norm": 0.4253605521821316,
|
| 1338 |
+
"learning_rate": 9.72752153746242e-05,
|
| 1339 |
+
"loss": 0.0304,
|
| 1340 |
+
"step": 1900
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.5694330320460148,
|
| 1344 |
+
"grad_norm": 0.44883217386433105,
|
| 1345 |
+
"learning_rate": 9.721653838174972e-05,
|
| 1346 |
+
"loss": 0.0292,
|
| 1347 |
+
"step": 1910
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 1.5776499589153656,
|
| 1351 |
+
"grad_norm": 0.42675407259402404,
|
| 1352 |
+
"learning_rate": 9.715725439134132e-05,
|
| 1353 |
+
"loss": 0.0286,
|
| 1354 |
+
"step": 1920
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 1.5858668857847165,
|
| 1358 |
+
"grad_norm": 0.4520177667187408,
|
| 1359 |
+
"learning_rate": 9.709736416553106e-05,
|
| 1360 |
+
"loss": 0.0283,
|
| 1361 |
+
"step": 1930
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 1.5940838126540675,
|
| 1365 |
+
"grad_norm": 0.41122780501579803,
|
| 1366 |
+
"learning_rate": 9.703686847424449e-05,
|
| 1367 |
+
"loss": 0.0282,
|
| 1368 |
+
"step": 1940
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 1.6023007395234181,
|
| 1372 |
+
"grad_norm": 0.4394266470273975,
|
| 1373 |
+
"learning_rate": 9.697576809519079e-05,
|
| 1374 |
+
"loss": 0.0286,
|
| 1375 |
+
"step": 1950
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.6105176663927692,
|
| 1379 |
+
"grad_norm": 0.4419576740762924,
|
| 1380 |
+
"learning_rate": 9.691406381385278e-05,
|
| 1381 |
+
"loss": 0.0279,
|
| 1382 |
+
"step": 1960
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.61873459326212,
|
| 1386 |
+
"grad_norm": 0.442978550961406,
|
| 1387 |
+
"learning_rate": 9.685175642347679e-05,
|
| 1388 |
+
"loss": 0.0271,
|
| 1389 |
+
"step": 1970
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 1.6269515201314708,
|
| 1393 |
+
"grad_norm": 0.4676149644951331,
|
| 1394 |
+
"learning_rate": 9.678884672506252e-05,
|
| 1395 |
+
"loss": 0.0274,
|
| 1396 |
+
"step": 1980
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 1.6351684470008216,
|
| 1400 |
+
"grad_norm": 0.41214249622744714,
|
| 1401 |
+
"learning_rate": 9.672533552735268e-05,
|
| 1402 |
+
"loss": 0.0276,
|
| 1403 |
+
"step": 1990
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 1.6433853738701725,
|
| 1407 |
+
"grad_norm": 0.43368213538774986,
|
| 1408 |
+
"learning_rate": 9.666122364682263e-05,
|
| 1409 |
+
"loss": 0.0268,
|
| 1410 |
+
"step": 2000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 1.6516023007395235,
|
| 1414 |
+
"grad_norm": 0.39587469514756823,
|
| 1415 |
+
"learning_rate": 9.659651190766992e-05,
|
| 1416 |
+
"loss": 0.0266,
|
| 1417 |
+
"step": 2010
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.6598192276088741,
|
| 1421 |
+
"grad_norm": 0.4263476346289595,
|
| 1422 |
+
"learning_rate": 9.653120114180356e-05,
|
| 1423 |
+
"loss": 0.0262,
|
| 1424 |
+
"step": 2020
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.6680361544782252,
|
| 1428 |
+
"grad_norm": 0.4196954209700085,
|
| 1429 |
+
"learning_rate": 9.646529218883349e-05,
|
| 1430 |
+
"loss": 0.0257,
|
| 1431 |
+
"step": 2030
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 1.676253081347576,
|
| 1435 |
+
"grad_norm": 0.36842453680103737,
|
| 1436 |
+
"learning_rate": 9.639878589605969e-05,
|
| 1437 |
+
"loss": 0.0263,
|
| 1438 |
+
"step": 2040
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 1.6844700082169268,
|
| 1442 |
+
"grad_norm": 0.4098763575060835,
|
| 1443 |
+
"learning_rate": 9.63316831184613e-05,
|
| 1444 |
+
"loss": 0.0247,
|
| 1445 |
+
"step": 2050
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 1.6926869350862779,
|
| 1449 |
+
"grad_norm": 0.36491073534209384,
|
| 1450 |
+
"learning_rate": 9.626398471868564e-05,
|
| 1451 |
+
"loss": 0.0249,
|
| 1452 |
+
"step": 2060
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 1.7009038619556285,
|
| 1456 |
+
"grad_norm": 0.3326235227359924,
|
| 1457 |
+
"learning_rate": 9.619569156703714e-05,
|
| 1458 |
+
"loss": 0.0243,
|
| 1459 |
+
"step": 2070
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.7091207888249795,
|
| 1463 |
+
"grad_norm": 0.4317864937663555,
|
| 1464 |
+
"learning_rate": 9.612680454146609e-05,
|
| 1465 |
+
"loss": 0.0249,
|
| 1466 |
+
"step": 2080
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.7173377156943304,
|
| 1470 |
+
"grad_norm": 0.3927870393510112,
|
| 1471 |
+
"learning_rate": 9.605732452755742e-05,
|
| 1472 |
+
"loss": 0.0238,
|
| 1473 |
+
"step": 2090
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 1.7255546425636812,
|
| 1477 |
+
"grad_norm": 0.3456898252191168,
|
| 1478 |
+
"learning_rate": 9.598725241851932e-05,
|
| 1479 |
+
"loss": 0.0251,
|
| 1480 |
+
"step": 2100
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 1.733771569433032,
|
| 1484 |
+
"grad_norm": 0.4148326323819897,
|
| 1485 |
+
"learning_rate": 9.591658911517162e-05,
|
| 1486 |
+
"loss": 0.0242,
|
| 1487 |
+
"step": 2110
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 1.7419884963023828,
|
| 1491 |
+
"grad_norm": 0.3742491564992171,
|
| 1492 |
+
"learning_rate": 9.584533552593447e-05,
|
| 1493 |
+
"loss": 0.0235,
|
| 1494 |
+
"step": 2120
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.7502054231717339,
|
| 1498 |
+
"grad_norm": 0.38217791470351015,
|
| 1499 |
+
"learning_rate": 9.577349256681636e-05,
|
| 1500 |
+
"loss": 0.0238,
|
| 1501 |
+
"step": 2130
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.7584223500410845,
|
| 1505 |
+
"grad_norm": 0.40231219986969197,
|
| 1506 |
+
"learning_rate": 9.570106116140256e-05,
|
| 1507 |
+
"loss": 0.0238,
|
| 1508 |
+
"step": 2140
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.7666392769104355,
|
| 1512 |
+
"grad_norm": 0.35604574469856265,
|
| 1513 |
+
"learning_rate": 9.56280422408432e-05,
|
| 1514 |
+
"loss": 0.0225,
|
| 1515 |
+
"step": 2150
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 1.7748562037797864,
|
| 1519 |
+
"grad_norm": 0.4359335650509276,
|
| 1520 |
+
"learning_rate": 9.555443674384122e-05,
|
| 1521 |
+
"loss": 0.0234,
|
| 1522 |
+
"step": 2160
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 1.7830731306491372,
|
| 1526 |
+
"grad_norm": 0.37099531274650005,
|
| 1527 |
+
"learning_rate": 9.548024561664039e-05,
|
| 1528 |
+
"loss": 0.0221,
|
| 1529 |
+
"step": 2170
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 1.7912900575184882,
|
| 1533 |
+
"grad_norm": 0.4042428161160167,
|
| 1534 |
+
"learning_rate": 9.540546981301313e-05,
|
| 1535 |
+
"loss": 0.0223,
|
| 1536 |
+
"step": 2180
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 1.7995069843878388,
|
| 1540 |
+
"grad_norm": 0.3933354408217013,
|
| 1541 |
+
"learning_rate": 9.53301102942482e-05,
|
| 1542 |
+
"loss": 0.0214,
|
| 1543 |
+
"step": 2190
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.80772391125719,
|
| 1547 |
+
"grad_norm": 0.40454570340041107,
|
| 1548 |
+
"learning_rate": 9.525416802913841e-05,
|
| 1549 |
+
"loss": 0.0218,
|
| 1550 |
+
"step": 2200
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 1.8159408381265407,
|
| 1554 |
+
"grad_norm": 0.44606137809155494,
|
| 1555 |
+
"learning_rate": 9.51776439939681e-05,
|
| 1556 |
+
"loss": 0.0216,
|
| 1557 |
+
"step": 2210
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 1.8241577649958916,
|
| 1561 |
+
"grad_norm": 0.3794919384637524,
|
| 1562 |
+
"learning_rate": 9.510053917250061e-05,
|
| 1563 |
+
"loss": 0.0206,
|
| 1564 |
+
"step": 2220
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 1.8323746918652424,
|
| 1568 |
+
"grad_norm": 0.4246386637316309,
|
| 1569 |
+
"learning_rate": 9.502285455596572e-05,
|
| 1570 |
+
"loss": 0.0214,
|
| 1571 |
+
"step": 2230
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.8405916187345932,
|
| 1575 |
+
"grad_norm": 0.3638287455568523,
|
| 1576 |
+
"learning_rate": 9.494459114304672e-05,
|
| 1577 |
+
"loss": 0.0213,
|
| 1578 |
+
"step": 2240
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 1.8488085456039443,
|
| 1582 |
+
"grad_norm": 0.37913499098378634,
|
| 1583 |
+
"learning_rate": 9.48657499398678e-05,
|
| 1584 |
+
"loss": 0.0212,
|
| 1585 |
+
"step": 2250
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.8570254724732949,
|
| 1589 |
+
"grad_norm": 0.3670647135686232,
|
| 1590 |
+
"learning_rate": 9.47863319599809e-05,
|
| 1591 |
+
"loss": 0.0208,
|
| 1592 |
+
"step": 2260
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 1.865242399342646,
|
| 1596 |
+
"grad_norm": 0.40206233339714315,
|
| 1597 |
+
"learning_rate": 9.470633822435282e-05,
|
| 1598 |
+
"loss": 0.0204,
|
| 1599 |
+
"step": 2270
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 1.8734593262119967,
|
| 1603 |
+
"grad_norm": 0.37239640663178286,
|
| 1604 |
+
"learning_rate": 9.462576976135207e-05,
|
| 1605 |
+
"loss": 0.0206,
|
| 1606 |
+
"step": 2280
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 1.8816762530813476,
|
| 1610 |
+
"grad_norm": 0.39727207631147365,
|
| 1611 |
+
"learning_rate": 9.454462760673559e-05,
|
| 1612 |
+
"loss": 0.0208,
|
| 1613 |
+
"step": 2290
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.8898931799506984,
|
| 1617 |
+
"grad_norm": 0.35764662278704595,
|
| 1618 |
+
"learning_rate": 9.446291280363551e-05,
|
| 1619 |
+
"loss": 0.0206,
|
| 1620 |
+
"step": 2300
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 1.8981101068200492,
|
| 1624 |
+
"grad_norm": 0.38375472594438204,
|
| 1625 |
+
"learning_rate": 9.43806264025457e-05,
|
| 1626 |
+
"loss": 0.0201,
|
| 1627 |
+
"step": 2310
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.9063270336894003,
|
| 1631 |
+
"grad_norm": 0.35816145402116345,
|
| 1632 |
+
"learning_rate": 9.429776946130826e-05,
|
| 1633 |
+
"loss": 0.0198,
|
| 1634 |
+
"step": 2320
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 1.9145439605587509,
|
| 1638 |
+
"grad_norm": 0.3582191092445618,
|
| 1639 |
+
"learning_rate": 9.421434304509992e-05,
|
| 1640 |
+
"loss": 0.019,
|
| 1641 |
+
"step": 2330
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 1.922760887428102,
|
| 1645 |
+
"grad_norm": 0.3284683434632605,
|
| 1646 |
+
"learning_rate": 9.413034822641844e-05,
|
| 1647 |
+
"loss": 0.019,
|
| 1648 |
+
"step": 2340
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 1.9309778142974527,
|
| 1652 |
+
"grad_norm": 0.3639213641523152,
|
| 1653 |
+
"learning_rate": 9.404578608506865e-05,
|
| 1654 |
+
"loss": 0.019,
|
| 1655 |
+
"step": 2350
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.9391947411668036,
|
| 1659 |
+
"grad_norm": 0.3491645108224867,
|
| 1660 |
+
"learning_rate": 9.396065770814867e-05,
|
| 1661 |
+
"loss": 0.0196,
|
| 1662 |
+
"step": 2360
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 1.9474116680361546,
|
| 1666 |
+
"grad_norm": 0.3911725317460492,
|
| 1667 |
+
"learning_rate": 9.387496419003599e-05,
|
| 1668 |
+
"loss": 0.0196,
|
| 1669 |
+
"step": 2370
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.9556285949055052,
|
| 1673 |
+
"grad_norm": 0.32174808087631224,
|
| 1674 |
+
"learning_rate": 9.378870663237328e-05,
|
| 1675 |
+
"loss": 0.0191,
|
| 1676 |
+
"step": 2380
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 1.9638455217748563,
|
| 1680 |
+
"grad_norm": 0.3583440347297591,
|
| 1681 |
+
"learning_rate": 9.37018861440543e-05,
|
| 1682 |
+
"loss": 0.019,
|
| 1683 |
+
"step": 2390
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 1.972062448644207,
|
| 1687 |
+
"grad_norm": 0.4078220177504098,
|
| 1688 |
+
"learning_rate": 9.361450384120963e-05,
|
| 1689 |
+
"loss": 0.0189,
|
| 1690 |
+
"step": 2400
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 1.980279375513558,
|
| 1694 |
+
"grad_norm": 0.32710238893529553,
|
| 1695 |
+
"learning_rate": 9.352656084719231e-05,
|
| 1696 |
+
"loss": 0.0184,
|
| 1697 |
+
"step": 2410
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.9884963023829088,
|
| 1701 |
+
"grad_norm": 0.31840080155890144,
|
| 1702 |
+
"learning_rate": 9.343805829256343e-05,
|
| 1703 |
+
"loss": 0.0185,
|
| 1704 |
+
"step": 2420
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 1.9967132292522596,
|
| 1708 |
+
"grad_norm": 0.35455068624210284,
|
| 1709 |
+
"learning_rate": 9.334899731507756e-05,
|
| 1710 |
+
"loss": 0.0181,
|
| 1711 |
+
"step": 2430
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.0049301561216106,
|
| 1715 |
+
"grad_norm": 0.3071054231037622,
|
| 1716 |
+
"learning_rate": 9.325937905966813e-05,
|
| 1717 |
+
"loss": 0.0144,
|
| 1718 |
+
"step": 2440
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 2.0131470829909612,
|
| 1722 |
+
"grad_norm": 0.26953847699923084,
|
| 1723 |
+
"learning_rate": 9.316920467843273e-05,
|
| 1724 |
+
"loss": 0.0118,
|
| 1725 |
+
"step": 2450
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 2.0213640098603123,
|
| 1729 |
+
"grad_norm": 0.2547464262612154,
|
| 1730 |
+
"learning_rate": 9.307847533061831e-05,
|
| 1731 |
+
"loss": 0.0109,
|
| 1732 |
+
"step": 2460
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 2.029580936729663,
|
| 1736 |
+
"grad_norm": 0.2758973898515017,
|
| 1737 |
+
"learning_rate": 9.29871921826062e-05,
|
| 1738 |
+
"loss": 0.0109,
|
| 1739 |
+
"step": 2470
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.037797863599014,
|
| 1743 |
+
"grad_norm": 0.2645333609854512,
|
| 1744 |
+
"learning_rate": 9.289535640789721e-05,
|
| 1745 |
+
"loss": 0.0102,
|
| 1746 |
+
"step": 2480
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 2.046014790468365,
|
| 1750 |
+
"grad_norm": 0.2605913577751368,
|
| 1751 |
+
"learning_rate": 9.280296918709653e-05,
|
| 1752 |
+
"loss": 0.0094,
|
| 1753 |
+
"step": 2490
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 2.0542317173377156,
|
| 1757 |
+
"grad_norm": 0.2352477044120282,
|
| 1758 |
+
"learning_rate": 9.271003170789846e-05,
|
| 1759 |
+
"loss": 0.0089,
|
| 1760 |
+
"step": 2500
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 2.0624486442070666,
|
| 1764 |
+
"grad_norm": 0.23955846077650936,
|
| 1765 |
+
"learning_rate": 9.261654516507126e-05,
|
| 1766 |
+
"loss": 0.0094,
|
| 1767 |
+
"step": 2510
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 2.0706655710764172,
|
| 1771 |
+
"grad_norm": 0.25860853720139226,
|
| 1772 |
+
"learning_rate": 9.252251076044172e-05,
|
| 1773 |
+
"loss": 0.0087,
|
| 1774 |
+
"step": 2520
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 2.0788824979457683,
|
| 1778 |
+
"grad_norm": 0.25940890635917274,
|
| 1779 |
+
"learning_rate": 9.242792970287972e-05,
|
| 1780 |
+
"loss": 0.0092,
|
| 1781 |
+
"step": 2530
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.0870994248151193,
|
| 1785 |
+
"grad_norm": 0.2818560862382246,
|
| 1786 |
+
"learning_rate": 9.233280320828274e-05,
|
| 1787 |
+
"loss": 0.0095,
|
| 1788 |
+
"step": 2540
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 2.09531635168447,
|
| 1792 |
+
"grad_norm": 0.2738556739938728,
|
| 1793 |
+
"learning_rate": 9.223713249956016e-05,
|
| 1794 |
+
"loss": 0.0095,
|
| 1795 |
+
"step": 2550
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 2.103533278553821,
|
| 1799 |
+
"grad_norm": 0.24521364436694368,
|
| 1800 |
+
"learning_rate": 9.214091880661752e-05,
|
| 1801 |
+
"loss": 0.0092,
|
| 1802 |
+
"step": 2560
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 2.1117502054231716,
|
| 1806 |
+
"grad_norm": 0.25699508253979375,
|
| 1807 |
+
"learning_rate": 9.204416336634084e-05,
|
| 1808 |
+
"loss": 0.0092,
|
| 1809 |
+
"step": 2570
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 2.1199671322925226,
|
| 1813 |
+
"grad_norm": 0.25530206616166484,
|
| 1814 |
+
"learning_rate": 9.194686742258059e-05,
|
| 1815 |
+
"loss": 0.0093,
|
| 1816 |
+
"step": 2580
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 2.1281840591618733,
|
| 1820 |
+
"grad_norm": 0.2576855814283002,
|
| 1821 |
+
"learning_rate": 9.184903222613574e-05,
|
| 1822 |
+
"loss": 0.0093,
|
| 1823 |
+
"step": 2590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.1364009860312243,
|
| 1827 |
+
"grad_norm": 0.22979520076080331,
|
| 1828 |
+
"learning_rate": 9.175065903473768e-05,
|
| 1829 |
+
"loss": 0.0085,
|
| 1830 |
+
"step": 2600
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 2.1446179129005754,
|
| 1834 |
+
"grad_norm": 0.2836426166631476,
|
| 1835 |
+
"learning_rate": 9.165174911303409e-05,
|
| 1836 |
+
"loss": 0.0088,
|
| 1837 |
+
"step": 2610
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 2.152834839769926,
|
| 1841 |
+
"grad_norm": 0.29495929086880596,
|
| 1842 |
+
"learning_rate": 9.15523037325726e-05,
|
| 1843 |
+
"loss": 0.0096,
|
| 1844 |
+
"step": 2620
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 2.161051766639277,
|
| 1848 |
+
"grad_norm": 0.22967580676824986,
|
| 1849 |
+
"learning_rate": 9.145232417178453e-05,
|
| 1850 |
+
"loss": 0.0092,
|
| 1851 |
+
"step": 2630
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 2.1692686935086276,
|
| 1855 |
+
"grad_norm": 0.26572615456213333,
|
| 1856 |
+
"learning_rate": 9.135181171596838e-05,
|
| 1857 |
+
"loss": 0.0094,
|
| 1858 |
+
"step": 2640
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 2.1774856203779787,
|
| 1862 |
+
"grad_norm": 0.2662740974356754,
|
| 1863 |
+
"learning_rate": 9.125076765727338e-05,
|
| 1864 |
+
"loss": 0.0087,
|
| 1865 |
+
"step": 2650
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.1857025472473293,
|
| 1869 |
+
"grad_norm": 0.30554630251156595,
|
| 1870 |
+
"learning_rate": 9.114919329468282e-05,
|
| 1871 |
+
"loss": 0.0088,
|
| 1872 |
+
"step": 2660
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 2.1939194741166803,
|
| 1876 |
+
"grad_norm": 0.2509783969525275,
|
| 1877 |
+
"learning_rate": 9.104708993399738e-05,
|
| 1878 |
+
"loss": 0.0085,
|
| 1879 |
+
"step": 2670
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 2.2021364009860314,
|
| 1883 |
+
"grad_norm": 0.29602083692919084,
|
| 1884 |
+
"learning_rate": 9.094445888781829e-05,
|
| 1885 |
+
"loss": 0.0093,
|
| 1886 |
+
"step": 2680
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.210353327855382,
|
| 1890 |
+
"grad_norm": 0.3041851856542598,
|
| 1891 |
+
"learning_rate": 9.084130147553057e-05,
|
| 1892 |
+
"loss": 0.0094,
|
| 1893 |
+
"step": 2690
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 2.218570254724733,
|
| 1897 |
+
"grad_norm": 0.2638783260630082,
|
| 1898 |
+
"learning_rate": 9.073761902328597e-05,
|
| 1899 |
+
"loss": 0.0092,
|
| 1900 |
+
"step": 2700
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 2.2267871815940836,
|
| 1904 |
+
"grad_norm": 0.2490709387523489,
|
| 1905 |
+
"learning_rate": 9.063341286398592e-05,
|
| 1906 |
+
"loss": 0.0091,
|
| 1907 |
+
"step": 2710
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.2350041084634347,
|
| 1911 |
+
"grad_norm": 0.3032383031195719,
|
| 1912 |
+
"learning_rate": 9.052868433726448e-05,
|
| 1913 |
+
"loss": 0.0089,
|
| 1914 |
+
"step": 2720
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 2.2432210353327857,
|
| 1918 |
+
"grad_norm": 0.252847381127495,
|
| 1919 |
+
"learning_rate": 9.042343478947103e-05,
|
| 1920 |
+
"loss": 0.0093,
|
| 1921 |
+
"step": 2730
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 2.2514379622021363,
|
| 1925 |
+
"grad_norm": 0.2620320939697099,
|
| 1926 |
+
"learning_rate": 9.0317665573653e-05,
|
| 1927 |
+
"loss": 0.0085,
|
| 1928 |
+
"step": 2740
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.2596548890714874,
|
| 1932 |
+
"grad_norm": 0.28243380866943507,
|
| 1933 |
+
"learning_rate": 9.021137804953846e-05,
|
| 1934 |
+
"loss": 0.0095,
|
| 1935 |
+
"step": 2750
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 2.267871815940838,
|
| 1939 |
+
"grad_norm": 0.2930988043769031,
|
| 1940 |
+
"learning_rate": 9.010457358351867e-05,
|
| 1941 |
+
"loss": 0.0089,
|
| 1942 |
+
"step": 2760
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 2.276088742810189,
|
| 1946 |
+
"grad_norm": 0.251699643820181,
|
| 1947 |
+
"learning_rate": 8.99972535486305e-05,
|
| 1948 |
+
"loss": 0.0088,
|
| 1949 |
+
"step": 2770
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.28430566967954,
|
| 1953 |
+
"grad_norm": 0.292931245753803,
|
| 1954 |
+
"learning_rate": 8.988941932453872e-05,
|
| 1955 |
+
"loss": 0.0092,
|
| 1956 |
+
"step": 2780
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 2.2925225965488907,
|
| 1960 |
+
"grad_norm": 0.2377722913294613,
|
| 1961 |
+
"learning_rate": 8.978107229751842e-05,
|
| 1962 |
+
"loss": 0.0095,
|
| 1963 |
+
"step": 2790
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 2.3007395234182417,
|
| 1967 |
+
"grad_norm": 0.25904938501329006,
|
| 1968 |
+
"learning_rate": 8.967221386043699e-05,
|
| 1969 |
+
"loss": 0.0088,
|
| 1970 |
+
"step": 2800
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.3089564502875923,
|
| 1974 |
+
"grad_norm": 0.27824310398994767,
|
| 1975 |
+
"learning_rate": 8.956284541273632e-05,
|
| 1976 |
+
"loss": 0.0099,
|
| 1977 |
+
"step": 2810
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 2.3171733771569434,
|
| 1981 |
+
"grad_norm": 0.2839375000972548,
|
| 1982 |
+
"learning_rate": 8.945296836041487e-05,
|
| 1983 |
+
"loss": 0.009,
|
| 1984 |
+
"step": 2820
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 2.325390304026294,
|
| 1988 |
+
"grad_norm": 0.28689394403734353,
|
| 1989 |
+
"learning_rate": 8.934258411600945e-05,
|
| 1990 |
+
"loss": 0.0092,
|
| 1991 |
+
"step": 2830
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.333607230895645,
|
| 1995 |
+
"grad_norm": 0.29946533063521097,
|
| 1996 |
+
"learning_rate": 8.923169409857718e-05,
|
| 1997 |
+
"loss": 0.0091,
|
| 1998 |
+
"step": 2840
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 2.341824157764996,
|
| 2002 |
+
"grad_norm": 0.25014388121553105,
|
| 2003 |
+
"learning_rate": 8.912029973367717e-05,
|
| 2004 |
+
"loss": 0.0094,
|
| 2005 |
+
"step": 2850
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 2.3500410846343467,
|
| 2009 |
+
"grad_norm": 0.26946873745003785,
|
| 2010 |
+
"learning_rate": 8.900840245335225e-05,
|
| 2011 |
+
"loss": 0.0089,
|
| 2012 |
+
"step": 2860
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.3582580115036977,
|
| 2016 |
+
"grad_norm": 0.2720064437159933,
|
| 2017 |
+
"learning_rate": 8.889600369611051e-05,
|
| 2018 |
+
"loss": 0.0094,
|
| 2019 |
+
"step": 2870
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 2.3664749383730483,
|
| 2023 |
+
"grad_norm": 0.26006919062307626,
|
| 2024 |
+
"learning_rate": 8.878310490690686e-05,
|
| 2025 |
+
"loss": 0.0096,
|
| 2026 |
+
"step": 2880
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 2.3746918652423994,
|
| 2030 |
+
"grad_norm": 0.2655250200457269,
|
| 2031 |
+
"learning_rate": 8.866970753712442e-05,
|
| 2032 |
+
"loss": 0.0097,
|
| 2033 |
+
"step": 2890
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.38290879211175,
|
| 2037 |
+
"grad_norm": 0.2500828943265464,
|
| 2038 |
+
"learning_rate": 8.855581304455584e-05,
|
| 2039 |
+
"loss": 0.0092,
|
| 2040 |
+
"step": 2900
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 2.391125718981101,
|
| 2044 |
+
"grad_norm": 0.2729008024818121,
|
| 2045 |
+
"learning_rate": 8.844142289338466e-05,
|
| 2046 |
+
"loss": 0.0089,
|
| 2047 |
+
"step": 2910
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 2.399342645850452,
|
| 2051 |
+
"grad_norm": 0.26949321217301414,
|
| 2052 |
+
"learning_rate": 8.832653855416634e-05,
|
| 2053 |
+
"loss": 0.0094,
|
| 2054 |
+
"step": 2920
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.4075595727198027,
|
| 2058 |
+
"grad_norm": 0.2563433496606597,
|
| 2059 |
+
"learning_rate": 8.821116150380946e-05,
|
| 2060 |
+
"loss": 0.0089,
|
| 2061 |
+
"step": 2930
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 2.4157764995891537,
|
| 2065 |
+
"grad_norm": 0.25430548752353105,
|
| 2066 |
+
"learning_rate": 8.80952932255567e-05,
|
| 2067 |
+
"loss": 0.0087,
|
| 2068 |
+
"step": 2940
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 2.4239934264585044,
|
| 2072 |
+
"grad_norm": 0.27239357241299744,
|
| 2073 |
+
"learning_rate": 8.797893520896579e-05,
|
| 2074 |
+
"loss": 0.0088,
|
| 2075 |
+
"step": 2950
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.4322103533278554,
|
| 2079 |
+
"grad_norm": 0.2904806280991979,
|
| 2080 |
+
"learning_rate": 8.786208894989032e-05,
|
| 2081 |
+
"loss": 0.0086,
|
| 2082 |
+
"step": 2960
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 2.4404272801972064,
|
| 2086 |
+
"grad_norm": 0.21052899106839745,
|
| 2087 |
+
"learning_rate": 8.774475595046055e-05,
|
| 2088 |
+
"loss": 0.009,
|
| 2089 |
+
"step": 2970
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 2.448644207066557,
|
| 2093 |
+
"grad_norm": 0.20821907134940407,
|
| 2094 |
+
"learning_rate": 8.762693771906409e-05,
|
| 2095 |
+
"loss": 0.0084,
|
| 2096 |
+
"step": 2980
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.456861133935908,
|
| 2100 |
+
"grad_norm": 0.24690866493061625,
|
| 2101 |
+
"learning_rate": 8.750863577032652e-05,
|
| 2102 |
+
"loss": 0.0086,
|
| 2103 |
+
"step": 2990
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 2.4650780608052587,
|
| 2107 |
+
"grad_norm": 0.27779361917607137,
|
| 2108 |
+
"learning_rate": 8.738985162509186e-05,
|
| 2109 |
+
"loss": 0.0091,
|
| 2110 |
+
"step": 3000
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 2.4732949876746098,
|
| 2114 |
+
"grad_norm": 0.2738741873571157,
|
| 2115 |
+
"learning_rate": 8.727058681040308e-05,
|
| 2116 |
+
"loss": 0.0087,
|
| 2117 |
+
"step": 3010
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.481511914543961,
|
| 2121 |
+
"grad_norm": 0.2306669670834553,
|
| 2122 |
+
"learning_rate": 8.715084285948248e-05,
|
| 2123 |
+
"loss": 0.0086,
|
| 2124 |
+
"step": 3020
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 2.4897288414133114,
|
| 2128 |
+
"grad_norm": 0.24305905366817937,
|
| 2129 |
+
"learning_rate": 8.703062131171192e-05,
|
| 2130 |
+
"loss": 0.009,
|
| 2131 |
+
"step": 3030
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 2.4979457682826625,
|
| 2135 |
+
"grad_norm": 0.25159950182400226,
|
| 2136 |
+
"learning_rate": 8.690992371261302e-05,
|
| 2137 |
+
"loss": 0.0092,
|
| 2138 |
+
"step": 3040
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.506162695152013,
|
| 2142 |
+
"grad_norm": 0.3208020474430245,
|
| 2143 |
+
"learning_rate": 8.678875161382745e-05,
|
| 2144 |
+
"loss": 0.0095,
|
| 2145 |
+
"step": 3050
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 2.514379622021364,
|
| 2149 |
+
"grad_norm": 0.24387207341884864,
|
| 2150 |
+
"learning_rate": 8.666710657309672e-05,
|
| 2151 |
+
"loss": 0.0088,
|
| 2152 |
+
"step": 3060
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 2.5225965488907147,
|
| 2156 |
+
"grad_norm": 0.22152256129132378,
|
| 2157 |
+
"learning_rate": 8.654499015424243e-05,
|
| 2158 |
+
"loss": 0.0089,
|
| 2159 |
+
"step": 3070
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.5308134757600658,
|
| 2163 |
+
"grad_norm": 0.32073953250921833,
|
| 2164 |
+
"learning_rate": 8.642240392714593e-05,
|
| 2165 |
+
"loss": 0.0084,
|
| 2166 |
+
"step": 3080
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 2.5390304026294164,
|
| 2170 |
+
"grad_norm": 0.2533741314967284,
|
| 2171 |
+
"learning_rate": 8.629934946772831e-05,
|
| 2172 |
+
"loss": 0.0093,
|
| 2173 |
+
"step": 3090
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 2.5472473294987674,
|
| 2177 |
+
"grad_norm": 0.28853651791957624,
|
| 2178 |
+
"learning_rate": 8.617582835793003e-05,
|
| 2179 |
+
"loss": 0.0094,
|
| 2180 |
+
"step": 3100
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 2.5554642563681185,
|
| 2184 |
+
"grad_norm": 0.2450454340740442,
|
| 2185 |
+
"learning_rate": 8.605184218569067e-05,
|
| 2186 |
+
"loss": 0.0087,
|
| 2187 |
+
"step": 3110
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 2.563681183237469,
|
| 2191 |
+
"grad_norm": 0.2564844184287705,
|
| 2192 |
+
"learning_rate": 8.592739254492844e-05,
|
| 2193 |
+
"loss": 0.0085,
|
| 2194 |
+
"step": 3120
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 2.57189811010682,
|
| 2198 |
+
"grad_norm": 0.22411359245516813,
|
| 2199 |
+
"learning_rate": 8.580248103551973e-05,
|
| 2200 |
+
"loss": 0.0085,
|
| 2201 |
+
"step": 3130
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.5801150369761707,
|
| 2205 |
+
"grad_norm": 0.2441583070658446,
|
| 2206 |
+
"learning_rate": 8.567710926327855e-05,
|
| 2207 |
+
"loss": 0.0084,
|
| 2208 |
+
"step": 3140
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 2.5883319638455218,
|
| 2212 |
+
"grad_norm": 0.3002781761942687,
|
| 2213 |
+
"learning_rate": 8.555127883993582e-05,
|
| 2214 |
+
"loss": 0.0083,
|
| 2215 |
+
"step": 3150
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 2.596548890714873,
|
| 2219 |
+
"grad_norm": 0.2613450003135346,
|
| 2220 |
+
"learning_rate": 8.542499138311877e-05,
|
| 2221 |
+
"loss": 0.0084,
|
| 2222 |
+
"step": 3160
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 2.6047658175842234,
|
| 2226 |
+
"grad_norm": 0.2827223351091325,
|
| 2227 |
+
"learning_rate": 8.529824851633004e-05,
|
| 2228 |
+
"loss": 0.0091,
|
| 2229 |
+
"step": 3170
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 2.6129827444535745,
|
| 2233 |
+
"grad_norm": 0.26781323505209714,
|
| 2234 |
+
"learning_rate": 8.517105186892683e-05,
|
| 2235 |
+
"loss": 0.0083,
|
| 2236 |
+
"step": 3180
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 2.621199671322925,
|
| 2240 |
+
"grad_norm": 0.22857572198441636,
|
| 2241 |
+
"learning_rate": 8.50434030761e-05,
|
| 2242 |
+
"loss": 0.0086,
|
| 2243 |
+
"step": 3190
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.629416598192276,
|
| 2247 |
+
"grad_norm": 0.25244081996610723,
|
| 2248 |
+
"learning_rate": 8.4915303778853e-05,
|
| 2249 |
+
"loss": 0.0083,
|
| 2250 |
+
"step": 3200
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 2.637633525061627,
|
| 2254 |
+
"grad_norm": 0.26276719839556406,
|
| 2255 |
+
"learning_rate": 8.478675562398077e-05,
|
| 2256 |
+
"loss": 0.0085,
|
| 2257 |
+
"step": 3210
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 2.645850451930978,
|
| 2261 |
+
"grad_norm": 0.2783615420618706,
|
| 2262 |
+
"learning_rate": 8.465776026404866e-05,
|
| 2263 |
+
"loss": 0.0085,
|
| 2264 |
+
"step": 3220
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 2.654067378800329,
|
| 2268 |
+
"grad_norm": 0.25331639937010203,
|
| 2269 |
+
"learning_rate": 8.452831935737102e-05,
|
| 2270 |
+
"loss": 0.0084,
|
| 2271 |
+
"step": 3230
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 2.6622843056696794,
|
| 2275 |
+
"grad_norm": 0.26160640757760606,
|
| 2276 |
+
"learning_rate": 8.439843456799005e-05,
|
| 2277 |
+
"loss": 0.0085,
|
| 2278 |
+
"step": 3240
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 2.6705012325390305,
|
| 2282 |
+
"grad_norm": 0.24658965460321536,
|
| 2283 |
+
"learning_rate": 8.426810756565429e-05,
|
| 2284 |
+
"loss": 0.0085,
|
| 2285 |
+
"step": 3250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.6787181594083815,
|
| 2289 |
+
"grad_norm": 0.27268220169048674,
|
| 2290 |
+
"learning_rate": 8.413734002579723e-05,
|
| 2291 |
+
"loss": 0.0089,
|
| 2292 |
+
"step": 3260
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 2.686935086277732,
|
| 2296 |
+
"grad_norm": 0.27608800070639516,
|
| 2297 |
+
"learning_rate": 8.40061336295157e-05,
|
| 2298 |
+
"loss": 0.0089,
|
| 2299 |
+
"step": 3270
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 2.6951520131470827,
|
| 2303 |
+
"grad_norm": 0.277674951643079,
|
| 2304 |
+
"learning_rate": 8.387449006354837e-05,
|
| 2305 |
+
"loss": 0.0089,
|
| 2306 |
+
"step": 3280
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 2.703368940016434,
|
| 2310 |
+
"grad_norm": 0.29362299478749754,
|
| 2311 |
+
"learning_rate": 8.374241102025391e-05,
|
| 2312 |
+
"loss": 0.0092,
|
| 2313 |
+
"step": 3290
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 2.711585866885785,
|
| 2317 |
+
"grad_norm": 0.24974090122124146,
|
| 2318 |
+
"learning_rate": 8.360989819758935e-05,
|
| 2319 |
+
"loss": 0.0089,
|
| 2320 |
+
"step": 3300
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 2.7198027937551355,
|
| 2324 |
+
"grad_norm": 0.24409808302370875,
|
| 2325 |
+
"learning_rate": 8.347695329908827e-05,
|
| 2326 |
+
"loss": 0.0093,
|
| 2327 |
+
"step": 3310
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.7280197206244865,
|
| 2331 |
+
"grad_norm": 0.24032759053052233,
|
| 2332 |
+
"learning_rate": 8.334357803383875e-05,
|
| 2333 |
+
"loss": 0.0085,
|
| 2334 |
+
"step": 3320
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 2.736236647493837,
|
| 2338 |
+
"grad_norm": 0.22024407549796368,
|
| 2339 |
+
"learning_rate": 8.32097741164616e-05,
|
| 2340 |
+
"loss": 0.0087,
|
| 2341 |
+
"step": 3330
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 2.744453574363188,
|
| 2345 |
+
"grad_norm": 0.2395884254526605,
|
| 2346 |
+
"learning_rate": 8.307554326708815e-05,
|
| 2347 |
+
"loss": 0.0083,
|
| 2348 |
+
"step": 3340
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 2.752670501232539,
|
| 2352 |
+
"grad_norm": 0.3016599666728437,
|
| 2353 |
+
"learning_rate": 8.294088721133821e-05,
|
| 2354 |
+
"loss": 0.0084,
|
| 2355 |
+
"step": 3350
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 2.76088742810189,
|
| 2359 |
+
"grad_norm": 0.25600020905060944,
|
| 2360 |
+
"learning_rate": 8.28058076802979e-05,
|
| 2361 |
+
"loss": 0.0087,
|
| 2362 |
+
"step": 3360
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 2.769104354971241,
|
| 2366 |
+
"grad_norm": 0.2649664553277115,
|
| 2367 |
+
"learning_rate": 8.267030641049738e-05,
|
| 2368 |
+
"loss": 0.0083,
|
| 2369 |
+
"step": 3370
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.7773212818405915,
|
| 2373 |
+
"grad_norm": 0.26462137135092484,
|
| 2374 |
+
"learning_rate": 8.25343851438885e-05,
|
| 2375 |
+
"loss": 0.0081,
|
| 2376 |
+
"step": 3380
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 2.7855382087099425,
|
| 2380 |
+
"grad_norm": 0.23318952147711133,
|
| 2381 |
+
"learning_rate": 8.239804562782241e-05,
|
| 2382 |
+
"loss": 0.0085,
|
| 2383 |
+
"step": 3390
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 2.7937551355792936,
|
| 2387 |
+
"grad_norm": 0.23450839984357066,
|
| 2388 |
+
"learning_rate": 8.226128961502715e-05,
|
| 2389 |
+
"loss": 0.0083,
|
| 2390 |
+
"step": 3400
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 2.801972062448644,
|
| 2394 |
+
"grad_norm": 0.2612729281193337,
|
| 2395 |
+
"learning_rate": 8.212411886358505e-05,
|
| 2396 |
+
"loss": 0.0085,
|
| 2397 |
+
"step": 3410
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 2.810188989317995,
|
| 2401 |
+
"grad_norm": 0.34263118420763916,
|
| 2402 |
+
"learning_rate": 8.198653513691019e-05,
|
| 2403 |
+
"loss": 0.0084,
|
| 2404 |
+
"step": 3420
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 2.818405916187346,
|
| 2408 |
+
"grad_norm": 0.22428156670157512,
|
| 2409 |
+
"learning_rate": 8.184854020372567e-05,
|
| 2410 |
+
"loss": 0.0082,
|
| 2411 |
+
"step": 3430
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.826622843056697,
|
| 2415 |
+
"grad_norm": 0.22173955007116375,
|
| 2416 |
+
"learning_rate": 8.171013583804095e-05,
|
| 2417 |
+
"loss": 0.0083,
|
| 2418 |
+
"step": 3440
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 2.834839769926048,
|
| 2422 |
+
"grad_norm": 0.2556051035140569,
|
| 2423 |
+
"learning_rate": 8.157132381912894e-05,
|
| 2424 |
+
"loss": 0.0082,
|
| 2425 |
+
"step": 3450
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 2.8430566967953985,
|
| 2429 |
+
"grad_norm": 0.25207338755160913,
|
| 2430 |
+
"learning_rate": 8.143210593150317e-05,
|
| 2431 |
+
"loss": 0.0081,
|
| 2432 |
+
"step": 3460
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 2.851273623664749,
|
| 2436 |
+
"grad_norm": 0.27270654388866217,
|
| 2437 |
+
"learning_rate": 8.129248396489493e-05,
|
| 2438 |
+
"loss": 0.0086,
|
| 2439 |
+
"step": 3470
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 2.8594905505341,
|
| 2443 |
+
"grad_norm": 0.2782218091570562,
|
| 2444 |
+
"learning_rate": 8.115245971423016e-05,
|
| 2445 |
+
"loss": 0.0085,
|
| 2446 |
+
"step": 3480
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 2.867707477403451,
|
| 2450 |
+
"grad_norm": 0.24242767736338208,
|
| 2451 |
+
"learning_rate": 8.101203497960639e-05,
|
| 2452 |
+
"loss": 0.0082,
|
| 2453 |
+
"step": 3490
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.875924404272802,
|
| 2457 |
+
"grad_norm": 0.20874710742285968,
|
| 2458 |
+
"learning_rate": 8.087121156626965e-05,
|
| 2459 |
+
"loss": 0.0081,
|
| 2460 |
+
"step": 3500
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 2.884141331142153,
|
| 2464 |
+
"grad_norm": 0.27306593621841513,
|
| 2465 |
+
"learning_rate": 8.072999128459119e-05,
|
| 2466 |
+
"loss": 0.0087,
|
| 2467 |
+
"step": 3510
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 2.8923582580115035,
|
| 2471 |
+
"grad_norm": 0.2911700894926946,
|
| 2472 |
+
"learning_rate": 8.05883759500443e-05,
|
| 2473 |
+
"loss": 0.0083,
|
| 2474 |
+
"step": 3520
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 2.9005751848808545,
|
| 2478 |
+
"grad_norm": 0.21963544354429723,
|
| 2479 |
+
"learning_rate": 8.044636738318082e-05,
|
| 2480 |
+
"loss": 0.0082,
|
| 2481 |
+
"step": 3530
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 2.9087921117502056,
|
| 2485 |
+
"grad_norm": 0.28438847429592207,
|
| 2486 |
+
"learning_rate": 8.030396740960795e-05,
|
| 2487 |
+
"loss": 0.0083,
|
| 2488 |
+
"step": 3540
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 2.917009038619556,
|
| 2492 |
+
"grad_norm": 0.2791357780384084,
|
| 2493 |
+
"learning_rate": 8.01611778599646e-05,
|
| 2494 |
+
"loss": 0.0089,
|
| 2495 |
+
"step": 3550
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 2.9252259654889072,
|
| 2499 |
+
"grad_norm": 0.270558213498406,
|
| 2500 |
+
"learning_rate": 8.001800056989792e-05,
|
| 2501 |
+
"loss": 0.0083,
|
| 2502 |
+
"step": 3560
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 2.933442892358258,
|
| 2506 |
+
"grad_norm": 0.24525793239252405,
|
| 2507 |
+
"learning_rate": 7.987443738003973e-05,
|
| 2508 |
+
"loss": 0.0079,
|
| 2509 |
+
"step": 3570
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 2.941659819227609,
|
| 2513 |
+
"grad_norm": 0.3195078999608411,
|
| 2514 |
+
"learning_rate": 7.973049013598277e-05,
|
| 2515 |
+
"loss": 0.0087,
|
| 2516 |
+
"step": 3580
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 2.94987674609696,
|
| 2520 |
+
"grad_norm": 0.25526324899137137,
|
| 2521 |
+
"learning_rate": 7.958616068825709e-05,
|
| 2522 |
+
"loss": 0.0086,
|
| 2523 |
+
"step": 3590
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 2.9580936729663105,
|
| 2527 |
+
"grad_norm": 0.25847558156104006,
|
| 2528 |
+
"learning_rate": 7.944145089230619e-05,
|
| 2529 |
+
"loss": 0.0081,
|
| 2530 |
+
"step": 3600
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 2.9663105998356616,
|
| 2534 |
+
"grad_norm": 0.24088162327594062,
|
| 2535 |
+
"learning_rate": 7.929636260846316e-05,
|
| 2536 |
+
"loss": 0.008,
|
| 2537 |
+
"step": 3610
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 2.974527526705012,
|
| 2541 |
+
"grad_norm": 0.27486387754064306,
|
| 2542 |
+
"learning_rate": 7.915089770192679e-05,
|
| 2543 |
+
"loss": 0.008,
|
| 2544 |
+
"step": 3620
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 2.9827444535743632,
|
| 2548 |
+
"grad_norm": 0.23754108351076786,
|
| 2549 |
+
"learning_rate": 7.900505804273759e-05,
|
| 2550 |
+
"loss": 0.0074,
|
| 2551 |
+
"step": 3630
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 2.9909613804437143,
|
| 2555 |
+
"grad_norm": 0.25789692279488435,
|
| 2556 |
+
"learning_rate": 7.885884550575376e-05,
|
| 2557 |
+
"loss": 0.0077,
|
| 2558 |
+
"step": 3640
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 2.999178307313065,
|
| 2562 |
+
"grad_norm": 0.23193412003095776,
|
| 2563 |
+
"learning_rate": 7.871226197062705e-05,
|
| 2564 |
+
"loss": 0.0077,
|
| 2565 |
+
"step": 3650
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 3.007395234182416,
|
| 2569 |
+
"grad_norm": 0.20611838311679784,
|
| 2570 |
+
"learning_rate": 7.856530932177862e-05,
|
| 2571 |
+
"loss": 0.005,
|
| 2572 |
+
"step": 3660
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 3.0156121610517665,
|
| 2576 |
+
"grad_norm": 0.20832054136973702,
|
| 2577 |
+
"learning_rate": 7.841798944837483e-05,
|
| 2578 |
+
"loss": 0.0048,
|
| 2579 |
+
"step": 3670
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 3.0238290879211176,
|
| 2583 |
+
"grad_norm": 0.1468814261205929,
|
| 2584 |
+
"learning_rate": 7.827030424430292e-05,
|
| 2585 |
+
"loss": 0.0044,
|
| 2586 |
+
"step": 3680
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 3.032046014790468,
|
| 2590 |
+
"grad_norm": 0.16265974747115644,
|
| 2591 |
+
"learning_rate": 7.812225560814669e-05,
|
| 2592 |
+
"loss": 0.004,
|
| 2593 |
+
"step": 3690
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 3.0402629416598193,
|
| 2597 |
+
"grad_norm": 0.1840058987828429,
|
| 2598 |
+
"learning_rate": 7.797384544316207e-05,
|
| 2599 |
+
"loss": 0.0038,
|
| 2600 |
+
"step": 3700
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 3.0484798685291703,
|
| 2604 |
+
"grad_norm": 0.13617101383458108,
|
| 2605 |
+
"learning_rate": 7.782507565725267e-05,
|
| 2606 |
+
"loss": 0.0036,
|
| 2607 |
+
"step": 3710
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 3.056696795398521,
|
| 2611 |
+
"grad_norm": 0.20684982155702297,
|
| 2612 |
+
"learning_rate": 7.767594816294528e-05,
|
| 2613 |
+
"loss": 0.0036,
|
| 2614 |
+
"step": 3720
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 3.064913722267872,
|
| 2618 |
+
"grad_norm": 0.16008088993018624,
|
| 2619 |
+
"learning_rate": 7.752646487736519e-05,
|
| 2620 |
+
"loss": 0.0033,
|
| 2621 |
+
"step": 3730
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 3.0731306491372226,
|
| 2625 |
+
"grad_norm": 0.1611386381907941,
|
| 2626 |
+
"learning_rate": 7.737662772221165e-05,
|
| 2627 |
+
"loss": 0.0035,
|
| 2628 |
+
"step": 3740
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 3.0813475760065736,
|
| 2632 |
+
"grad_norm": 0.1513272084012025,
|
| 2633 |
+
"learning_rate": 7.722643862373309e-05,
|
| 2634 |
+
"loss": 0.003,
|
| 2635 |
+
"step": 3750
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 3.089564502875924,
|
| 2639 |
+
"grad_norm": 0.14684151838561355,
|
| 2640 |
+
"learning_rate": 7.707589951270242e-05,
|
| 2641 |
+
"loss": 0.0032,
|
| 2642 |
+
"step": 3760
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 3.0977814297452753,
|
| 2646 |
+
"grad_norm": 0.1651386784320631,
|
| 2647 |
+
"learning_rate": 7.692501232439214e-05,
|
| 2648 |
+
"loss": 0.0032,
|
| 2649 |
+
"step": 3770
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 3.1059983566146263,
|
| 2653 |
+
"grad_norm": 0.14390663295343298,
|
| 2654 |
+
"learning_rate": 7.677377899854954e-05,
|
| 2655 |
+
"loss": 0.003,
|
| 2656 |
+
"step": 3780
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 3.114215283483977,
|
| 2660 |
+
"grad_norm": 0.11462029526186665,
|
| 2661 |
+
"learning_rate": 7.662220147937167e-05,
|
| 2662 |
+
"loss": 0.0032,
|
| 2663 |
+
"step": 3790
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 3.122432210353328,
|
| 2667 |
+
"grad_norm": 0.14646645133362796,
|
| 2668 |
+
"learning_rate": 7.647028171548043e-05,
|
| 2669 |
+
"loss": 0.0035,
|
| 2670 |
+
"step": 3800
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 3.1306491372226786,
|
| 2674 |
+
"grad_norm": 0.1623310045169937,
|
| 2675 |
+
"learning_rate": 7.631802165989749e-05,
|
| 2676 |
+
"loss": 0.003,
|
| 2677 |
+
"step": 3810
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 3.1388660640920296,
|
| 2681 |
+
"grad_norm": 0.14085603049103712,
|
| 2682 |
+
"learning_rate": 7.616542327001915e-05,
|
| 2683 |
+
"loss": 0.0032,
|
| 2684 |
+
"step": 3820
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 3.1470829909613807,
|
| 2688 |
+
"grad_norm": 0.19472246423837264,
|
| 2689 |
+
"learning_rate": 7.601248850759122e-05,
|
| 2690 |
+
"loss": 0.0032,
|
| 2691 |
+
"step": 3830
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 3.1552999178307313,
|
| 2695 |
+
"grad_norm": 0.16673368948205547,
|
| 2696 |
+
"learning_rate": 7.585921933868379e-05,
|
| 2697 |
+
"loss": 0.0032,
|
| 2698 |
+
"step": 3840
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 3.1635168447000823,
|
| 2702 |
+
"grad_norm": 0.15604748453875794,
|
| 2703 |
+
"learning_rate": 7.570561773366596e-05,
|
| 2704 |
+
"loss": 0.003,
|
| 2705 |
+
"step": 3850
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 3.171733771569433,
|
| 2709 |
+
"grad_norm": 0.16402585274140233,
|
| 2710 |
+
"learning_rate": 7.555168566718046e-05,
|
| 2711 |
+
"loss": 0.0028,
|
| 2712 |
+
"step": 3860
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 3.179950698438784,
|
| 2716 |
+
"grad_norm": 0.17232716930776026,
|
| 2717 |
+
"learning_rate": 7.539742511811833e-05,
|
| 2718 |
+
"loss": 0.003,
|
| 2719 |
+
"step": 3870
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 3.1881676253081346,
|
| 2723 |
+
"grad_norm": 0.14615636398793994,
|
| 2724 |
+
"learning_rate": 7.52428380695935e-05,
|
| 2725 |
+
"loss": 0.0029,
|
| 2726 |
+
"step": 3880
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 3.1963845521774856,
|
| 2730 |
+
"grad_norm": 0.15534493590729612,
|
| 2731 |
+
"learning_rate": 7.508792650891713e-05,
|
| 2732 |
+
"loss": 0.0031,
|
| 2733 |
+
"step": 3890
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 3.2046014790468367,
|
| 2737 |
+
"grad_norm": 0.23658113749842286,
|
| 2738 |
+
"learning_rate": 7.493269242757233e-05,
|
| 2739 |
+
"loss": 0.0028,
|
| 2740 |
+
"step": 3900
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 3.2128184059161873,
|
| 2744 |
+
"grad_norm": 0.18015077405140434,
|
| 2745 |
+
"learning_rate": 7.477713782118829e-05,
|
| 2746 |
+
"loss": 0.0031,
|
| 2747 |
+
"step": 3910
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 3.2210353327855383,
|
| 2751 |
+
"grad_norm": 0.17219078522340617,
|
| 2752 |
+
"learning_rate": 7.462126468951483e-05,
|
| 2753 |
+
"loss": 0.0028,
|
| 2754 |
+
"step": 3920
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 3.229252259654889,
|
| 2758 |
+
"grad_norm": 0.15175809799598178,
|
| 2759 |
+
"learning_rate": 7.446507503639654e-05,
|
| 2760 |
+
"loss": 0.003,
|
| 2761 |
+
"step": 3930
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 3.23746918652424,
|
| 2765 |
+
"grad_norm": 0.14898027368187575,
|
| 2766 |
+
"learning_rate": 7.430857086974714e-05,
|
| 2767 |
+
"loss": 0.0029,
|
| 2768 |
+
"step": 3940
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 3.2456861133935906,
|
| 2772 |
+
"grad_norm": 0.15537077187243506,
|
| 2773 |
+
"learning_rate": 7.415175420152358e-05,
|
| 2774 |
+
"loss": 0.0029,
|
| 2775 |
+
"step": 3950
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 3.2539030402629416,
|
| 2779 |
+
"grad_norm": 0.19946700618794544,
|
| 2780 |
+
"learning_rate": 7.399462704770023e-05,
|
| 2781 |
+
"loss": 0.0032,
|
| 2782 |
+
"step": 3960
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 3.2621199671322927,
|
| 2786 |
+
"grad_norm": 0.20987038286480436,
|
| 2787 |
+
"learning_rate": 7.383719142824296e-05,
|
| 2788 |
+
"loss": 0.0033,
|
| 2789 |
+
"step": 3970
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 3.2703368940016433,
|
| 2793 |
+
"grad_norm": 0.16617324816598264,
|
| 2794 |
+
"learning_rate": 7.367944936708309e-05,
|
| 2795 |
+
"loss": 0.0031,
|
| 2796 |
+
"step": 3980
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 3.2785538208709943,
|
| 2800 |
+
"grad_norm": 0.14705169237462418,
|
| 2801 |
+
"learning_rate": 7.352140289209149e-05,
|
| 2802 |
+
"loss": 0.0031,
|
| 2803 |
+
"step": 3990
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 3.286770747740345,
|
| 2807 |
+
"grad_norm": 0.23591449068985054,
|
| 2808 |
+
"learning_rate": 7.336305403505246e-05,
|
| 2809 |
+
"loss": 0.0034,
|
| 2810 |
+
"step": 4000
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 3.294987674609696,
|
| 2814 |
+
"grad_norm": 0.16990351384549035,
|
| 2815 |
+
"learning_rate": 7.320440483163757e-05,
|
| 2816 |
+
"loss": 0.0031,
|
| 2817 |
+
"step": 4010
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 3.303204601479047,
|
| 2821 |
+
"grad_norm": 0.14571315806011886,
|
| 2822 |
+
"learning_rate": 7.304545732137955e-05,
|
| 2823 |
+
"loss": 0.0037,
|
| 2824 |
+
"step": 4020
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 3.3114215283483976,
|
| 2828 |
+
"grad_norm": 0.20408138805569445,
|
| 2829 |
+
"learning_rate": 7.288621354764604e-05,
|
| 2830 |
+
"loss": 0.0033,
|
| 2831 |
+
"step": 4030
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 3.3196384552177487,
|
| 2835 |
+
"grad_norm": 0.1529759495306251,
|
| 2836 |
+
"learning_rate": 7.272667555761332e-05,
|
| 2837 |
+
"loss": 0.0033,
|
| 2838 |
+
"step": 4040
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 3.3278553820870993,
|
| 2842 |
+
"grad_norm": 0.16613488261024892,
|
| 2843 |
+
"learning_rate": 7.256684540224006e-05,
|
| 2844 |
+
"loss": 0.0035,
|
| 2845 |
+
"step": 4050
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 3.3360723089564503,
|
| 2849 |
+
"grad_norm": 0.216138731345594,
|
| 2850 |
+
"learning_rate": 7.24067251362408e-05,
|
| 2851 |
+
"loss": 0.0035,
|
| 2852 |
+
"step": 4060
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 3.3442892358258014,
|
| 2856 |
+
"grad_norm": 0.20249585543994236,
|
| 2857 |
+
"learning_rate": 7.224631681805969e-05,
|
| 2858 |
+
"loss": 0.0032,
|
| 2859 |
+
"step": 4070
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 3.352506162695152,
|
| 2863 |
+
"grad_norm": 0.1523713201485658,
|
| 2864 |
+
"learning_rate": 7.208562250984395e-05,
|
| 2865 |
+
"loss": 0.0032,
|
| 2866 |
+
"step": 4080
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 3.360723089564503,
|
| 2870 |
+
"grad_norm": 0.16021202587349764,
|
| 2871 |
+
"learning_rate": 7.19246442774174e-05,
|
| 2872 |
+
"loss": 0.0034,
|
| 2873 |
+
"step": 4090
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 3.3689400164338537,
|
| 2877 |
+
"grad_norm": 0.15798125172922542,
|
| 2878 |
+
"learning_rate": 7.176338419025384e-05,
|
| 2879 |
+
"loss": 0.0036,
|
| 2880 |
+
"step": 4100
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 3.3771569433032047,
|
| 2884 |
+
"grad_norm": 0.18322966397347049,
|
| 2885 |
+
"learning_rate": 7.160184432145049e-05,
|
| 2886 |
+
"loss": 0.0035,
|
| 2887 |
+
"step": 4110
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 3.3853738701725553,
|
| 2891 |
+
"grad_norm": 0.1568420066193485,
|
| 2892 |
+
"learning_rate": 7.144002674770142e-05,
|
| 2893 |
+
"loss": 0.0032,
|
| 2894 |
+
"step": 4120
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 3.3935907970419064,
|
| 2898 |
+
"grad_norm": 0.17611676847589813,
|
| 2899 |
+
"learning_rate": 7.127793354927062e-05,
|
| 2900 |
+
"loss": 0.0036,
|
| 2901 |
+
"step": 4130
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 3.401807723911257,
|
| 2905 |
+
"grad_norm": 0.1844863081048722,
|
| 2906 |
+
"learning_rate": 7.111556680996551e-05,
|
| 2907 |
+
"loss": 0.0034,
|
| 2908 |
+
"step": 4140
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 3.410024650780608,
|
| 2912 |
+
"grad_norm": 0.21600956572661287,
|
| 2913 |
+
"learning_rate": 7.095292861711005e-05,
|
| 2914 |
+
"loss": 0.0035,
|
| 2915 |
+
"step": 4150
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 3.418241577649959,
|
| 2919 |
+
"grad_norm": 0.18267255165799348,
|
| 2920 |
+
"learning_rate": 7.079002106151782e-05,
|
| 2921 |
+
"loss": 0.0034,
|
| 2922 |
+
"step": 4160
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 3.4264585045193097,
|
| 2926 |
+
"grad_norm": 0.17520388258678823,
|
| 2927 |
+
"learning_rate": 7.06268462374653e-05,
|
| 2928 |
+
"loss": 0.0039,
|
| 2929 |
+
"step": 4170
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 3.4346754313886607,
|
| 2933 |
+
"grad_norm": 0.15959064628289352,
|
| 2934 |
+
"learning_rate": 7.046340624266483e-05,
|
| 2935 |
+
"loss": 0.0033,
|
| 2936 |
+
"step": 4180
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 3.4428923582580113,
|
| 2940 |
+
"grad_norm": 0.1859360716007541,
|
| 2941 |
+
"learning_rate": 7.029970317823771e-05,
|
| 2942 |
+
"loss": 0.0034,
|
| 2943 |
+
"step": 4190
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 3.4511092851273624,
|
| 2947 |
+
"grad_norm": 0.18253730889004086,
|
| 2948 |
+
"learning_rate": 7.013573914868711e-05,
|
| 2949 |
+
"loss": 0.0036,
|
| 2950 |
+
"step": 4200
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 3.4593262119967134,
|
| 2954 |
+
"grad_norm": 0.202833868358773,
|
| 2955 |
+
"learning_rate": 6.997151626187116e-05,
|
| 2956 |
+
"loss": 0.0038,
|
| 2957 |
+
"step": 4210
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 3.467543138866064,
|
| 2961 |
+
"grad_norm": 0.16620497548550717,
|
| 2962 |
+
"learning_rate": 6.980703662897565e-05,
|
| 2963 |
+
"loss": 0.0037,
|
| 2964 |
+
"step": 4220
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 3.475760065735415,
|
| 2968 |
+
"grad_norm": 0.178575028828985,
|
| 2969 |
+
"learning_rate": 6.96423023644871e-05,
|
| 2970 |
+
"loss": 0.0037,
|
| 2971 |
+
"step": 4230
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 3.4839769926047657,
|
| 2975 |
+
"grad_norm": 0.1655444030360919,
|
| 2976 |
+
"learning_rate": 6.947731558616535e-05,
|
| 2977 |
+
"loss": 0.0029,
|
| 2978 |
+
"step": 4240
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 3.4921939194741167,
|
| 2982 |
+
"grad_norm": 0.16712614649679908,
|
| 2983 |
+
"learning_rate": 6.93120784150166e-05,
|
| 2984 |
+
"loss": 0.0039,
|
| 2985 |
+
"step": 4250
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 3.5004108463434678,
|
| 2989 |
+
"grad_norm": 0.2365818753770848,
|
| 2990 |
+
"learning_rate": 6.914659297526593e-05,
|
| 2991 |
+
"loss": 0.0039,
|
| 2992 |
+
"step": 4260
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 3.5086277732128184,
|
| 2996 |
+
"grad_norm": 0.1657733694924109,
|
| 2997 |
+
"learning_rate": 6.898086139433003e-05,
|
| 2998 |
+
"loss": 0.0035,
|
| 2999 |
+
"step": 4270
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 3.5168447000821694,
|
| 3003 |
+
"grad_norm": 0.13765746085497182,
|
| 3004 |
+
"learning_rate": 6.881488580278999e-05,
|
| 3005 |
+
"loss": 0.0032,
|
| 3006 |
+
"step": 4280
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 3.52506162695152,
|
| 3010 |
+
"grad_norm": 0.22175413809769864,
|
| 3011 |
+
"learning_rate": 6.864866833436368e-05,
|
| 3012 |
+
"loss": 0.0033,
|
| 3013 |
+
"step": 4290
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 3.533278553820871,
|
| 3017 |
+
"grad_norm": 0.14911731186659907,
|
| 3018 |
+
"learning_rate": 6.848221112587857e-05,
|
| 3019 |
+
"loss": 0.0037,
|
| 3020 |
+
"step": 4300
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 3.541495480690222,
|
| 3024 |
+
"grad_norm": 0.23023295249240697,
|
| 3025 |
+
"learning_rate": 6.831551631724402e-05,
|
| 3026 |
+
"loss": 0.0033,
|
| 3027 |
+
"step": 4310
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 3.5497124075595727,
|
| 3031 |
+
"grad_norm": 0.1720737753577987,
|
| 3032 |
+
"learning_rate": 6.8148586051424e-05,
|
| 3033 |
+
"loss": 0.0034,
|
| 3034 |
+
"step": 4320
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 3.5579293344289233,
|
| 3038 |
+
"grad_norm": 0.16406140755048426,
|
| 3039 |
+
"learning_rate": 6.798142247440932e-05,
|
| 3040 |
+
"loss": 0.0033,
|
| 3041 |
+
"step": 4330
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 3.5661462612982744,
|
| 3045 |
+
"grad_norm": 0.19878160507867879,
|
| 3046 |
+
"learning_rate": 6.781402773519021e-05,
|
| 3047 |
+
"loss": 0.0032,
|
| 3048 |
+
"step": 4340
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 3.5743631881676254,
|
| 3052 |
+
"grad_norm": 0.15835399871437064,
|
| 3053 |
+
"learning_rate": 6.764640398572864e-05,
|
| 3054 |
+
"loss": 0.0034,
|
| 3055 |
+
"step": 4350
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 3.582580115036976,
|
| 3059 |
+
"grad_norm": 0.14796838742423024,
|
| 3060 |
+
"learning_rate": 6.747855338093058e-05,
|
| 3061 |
+
"loss": 0.0034,
|
| 3062 |
+
"step": 4360
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 3.590797041906327,
|
| 3066 |
+
"grad_norm": 0.21938471856353853,
|
| 3067 |
+
"learning_rate": 6.731047807861846e-05,
|
| 3068 |
+
"loss": 0.0036,
|
| 3069 |
+
"step": 4370
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 3.5990139687756777,
|
| 3073 |
+
"grad_norm": 0.17971078316387218,
|
| 3074 |
+
"learning_rate": 6.714218023950323e-05,
|
| 3075 |
+
"loss": 0.0039,
|
| 3076 |
+
"step": 4380
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 3.6072308956450287,
|
| 3080 |
+
"grad_norm": 0.1745794123526125,
|
| 3081 |
+
"learning_rate": 6.697366202715677e-05,
|
| 3082 |
+
"loss": 0.0035,
|
| 3083 |
+
"step": 4390
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 3.61544782251438,
|
| 3087 |
+
"grad_norm": 0.17105334457637242,
|
| 3088 |
+
"learning_rate": 6.680492560798396e-05,
|
| 3089 |
+
"loss": 0.0037,
|
| 3090 |
+
"step": 4400
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 3.6236647493837304,
|
| 3094 |
+
"grad_norm": 0.2547405991108378,
|
| 3095 |
+
"learning_rate": 6.663597315119487e-05,
|
| 3096 |
+
"loss": 0.0037,
|
| 3097 |
+
"step": 4410
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 3.6318816762530814,
|
| 3101 |
+
"grad_norm": 0.1956208164605495,
|
| 3102 |
+
"learning_rate": 6.646680682877687e-05,
|
| 3103 |
+
"loss": 0.0039,
|
| 3104 |
+
"step": 4420
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 3.640098603122432,
|
| 3108 |
+
"grad_norm": 0.21528211446344578,
|
| 3109 |
+
"learning_rate": 6.629742881546666e-05,
|
| 3110 |
+
"loss": 0.0038,
|
| 3111 |
+
"step": 4430
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 3.648315529991783,
|
| 3115 |
+
"grad_norm": 0.17250519130965966,
|
| 3116 |
+
"learning_rate": 6.612784128872243e-05,
|
| 3117 |
+
"loss": 0.0037,
|
| 3118 |
+
"step": 4440
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 3.656532456861134,
|
| 3122 |
+
"grad_norm": 0.16141425415599522,
|
| 3123 |
+
"learning_rate": 6.595804642869572e-05,
|
| 3124 |
+
"loss": 0.0036,
|
| 3125 |
+
"step": 4450
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 3.6647493837304848,
|
| 3129 |
+
"grad_norm": 0.19769037842671158,
|
| 3130 |
+
"learning_rate": 6.578804641820352e-05,
|
| 3131 |
+
"loss": 0.0036,
|
| 3132 |
+
"step": 4460
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 3.672966310599836,
|
| 3136 |
+
"grad_norm": 0.1593671340462,
|
| 3137 |
+
"learning_rate": 6.561784344270014e-05,
|
| 3138 |
+
"loss": 0.0037,
|
| 3139 |
+
"step": 4470
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 3.6811832374691864,
|
| 3143 |
+
"grad_norm": 0.1948566085290808,
|
| 3144 |
+
"learning_rate": 6.544743969024912e-05,
|
| 3145 |
+
"loss": 0.0039,
|
| 3146 |
+
"step": 4480
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 3.6894001643385375,
|
| 3150 |
+
"grad_norm": 0.16021867166562243,
|
| 3151 |
+
"learning_rate": 6.527683735149512e-05,
|
| 3152 |
+
"loss": 0.004,
|
| 3153 |
+
"step": 4490
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 3.6976170912078885,
|
| 3157 |
+
"grad_norm": 0.224839451922574,
|
| 3158 |
+
"learning_rate": 6.510603861963575e-05,
|
| 3159 |
+
"loss": 0.0035,
|
| 3160 |
+
"step": 4500
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 3.705834018077239,
|
| 3164 |
+
"grad_norm": 0.191026901013589,
|
| 3165 |
+
"learning_rate": 6.493504569039336e-05,
|
| 3166 |
+
"loss": 0.0037,
|
| 3167 |
+
"step": 4510
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 3.7140509449465897,
|
| 3171 |
+
"grad_norm": 0.1980479101601017,
|
| 3172 |
+
"learning_rate": 6.476386076198682e-05,
|
| 3173 |
+
"loss": 0.0038,
|
| 3174 |
+
"step": 4520
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 3.7222678718159408,
|
| 3178 |
+
"grad_norm": 0.14276908599243227,
|
| 3179 |
+
"learning_rate": 6.459248603510331e-05,
|
| 3180 |
+
"loss": 0.0036,
|
| 3181 |
+
"step": 4530
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 3.730484798685292,
|
| 3185 |
+
"grad_norm": 0.1623914427970986,
|
| 3186 |
+
"learning_rate": 6.442092371286996e-05,
|
| 3187 |
+
"loss": 0.0037,
|
| 3188 |
+
"step": 4540
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 3.7387017255546424,
|
| 3192 |
+
"grad_norm": 0.17030799535465974,
|
| 3193 |
+
"learning_rate": 6.424917600082551e-05,
|
| 3194 |
+
"loss": 0.0035,
|
| 3195 |
+
"step": 4550
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 3.7469186524239935,
|
| 3199 |
+
"grad_norm": 0.19443118010484106,
|
| 3200 |
+
"learning_rate": 6.407724510689208e-05,
|
| 3201 |
+
"loss": 0.0035,
|
| 3202 |
+
"step": 4560
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 3.755135579293344,
|
| 3206 |
+
"grad_norm": 0.16883077758634854,
|
| 3207 |
+
"learning_rate": 6.390513324134666e-05,
|
| 3208 |
+
"loss": 0.0032,
|
| 3209 |
+
"step": 4570
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 3.763352506162695,
|
| 3213 |
+
"grad_norm": 0.1446254144439802,
|
| 3214 |
+
"learning_rate": 6.373284261679274e-05,
|
| 3215 |
+
"loss": 0.0037,
|
| 3216 |
+
"step": 4580
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 3.771569433032046,
|
| 3220 |
+
"grad_norm": 0.17807271909664887,
|
| 3221 |
+
"learning_rate": 6.356037544813186e-05,
|
| 3222 |
+
"loss": 0.0034,
|
| 3223 |
+
"step": 4590
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 3.7797863599013968,
|
| 3227 |
+
"grad_norm": 0.17167078258678709,
|
| 3228 |
+
"learning_rate": 6.338773395253518e-05,
|
| 3229 |
+
"loss": 0.0036,
|
| 3230 |
+
"step": 4600
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 3.788003286770748,
|
| 3234 |
+
"grad_norm": 0.17778471772279328,
|
| 3235 |
+
"learning_rate": 6.32149203494149e-05,
|
| 3236 |
+
"loss": 0.0036,
|
| 3237 |
+
"step": 4610
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 3.7962202136400984,
|
| 3241 |
+
"grad_norm": 0.17582409162117943,
|
| 3242 |
+
"learning_rate": 6.30419368603958e-05,
|
| 3243 |
+
"loss": 0.0035,
|
| 3244 |
+
"step": 4620
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 3.8044371405094495,
|
| 3248 |
+
"grad_norm": 0.18125539367559815,
|
| 3249 |
+
"learning_rate": 6.286878570928661e-05,
|
| 3250 |
+
"loss": 0.0035,
|
| 3251 |
+
"step": 4630
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 3.8126540673788005,
|
| 3255 |
+
"grad_norm": 0.21154767131744387,
|
| 3256 |
+
"learning_rate": 6.26954691220515e-05,
|
| 3257 |
+
"loss": 0.004,
|
| 3258 |
+
"step": 4640
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 3.820870994248151,
|
| 3262 |
+
"grad_norm": 0.1463818425284848,
|
| 3263 |
+
"learning_rate": 6.252198932678139e-05,
|
| 3264 |
+
"loss": 0.0034,
|
| 3265 |
+
"step": 4650
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 3.829087921117502,
|
| 3269 |
+
"grad_norm": 0.15967244918901152,
|
| 3270 |
+
"learning_rate": 6.234834855366536e-05,
|
| 3271 |
+
"loss": 0.0039,
|
| 3272 |
+
"step": 4660
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 3.837304847986853,
|
| 3276 |
+
"grad_norm": 0.1857473122185739,
|
| 3277 |
+
"learning_rate": 6.217454903496194e-05,
|
| 3278 |
+
"loss": 0.0035,
|
| 3279 |
+
"step": 4670
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 3.845521774856204,
|
| 3283 |
+
"grad_norm": 0.18521807360624734,
|
| 3284 |
+
"learning_rate": 6.200059300497045e-05,
|
| 3285 |
+
"loss": 0.0039,
|
| 3286 |
+
"step": 4680
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 3.853738701725555,
|
| 3290 |
+
"grad_norm": 0.2421664768320646,
|
| 3291 |
+
"learning_rate": 6.182648270000226e-05,
|
| 3292 |
+
"loss": 0.0037,
|
| 3293 |
+
"step": 4690
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 3.8619556285949055,
|
| 3297 |
+
"grad_norm": 0.19545607049185154,
|
| 3298 |
+
"learning_rate": 6.1652220358352e-05,
|
| 3299 |
+
"loss": 0.0036,
|
| 3300 |
+
"step": 4700
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 3.870172555464256,
|
| 3304 |
+
"grad_norm": 0.20721071787740614,
|
| 3305 |
+
"learning_rate": 6.147780822026888e-05,
|
| 3306 |
+
"loss": 0.0037,
|
| 3307 |
+
"step": 4710
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 3.878389482333607,
|
| 3311 |
+
"grad_norm": 0.18450330160048478,
|
| 3312 |
+
"learning_rate": 6.13032485279278e-05,
|
| 3313 |
+
"loss": 0.0034,
|
| 3314 |
+
"step": 4720
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 3.886606409202958,
|
| 3318 |
+
"grad_norm": 0.1898947322525971,
|
| 3319 |
+
"learning_rate": 6.112854352540058e-05,
|
| 3320 |
+
"loss": 0.0036,
|
| 3321 |
+
"step": 4730
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 3.894823336072309,
|
| 3325 |
+
"grad_norm": 0.1636533221065079,
|
| 3326 |
+
"learning_rate": 6.095369545862708e-05,
|
| 3327 |
+
"loss": 0.0034,
|
| 3328 |
+
"step": 4740
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 3.90304026294166,
|
| 3332 |
+
"grad_norm": 0.16619525328679463,
|
| 3333 |
+
"learning_rate": 6.0778706575386335e-05,
|
| 3334 |
+
"loss": 0.004,
|
| 3335 |
+
"step": 4750
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"epoch": 3.9112571898110104,
|
| 3339 |
+
"grad_norm": 0.20067592543633048,
|
| 3340 |
+
"learning_rate": 6.060357912526766e-05,
|
| 3341 |
+
"loss": 0.0038,
|
| 3342 |
+
"step": 4760
|
| 3343 |
+
},
|
| 3344 |
+
{
|
| 3345 |
+
"epoch": 3.9194741166803615,
|
| 3346 |
+
"grad_norm": 0.22596168198940297,
|
| 3347 |
+
"learning_rate": 6.042831535964177e-05,
|
| 3348 |
+
"loss": 0.0034,
|
| 3349 |
+
"step": 4770
|
| 3350 |
+
},
|
| 3351 |
+
{
|
| 3352 |
+
"epoch": 3.9276910435497125,
|
| 3353 |
+
"grad_norm": 0.14817899044077412,
|
| 3354 |
+
"learning_rate": 6.025291753163174e-05,
|
| 3355 |
+
"loss": 0.0035,
|
| 3356 |
+
"step": 4780
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 3.935907970419063,
|
| 3360 |
+
"grad_norm": 0.2423064027524078,
|
| 3361 |
+
"learning_rate": 6.007738789608414e-05,
|
| 3362 |
+
"loss": 0.0037,
|
| 3363 |
+
"step": 4790
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 3.944124897288414,
|
| 3367 |
+
"grad_norm": 0.21556181623439455,
|
| 3368 |
+
"learning_rate": 5.990172870953997e-05,
|
| 3369 |
+
"loss": 0.0038,
|
| 3370 |
+
"step": 4800
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 3.952341824157765,
|
| 3374 |
+
"grad_norm": 0.1950354253714728,
|
| 3375 |
+
"learning_rate": 5.972594223020576e-05,
|
| 3376 |
+
"loss": 0.0037,
|
| 3377 |
+
"step": 4810
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 3.960558751027116,
|
| 3381 |
+
"grad_norm": 0.17030134203570338,
|
| 3382 |
+
"learning_rate": 5.955003071792438e-05,
|
| 3383 |
+
"loss": 0.0031,
|
| 3384 |
+
"step": 4820
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 3.968775677896467,
|
| 3388 |
+
"grad_norm": 0.21912264273169446,
|
| 3389 |
+
"learning_rate": 5.937399643414611e-05,
|
| 3390 |
+
"loss": 0.0032,
|
| 3391 |
+
"step": 4830
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 3.9769926047658175,
|
| 3395 |
+
"grad_norm": 0.16893916392414404,
|
| 3396 |
+
"learning_rate": 5.9197841641899575e-05,
|
| 3397 |
+
"loss": 0.0035,
|
| 3398 |
+
"step": 4840
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 3.9852095316351686,
|
| 3402 |
+
"grad_norm": 0.18413453070171962,
|
| 3403 |
+
"learning_rate": 5.9021568605762534e-05,
|
| 3404 |
+
"loss": 0.0039,
|
| 3405 |
+
"step": 4850
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 3.993426458504519,
|
| 3409 |
+
"grad_norm": 0.18940235793711632,
|
| 3410 |
+
"learning_rate": 5.884517959183291e-05,
|
| 3411 |
+
"loss": 0.0037,
|
| 3412 |
+
"step": 4860
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 4.00164338537387,
|
| 3416 |
+
"grad_norm": 0.1393221881449365,
|
| 3417 |
+
"learning_rate": 5.8668676867699545e-05,
|
| 3418 |
+
"loss": 0.0034,
|
| 3419 |
+
"step": 4870
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 4.009860312243221,
|
| 3423 |
+
"grad_norm": 0.1603247488228045,
|
| 3424 |
+
"learning_rate": 5.849206270241313e-05,
|
| 3425 |
+
"loss": 0.0024,
|
| 3426 |
+
"step": 4880
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 4.018077239112572,
|
| 3430 |
+
"grad_norm": 0.16674488753226002,
|
| 3431 |
+
"learning_rate": 5.831533936645697e-05,
|
| 3432 |
+
"loss": 0.0021,
|
| 3433 |
+
"step": 4890
|
| 3434 |
+
},
|
| 3435 |
+
{
|
| 3436 |
+
"epoch": 4.0262941659819225,
|
| 3437 |
+
"grad_norm": 0.15020824654412976,
|
| 3438 |
+
"learning_rate": 5.813850913171782e-05,
|
| 3439 |
+
"loss": 0.002,
|
| 3440 |
+
"step": 4900
|
| 3441 |
+
},
|
| 3442 |
+
{
|
| 3443 |
+
"epoch": 4.0345110928512735,
|
| 3444 |
+
"grad_norm": 0.10482588507029508,
|
| 3445 |
+
"learning_rate": 5.7961574271456766e-05,
|
| 3446 |
+
"loss": 0.0016,
|
| 3447 |
+
"step": 4910
|
| 3448 |
+
},
|
| 3449 |
+
{
|
| 3450 |
+
"epoch": 4.042728019720625,
|
| 3451 |
+
"grad_norm": 0.14928575717348813,
|
| 3452 |
+
"learning_rate": 5.778453706027979e-05,
|
| 3453 |
+
"loss": 0.0019,
|
| 3454 |
+
"step": 4920
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 4.050944946589976,
|
| 3458 |
+
"grad_norm": 0.1150793056715666,
|
| 3459 |
+
"learning_rate": 5.7607399774108775e-05,
|
| 3460 |
+
"loss": 0.0017,
|
| 3461 |
+
"step": 4930
|
| 3462 |
+
},
|
| 3463 |
+
{
|
| 3464 |
+
"epoch": 4.059161873459326,
|
| 3465 |
+
"grad_norm": 0.12505382928693612,
|
| 3466 |
+
"learning_rate": 5.743016469015204e-05,
|
| 3467 |
+
"loss": 0.0015,
|
| 3468 |
+
"step": 4940
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"epoch": 4.067378800328677,
|
| 3472 |
+
"grad_norm": 0.11790326542154293,
|
| 3473 |
+
"learning_rate": 5.725283408687524e-05,
|
| 3474 |
+
"loss": 0.0015,
|
| 3475 |
+
"step": 4950
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 4.075595727198028,
|
| 3479 |
+
"grad_norm": 0.1381199290658754,
|
| 3480 |
+
"learning_rate": 5.7075410243971896e-05,
|
| 3481 |
+
"loss": 0.0016,
|
| 3482 |
+
"step": 4960
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 4.083812654067379,
|
| 3486 |
+
"grad_norm": 0.05824900643358759,
|
| 3487 |
+
"learning_rate": 5.689789544233425e-05,
|
| 3488 |
+
"loss": 0.0015,
|
| 3489 |
+
"step": 4970
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 4.09202958093673,
|
| 3493 |
+
"grad_norm": 0.0891691142827222,
|
| 3494 |
+
"learning_rate": 5.6720291964023866e-05,
|
| 3495 |
+
"loss": 0.0015,
|
| 3496 |
+
"step": 4980
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 4.10024650780608,
|
| 3500 |
+
"grad_norm": 0.085824973416845,
|
| 3501 |
+
"learning_rate": 5.654260209224225e-05,
|
| 3502 |
+
"loss": 0.0014,
|
| 3503 |
+
"step": 4990
|
| 3504 |
+
},
|
| 3505 |
+
{
|
| 3506 |
+
"epoch": 4.108463434675431,
|
| 3507 |
+
"grad_norm": 0.11834543603170807,
|
| 3508 |
+
"learning_rate": 5.6364828111301636e-05,
|
| 3509 |
+
"loss": 0.0013,
|
| 3510 |
+
"step": 5000
|
| 3511 |
+
},
|
| 3512 |
+
{
|
| 3513 |
+
"epoch": 4.116680361544782,
|
| 3514 |
+
"grad_norm": 0.14017021539733548,
|
| 3515 |
+
"learning_rate": 5.618697230659545e-05,
|
| 3516 |
+
"loss": 0.0015,
|
| 3517 |
+
"step": 5010
|
| 3518 |
+
},
|
| 3519 |
+
{
|
| 3520 |
+
"epoch": 4.124897288414133,
|
| 3521 |
+
"grad_norm": 0.08538943269149613,
|
| 3522 |
+
"learning_rate": 5.600903696456907e-05,
|
| 3523 |
+
"loss": 0.0014,
|
| 3524 |
+
"step": 5020
|
| 3525 |
+
},
|
| 3526 |
+
{
|
| 3527 |
+
"epoch": 4.133114215283484,
|
| 3528 |
+
"grad_norm": 0.1586537799394671,
|
| 3529 |
+
"learning_rate": 5.583102437269032e-05,
|
| 3530 |
+
"loss": 0.0014,
|
| 3531 |
+
"step": 5030
|
| 3532 |
+
},
|
| 3533 |
+
{
|
| 3534 |
+
"epoch": 4.1413311421528345,
|
| 3535 |
+
"grad_norm": 0.08701617195106857,
|
| 3536 |
+
"learning_rate": 5.565293681942018e-05,
|
| 3537 |
+
"loss": 0.0014,
|
| 3538 |
+
"step": 5040
|
| 3539 |
+
},
|
| 3540 |
+
{
|
| 3541 |
+
"epoch": 4.1495480690221855,
|
| 3542 |
+
"grad_norm": 0.09634033751054992,
|
| 3543 |
+
"learning_rate": 5.5474776594183266e-05,
|
| 3544 |
+
"loss": 0.0015,
|
| 3545 |
+
"step": 5050
|
| 3546 |
+
},
|
| 3547 |
+
{
|
| 3548 |
+
"epoch": 4.157764995891537,
|
| 3549 |
+
"grad_norm": 0.10125401275989966,
|
| 3550 |
+
"learning_rate": 5.5296545987338423e-05,
|
| 3551 |
+
"loss": 0.0013,
|
| 3552 |
+
"step": 5060
|
| 3553 |
+
},
|
| 3554 |
+
{
|
| 3555 |
+
"epoch": 4.165981922760888,
|
| 3556 |
+
"grad_norm": 0.07335899116635244,
|
| 3557 |
+
"learning_rate": 5.511824729014936e-05,
|
| 3558 |
+
"loss": 0.0013,
|
| 3559 |
+
"step": 5070
|
| 3560 |
+
},
|
| 3561 |
+
{
|
| 3562 |
+
"epoch": 4.174198849630239,
|
| 3563 |
+
"grad_norm": 0.10446876517102897,
|
| 3564 |
+
"learning_rate": 5.493988279475504e-05,
|
| 3565 |
+
"loss": 0.0013,
|
| 3566 |
+
"step": 5080
|
| 3567 |
+
},
|
| 3568 |
+
{
|
| 3569 |
+
"epoch": 4.182415776499589,
|
| 3570 |
+
"grad_norm": 0.08071845171006475,
|
| 3571 |
+
"learning_rate": 5.476145479414039e-05,
|
| 3572 |
+
"loss": 0.0014,
|
| 3573 |
+
"step": 5090
|
| 3574 |
+
},
|
| 3575 |
+
{
|
| 3576 |
+
"epoch": 4.19063270336894,
|
| 3577 |
+
"grad_norm": 0.1474918516996669,
|
| 3578 |
+
"learning_rate": 5.458296558210665e-05,
|
| 3579 |
+
"loss": 0.0012,
|
| 3580 |
+
"step": 5100
|
| 3581 |
+
},
|
| 3582 |
+
{
|
| 3583 |
+
"epoch": 4.198849630238291,
|
| 3584 |
+
"grad_norm": 0.12475171586161964,
|
| 3585 |
+
"learning_rate": 5.440441745324205e-05,
|
| 3586 |
+
"loss": 0.0015,
|
| 3587 |
+
"step": 5110
|
| 3588 |
+
},
|
| 3589 |
+
{
|
| 3590 |
+
"epoch": 4.207066557107642,
|
| 3591 |
+
"grad_norm": 0.0820426621270053,
|
| 3592 |
+
"learning_rate": 5.422581270289219e-05,
|
| 3593 |
+
"loss": 0.0014,
|
| 3594 |
+
"step": 5120
|
| 3595 |
+
},
|
| 3596 |
+
{
|
| 3597 |
+
"epoch": 4.215283483976993,
|
| 3598 |
+
"grad_norm": 0.09495918927284339,
|
| 3599 |
+
"learning_rate": 5.4047153627130554e-05,
|
| 3600 |
+
"loss": 0.0016,
|
| 3601 |
+
"step": 5130
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"epoch": 4.223500410846343,
|
| 3605 |
+
"grad_norm": 0.13207275196758664,
|
| 3606 |
+
"learning_rate": 5.386844252272909e-05,
|
| 3607 |
+
"loss": 0.0016,
|
| 3608 |
+
"step": 5140
|
| 3609 |
+
},
|
| 3610 |
+
{
|
| 3611 |
+
"epoch": 4.231717337715694,
|
| 3612 |
+
"grad_norm": 0.11877585926655844,
|
| 3613 |
+
"learning_rate": 5.368968168712849e-05,
|
| 3614 |
+
"loss": 0.0013,
|
| 3615 |
+
"step": 5150
|
| 3616 |
+
},
|
| 3617 |
+
{
|
| 3618 |
+
"epoch": 4.239934264585045,
|
| 3619 |
+
"grad_norm": 0.1070403964394453,
|
| 3620 |
+
"learning_rate": 5.351087341840888e-05,
|
| 3621 |
+
"loss": 0.0014,
|
| 3622 |
+
"step": 5160
|
| 3623 |
+
},
|
| 3624 |
+
{
|
| 3625 |
+
"epoch": 4.248151191454396,
|
| 3626 |
+
"grad_norm": 0.09820763963218146,
|
| 3627 |
+
"learning_rate": 5.3332020015260096e-05,
|
| 3628 |
+
"loss": 0.0014,
|
| 3629 |
+
"step": 5170
|
| 3630 |
+
},
|
| 3631 |
+
{
|
| 3632 |
+
"epoch": 4.2563681183237465,
|
| 3633 |
+
"grad_norm": 0.123107123751968,
|
| 3634 |
+
"learning_rate": 5.315312377695226e-05,
|
| 3635 |
+
"loss": 0.0014,
|
| 3636 |
+
"step": 5180
|
| 3637 |
+
},
|
| 3638 |
+
{
|
| 3639 |
+
"epoch": 4.264585045193098,
|
| 3640 |
+
"grad_norm": 0.11243233131026976,
|
| 3641 |
+
"learning_rate": 5.297418700330612e-05,
|
| 3642 |
+
"loss": 0.0013,
|
| 3643 |
+
"step": 5190
|
| 3644 |
+
},
|
| 3645 |
+
{
|
| 3646 |
+
"epoch": 4.272801972062449,
|
| 3647 |
+
"grad_norm": 0.11340801246445088,
|
| 3648 |
+
"learning_rate": 5.279521199466356e-05,
|
| 3649 |
+
"loss": 0.0014,
|
| 3650 |
+
"step": 5200
|
| 3651 |
+
},
|
| 3652 |
+
{
|
| 3653 |
+
"epoch": 4.2810188989318,
|
| 3654 |
+
"grad_norm": 0.12876858200295097,
|
| 3655 |
+
"learning_rate": 5.2616201051857975e-05,
|
| 3656 |
+
"loss": 0.0017,
|
| 3657 |
+
"step": 5210
|
| 3658 |
+
},
|
| 3659 |
+
{
|
| 3660 |
+
"epoch": 4.289235825801151,
|
| 3661 |
+
"grad_norm": 0.12108404998878051,
|
| 3662 |
+
"learning_rate": 5.243715647618473e-05,
|
| 3663 |
+
"loss": 0.0015,
|
| 3664 |
+
"step": 5220
|
| 3665 |
+
},
|
| 3666 |
+
{
|
| 3667 |
+
"epoch": 4.297452752670501,
|
| 3668 |
+
"grad_norm": 0.11813429448747585,
|
| 3669 |
+
"learning_rate": 5.225808056937156e-05,
|
| 3670 |
+
"loss": 0.0014,
|
| 3671 |
+
"step": 5230
|
| 3672 |
+
},
|
| 3673 |
+
{
|
| 3674 |
+
"epoch": 4.305669679539852,
|
| 3675 |
+
"grad_norm": 0.09506864360769514,
|
| 3676 |
+
"learning_rate": 5.2078975633548985e-05,
|
| 3677 |
+
"loss": 0.0016,
|
| 3678 |
+
"step": 5240
|
| 3679 |
+
},
|
| 3680 |
+
{
|
| 3681 |
+
"epoch": 4.313886606409203,
|
| 3682 |
+
"grad_norm": 0.10922426672437859,
|
| 3683 |
+
"learning_rate": 5.189984397122071e-05,
|
| 3684 |
+
"loss": 0.0014,
|
| 3685 |
+
"step": 5250
|
| 3686 |
+
},
|
| 3687 |
+
{
|
| 3688 |
+
"epoch": 4.322103533278554,
|
| 3689 |
+
"grad_norm": 0.08415452021791223,
|
| 3690 |
+
"learning_rate": 5.1720687885233996e-05,
|
| 3691 |
+
"loss": 0.0013,
|
| 3692 |
+
"step": 5260
|
| 3693 |
+
},
|
| 3694 |
+
{
|
| 3695 |
+
"epoch": 4.330320460147905,
|
| 3696 |
+
"grad_norm": 0.13308743766878417,
|
| 3697 |
+
"learning_rate": 5.1541509678750146e-05,
|
| 3698 |
+
"loss": 0.0012,
|
| 3699 |
+
"step": 5270
|
| 3700 |
+
},
|
| 3701 |
+
{
|
| 3702 |
+
"epoch": 4.338537387017255,
|
| 3703 |
+
"grad_norm": 0.11579490968423159,
|
| 3704 |
+
"learning_rate": 5.1362311655214765e-05,
|
| 3705 |
+
"loss": 0.0013,
|
| 3706 |
+
"step": 5280
|
| 3707 |
+
},
|
| 3708 |
+
{
|
| 3709 |
+
"epoch": 4.346754313886606,
|
| 3710 |
+
"grad_norm": 0.07822010831326952,
|
| 3711 |
+
"learning_rate": 5.118309611832828e-05,
|
| 3712 |
+
"loss": 0.0013,
|
| 3713 |
+
"step": 5290
|
| 3714 |
+
},
|
| 3715 |
+
{
|
| 3716 |
+
"epoch": 4.354971240755957,
|
| 3717 |
+
"grad_norm": 0.10449415892810263,
|
| 3718 |
+
"learning_rate": 5.100386537201621e-05,
|
| 3719 |
+
"loss": 0.0013,
|
| 3720 |
+
"step": 5300
|
| 3721 |
+
},
|
| 3722 |
+
{
|
| 3723 |
+
"epoch": 4.363188167625308,
|
| 3724 |
+
"grad_norm": 0.06674402491842424,
|
| 3725 |
+
"learning_rate": 5.082462172039966e-05,
|
| 3726 |
+
"loss": 0.0014,
|
| 3727 |
+
"step": 5310
|
| 3728 |
+
},
|
| 3729 |
+
{
|
| 3730 |
+
"epoch": 4.3714050944946585,
|
| 3731 |
+
"grad_norm": 0.10034554995156804,
|
| 3732 |
+
"learning_rate": 5.064536746776558e-05,
|
| 3733 |
+
"loss": 0.0015,
|
| 3734 |
+
"step": 5320
|
| 3735 |
+
},
|
| 3736 |
+
{
|
| 3737 |
+
"epoch": 4.37962202136401,
|
| 3738 |
+
"grad_norm": 0.09959820654455684,
|
| 3739 |
+
"learning_rate": 5.046610491853724e-05,
|
| 3740 |
+
"loss": 0.0014,
|
| 3741 |
+
"step": 5330
|
| 3742 |
+
},
|
| 3743 |
+
{
|
| 3744 |
+
"epoch": 4.387838948233361,
|
| 3745 |
+
"grad_norm": 0.10714573521331806,
|
| 3746 |
+
"learning_rate": 5.028683637724455e-05,
|
| 3747 |
+
"loss": 0.0014,
|
| 3748 |
+
"step": 5340
|
| 3749 |
+
},
|
| 3750 |
+
{
|
| 3751 |
+
"epoch": 4.396055875102712,
|
| 3752 |
+
"grad_norm": 0.11368484767214325,
|
| 3753 |
+
"learning_rate": 5.010756414849448e-05,
|
| 3754 |
+
"loss": 0.0014,
|
| 3755 |
+
"step": 5350
|
| 3756 |
+
},
|
| 3757 |
+
{
|
| 3758 |
+
"epoch": 4.404272801972063,
|
| 3759 |
+
"grad_norm": 0.15147407334313445,
|
| 3760 |
+
"learning_rate": 4.9928290536941355e-05,
|
| 3761 |
+
"loss": 0.0013,
|
| 3762 |
+
"step": 5360
|
| 3763 |
+
},
|
| 3764 |
+
{
|
| 3765 |
+
"epoch": 4.412489728841413,
|
| 3766 |
+
"grad_norm": 0.1304243052575926,
|
| 3767 |
+
"learning_rate": 4.9749017847257324e-05,
|
| 3768 |
+
"loss": 0.0014,
|
| 3769 |
+
"step": 5370
|
| 3770 |
+
},
|
| 3771 |
+
{
|
| 3772 |
+
"epoch": 4.420706655710764,
|
| 3773 |
+
"grad_norm": 0.1569598323839212,
|
| 3774 |
+
"learning_rate": 4.956974838410265e-05,
|
| 3775 |
+
"loss": 0.0014,
|
| 3776 |
+
"step": 5380
|
| 3777 |
+
},
|
| 3778 |
+
{
|
| 3779 |
+
"epoch": 4.428923582580115,
|
| 3780 |
+
"grad_norm": 0.11556205783668044,
|
| 3781 |
+
"learning_rate": 4.939048445209617e-05,
|
| 3782 |
+
"loss": 0.0013,
|
| 3783 |
+
"step": 5390
|
| 3784 |
+
},
|
| 3785 |
+
{
|
| 3786 |
+
"epoch": 4.437140509449466,
|
| 3787 |
+
"grad_norm": 0.16036602462798313,
|
| 3788 |
+
"learning_rate": 4.921122835578554e-05,
|
| 3789 |
+
"loss": 0.0014,
|
| 3790 |
+
"step": 5400
|
| 3791 |
+
}
|
| 3792 |
+
],
|
| 3793 |
+
"logging_steps": 10,
|
| 3794 |
+
"max_steps": 9736,
|
| 3795 |
+
"num_input_tokens_seen": 0,
|
| 3796 |
+
"num_train_epochs": 8,
|
| 3797 |
+
"save_steps": 1800,
|
| 3798 |
+
"stateful_callbacks": {
|
| 3799 |
+
"TrainerControl": {
|
| 3800 |
+
"args": {
|
| 3801 |
+
"should_epoch_stop": false,
|
| 3802 |
+
"should_evaluate": false,
|
| 3803 |
+
"should_log": false,
|
| 3804 |
+
"should_save": true,
|
| 3805 |
+
"should_training_stop": false
|
| 3806 |
+
},
|
| 3807 |
+
"attributes": {}
|
| 3808 |
+
}
|
| 3809 |
+
},
|
| 3810 |
+
"total_flos": 354942328504320.0,
|
| 3811 |
+
"train_batch_size": 128,
|
| 3812 |
+
"trial_name": null,
|
| 3813 |
+
"trial_params": null
|
| 3814 |
+
}
|