Upload checkpoint-5000/trainer_state.json with huggingface_hub
Browse files- checkpoint-5000/trainer_state.json +3534 -0
checkpoint-5000/trainer_state.json
ADDED
|
@@ -0,0 +1,3534 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.1902003570720094,
|
| 6 |
+
"eval_steps": 50,
|
| 7 |
+
"global_step": 5000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0023804800634794686,
|
| 14 |
+
"grad_norm": 43.0,
|
| 15 |
+
"learning_rate": 1.6875e-07,
|
| 16 |
+
"loss": 1.6633,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.004760960126958937,
|
| 21 |
+
"grad_norm": 63.5,
|
| 22 |
+
"learning_rate": 3.5625000000000003e-07,
|
| 23 |
+
"loss": 1.6375,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.007141440190438405,
|
| 28 |
+
"grad_norm": 49.25,
|
| 29 |
+
"learning_rate": 5.4375e-07,
|
| 30 |
+
"loss": 1.6325,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.009521920253917874,
|
| 35 |
+
"grad_norm": 43.25,
|
| 36 |
+
"learning_rate": 7.312500000000001e-07,
|
| 37 |
+
"loss": 1.6621,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.011902400317397342,
|
| 42 |
+
"grad_norm": 39.75,
|
| 43 |
+
"learning_rate": 9.1875e-07,
|
| 44 |
+
"loss": 1.5987,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.01428288038087681,
|
| 49 |
+
"grad_norm": 35.25,
|
| 50 |
+
"learning_rate": 1.10625e-06,
|
| 51 |
+
"loss": 1.6187,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.01666336044435628,
|
| 56 |
+
"grad_norm": 37.25,
|
| 57 |
+
"learning_rate": 1.29375e-06,
|
| 58 |
+
"loss": 1.5005,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.01904384050783575,
|
| 63 |
+
"grad_norm": 23.875,
|
| 64 |
+
"learning_rate": 1.48125e-06,
|
| 65 |
+
"loss": 1.4614,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.021424320571315214,
|
| 70 |
+
"grad_norm": 15.5625,
|
| 71 |
+
"learning_rate": 1.6687500000000001e-06,
|
| 72 |
+
"loss": 1.3327,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.023804800634794683,
|
| 77 |
+
"grad_norm": 17.0,
|
| 78 |
+
"learning_rate": 1.85625e-06,
|
| 79 |
+
"loss": 1.3133,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.026185280698274152,
|
| 84 |
+
"grad_norm": 14.0625,
|
| 85 |
+
"learning_rate": 2.04375e-06,
|
| 86 |
+
"loss": 1.2393,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.02856576076175362,
|
| 91 |
+
"grad_norm": 10.6875,
|
| 92 |
+
"learning_rate": 2.23125e-06,
|
| 93 |
+
"loss": 1.2397,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.03094624082523309,
|
| 98 |
+
"grad_norm": 8.9375,
|
| 99 |
+
"learning_rate": 2.41875e-06,
|
| 100 |
+
"loss": 1.1328,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.03332672088871256,
|
| 105 |
+
"grad_norm": 7.75,
|
| 106 |
+
"learning_rate": 2.60625e-06,
|
| 107 |
+
"loss": 1.1205,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.03570720095219203,
|
| 112 |
+
"grad_norm": 10.3125,
|
| 113 |
+
"learning_rate": 2.79375e-06,
|
| 114 |
+
"loss": 1.0921,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.0380876810156715,
|
| 119 |
+
"grad_norm": 7.90625,
|
| 120 |
+
"learning_rate": 2.98125e-06,
|
| 121 |
+
"loss": 1.0515,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.04046816107915096,
|
| 126 |
+
"grad_norm": 6.90625,
|
| 127 |
+
"learning_rate": 3.16875e-06,
|
| 128 |
+
"loss": 1.0178,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.04284864114263043,
|
| 133 |
+
"grad_norm": 7.78125,
|
| 134 |
+
"learning_rate": 3.3562500000000003e-06,
|
| 135 |
+
"loss": 0.9527,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.0452291212061099,
|
| 140 |
+
"grad_norm": 9.8125,
|
| 141 |
+
"learning_rate": 3.5437499999999997e-06,
|
| 142 |
+
"loss": 0.9037,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.047609601269589366,
|
| 147 |
+
"grad_norm": 7.03125,
|
| 148 |
+
"learning_rate": 3.73125e-06,
|
| 149 |
+
"loss": 0.8909,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.049990081333068835,
|
| 154 |
+
"grad_norm": 7.34375,
|
| 155 |
+
"learning_rate": 3.9187499999999995e-06,
|
| 156 |
+
"loss": 0.8759,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.052370561396548304,
|
| 161 |
+
"grad_norm": 8.9375,
|
| 162 |
+
"learning_rate": 4.10625e-06,
|
| 163 |
+
"loss": 0.8669,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.05475104146002777,
|
| 168 |
+
"grad_norm": 9.5625,
|
| 169 |
+
"learning_rate": 4.29375e-06,
|
| 170 |
+
"loss": 0.8139,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.05713152152350724,
|
| 175 |
+
"grad_norm": 6.0625,
|
| 176 |
+
"learning_rate": 4.4812500000000005e-06,
|
| 177 |
+
"loss": 0.813,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.05951200158698671,
|
| 182 |
+
"grad_norm": 9.875,
|
| 183 |
+
"learning_rate": 4.668750000000001e-06,
|
| 184 |
+
"loss": 0.7689,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.06189248165046618,
|
| 189 |
+
"grad_norm": 40.0,
|
| 190 |
+
"learning_rate": 4.8562499999999995e-06,
|
| 191 |
+
"loss": 0.7372,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.06427296171394564,
|
| 196 |
+
"grad_norm": 7.34375,
|
| 197 |
+
"learning_rate": 5.04375e-06,
|
| 198 |
+
"loss": 0.7333,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.06665344177742512,
|
| 203 |
+
"grad_norm": 7.15625,
|
| 204 |
+
"learning_rate": 5.23125e-06,
|
| 205 |
+
"loss": 0.706,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.06903392184090458,
|
| 210 |
+
"grad_norm": 10.8125,
|
| 211 |
+
"learning_rate": 5.41875e-06,
|
| 212 |
+
"loss": 0.6806,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.07141440190438406,
|
| 217 |
+
"grad_norm": 10.75,
|
| 218 |
+
"learning_rate": 5.606250000000001e-06,
|
| 219 |
+
"loss": 0.6685,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.07379488196786352,
|
| 224 |
+
"grad_norm": 7.59375,
|
| 225 |
+
"learning_rate": 5.79375e-06,
|
| 226 |
+
"loss": 0.6782,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.076175362031343,
|
| 231 |
+
"grad_norm": 10.625,
|
| 232 |
+
"learning_rate": 5.98125e-06,
|
| 233 |
+
"loss": 0.6117,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.07855584209482246,
|
| 238 |
+
"grad_norm": 11.0625,
|
| 239 |
+
"learning_rate": 6.16875e-06,
|
| 240 |
+
"loss": 0.6061,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.08093632215830192,
|
| 245 |
+
"grad_norm": 11.125,
|
| 246 |
+
"learning_rate": 6.35625e-06,
|
| 247 |
+
"loss": 0.5569,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.0833168022217814,
|
| 252 |
+
"grad_norm": 12.375,
|
| 253 |
+
"learning_rate": 6.543750000000001e-06,
|
| 254 |
+
"loss": 0.5565,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.08569728228526086,
|
| 259 |
+
"grad_norm": 9.875,
|
| 260 |
+
"learning_rate": 6.73125e-06,
|
| 261 |
+
"loss": 0.5461,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.08807776234874033,
|
| 266 |
+
"grad_norm": 13.0625,
|
| 267 |
+
"learning_rate": 6.9187500000000005e-06,
|
| 268 |
+
"loss": 0.5104,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.0904582424122198,
|
| 273 |
+
"grad_norm": 12.375,
|
| 274 |
+
"learning_rate": 7.10625e-06,
|
| 275 |
+
"loss": 0.4818,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.09283872247569927,
|
| 280 |
+
"grad_norm": 12.75,
|
| 281 |
+
"learning_rate": 7.29375e-06,
|
| 282 |
+
"loss": 0.4644,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.09521920253917873,
|
| 287 |
+
"grad_norm": 23.375,
|
| 288 |
+
"learning_rate": 7.481250000000001e-06,
|
| 289 |
+
"loss": 0.4571,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.09759968260265821,
|
| 294 |
+
"grad_norm": 11.0625,
|
| 295 |
+
"learning_rate": 7.66875e-06,
|
| 296 |
+
"loss": 0.4104,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.09998016266613767,
|
| 301 |
+
"grad_norm": 12.125,
|
| 302 |
+
"learning_rate": 7.856250000000001e-06,
|
| 303 |
+
"loss": 0.3985,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.10236064272961715,
|
| 308 |
+
"grad_norm": 16.0,
|
| 309 |
+
"learning_rate": 8.04375e-06,
|
| 310 |
+
"loss": 0.3792,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.10474112279309661,
|
| 315 |
+
"grad_norm": 15.3125,
|
| 316 |
+
"learning_rate": 8.23125e-06,
|
| 317 |
+
"loss": 0.3884,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.10712160285657607,
|
| 322 |
+
"grad_norm": 13.4375,
|
| 323 |
+
"learning_rate": 8.418750000000001e-06,
|
| 324 |
+
"loss": 0.3524,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.10950208292005555,
|
| 329 |
+
"grad_norm": 10.625,
|
| 330 |
+
"learning_rate": 8.606249999999999e-06,
|
| 331 |
+
"loss": 0.3237,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.11188256298353501,
|
| 336 |
+
"grad_norm": 13.5,
|
| 337 |
+
"learning_rate": 8.79375e-06,
|
| 338 |
+
"loss": 0.3263,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.11426304304701448,
|
| 343 |
+
"grad_norm": 11.5625,
|
| 344 |
+
"learning_rate": 8.98125e-06,
|
| 345 |
+
"loss": 0.2895,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.11664352311049395,
|
| 350 |
+
"grad_norm": 13.75,
|
| 351 |
+
"learning_rate": 9.16875e-06,
|
| 352 |
+
"loss": 0.2743,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.11902400317397342,
|
| 357 |
+
"grad_norm": 8.75,
|
| 358 |
+
"learning_rate": 9.35625e-06,
|
| 359 |
+
"loss": 0.2516,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.12140448323745288,
|
| 364 |
+
"grad_norm": 13.4375,
|
| 365 |
+
"learning_rate": 9.54375e-06,
|
| 366 |
+
"loss": 0.2555,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.12378496330093236,
|
| 371 |
+
"grad_norm": 16.125,
|
| 372 |
+
"learning_rate": 9.731250000000001e-06,
|
| 373 |
+
"loss": 0.2468,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.12616544336441182,
|
| 378 |
+
"grad_norm": 18.125,
|
| 379 |
+
"learning_rate": 9.91875e-06,
|
| 380 |
+
"loss": 0.2166,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.12854592342789128,
|
| 385 |
+
"grad_norm": 13.5625,
|
| 386 |
+
"learning_rate": 1.010625e-05,
|
| 387 |
+
"loss": 0.2262,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.13092640349137075,
|
| 392 |
+
"grad_norm": 16.625,
|
| 393 |
+
"learning_rate": 1.0293750000000001e-05,
|
| 394 |
+
"loss": 0.2049,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.13330688355485024,
|
| 399 |
+
"grad_norm": 10.8125,
|
| 400 |
+
"learning_rate": 1.048125e-05,
|
| 401 |
+
"loss": 0.1753,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.1356873636183297,
|
| 406 |
+
"grad_norm": 9.0625,
|
| 407 |
+
"learning_rate": 1.0668750000000002e-05,
|
| 408 |
+
"loss": 0.1566,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.13806784368180916,
|
| 413 |
+
"grad_norm": 10.6875,
|
| 414 |
+
"learning_rate": 1.085625e-05,
|
| 415 |
+
"loss": 0.19,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.14044832374528862,
|
| 420 |
+
"grad_norm": 8.8125,
|
| 421 |
+
"learning_rate": 1.1043749999999999e-05,
|
| 422 |
+
"loss": 0.1594,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.1428288038087681,
|
| 427 |
+
"grad_norm": 10.875,
|
| 428 |
+
"learning_rate": 1.123125e-05,
|
| 429 |
+
"loss": 0.1735,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.14520928387224757,
|
| 434 |
+
"grad_norm": 8.625,
|
| 435 |
+
"learning_rate": 1.141875e-05,
|
| 436 |
+
"loss": 0.1394,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.14758976393572704,
|
| 441 |
+
"grad_norm": 12.0625,
|
| 442 |
+
"learning_rate": 1.1606250000000001e-05,
|
| 443 |
+
"loss": 0.1358,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.1499702439992065,
|
| 448 |
+
"grad_norm": 8.3125,
|
| 449 |
+
"learning_rate": 1.179375e-05,
|
| 450 |
+
"loss": 0.1099,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.152350724062686,
|
| 455 |
+
"grad_norm": 8.25,
|
| 456 |
+
"learning_rate": 1.198125e-05,
|
| 457 |
+
"loss": 0.1155,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.15473120412616545,
|
| 462 |
+
"grad_norm": 12.5,
|
| 463 |
+
"learning_rate": 1.2168750000000001e-05,
|
| 464 |
+
"loss": 0.1133,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.1571116841896449,
|
| 469 |
+
"grad_norm": 11.125,
|
| 470 |
+
"learning_rate": 1.235625e-05,
|
| 471 |
+
"loss": 0.0946,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.15949216425312437,
|
| 476 |
+
"grad_norm": 9.875,
|
| 477 |
+
"learning_rate": 1.2543750000000002e-05,
|
| 478 |
+
"loss": 0.1233,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.16187264431660384,
|
| 483 |
+
"grad_norm": 10.8125,
|
| 484 |
+
"learning_rate": 1.2731250000000001e-05,
|
| 485 |
+
"loss": 0.1115,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.16425312438008333,
|
| 490 |
+
"grad_norm": 7.9375,
|
| 491 |
+
"learning_rate": 1.2918749999999999e-05,
|
| 492 |
+
"loss": 0.1002,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.1666336044435628,
|
| 497 |
+
"grad_norm": 9.1875,
|
| 498 |
+
"learning_rate": 1.310625e-05,
|
| 499 |
+
"loss": 0.0825,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.16901408450704225,
|
| 504 |
+
"grad_norm": 22.0,
|
| 505 |
+
"learning_rate": 1.329375e-05,
|
| 506 |
+
"loss": 0.087,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.1713945645705217,
|
| 511 |
+
"grad_norm": 6.84375,
|
| 512 |
+
"learning_rate": 1.348125e-05,
|
| 513 |
+
"loss": 0.0843,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.1737750446340012,
|
| 518 |
+
"grad_norm": 10.625,
|
| 519 |
+
"learning_rate": 1.366875e-05,
|
| 520 |
+
"loss": 0.068,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.17615552469748066,
|
| 525 |
+
"grad_norm": 6.75,
|
| 526 |
+
"learning_rate": 1.385625e-05,
|
| 527 |
+
"loss": 0.0756,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.17853600476096013,
|
| 532 |
+
"grad_norm": 8.25,
|
| 533 |
+
"learning_rate": 1.4043750000000001e-05,
|
| 534 |
+
"loss": 0.0678,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.1809164848244396,
|
| 539 |
+
"grad_norm": 6.71875,
|
| 540 |
+
"learning_rate": 1.423125e-05,
|
| 541 |
+
"loss": 0.069,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.18329696488791905,
|
| 546 |
+
"grad_norm": 7.625,
|
| 547 |
+
"learning_rate": 1.4418750000000001e-05,
|
| 548 |
+
"loss": 0.0677,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.18567744495139854,
|
| 553 |
+
"grad_norm": 6.03125,
|
| 554 |
+
"learning_rate": 1.4606250000000001e-05,
|
| 555 |
+
"loss": 0.0644,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.188057925014878,
|
| 560 |
+
"grad_norm": 5.40625,
|
| 561 |
+
"learning_rate": 1.479375e-05,
|
| 562 |
+
"loss": 0.0569,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.19043840507835746,
|
| 567 |
+
"grad_norm": 5.625,
|
| 568 |
+
"learning_rate": 1.4981250000000002e-05,
|
| 569 |
+
"loss": 0.043,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.19281888514183693,
|
| 574 |
+
"grad_norm": 6.0,
|
| 575 |
+
"learning_rate": 1.4960305792413996e-05,
|
| 576 |
+
"loss": 0.0619,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.19519936520531642,
|
| 581 |
+
"grad_norm": 7.15625,
|
| 582 |
+
"learning_rate": 1.4916201117318435e-05,
|
| 583 |
+
"loss": 0.057,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.19757984526879588,
|
| 588 |
+
"grad_norm": 7.0625,
|
| 589 |
+
"learning_rate": 1.4872096442222876e-05,
|
| 590 |
+
"loss": 0.0448,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.19996032533227534,
|
| 595 |
+
"grad_norm": 7.25,
|
| 596 |
+
"learning_rate": 1.4827991767127316e-05,
|
| 597 |
+
"loss": 0.0473,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.2023408053957548,
|
| 602 |
+
"grad_norm": 5.28125,
|
| 603 |
+
"learning_rate": 1.4783887092031756e-05,
|
| 604 |
+
"loss": 0.0494,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.2047212854592343,
|
| 609 |
+
"grad_norm": 6.6875,
|
| 610 |
+
"learning_rate": 1.4739782416936196e-05,
|
| 611 |
+
"loss": 0.0503,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.20710176552271375,
|
| 616 |
+
"grad_norm": 5.25,
|
| 617 |
+
"learning_rate": 1.4695677741840636e-05,
|
| 618 |
+
"loss": 0.0387,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.20948224558619322,
|
| 623 |
+
"grad_norm": 10.875,
|
| 624 |
+
"learning_rate": 1.4651573066745075e-05,
|
| 625 |
+
"loss": 0.0414,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.21186272564967268,
|
| 630 |
+
"grad_norm": 6.5625,
|
| 631 |
+
"learning_rate": 1.4607468391649515e-05,
|
| 632 |
+
"loss": 0.0517,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.21424320571315214,
|
| 637 |
+
"grad_norm": 3.9375,
|
| 638 |
+
"learning_rate": 1.4563363716553955e-05,
|
| 639 |
+
"loss": 0.0319,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.21662368577663163,
|
| 644 |
+
"grad_norm": 6.28125,
|
| 645 |
+
"learning_rate": 1.4519259041458395e-05,
|
| 646 |
+
"loss": 0.0413,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.2190041658401111,
|
| 651 |
+
"grad_norm": 4.6875,
|
| 652 |
+
"learning_rate": 1.4475154366362835e-05,
|
| 653 |
+
"loss": 0.0399,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.22138464590359055,
|
| 658 |
+
"grad_norm": 6.28125,
|
| 659 |
+
"learning_rate": 1.4431049691267275e-05,
|
| 660 |
+
"loss": 0.0304,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.22376512596707002,
|
| 665 |
+
"grad_norm": 4.09375,
|
| 666 |
+
"learning_rate": 1.4386945016171715e-05,
|
| 667 |
+
"loss": 0.0295,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.2261456060305495,
|
| 672 |
+
"grad_norm": 5.5,
|
| 673 |
+
"learning_rate": 1.4342840341076155e-05,
|
| 674 |
+
"loss": 0.0329,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.22852608609402897,
|
| 679 |
+
"grad_norm": 4.75,
|
| 680 |
+
"learning_rate": 1.4298735665980595e-05,
|
| 681 |
+
"loss": 0.0313,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.23090656615750843,
|
| 686 |
+
"grad_norm": 3.796875,
|
| 687 |
+
"learning_rate": 1.4254630990885035e-05,
|
| 688 |
+
"loss": 0.0309,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.2332870462209879,
|
| 693 |
+
"grad_norm": 7.875,
|
| 694 |
+
"learning_rate": 1.4210526315789473e-05,
|
| 695 |
+
"loss": 0.0373,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.23566752628446735,
|
| 700 |
+
"grad_norm": 2.8125,
|
| 701 |
+
"learning_rate": 1.4166421640693913e-05,
|
| 702 |
+
"loss": 0.0294,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.23804800634794684,
|
| 707 |
+
"grad_norm": 3.8125,
|
| 708 |
+
"learning_rate": 1.4122316965598353e-05,
|
| 709 |
+
"loss": 0.0266,
|
| 710 |
+
"step": 1000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.2404284864114263,
|
| 714 |
+
"grad_norm": 5.875,
|
| 715 |
+
"learning_rate": 1.4078212290502793e-05,
|
| 716 |
+
"loss": 0.0251,
|
| 717 |
+
"step": 1010
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.24280896647490577,
|
| 721 |
+
"grad_norm": 6.15625,
|
| 722 |
+
"learning_rate": 1.4034107615407233e-05,
|
| 723 |
+
"loss": 0.0216,
|
| 724 |
+
"step": 1020
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.24518944653838523,
|
| 728 |
+
"grad_norm": 3.03125,
|
| 729 |
+
"learning_rate": 1.3990002940311673e-05,
|
| 730 |
+
"loss": 0.0276,
|
| 731 |
+
"step": 1030
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.24756992660186472,
|
| 735 |
+
"grad_norm": 3.0625,
|
| 736 |
+
"learning_rate": 1.3945898265216113e-05,
|
| 737 |
+
"loss": 0.0203,
|
| 738 |
+
"step": 1040
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.24995040666534418,
|
| 742 |
+
"grad_norm": 6.3125,
|
| 743 |
+
"learning_rate": 1.3901793590120553e-05,
|
| 744 |
+
"loss": 0.0209,
|
| 745 |
+
"step": 1050
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.25233088672882364,
|
| 749 |
+
"grad_norm": 4.125,
|
| 750 |
+
"learning_rate": 1.3857688915024993e-05,
|
| 751 |
+
"loss": 0.0189,
|
| 752 |
+
"step": 1060
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.2547113667923031,
|
| 756 |
+
"grad_norm": 2.1875,
|
| 757 |
+
"learning_rate": 1.3813584239929433e-05,
|
| 758 |
+
"loss": 0.027,
|
| 759 |
+
"step": 1070
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.25709184685578257,
|
| 763 |
+
"grad_norm": 2.734375,
|
| 764 |
+
"learning_rate": 1.3769479564833873e-05,
|
| 765 |
+
"loss": 0.0206,
|
| 766 |
+
"step": 1080
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.25947232691926203,
|
| 770 |
+
"grad_norm": 3.125,
|
| 771 |
+
"learning_rate": 1.3725374889738312e-05,
|
| 772 |
+
"loss": 0.028,
|
| 773 |
+
"step": 1090
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.2618528069827415,
|
| 777 |
+
"grad_norm": 3.109375,
|
| 778 |
+
"learning_rate": 1.3681270214642752e-05,
|
| 779 |
+
"loss": 0.0182,
|
| 780 |
+
"step": 1100
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.264233287046221,
|
| 784 |
+
"grad_norm": 2.65625,
|
| 785 |
+
"learning_rate": 1.3637165539547192e-05,
|
| 786 |
+
"loss": 0.0247,
|
| 787 |
+
"step": 1110
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.2666137671097005,
|
| 791 |
+
"grad_norm": 2.46875,
|
| 792 |
+
"learning_rate": 1.3593060864451632e-05,
|
| 793 |
+
"loss": 0.0195,
|
| 794 |
+
"step": 1120
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.26899424717317993,
|
| 798 |
+
"grad_norm": 2.609375,
|
| 799 |
+
"learning_rate": 1.3548956189356072e-05,
|
| 800 |
+
"loss": 0.0185,
|
| 801 |
+
"step": 1130
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.2713747272366594,
|
| 805 |
+
"grad_norm": 2.546875,
|
| 806 |
+
"learning_rate": 1.3504851514260514e-05,
|
| 807 |
+
"loss": 0.0188,
|
| 808 |
+
"step": 1140
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.27375520730013886,
|
| 812 |
+
"grad_norm": 2.53125,
|
| 813 |
+
"learning_rate": 1.3460746839164952e-05,
|
| 814 |
+
"loss": 0.0171,
|
| 815 |
+
"step": 1150
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.2761356873636183,
|
| 819 |
+
"grad_norm": 4.15625,
|
| 820 |
+
"learning_rate": 1.3416642164069392e-05,
|
| 821 |
+
"loss": 0.0208,
|
| 822 |
+
"step": 1160
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.2785161674270978,
|
| 826 |
+
"grad_norm": 3.328125,
|
| 827 |
+
"learning_rate": 1.3372537488973832e-05,
|
| 828 |
+
"loss": 0.0162,
|
| 829 |
+
"step": 1170
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.28089664749057724,
|
| 833 |
+
"grad_norm": 3.984375,
|
| 834 |
+
"learning_rate": 1.3328432813878272e-05,
|
| 835 |
+
"loss": 0.0196,
|
| 836 |
+
"step": 1180
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.28327712755405676,
|
| 840 |
+
"grad_norm": 3.1875,
|
| 841 |
+
"learning_rate": 1.3284328138782712e-05,
|
| 842 |
+
"loss": 0.0166,
|
| 843 |
+
"step": 1190
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.2856576076175362,
|
| 847 |
+
"grad_norm": 2.28125,
|
| 848 |
+
"learning_rate": 1.324022346368715e-05,
|
| 849 |
+
"loss": 0.016,
|
| 850 |
+
"step": 1200
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.2880380876810157,
|
| 854 |
+
"grad_norm": 3.03125,
|
| 855 |
+
"learning_rate": 1.319611878859159e-05,
|
| 856 |
+
"loss": 0.0148,
|
| 857 |
+
"step": 1210
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.29041856774449515,
|
| 861 |
+
"grad_norm": 1.90625,
|
| 862 |
+
"learning_rate": 1.315201411349603e-05,
|
| 863 |
+
"loss": 0.014,
|
| 864 |
+
"step": 1220
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.2927990478079746,
|
| 868 |
+
"grad_norm": 3.078125,
|
| 869 |
+
"learning_rate": 1.310790943840047e-05,
|
| 870 |
+
"loss": 0.0129,
|
| 871 |
+
"step": 1230
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.2951795278714541,
|
| 875 |
+
"grad_norm": 2.234375,
|
| 876 |
+
"learning_rate": 1.306380476330491e-05,
|
| 877 |
+
"loss": 0.0185,
|
| 878 |
+
"step": 1240
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.29756000793493353,
|
| 882 |
+
"grad_norm": 4.125,
|
| 883 |
+
"learning_rate": 1.301970008820935e-05,
|
| 884 |
+
"loss": 0.0134,
|
| 885 |
+
"step": 1250
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.299940487998413,
|
| 889 |
+
"grad_norm": 5.875,
|
| 890 |
+
"learning_rate": 1.297559541311379e-05,
|
| 891 |
+
"loss": 0.0166,
|
| 892 |
+
"step": 1260
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.30232096806189246,
|
| 896 |
+
"grad_norm": 3.140625,
|
| 897 |
+
"learning_rate": 1.293149073801823e-05,
|
| 898 |
+
"loss": 0.0147,
|
| 899 |
+
"step": 1270
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.304701448125372,
|
| 903 |
+
"grad_norm": 15.375,
|
| 904 |
+
"learning_rate": 1.288738606292267e-05,
|
| 905 |
+
"loss": 0.0272,
|
| 906 |
+
"step": 1280
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.30708192818885144,
|
| 910 |
+
"grad_norm": 3.046875,
|
| 911 |
+
"learning_rate": 1.284328138782711e-05,
|
| 912 |
+
"loss": 0.0152,
|
| 913 |
+
"step": 1290
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.3094624082523309,
|
| 917 |
+
"grad_norm": 1.9375,
|
| 918 |
+
"learning_rate": 1.279917671273155e-05,
|
| 919 |
+
"loss": 0.0127,
|
| 920 |
+
"step": 1300
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.31184288831581036,
|
| 924 |
+
"grad_norm": 3.484375,
|
| 925 |
+
"learning_rate": 1.2755072037635989e-05,
|
| 926 |
+
"loss": 0.0117,
|
| 927 |
+
"step": 1310
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.3142233683792898,
|
| 931 |
+
"grad_norm": 8.75,
|
| 932 |
+
"learning_rate": 1.2710967362540429e-05,
|
| 933 |
+
"loss": 0.0152,
|
| 934 |
+
"step": 1320
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.3166038484427693,
|
| 938 |
+
"grad_norm": 2.640625,
|
| 939 |
+
"learning_rate": 1.2666862687444869e-05,
|
| 940 |
+
"loss": 0.0116,
|
| 941 |
+
"step": 1330
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.31898432850624875,
|
| 945 |
+
"grad_norm": 2.203125,
|
| 946 |
+
"learning_rate": 1.2622758012349309e-05,
|
| 947 |
+
"loss": 0.0098,
|
| 948 |
+
"step": 1340
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 0.3213648085697282,
|
| 952 |
+
"grad_norm": 2.59375,
|
| 953 |
+
"learning_rate": 1.257865333725375e-05,
|
| 954 |
+
"loss": 0.0116,
|
| 955 |
+
"step": 1350
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.3237452886332077,
|
| 959 |
+
"grad_norm": 4.21875,
|
| 960 |
+
"learning_rate": 1.2534548662158189e-05,
|
| 961 |
+
"loss": 0.0111,
|
| 962 |
+
"step": 1360
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.3261257686966872,
|
| 966 |
+
"grad_norm": 1.9453125,
|
| 967 |
+
"learning_rate": 1.2490443987062629e-05,
|
| 968 |
+
"loss": 0.0137,
|
| 969 |
+
"step": 1370
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.32850624876016665,
|
| 973 |
+
"grad_norm": 2.546875,
|
| 974 |
+
"learning_rate": 1.2446339311967069e-05,
|
| 975 |
+
"loss": 0.009,
|
| 976 |
+
"step": 1380
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.3308867288236461,
|
| 980 |
+
"grad_norm": 1.5,
|
| 981 |
+
"learning_rate": 1.2402234636871509e-05,
|
| 982 |
+
"loss": 0.0089,
|
| 983 |
+
"step": 1390
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.3332672088871256,
|
| 987 |
+
"grad_norm": 1.828125,
|
| 988 |
+
"learning_rate": 1.2358129961775949e-05,
|
| 989 |
+
"loss": 0.0184,
|
| 990 |
+
"step": 1400
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.33564768895060504,
|
| 994 |
+
"grad_norm": 2.59375,
|
| 995 |
+
"learning_rate": 1.2314025286680387e-05,
|
| 996 |
+
"loss": 0.0115,
|
| 997 |
+
"step": 1410
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.3380281690140845,
|
| 1001 |
+
"grad_norm": 2.828125,
|
| 1002 |
+
"learning_rate": 1.2269920611584827e-05,
|
| 1003 |
+
"loss": 0.0155,
|
| 1004 |
+
"step": 1420
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.34040864907756396,
|
| 1008 |
+
"grad_norm": 3.0625,
|
| 1009 |
+
"learning_rate": 1.2225815936489268e-05,
|
| 1010 |
+
"loss": 0.0147,
|
| 1011 |
+
"step": 1430
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.3427891291410434,
|
| 1015 |
+
"grad_norm": 2.03125,
|
| 1016 |
+
"learning_rate": 1.2181711261393708e-05,
|
| 1017 |
+
"loss": 0.0085,
|
| 1018 |
+
"step": 1440
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.3451696092045229,
|
| 1022 |
+
"grad_norm": 2.84375,
|
| 1023 |
+
"learning_rate": 1.2137606586298148e-05,
|
| 1024 |
+
"loss": 0.0118,
|
| 1025 |
+
"step": 1450
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.3475500892680024,
|
| 1029 |
+
"grad_norm": 2.203125,
|
| 1030 |
+
"learning_rate": 1.209350191120259e-05,
|
| 1031 |
+
"loss": 0.0102,
|
| 1032 |
+
"step": 1460
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.34993056933148187,
|
| 1036 |
+
"grad_norm": 2.359375,
|
| 1037 |
+
"learning_rate": 1.2049397236107028e-05,
|
| 1038 |
+
"loss": 0.0113,
|
| 1039 |
+
"step": 1470
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.35231104939496133,
|
| 1043 |
+
"grad_norm": 2.0,
|
| 1044 |
+
"learning_rate": 1.2005292561011468e-05,
|
| 1045 |
+
"loss": 0.0096,
|
| 1046 |
+
"step": 1480
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.3546915294584408,
|
| 1050 |
+
"grad_norm": 2.03125,
|
| 1051 |
+
"learning_rate": 1.1961187885915908e-05,
|
| 1052 |
+
"loss": 0.0103,
|
| 1053 |
+
"step": 1490
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.35707200952192025,
|
| 1057 |
+
"grad_norm": 3.28125,
|
| 1058 |
+
"learning_rate": 1.1917083210820348e-05,
|
| 1059 |
+
"loss": 0.0146,
|
| 1060 |
+
"step": 1500
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.3594524895853997,
|
| 1064 |
+
"grad_norm": 2.53125,
|
| 1065 |
+
"learning_rate": 1.1872978535724788e-05,
|
| 1066 |
+
"loss": 0.0079,
|
| 1067 |
+
"step": 1510
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.3618329696488792,
|
| 1071 |
+
"grad_norm": 2.578125,
|
| 1072 |
+
"learning_rate": 1.1828873860629226e-05,
|
| 1073 |
+
"loss": 0.0103,
|
| 1074 |
+
"step": 1520
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.36421344971235864,
|
| 1078 |
+
"grad_norm": 2.75,
|
| 1079 |
+
"learning_rate": 1.1784769185533666e-05,
|
| 1080 |
+
"loss": 0.0066,
|
| 1081 |
+
"step": 1530
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.3665939297758381,
|
| 1085 |
+
"grad_norm": 2.921875,
|
| 1086 |
+
"learning_rate": 1.1740664510438106e-05,
|
| 1087 |
+
"loss": 0.0081,
|
| 1088 |
+
"step": 1540
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.3689744098393176,
|
| 1092 |
+
"grad_norm": 3.1875,
|
| 1093 |
+
"learning_rate": 1.1696559835342546e-05,
|
| 1094 |
+
"loss": 0.0102,
|
| 1095 |
+
"step": 1550
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.3713548899027971,
|
| 1099 |
+
"grad_norm": 3.6875,
|
| 1100 |
+
"learning_rate": 1.1652455160246988e-05,
|
| 1101 |
+
"loss": 0.0077,
|
| 1102 |
+
"step": 1560
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.37373536996627654,
|
| 1106 |
+
"grad_norm": 1.1953125,
|
| 1107 |
+
"learning_rate": 1.1608350485151428e-05,
|
| 1108 |
+
"loss": 0.0091,
|
| 1109 |
+
"step": 1570
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.376115850029756,
|
| 1113 |
+
"grad_norm": 3.515625,
|
| 1114 |
+
"learning_rate": 1.1564245810055866e-05,
|
| 1115 |
+
"loss": 0.0111,
|
| 1116 |
+
"step": 1580
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.37849633009323547,
|
| 1120 |
+
"grad_norm": 1.890625,
|
| 1121 |
+
"learning_rate": 1.1520141134960306e-05,
|
| 1122 |
+
"loss": 0.0086,
|
| 1123 |
+
"step": 1590
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.38087681015671493,
|
| 1127 |
+
"grad_norm": 2.140625,
|
| 1128 |
+
"learning_rate": 1.1476036459864746e-05,
|
| 1129 |
+
"loss": 0.0082,
|
| 1130 |
+
"step": 1600
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.3832572902201944,
|
| 1134 |
+
"grad_norm": 1.828125,
|
| 1135 |
+
"learning_rate": 1.1431931784769186e-05,
|
| 1136 |
+
"loss": 0.0073,
|
| 1137 |
+
"step": 1610
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.38563777028367385,
|
| 1141 |
+
"grad_norm": 2.921875,
|
| 1142 |
+
"learning_rate": 1.1387827109673626e-05,
|
| 1143 |
+
"loss": 0.0105,
|
| 1144 |
+
"step": 1620
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.3880182503471533,
|
| 1148 |
+
"grad_norm": 1.6953125,
|
| 1149 |
+
"learning_rate": 1.1343722434578065e-05,
|
| 1150 |
+
"loss": 0.0074,
|
| 1151 |
+
"step": 1630
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.39039873041063283,
|
| 1155 |
+
"grad_norm": 4.09375,
|
| 1156 |
+
"learning_rate": 1.1299617759482505e-05,
|
| 1157 |
+
"loss": 0.0062,
|
| 1158 |
+
"step": 1640
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.3927792104741123,
|
| 1162 |
+
"grad_norm": 1.7734375,
|
| 1163 |
+
"learning_rate": 1.1255513084386945e-05,
|
| 1164 |
+
"loss": 0.0076,
|
| 1165 |
+
"step": 1650
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.39515969053759176,
|
| 1169 |
+
"grad_norm": 2.078125,
|
| 1170 |
+
"learning_rate": 1.1211408409291385e-05,
|
| 1171 |
+
"loss": 0.0074,
|
| 1172 |
+
"step": 1660
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.3975401706010712,
|
| 1176 |
+
"grad_norm": 2.359375,
|
| 1177 |
+
"learning_rate": 1.1167303734195826e-05,
|
| 1178 |
+
"loss": 0.0088,
|
| 1179 |
+
"step": 1670
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.3999206506645507,
|
| 1183 |
+
"grad_norm": 2.203125,
|
| 1184 |
+
"learning_rate": 1.1123199059100265e-05,
|
| 1185 |
+
"loss": 0.0083,
|
| 1186 |
+
"step": 1680
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.40230113072803014,
|
| 1190 |
+
"grad_norm": 4.5,
|
| 1191 |
+
"learning_rate": 1.1079094384004705e-05,
|
| 1192 |
+
"loss": 0.0067,
|
| 1193 |
+
"step": 1690
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.4046816107915096,
|
| 1197 |
+
"grad_norm": 1.1953125,
|
| 1198 |
+
"learning_rate": 1.1034989708909145e-05,
|
| 1199 |
+
"loss": 0.0121,
|
| 1200 |
+
"step": 1700
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.40706209085498907,
|
| 1204 |
+
"grad_norm": 2.03125,
|
| 1205 |
+
"learning_rate": 1.0990885033813585e-05,
|
| 1206 |
+
"loss": 0.0061,
|
| 1207 |
+
"step": 1710
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.4094425709184686,
|
| 1211 |
+
"grad_norm": 2.078125,
|
| 1212 |
+
"learning_rate": 1.0946780358718025e-05,
|
| 1213 |
+
"loss": 0.0065,
|
| 1214 |
+
"step": 1720
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.41182305098194805,
|
| 1218 |
+
"grad_norm": 1.8125,
|
| 1219 |
+
"learning_rate": 1.0902675683622465e-05,
|
| 1220 |
+
"loss": 0.0092,
|
| 1221 |
+
"step": 1730
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.4142035310454275,
|
| 1225 |
+
"grad_norm": 1.640625,
|
| 1226 |
+
"learning_rate": 1.0858571008526903e-05,
|
| 1227 |
+
"loss": 0.0066,
|
| 1228 |
+
"step": 1740
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.41658401110890697,
|
| 1232 |
+
"grad_norm": 1.6640625,
|
| 1233 |
+
"learning_rate": 1.0814466333431343e-05,
|
| 1234 |
+
"loss": 0.0062,
|
| 1235 |
+
"step": 1750
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.41896449117238643,
|
| 1239 |
+
"grad_norm": 1.125,
|
| 1240 |
+
"learning_rate": 1.0770361658335783e-05,
|
| 1241 |
+
"loss": 0.0054,
|
| 1242 |
+
"step": 1760
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.4213449712358659,
|
| 1246 |
+
"grad_norm": 2.03125,
|
| 1247 |
+
"learning_rate": 1.0726256983240223e-05,
|
| 1248 |
+
"loss": 0.0077,
|
| 1249 |
+
"step": 1770
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.42372545129934536,
|
| 1253 |
+
"grad_norm": 1.8515625,
|
| 1254 |
+
"learning_rate": 1.0682152308144665e-05,
|
| 1255 |
+
"loss": 0.0073,
|
| 1256 |
+
"step": 1780
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.4261059313628248,
|
| 1260 |
+
"grad_norm": 1.59375,
|
| 1261 |
+
"learning_rate": 1.0638047633049103e-05,
|
| 1262 |
+
"loss": 0.0065,
|
| 1263 |
+
"step": 1790
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.4284864114263043,
|
| 1267 |
+
"grad_norm": 1.9609375,
|
| 1268 |
+
"learning_rate": 1.0593942957953543e-05,
|
| 1269 |
+
"loss": 0.0054,
|
| 1270 |
+
"step": 1800
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.4308668914897838,
|
| 1274 |
+
"grad_norm": 2.53125,
|
| 1275 |
+
"learning_rate": 1.0549838282857983e-05,
|
| 1276 |
+
"loss": 0.0072,
|
| 1277 |
+
"step": 1810
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.43324737155326326,
|
| 1281 |
+
"grad_norm": 1.515625,
|
| 1282 |
+
"learning_rate": 1.0505733607762423e-05,
|
| 1283 |
+
"loss": 0.006,
|
| 1284 |
+
"step": 1820
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.4356278516167427,
|
| 1288 |
+
"grad_norm": 2.46875,
|
| 1289 |
+
"learning_rate": 1.0461628932666863e-05,
|
| 1290 |
+
"loss": 0.0086,
|
| 1291 |
+
"step": 1830
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.4380083316802222,
|
| 1295 |
+
"grad_norm": 1.2265625,
|
| 1296 |
+
"learning_rate": 1.0417524257571302e-05,
|
| 1297 |
+
"loss": 0.0057,
|
| 1298 |
+
"step": 1840
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.44038881174370165,
|
| 1302 |
+
"grad_norm": 1.5390625,
|
| 1303 |
+
"learning_rate": 1.0373419582475742e-05,
|
| 1304 |
+
"loss": 0.0062,
|
| 1305 |
+
"step": 1850
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.4427692918071811,
|
| 1309 |
+
"grad_norm": 0.9140625,
|
| 1310 |
+
"learning_rate": 1.0329314907380182e-05,
|
| 1311 |
+
"loss": 0.0059,
|
| 1312 |
+
"step": 1860
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.44514977187066057,
|
| 1316 |
+
"grad_norm": 2.421875,
|
| 1317 |
+
"learning_rate": 1.0285210232284622e-05,
|
| 1318 |
+
"loss": 0.006,
|
| 1319 |
+
"step": 1870
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.44753025193414003,
|
| 1323 |
+
"grad_norm": 2.515625,
|
| 1324 |
+
"learning_rate": 1.0241105557189063e-05,
|
| 1325 |
+
"loss": 0.0061,
|
| 1326 |
+
"step": 1880
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.4499107319976195,
|
| 1330 |
+
"grad_norm": 2.75,
|
| 1331 |
+
"learning_rate": 1.0197000882093504e-05,
|
| 1332 |
+
"loss": 0.0056,
|
| 1333 |
+
"step": 1890
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.452291212061099,
|
| 1337 |
+
"grad_norm": 2.8125,
|
| 1338 |
+
"learning_rate": 1.0152896206997942e-05,
|
| 1339 |
+
"loss": 0.0057,
|
| 1340 |
+
"step": 1900
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.4546716921245785,
|
| 1344 |
+
"grad_norm": 2.09375,
|
| 1345 |
+
"learning_rate": 1.0108791531902382e-05,
|
| 1346 |
+
"loss": 0.0045,
|
| 1347 |
+
"step": 1910
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.45705217218805794,
|
| 1351 |
+
"grad_norm": 1.2890625,
|
| 1352 |
+
"learning_rate": 1.0064686856806822e-05,
|
| 1353 |
+
"loss": 0.0061,
|
| 1354 |
+
"step": 1920
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.4594326522515374,
|
| 1358 |
+
"grad_norm": 1.5078125,
|
| 1359 |
+
"learning_rate": 1.0020582181711262e-05,
|
| 1360 |
+
"loss": 0.0048,
|
| 1361 |
+
"step": 1930
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.46181313231501686,
|
| 1365 |
+
"grad_norm": 0.9921875,
|
| 1366 |
+
"learning_rate": 9.976477506615702e-06,
|
| 1367 |
+
"loss": 0.0134,
|
| 1368 |
+
"step": 1940
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.4641936123784963,
|
| 1372 |
+
"grad_norm": 3.1875,
|
| 1373 |
+
"learning_rate": 9.93237283152014e-06,
|
| 1374 |
+
"loss": 0.0051,
|
| 1375 |
+
"step": 1950
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.4665740924419758,
|
| 1379 |
+
"grad_norm": 1.59375,
|
| 1380 |
+
"learning_rate": 9.88826815642458e-06,
|
| 1381 |
+
"loss": 0.0049,
|
| 1382 |
+
"step": 1960
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.46895457250545525,
|
| 1386 |
+
"grad_norm": 1.9453125,
|
| 1387 |
+
"learning_rate": 9.84416348132902e-06,
|
| 1388 |
+
"loss": 0.0048,
|
| 1389 |
+
"step": 1970
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.4713350525689347,
|
| 1393 |
+
"grad_norm": 1.7578125,
|
| 1394 |
+
"learning_rate": 9.80005880623346e-06,
|
| 1395 |
+
"loss": 0.0163,
|
| 1396 |
+
"step": 1980
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.4737155326324142,
|
| 1400 |
+
"grad_norm": 1.5234375,
|
| 1401 |
+
"learning_rate": 9.755954131137902e-06,
|
| 1402 |
+
"loss": 0.0057,
|
| 1403 |
+
"step": 1990
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.4760960126958937,
|
| 1407 |
+
"grad_norm": 0.51953125,
|
| 1408 |
+
"learning_rate": 9.711849456042342e-06,
|
| 1409 |
+
"loss": 0.0047,
|
| 1410 |
+
"step": 2000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.47847649275937315,
|
| 1414 |
+
"grad_norm": 2.59375,
|
| 1415 |
+
"learning_rate": 9.66774478094678e-06,
|
| 1416 |
+
"loss": 0.0054,
|
| 1417 |
+
"step": 2010
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.4808569728228526,
|
| 1421 |
+
"grad_norm": 2.484375,
|
| 1422 |
+
"learning_rate": 9.62364010585122e-06,
|
| 1423 |
+
"loss": 0.0049,
|
| 1424 |
+
"step": 2020
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.4832374528863321,
|
| 1428 |
+
"grad_norm": 1.3125,
|
| 1429 |
+
"learning_rate": 9.57953543075566e-06,
|
| 1430 |
+
"loss": 0.0048,
|
| 1431 |
+
"step": 2030
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.48561793294981154,
|
| 1435 |
+
"grad_norm": 1.71875,
|
| 1436 |
+
"learning_rate": 9.5354307556601e-06,
|
| 1437 |
+
"loss": 0.006,
|
| 1438 |
+
"step": 2040
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.487998413013291,
|
| 1442 |
+
"grad_norm": 1.703125,
|
| 1443 |
+
"learning_rate": 9.49132608056454e-06,
|
| 1444 |
+
"loss": 0.0051,
|
| 1445 |
+
"step": 2050
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.49037889307677046,
|
| 1449 |
+
"grad_norm": 1.3359375,
|
| 1450 |
+
"learning_rate": 9.447221405468979e-06,
|
| 1451 |
+
"loss": 0.0045,
|
| 1452 |
+
"step": 2060
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.4927593731402499,
|
| 1456 |
+
"grad_norm": 0.74609375,
|
| 1457 |
+
"learning_rate": 9.403116730373419e-06,
|
| 1458 |
+
"loss": 0.0052,
|
| 1459 |
+
"step": 2070
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.49513985320372944,
|
| 1463 |
+
"grad_norm": 1.28125,
|
| 1464 |
+
"learning_rate": 9.359012055277859e-06,
|
| 1465 |
+
"loss": 0.0042,
|
| 1466 |
+
"step": 2080
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.4975203332672089,
|
| 1470 |
+
"grad_norm": 0.578125,
|
| 1471 |
+
"learning_rate": 9.3149073801823e-06,
|
| 1472 |
+
"loss": 0.0045,
|
| 1473 |
+
"step": 2090
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.49990081333068836,
|
| 1477 |
+
"grad_norm": 1.9140625,
|
| 1478 |
+
"learning_rate": 9.27080270508674e-06,
|
| 1479 |
+
"loss": 0.0063,
|
| 1480 |
+
"step": 2100
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.5022812933941678,
|
| 1484 |
+
"grad_norm": 3.609375,
|
| 1485 |
+
"learning_rate": 9.226698029991179e-06,
|
| 1486 |
+
"loss": 0.0038,
|
| 1487 |
+
"step": 2110
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.5046617734576473,
|
| 1491 |
+
"grad_norm": 0.8359375,
|
| 1492 |
+
"learning_rate": 9.182593354895619e-06,
|
| 1493 |
+
"loss": 0.0045,
|
| 1494 |
+
"step": 2120
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.5070422535211268,
|
| 1498 |
+
"grad_norm": 2.40625,
|
| 1499 |
+
"learning_rate": 9.138488679800059e-06,
|
| 1500 |
+
"loss": 0.0053,
|
| 1501 |
+
"step": 2130
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.5094227335846062,
|
| 1505 |
+
"grad_norm": 0.9140625,
|
| 1506 |
+
"learning_rate": 9.094384004704499e-06,
|
| 1507 |
+
"loss": 0.0054,
|
| 1508 |
+
"step": 2140
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.5118032136480857,
|
| 1512 |
+
"grad_norm": 1.5,
|
| 1513 |
+
"learning_rate": 9.050279329608939e-06,
|
| 1514 |
+
"loss": 0.0035,
|
| 1515 |
+
"step": 2150
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.5141836937115651,
|
| 1519 |
+
"grad_norm": 1.1796875,
|
| 1520 |
+
"learning_rate": 9.006174654513379e-06,
|
| 1521 |
+
"loss": 0.0036,
|
| 1522 |
+
"step": 2160
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.5165641737750446,
|
| 1526 |
+
"grad_norm": 0.9453125,
|
| 1527 |
+
"learning_rate": 8.962069979417817e-06,
|
| 1528 |
+
"loss": 0.0043,
|
| 1529 |
+
"step": 2170
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.5189446538385241,
|
| 1533 |
+
"grad_norm": 1.7265625,
|
| 1534 |
+
"learning_rate": 8.917965304322257e-06,
|
| 1535 |
+
"loss": 0.0119,
|
| 1536 |
+
"step": 2180
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.5213251339020035,
|
| 1540 |
+
"grad_norm": 2.09375,
|
| 1541 |
+
"learning_rate": 8.873860629226697e-06,
|
| 1542 |
+
"loss": 0.0057,
|
| 1543 |
+
"step": 2190
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.523705613965483,
|
| 1547 |
+
"grad_norm": 0.9140625,
|
| 1548 |
+
"learning_rate": 8.82975595413114e-06,
|
| 1549 |
+
"loss": 0.0045,
|
| 1550 |
+
"step": 2200
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.5260860940289626,
|
| 1554 |
+
"grad_norm": 1.203125,
|
| 1555 |
+
"learning_rate": 1.2219810576164167e-05,
|
| 1556 |
+
"loss": 0.005,
|
| 1557 |
+
"step": 2210
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.528466574092442,
|
| 1561 |
+
"grad_norm": 1.4453125,
|
| 1562 |
+
"learning_rate": 1.2200078926598264e-05,
|
| 1563 |
+
"loss": 0.0068,
|
| 1564 |
+
"step": 2220
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 0.5308470541559215,
|
| 1568 |
+
"grad_norm": 1.9609375,
|
| 1569 |
+
"learning_rate": 1.218034727703236e-05,
|
| 1570 |
+
"loss": 0.0048,
|
| 1571 |
+
"step": 2230
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.533227534219401,
|
| 1575 |
+
"grad_norm": 1.6640625,
|
| 1576 |
+
"learning_rate": 1.2160615627466457e-05,
|
| 1577 |
+
"loss": 0.0052,
|
| 1578 |
+
"step": 2240
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.5356080142828804,
|
| 1582 |
+
"grad_norm": 1.65625,
|
| 1583 |
+
"learning_rate": 1.2140883977900552e-05,
|
| 1584 |
+
"loss": 0.0036,
|
| 1585 |
+
"step": 2250
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 0.5379884943463599,
|
| 1589 |
+
"grad_norm": 1.4140625,
|
| 1590 |
+
"learning_rate": 1.2121152328334648e-05,
|
| 1591 |
+
"loss": 0.0046,
|
| 1592 |
+
"step": 2260
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.5403689744098393,
|
| 1596 |
+
"grad_norm": 1.859375,
|
| 1597 |
+
"learning_rate": 1.2101420678768745e-05,
|
| 1598 |
+
"loss": 0.0065,
|
| 1599 |
+
"step": 2270
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.5427494544733188,
|
| 1603 |
+
"grad_norm": 1.640625,
|
| 1604 |
+
"learning_rate": 1.2081689029202841e-05,
|
| 1605 |
+
"loss": 0.0052,
|
| 1606 |
+
"step": 2280
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 0.5451299345367983,
|
| 1610 |
+
"grad_norm": 1.25,
|
| 1611 |
+
"learning_rate": 1.206195737963694e-05,
|
| 1612 |
+
"loss": 0.0053,
|
| 1613 |
+
"step": 2290
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 0.5475104146002777,
|
| 1617 |
+
"grad_norm": 2.125,
|
| 1618 |
+
"learning_rate": 1.2042225730071034e-05,
|
| 1619 |
+
"loss": 0.0058,
|
| 1620 |
+
"step": 2300
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.5498908946637572,
|
| 1624 |
+
"grad_norm": 1.3671875,
|
| 1625 |
+
"learning_rate": 1.2022494080505131e-05,
|
| 1626 |
+
"loss": 0.0047,
|
| 1627 |
+
"step": 2310
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 0.5522713747272366,
|
| 1631 |
+
"grad_norm": 1.0546875,
|
| 1632 |
+
"learning_rate": 1.2002762430939227e-05,
|
| 1633 |
+
"loss": 0.0048,
|
| 1634 |
+
"step": 2320
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.5546518547907161,
|
| 1638 |
+
"grad_norm": 0.431640625,
|
| 1639 |
+
"learning_rate": 1.1983030781373324e-05,
|
| 1640 |
+
"loss": 0.007,
|
| 1641 |
+
"step": 2330
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.5570323348541956,
|
| 1645 |
+
"grad_norm": 1.28125,
|
| 1646 |
+
"learning_rate": 1.1963299131807419e-05,
|
| 1647 |
+
"loss": 0.0043,
|
| 1648 |
+
"step": 2340
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.559412814917675,
|
| 1652 |
+
"grad_norm": 1.34375,
|
| 1653 |
+
"learning_rate": 1.1943567482241515e-05,
|
| 1654 |
+
"loss": 0.0035,
|
| 1655 |
+
"step": 2350
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 0.5617932949811545,
|
| 1659 |
+
"grad_norm": 1.6484375,
|
| 1660 |
+
"learning_rate": 1.1923835832675612e-05,
|
| 1661 |
+
"loss": 0.0056,
|
| 1662 |
+
"step": 2360
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.564173775044634,
|
| 1666 |
+
"grad_norm": 2.359375,
|
| 1667 |
+
"learning_rate": 1.1904104183109708e-05,
|
| 1668 |
+
"loss": 0.0047,
|
| 1669 |
+
"step": 2370
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 0.5665542551081135,
|
| 1673 |
+
"grad_norm": 1.90625,
|
| 1674 |
+
"learning_rate": 1.1884372533543805e-05,
|
| 1675 |
+
"loss": 0.0051,
|
| 1676 |
+
"step": 2380
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.568934735171593,
|
| 1680 |
+
"grad_norm": 1.6796875,
|
| 1681 |
+
"learning_rate": 1.1864640883977901e-05,
|
| 1682 |
+
"loss": 0.0038,
|
| 1683 |
+
"step": 2390
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.5713152152350724,
|
| 1687 |
+
"grad_norm": 1.7265625,
|
| 1688 |
+
"learning_rate": 1.1844909234411998e-05,
|
| 1689 |
+
"loss": 0.0043,
|
| 1690 |
+
"step": 2400
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 0.5736956952985519,
|
| 1694 |
+
"grad_norm": 2.09375,
|
| 1695 |
+
"learning_rate": 1.1825177584846094e-05,
|
| 1696 |
+
"loss": 0.0041,
|
| 1697 |
+
"step": 2410
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 0.5760761753620314,
|
| 1701 |
+
"grad_norm": 1.296875,
|
| 1702 |
+
"learning_rate": 1.1805445935280189e-05,
|
| 1703 |
+
"loss": 0.0064,
|
| 1704 |
+
"step": 2420
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 0.5784566554255108,
|
| 1708 |
+
"grad_norm": 1.296875,
|
| 1709 |
+
"learning_rate": 1.1785714285714286e-05,
|
| 1710 |
+
"loss": 0.0043,
|
| 1711 |
+
"step": 2430
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 0.5808371354889903,
|
| 1715 |
+
"grad_norm": 0.92578125,
|
| 1716 |
+
"learning_rate": 1.1765982636148382e-05,
|
| 1717 |
+
"loss": 0.0042,
|
| 1718 |
+
"step": 2440
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.5832176155524698,
|
| 1722 |
+
"grad_norm": 1.7421875,
|
| 1723 |
+
"learning_rate": 1.1746250986582479e-05,
|
| 1724 |
+
"loss": 0.0053,
|
| 1725 |
+
"step": 2450
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.5855980956159492,
|
| 1729 |
+
"grad_norm": 1.3828125,
|
| 1730 |
+
"learning_rate": 1.1726519337016574e-05,
|
| 1731 |
+
"loss": 0.0041,
|
| 1732 |
+
"step": 2460
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 0.5879785756794287,
|
| 1736 |
+
"grad_norm": 1.1796875,
|
| 1737 |
+
"learning_rate": 1.1706787687450672e-05,
|
| 1738 |
+
"loss": 0.0049,
|
| 1739 |
+
"step": 2470
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 0.5903590557429081,
|
| 1743 |
+
"grad_norm": 0.69921875,
|
| 1744 |
+
"learning_rate": 1.1687056037884768e-05,
|
| 1745 |
+
"loss": 0.0051,
|
| 1746 |
+
"step": 2480
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 0.5927395358063876,
|
| 1750 |
+
"grad_norm": 1.8984375,
|
| 1751 |
+
"learning_rate": 1.1667324388318865e-05,
|
| 1752 |
+
"loss": 0.0045,
|
| 1753 |
+
"step": 2490
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 0.5951200158698671,
|
| 1757 |
+
"grad_norm": 1.359375,
|
| 1758 |
+
"learning_rate": 1.164759273875296e-05,
|
| 1759 |
+
"loss": 0.0043,
|
| 1760 |
+
"step": 2500
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.5975004959333465,
|
| 1764 |
+
"grad_norm": 1.0078125,
|
| 1765 |
+
"learning_rate": 1.1627861089187056e-05,
|
| 1766 |
+
"loss": 0.0037,
|
| 1767 |
+
"step": 2510
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.599880975996826,
|
| 1771 |
+
"grad_norm": 0.400390625,
|
| 1772 |
+
"learning_rate": 1.1608129439621153e-05,
|
| 1773 |
+
"loss": 0.0052,
|
| 1774 |
+
"step": 2520
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 0.6022614560603055,
|
| 1778 |
+
"grad_norm": 1.078125,
|
| 1779 |
+
"learning_rate": 1.1588397790055249e-05,
|
| 1780 |
+
"loss": 0.0055,
|
| 1781 |
+
"step": 2530
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 0.6046419361237849,
|
| 1785 |
+
"grad_norm": 0.87109375,
|
| 1786 |
+
"learning_rate": 1.1568666140489344e-05,
|
| 1787 |
+
"loss": 0.0032,
|
| 1788 |
+
"step": 2540
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 0.6070224161872644,
|
| 1792 |
+
"grad_norm": 2.3125,
|
| 1793 |
+
"learning_rate": 1.154893449092344e-05,
|
| 1794 |
+
"loss": 0.0046,
|
| 1795 |
+
"step": 2550
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 0.609402896250744,
|
| 1799 |
+
"grad_norm": 0.94921875,
|
| 1800 |
+
"learning_rate": 1.1529202841357539e-05,
|
| 1801 |
+
"loss": 0.0039,
|
| 1802 |
+
"step": 2560
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.6117833763142234,
|
| 1806 |
+
"grad_norm": 0.42578125,
|
| 1807 |
+
"learning_rate": 1.1509471191791635e-05,
|
| 1808 |
+
"loss": 0.0032,
|
| 1809 |
+
"step": 2570
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.6141638563777029,
|
| 1813 |
+
"grad_norm": 0.83984375,
|
| 1814 |
+
"learning_rate": 1.148973954222573e-05,
|
| 1815 |
+
"loss": 0.0036,
|
| 1816 |
+
"step": 2580
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 0.6165443364411823,
|
| 1820 |
+
"grad_norm": 1.8046875,
|
| 1821 |
+
"learning_rate": 1.1470007892659827e-05,
|
| 1822 |
+
"loss": 0.0054,
|
| 1823 |
+
"step": 2590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 0.6189248165046618,
|
| 1827 |
+
"grad_norm": 0.78515625,
|
| 1828 |
+
"learning_rate": 1.1450276243093923e-05,
|
| 1829 |
+
"loss": 0.004,
|
| 1830 |
+
"step": 2600
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 0.6213052965681413,
|
| 1834 |
+
"grad_norm": 1.0625,
|
| 1835 |
+
"learning_rate": 1.143054459352802e-05,
|
| 1836 |
+
"loss": 0.0049,
|
| 1837 |
+
"step": 2610
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 0.6236857766316207,
|
| 1841 |
+
"grad_norm": 1.96875,
|
| 1842 |
+
"learning_rate": 1.1410812943962114e-05,
|
| 1843 |
+
"loss": 0.0036,
|
| 1844 |
+
"step": 2620
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.6260662566951002,
|
| 1848 |
+
"grad_norm": 4.03125,
|
| 1849 |
+
"learning_rate": 1.1391081294396211e-05,
|
| 1850 |
+
"loss": 0.0049,
|
| 1851 |
+
"step": 2630
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.6284467367585796,
|
| 1855 |
+
"grad_norm": 1.3203125,
|
| 1856 |
+
"learning_rate": 1.1371349644830309e-05,
|
| 1857 |
+
"loss": 0.0036,
|
| 1858 |
+
"step": 2640
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 0.6308272168220591,
|
| 1862 |
+
"grad_norm": 1.5703125,
|
| 1863 |
+
"learning_rate": 1.1351617995264406e-05,
|
| 1864 |
+
"loss": 0.0031,
|
| 1865 |
+
"step": 2650
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 0.6332076968855386,
|
| 1869 |
+
"grad_norm": 1.078125,
|
| 1870 |
+
"learning_rate": 1.13318863456985e-05,
|
| 1871 |
+
"loss": 0.004,
|
| 1872 |
+
"step": 2660
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 0.635588176949018,
|
| 1876 |
+
"grad_norm": 0.875,
|
| 1877 |
+
"learning_rate": 1.1312154696132597e-05,
|
| 1878 |
+
"loss": 0.0043,
|
| 1879 |
+
"step": 2670
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 0.6379686570124975,
|
| 1883 |
+
"grad_norm": 0.8046875,
|
| 1884 |
+
"learning_rate": 1.1292423046566693e-05,
|
| 1885 |
+
"loss": 0.0044,
|
| 1886 |
+
"step": 2680
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 0.640349137075977,
|
| 1890 |
+
"grad_norm": 1.8984375,
|
| 1891 |
+
"learning_rate": 1.127269139700079e-05,
|
| 1892 |
+
"loss": 0.003,
|
| 1893 |
+
"step": 2690
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 0.6427296171394564,
|
| 1897 |
+
"grad_norm": 1.6484375,
|
| 1898 |
+
"learning_rate": 1.1252959747434885e-05,
|
| 1899 |
+
"loss": 0.0033,
|
| 1900 |
+
"step": 2700
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 0.6451100972029359,
|
| 1904 |
+
"grad_norm": 2.796875,
|
| 1905 |
+
"learning_rate": 1.1233228097868981e-05,
|
| 1906 |
+
"loss": 0.0038,
|
| 1907 |
+
"step": 2710
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 0.6474905772664153,
|
| 1911 |
+
"grad_norm": 1.4453125,
|
| 1912 |
+
"learning_rate": 1.1213496448303078e-05,
|
| 1913 |
+
"loss": 0.0041,
|
| 1914 |
+
"step": 2720
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 0.6498710573298948,
|
| 1918 |
+
"grad_norm": 1.3828125,
|
| 1919 |
+
"learning_rate": 1.1193764798737176e-05,
|
| 1920 |
+
"loss": 0.0034,
|
| 1921 |
+
"step": 2730
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 0.6522515373933744,
|
| 1925 |
+
"grad_norm": 1.2890625,
|
| 1926 |
+
"learning_rate": 1.1174033149171271e-05,
|
| 1927 |
+
"loss": 0.0033,
|
| 1928 |
+
"step": 2740
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 0.6546320174568538,
|
| 1932 |
+
"grad_norm": 1.1953125,
|
| 1933 |
+
"learning_rate": 1.1154301499605367e-05,
|
| 1934 |
+
"loss": 0.0032,
|
| 1935 |
+
"step": 2750
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 0.6570124975203333,
|
| 1939 |
+
"grad_norm": 1.4296875,
|
| 1940 |
+
"learning_rate": 1.1134569850039464e-05,
|
| 1941 |
+
"loss": 0.0037,
|
| 1942 |
+
"step": 2760
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 0.6593929775838128,
|
| 1946 |
+
"grad_norm": 1.6484375,
|
| 1947 |
+
"learning_rate": 1.111483820047356e-05,
|
| 1948 |
+
"loss": 0.0053,
|
| 1949 |
+
"step": 2770
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 0.6617734576472922,
|
| 1953 |
+
"grad_norm": 1.296875,
|
| 1954 |
+
"learning_rate": 1.1095106550907655e-05,
|
| 1955 |
+
"loss": 0.0041,
|
| 1956 |
+
"step": 2780
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 0.6641539377107717,
|
| 1960 |
+
"grad_norm": 1.1796875,
|
| 1961 |
+
"learning_rate": 1.1075374901341752e-05,
|
| 1962 |
+
"loss": 0.0037,
|
| 1963 |
+
"step": 2790
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 0.6665344177742512,
|
| 1967 |
+
"grad_norm": 0.94921875,
|
| 1968 |
+
"learning_rate": 1.1055643251775848e-05,
|
| 1969 |
+
"loss": 0.0126,
|
| 1970 |
+
"step": 2800
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 0.6689148978377306,
|
| 1974 |
+
"grad_norm": 1.96875,
|
| 1975 |
+
"learning_rate": 1.1035911602209945e-05,
|
| 1976 |
+
"loss": 0.003,
|
| 1977 |
+
"step": 2810
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 0.6712953779012101,
|
| 1981 |
+
"grad_norm": 2.15625,
|
| 1982 |
+
"learning_rate": 1.1016179952644041e-05,
|
| 1983 |
+
"loss": 0.0046,
|
| 1984 |
+
"step": 2820
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 0.6736758579646895,
|
| 1988 |
+
"grad_norm": 1.1953125,
|
| 1989 |
+
"learning_rate": 1.0996448303078138e-05,
|
| 1990 |
+
"loss": 0.003,
|
| 1991 |
+
"step": 2830
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 0.676056338028169,
|
| 1995 |
+
"grad_norm": 1.859375,
|
| 1996 |
+
"learning_rate": 1.0976716653512234e-05,
|
| 1997 |
+
"loss": 0.0032,
|
| 1998 |
+
"step": 2840
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 0.6784368180916485,
|
| 2002 |
+
"grad_norm": 1.140625,
|
| 2003 |
+
"learning_rate": 1.095698500394633e-05,
|
| 2004 |
+
"loss": 0.0029,
|
| 2005 |
+
"step": 2850
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 0.6808172981551279,
|
| 2009 |
+
"grad_norm": 1.484375,
|
| 2010 |
+
"learning_rate": 1.0937253354380427e-05,
|
| 2011 |
+
"loss": 0.0026,
|
| 2012 |
+
"step": 2860
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 0.6831977782186074,
|
| 2016 |
+
"grad_norm": 1.828125,
|
| 2017 |
+
"learning_rate": 1.0917521704814522e-05,
|
| 2018 |
+
"loss": 0.0035,
|
| 2019 |
+
"step": 2870
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 0.6855782582820868,
|
| 2023 |
+
"grad_norm": 0.44140625,
|
| 2024 |
+
"learning_rate": 1.0897790055248619e-05,
|
| 2025 |
+
"loss": 0.0031,
|
| 2026 |
+
"step": 2880
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 0.6879587383455663,
|
| 2030 |
+
"grad_norm": 1.15625,
|
| 2031 |
+
"learning_rate": 1.0878058405682715e-05,
|
| 2032 |
+
"loss": 0.0034,
|
| 2033 |
+
"step": 2890
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 0.6903392184090458,
|
| 2037 |
+
"grad_norm": 1.3515625,
|
| 2038 |
+
"learning_rate": 1.0858326756116812e-05,
|
| 2039 |
+
"loss": 0.0048,
|
| 2040 |
+
"step": 2900
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 0.6927196984725253,
|
| 2044 |
+
"grad_norm": 1.09375,
|
| 2045 |
+
"learning_rate": 1.0838595106550908e-05,
|
| 2046 |
+
"loss": 0.0028,
|
| 2047 |
+
"step": 2910
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 0.6951001785360048,
|
| 2051 |
+
"grad_norm": 0.41796875,
|
| 2052 |
+
"learning_rate": 1.0818863456985005e-05,
|
| 2053 |
+
"loss": 0.0028,
|
| 2054 |
+
"step": 2920
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 0.6974806585994843,
|
| 2058 |
+
"grad_norm": 0.40625,
|
| 2059 |
+
"learning_rate": 1.0799131807419101e-05,
|
| 2060 |
+
"loss": 0.0028,
|
| 2061 |
+
"step": 2930
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 0.6998611386629637,
|
| 2065 |
+
"grad_norm": 0.94140625,
|
| 2066 |
+
"learning_rate": 1.0779400157853198e-05,
|
| 2067 |
+
"loss": 0.004,
|
| 2068 |
+
"step": 2940
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 0.7022416187264432,
|
| 2072 |
+
"grad_norm": 1.234375,
|
| 2073 |
+
"learning_rate": 1.0759668508287293e-05,
|
| 2074 |
+
"loss": 0.0108,
|
| 2075 |
+
"step": 2950
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 0.7046220987899227,
|
| 2079 |
+
"grad_norm": 1.109375,
|
| 2080 |
+
"learning_rate": 1.0739936858721389e-05,
|
| 2081 |
+
"loss": 0.0027,
|
| 2082 |
+
"step": 2960
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 0.7070025788534021,
|
| 2086 |
+
"grad_norm": 0.96484375,
|
| 2087 |
+
"learning_rate": 1.0720205209155486e-05,
|
| 2088 |
+
"loss": 0.0039,
|
| 2089 |
+
"step": 2970
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 0.7093830589168816,
|
| 2093 |
+
"grad_norm": 4.5625,
|
| 2094 |
+
"learning_rate": 1.0700473559589582e-05,
|
| 2095 |
+
"loss": 0.0027,
|
| 2096 |
+
"step": 2980
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 0.711763538980361,
|
| 2100 |
+
"grad_norm": 2.1875,
|
| 2101 |
+
"learning_rate": 1.0680741910023677e-05,
|
| 2102 |
+
"loss": 0.0028,
|
| 2103 |
+
"step": 2990
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 0.7141440190438405,
|
| 2107 |
+
"grad_norm": 1.15625,
|
| 2108 |
+
"learning_rate": 1.0661010260457775e-05,
|
| 2109 |
+
"loss": 0.0026,
|
| 2110 |
+
"step": 3000
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 0.71652449910732,
|
| 2114 |
+
"grad_norm": 0.67578125,
|
| 2115 |
+
"learning_rate": 1.0641278610891872e-05,
|
| 2116 |
+
"loss": 0.003,
|
| 2117 |
+
"step": 3010
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 0.7189049791707994,
|
| 2121 |
+
"grad_norm": 2.484375,
|
| 2122 |
+
"learning_rate": 1.0621546961325968e-05,
|
| 2123 |
+
"loss": 0.0044,
|
| 2124 |
+
"step": 3020
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 0.7212854592342789,
|
| 2128 |
+
"grad_norm": 1.203125,
|
| 2129 |
+
"learning_rate": 1.0601815311760063e-05,
|
| 2130 |
+
"loss": 0.0034,
|
| 2131 |
+
"step": 3030
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 0.7236659392977584,
|
| 2135 |
+
"grad_norm": 1.0234375,
|
| 2136 |
+
"learning_rate": 1.058208366219416e-05,
|
| 2137 |
+
"loss": 0.0033,
|
| 2138 |
+
"step": 3040
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 0.7260464193612378,
|
| 2142 |
+
"grad_norm": 1.1328125,
|
| 2143 |
+
"learning_rate": 1.0562352012628256e-05,
|
| 2144 |
+
"loss": 0.0028,
|
| 2145 |
+
"step": 3050
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 0.7284268994247173,
|
| 2149 |
+
"grad_norm": 0.89453125,
|
| 2150 |
+
"learning_rate": 1.0542620363062353e-05,
|
| 2151 |
+
"loss": 0.0033,
|
| 2152 |
+
"step": 3060
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 0.7308073794881967,
|
| 2156 |
+
"grad_norm": 4.5,
|
| 2157 |
+
"learning_rate": 1.0522888713496447e-05,
|
| 2158 |
+
"loss": 0.0036,
|
| 2159 |
+
"step": 3070
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 0.7331878595516762,
|
| 2163 |
+
"grad_norm": 2.0,
|
| 2164 |
+
"learning_rate": 1.0503157063930544e-05,
|
| 2165 |
+
"loss": 0.0031,
|
| 2166 |
+
"step": 3080
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 0.7355683396151558,
|
| 2170 |
+
"grad_norm": 1.0234375,
|
| 2171 |
+
"learning_rate": 1.0483425414364642e-05,
|
| 2172 |
+
"loss": 0.0029,
|
| 2173 |
+
"step": 3090
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 0.7379488196786352,
|
| 2177 |
+
"grad_norm": 0.41796875,
|
| 2178 |
+
"learning_rate": 1.0463693764798739e-05,
|
| 2179 |
+
"loss": 0.0031,
|
| 2180 |
+
"step": 3100
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 0.7403292997421147,
|
| 2184 |
+
"grad_norm": 1.921875,
|
| 2185 |
+
"learning_rate": 1.0443962115232833e-05,
|
| 2186 |
+
"loss": 0.0021,
|
| 2187 |
+
"step": 3110
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 0.7427097798055942,
|
| 2191 |
+
"grad_norm": 1.203125,
|
| 2192 |
+
"learning_rate": 1.042423046566693e-05,
|
| 2193 |
+
"loss": 0.0028,
|
| 2194 |
+
"step": 3120
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 0.7450902598690736,
|
| 2198 |
+
"grad_norm": 1.1640625,
|
| 2199 |
+
"learning_rate": 1.0404498816101026e-05,
|
| 2200 |
+
"loss": 0.0033,
|
| 2201 |
+
"step": 3130
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 0.7474707399325531,
|
| 2205 |
+
"grad_norm": 2.265625,
|
| 2206 |
+
"learning_rate": 1.0384767166535123e-05,
|
| 2207 |
+
"loss": 0.0024,
|
| 2208 |
+
"step": 3140
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 0.7498512199960325,
|
| 2212 |
+
"grad_norm": 0.9375,
|
| 2213 |
+
"learning_rate": 1.0365035516969218e-05,
|
| 2214 |
+
"loss": 0.004,
|
| 2215 |
+
"step": 3150
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 0.752231700059512,
|
| 2219 |
+
"grad_norm": 1.4140625,
|
| 2220 |
+
"learning_rate": 1.0345303867403314e-05,
|
| 2221 |
+
"loss": 0.0031,
|
| 2222 |
+
"step": 3160
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 0.7546121801229915,
|
| 2226 |
+
"grad_norm": 0.84765625,
|
| 2227 |
+
"learning_rate": 1.0325572217837411e-05,
|
| 2228 |
+
"loss": 0.0025,
|
| 2229 |
+
"step": 3170
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 0.7569926601864709,
|
| 2233 |
+
"grad_norm": 1.1640625,
|
| 2234 |
+
"learning_rate": 1.0305840568271509e-05,
|
| 2235 |
+
"loss": 0.0027,
|
| 2236 |
+
"step": 3180
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 0.7593731402499504,
|
| 2240 |
+
"grad_norm": 1.65625,
|
| 2241 |
+
"learning_rate": 1.0286108918705604e-05,
|
| 2242 |
+
"loss": 0.0035,
|
| 2243 |
+
"step": 3190
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 0.7617536203134299,
|
| 2247 |
+
"grad_norm": 0.765625,
|
| 2248 |
+
"learning_rate": 1.02663772691397e-05,
|
| 2249 |
+
"loss": 0.0027,
|
| 2250 |
+
"step": 3200
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 0.7641341003769093,
|
| 2254 |
+
"grad_norm": 1.1796875,
|
| 2255 |
+
"learning_rate": 1.0246645619573797e-05,
|
| 2256 |
+
"loss": 0.0037,
|
| 2257 |
+
"step": 3210
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 0.7665145804403888,
|
| 2261 |
+
"grad_norm": 3.671875,
|
| 2262 |
+
"learning_rate": 1.0226913970007893e-05,
|
| 2263 |
+
"loss": 0.0028,
|
| 2264 |
+
"step": 3220
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 0.7688950605038682,
|
| 2268 |
+
"grad_norm": 1.1953125,
|
| 2269 |
+
"learning_rate": 1.0207182320441988e-05,
|
| 2270 |
+
"loss": 0.0023,
|
| 2271 |
+
"step": 3230
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 0.7712755405673477,
|
| 2275 |
+
"grad_norm": 1.671875,
|
| 2276 |
+
"learning_rate": 1.0187450670876085e-05,
|
| 2277 |
+
"loss": 0.0026,
|
| 2278 |
+
"step": 3240
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 0.7736560206308272,
|
| 2282 |
+
"grad_norm": 0.859375,
|
| 2283 |
+
"learning_rate": 1.0167719021310181e-05,
|
| 2284 |
+
"loss": 0.0022,
|
| 2285 |
+
"step": 3250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 0.7760365006943066,
|
| 2289 |
+
"grad_norm": 1.484375,
|
| 2290 |
+
"learning_rate": 1.0147987371744278e-05,
|
| 2291 |
+
"loss": 0.0023,
|
| 2292 |
+
"step": 3260
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 0.7784169807577862,
|
| 2296 |
+
"grad_norm": 1.1484375,
|
| 2297 |
+
"learning_rate": 1.0128255722178374e-05,
|
| 2298 |
+
"loss": 0.0018,
|
| 2299 |
+
"step": 3270
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 0.7807974608212657,
|
| 2303 |
+
"grad_norm": 0.734375,
|
| 2304 |
+
"learning_rate": 1.010852407261247e-05,
|
| 2305 |
+
"loss": 0.0029,
|
| 2306 |
+
"step": 3280
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 0.7831779408847451,
|
| 2310 |
+
"grad_norm": 2.0,
|
| 2311 |
+
"learning_rate": 1.0088792423046567e-05,
|
| 2312 |
+
"loss": 0.0023,
|
| 2313 |
+
"step": 3290
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 0.7855584209482246,
|
| 2317 |
+
"grad_norm": 0.498046875,
|
| 2318 |
+
"learning_rate": 1.0069060773480664e-05,
|
| 2319 |
+
"loss": 0.0035,
|
| 2320 |
+
"step": 3300
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 0.787938901011704,
|
| 2324 |
+
"grad_norm": 1.0625,
|
| 2325 |
+
"learning_rate": 1.0049329123914759e-05,
|
| 2326 |
+
"loss": 0.0031,
|
| 2327 |
+
"step": 3310
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 0.7903193810751835,
|
| 2331 |
+
"grad_norm": 3.6875,
|
| 2332 |
+
"learning_rate": 1.0029597474348855e-05,
|
| 2333 |
+
"loss": 0.0025,
|
| 2334 |
+
"step": 3320
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 0.792699861138663,
|
| 2338 |
+
"grad_norm": 0.68359375,
|
| 2339 |
+
"learning_rate": 1.0009865824782952e-05,
|
| 2340 |
+
"loss": 0.0032,
|
| 2341 |
+
"step": 3330
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 0.7950803412021424,
|
| 2345 |
+
"grad_norm": 1.7421875,
|
| 2346 |
+
"learning_rate": 9.990134175217048e-06,
|
| 2347 |
+
"loss": 0.0028,
|
| 2348 |
+
"step": 3340
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 0.7974608212656219,
|
| 2352 |
+
"grad_norm": 0.765625,
|
| 2353 |
+
"learning_rate": 9.970402525651145e-06,
|
| 2354 |
+
"loss": 0.0023,
|
| 2355 |
+
"step": 3350
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 0.7998413013291014,
|
| 2359 |
+
"grad_norm": 1.7109375,
|
| 2360 |
+
"learning_rate": 9.950670876085241e-06,
|
| 2361 |
+
"loss": 0.0023,
|
| 2362 |
+
"step": 3360
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 0.8022217813925808,
|
| 2366 |
+
"grad_norm": 1.328125,
|
| 2367 |
+
"learning_rate": 9.930939226519338e-06,
|
| 2368 |
+
"loss": 0.002,
|
| 2369 |
+
"step": 3370
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 0.8046022614560603,
|
| 2373 |
+
"grad_norm": 0.8828125,
|
| 2374 |
+
"learning_rate": 9.911207576953434e-06,
|
| 2375 |
+
"loss": 0.0022,
|
| 2376 |
+
"step": 3380
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 0.8069827415195397,
|
| 2380 |
+
"grad_norm": 2.3125,
|
| 2381 |
+
"learning_rate": 9.891475927387529e-06,
|
| 2382 |
+
"loss": 0.0038,
|
| 2383 |
+
"step": 3390
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 0.8093632215830192,
|
| 2387 |
+
"grad_norm": 1.2734375,
|
| 2388 |
+
"learning_rate": 9.871744277821626e-06,
|
| 2389 |
+
"loss": 0.0023,
|
| 2390 |
+
"step": 3400
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 0.8117437016464987,
|
| 2394 |
+
"grad_norm": 1.1015625,
|
| 2395 |
+
"learning_rate": 9.852012628255722e-06,
|
| 2396 |
+
"loss": 0.0019,
|
| 2397 |
+
"step": 3410
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 0.8141241817099781,
|
| 2401 |
+
"grad_norm": 0.7890625,
|
| 2402 |
+
"learning_rate": 9.832280978689819e-06,
|
| 2403 |
+
"loss": 0.0037,
|
| 2404 |
+
"step": 3420
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 0.8165046617734576,
|
| 2408 |
+
"grad_norm": 1.0546875,
|
| 2409 |
+
"learning_rate": 9.812549329123915e-06,
|
| 2410 |
+
"loss": 0.0018,
|
| 2411 |
+
"step": 3430
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 0.8188851418369372,
|
| 2415 |
+
"grad_norm": 1.6484375,
|
| 2416 |
+
"learning_rate": 9.792817679558012e-06,
|
| 2417 |
+
"loss": 0.0022,
|
| 2418 |
+
"step": 3440
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 0.8212656219004166,
|
| 2422 |
+
"grad_norm": 1.6328125,
|
| 2423 |
+
"learning_rate": 9.773086029992108e-06,
|
| 2424 |
+
"loss": 0.0026,
|
| 2425 |
+
"step": 3450
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 0.8236461019638961,
|
| 2429 |
+
"grad_norm": 1.1015625,
|
| 2430 |
+
"learning_rate": 9.753354380426205e-06,
|
| 2431 |
+
"loss": 0.0022,
|
| 2432 |
+
"step": 3460
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 0.8260265820273756,
|
| 2436 |
+
"grad_norm": 1.953125,
|
| 2437 |
+
"learning_rate": 9.733622730860301e-06,
|
| 2438 |
+
"loss": 0.0028,
|
| 2439 |
+
"step": 3470
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 0.828407062090855,
|
| 2443 |
+
"grad_norm": 0.49609375,
|
| 2444 |
+
"learning_rate": 9.713891081294396e-06,
|
| 2445 |
+
"loss": 0.0019,
|
| 2446 |
+
"step": 3480
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 0.8307875421543345,
|
| 2450 |
+
"grad_norm": 0.609375,
|
| 2451 |
+
"learning_rate": 9.694159431728493e-06,
|
| 2452 |
+
"loss": 0.0107,
|
| 2453 |
+
"step": 3490
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 0.8331680222178139,
|
| 2457 |
+
"grad_norm": 0.92578125,
|
| 2458 |
+
"learning_rate": 9.674427782162589e-06,
|
| 2459 |
+
"loss": 0.002,
|
| 2460 |
+
"step": 3500
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 0.8355485022812934,
|
| 2464 |
+
"grad_norm": 1.375,
|
| 2465 |
+
"learning_rate": 9.654696132596686e-06,
|
| 2466 |
+
"loss": 0.0026,
|
| 2467 |
+
"step": 3510
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 0.8379289823447729,
|
| 2471 |
+
"grad_norm": 0.6953125,
|
| 2472 |
+
"learning_rate": 9.63496448303078e-06,
|
| 2473 |
+
"loss": 0.0029,
|
| 2474 |
+
"step": 3520
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 0.8403094624082523,
|
| 2478 |
+
"grad_norm": 0.7734375,
|
| 2479 |
+
"learning_rate": 9.615232833464879e-06,
|
| 2480 |
+
"loss": 0.002,
|
| 2481 |
+
"step": 3530
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 0.8426899424717318,
|
| 2485 |
+
"grad_norm": 0.3828125,
|
| 2486 |
+
"learning_rate": 9.595501183898975e-06,
|
| 2487 |
+
"loss": 0.0023,
|
| 2488 |
+
"step": 3540
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 0.8450704225352113,
|
| 2492 |
+
"grad_norm": 1.1484375,
|
| 2493 |
+
"learning_rate": 9.575769534333072e-06,
|
| 2494 |
+
"loss": 0.0022,
|
| 2495 |
+
"step": 3550
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 0.8474509025986907,
|
| 2499 |
+
"grad_norm": 1.0859375,
|
| 2500 |
+
"learning_rate": 9.556037884767166e-06,
|
| 2501 |
+
"loss": 0.002,
|
| 2502 |
+
"step": 3560
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 0.8498313826621702,
|
| 2506 |
+
"grad_norm": 0.58203125,
|
| 2507 |
+
"learning_rate": 9.536306235201263e-06,
|
| 2508 |
+
"loss": 0.0028,
|
| 2509 |
+
"step": 3570
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 0.8522118627256496,
|
| 2513 |
+
"grad_norm": 1.6640625,
|
| 2514 |
+
"learning_rate": 9.51657458563536e-06,
|
| 2515 |
+
"loss": 0.0027,
|
| 2516 |
+
"step": 3580
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 0.8545923427891291,
|
| 2520 |
+
"grad_norm": 2.21875,
|
| 2521 |
+
"learning_rate": 9.496842936069456e-06,
|
| 2522 |
+
"loss": 0.0027,
|
| 2523 |
+
"step": 3590
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 0.8569728228526086,
|
| 2527 |
+
"grad_norm": 1.0546875,
|
| 2528 |
+
"learning_rate": 9.477111286503551e-06,
|
| 2529 |
+
"loss": 0.0024,
|
| 2530 |
+
"step": 3600
|
| 2531 |
+
},
|
| 2532 |
+
{
|
| 2533 |
+
"epoch": 0.859353302916088,
|
| 2534 |
+
"grad_norm": 1.9296875,
|
| 2535 |
+
"learning_rate": 9.457379636937647e-06,
|
| 2536 |
+
"loss": 0.0029,
|
| 2537 |
+
"step": 3610
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 0.8617337829795676,
|
| 2541 |
+
"grad_norm": 0.71875,
|
| 2542 |
+
"learning_rate": 9.437647987371746e-06,
|
| 2543 |
+
"loss": 0.0026,
|
| 2544 |
+
"step": 3620
|
| 2545 |
+
},
|
| 2546 |
+
{
|
| 2547 |
+
"epoch": 0.8641142630430471,
|
| 2548 |
+
"grad_norm": 0.96484375,
|
| 2549 |
+
"learning_rate": 9.417916337805842e-06,
|
| 2550 |
+
"loss": 0.0021,
|
| 2551 |
+
"step": 3630
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 0.8664947431065265,
|
| 2555 |
+
"grad_norm": 0.58203125,
|
| 2556 |
+
"learning_rate": 9.398184688239937e-06,
|
| 2557 |
+
"loss": 0.0019,
|
| 2558 |
+
"step": 3640
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 0.868875223170006,
|
| 2562 |
+
"grad_norm": 0.87109375,
|
| 2563 |
+
"learning_rate": 9.378453038674033e-06,
|
| 2564 |
+
"loss": 0.0017,
|
| 2565 |
+
"step": 3650
|
| 2566 |
+
},
|
| 2567 |
+
{
|
| 2568 |
+
"epoch": 0.8712557032334854,
|
| 2569 |
+
"grad_norm": 1.28125,
|
| 2570 |
+
"learning_rate": 9.35872138910813e-06,
|
| 2571 |
+
"loss": 0.0017,
|
| 2572 |
+
"step": 3660
|
| 2573 |
+
},
|
| 2574 |
+
{
|
| 2575 |
+
"epoch": 0.8736361832969649,
|
| 2576 |
+
"grad_norm": 0.54296875,
|
| 2577 |
+
"learning_rate": 9.338989739542226e-06,
|
| 2578 |
+
"loss": 0.0022,
|
| 2579 |
+
"step": 3670
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 0.8760166633604444,
|
| 2583 |
+
"grad_norm": 0.35546875,
|
| 2584 |
+
"learning_rate": 9.319258089976321e-06,
|
| 2585 |
+
"loss": 0.0019,
|
| 2586 |
+
"step": 3680
|
| 2587 |
+
},
|
| 2588 |
+
{
|
| 2589 |
+
"epoch": 0.8783971434239238,
|
| 2590 |
+
"grad_norm": 1.109375,
|
| 2591 |
+
"learning_rate": 9.299526440410418e-06,
|
| 2592 |
+
"loss": 0.0023,
|
| 2593 |
+
"step": 3690
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 0.8807776234874033,
|
| 2597 |
+
"grad_norm": 1.1171875,
|
| 2598 |
+
"learning_rate": 9.279794790844514e-06,
|
| 2599 |
+
"loss": 0.0021,
|
| 2600 |
+
"step": 3700
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 0.8831581035508828,
|
| 2604 |
+
"grad_norm": 0.462890625,
|
| 2605 |
+
"learning_rate": 9.260063141278613e-06,
|
| 2606 |
+
"loss": 0.0026,
|
| 2607 |
+
"step": 3710
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 0.8855385836143622,
|
| 2611 |
+
"grad_norm": 1.828125,
|
| 2612 |
+
"learning_rate": 9.240331491712707e-06,
|
| 2613 |
+
"loss": 0.0022,
|
| 2614 |
+
"step": 3720
|
| 2615 |
+
},
|
| 2616 |
+
{
|
| 2617 |
+
"epoch": 0.8879190636778417,
|
| 2618 |
+
"grad_norm": 0.90625,
|
| 2619 |
+
"learning_rate": 9.220599842146804e-06,
|
| 2620 |
+
"loss": 0.0026,
|
| 2621 |
+
"step": 3730
|
| 2622 |
+
},
|
| 2623 |
+
{
|
| 2624 |
+
"epoch": 0.8902995437413211,
|
| 2625 |
+
"grad_norm": 1.828125,
|
| 2626 |
+
"learning_rate": 9.2008681925809e-06,
|
| 2627 |
+
"loss": 0.0017,
|
| 2628 |
+
"step": 3740
|
| 2629 |
+
},
|
| 2630 |
+
{
|
| 2631 |
+
"epoch": 0.8926800238048006,
|
| 2632 |
+
"grad_norm": 1.5703125,
|
| 2633 |
+
"learning_rate": 9.181136543014997e-06,
|
| 2634 |
+
"loss": 0.0026,
|
| 2635 |
+
"step": 3750
|
| 2636 |
+
},
|
| 2637 |
+
{
|
| 2638 |
+
"epoch": 0.8950605038682801,
|
| 2639 |
+
"grad_norm": 1.125,
|
| 2640 |
+
"learning_rate": 9.161404893449092e-06,
|
| 2641 |
+
"loss": 0.0016,
|
| 2642 |
+
"step": 3760
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 0.8974409839317595,
|
| 2646 |
+
"grad_norm": 1.3125,
|
| 2647 |
+
"learning_rate": 9.141673243883188e-06,
|
| 2648 |
+
"loss": 0.0032,
|
| 2649 |
+
"step": 3770
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 0.899821463995239,
|
| 2653 |
+
"grad_norm": 0.9296875,
|
| 2654 |
+
"learning_rate": 9.121941594317285e-06,
|
| 2655 |
+
"loss": 0.0021,
|
| 2656 |
+
"step": 3780
|
| 2657 |
+
},
|
| 2658 |
+
{
|
| 2659 |
+
"epoch": 0.9022019440587186,
|
| 2660 |
+
"grad_norm": 0.30078125,
|
| 2661 |
+
"learning_rate": 9.102209944751381e-06,
|
| 2662 |
+
"loss": 0.0018,
|
| 2663 |
+
"step": 3790
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 0.904582424122198,
|
| 2667 |
+
"grad_norm": 0.98828125,
|
| 2668 |
+
"learning_rate": 9.082478295185478e-06,
|
| 2669 |
+
"loss": 0.0022,
|
| 2670 |
+
"step": 3800
|
| 2671 |
+
},
|
| 2672 |
+
{
|
| 2673 |
+
"epoch": 0.9069629041856775,
|
| 2674 |
+
"grad_norm": 0.44140625,
|
| 2675 |
+
"learning_rate": 9.062746645619574e-06,
|
| 2676 |
+
"loss": 0.0027,
|
| 2677 |
+
"step": 3810
|
| 2678 |
+
},
|
| 2679 |
+
{
|
| 2680 |
+
"epoch": 0.909343384249157,
|
| 2681 |
+
"grad_norm": 1.1875,
|
| 2682 |
+
"learning_rate": 9.04301499605367e-06,
|
| 2683 |
+
"loss": 0.0027,
|
| 2684 |
+
"step": 3820
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 0.9117238643126364,
|
| 2688 |
+
"grad_norm": 1.2578125,
|
| 2689 |
+
"learning_rate": 9.023283346487767e-06,
|
| 2690 |
+
"loss": 0.0018,
|
| 2691 |
+
"step": 3830
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 0.9141043443761159,
|
| 2695 |
+
"grad_norm": 0.82421875,
|
| 2696 |
+
"learning_rate": 9.003551696921862e-06,
|
| 2697 |
+
"loss": 0.0017,
|
| 2698 |
+
"step": 3840
|
| 2699 |
+
},
|
| 2700 |
+
{
|
| 2701 |
+
"epoch": 0.9164848244395953,
|
| 2702 |
+
"grad_norm": 1.46875,
|
| 2703 |
+
"learning_rate": 8.983820047355959e-06,
|
| 2704 |
+
"loss": 0.0039,
|
| 2705 |
+
"step": 3850
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 0.9188653045030748,
|
| 2709 |
+
"grad_norm": 0.51171875,
|
| 2710 |
+
"learning_rate": 8.964088397790055e-06,
|
| 2711 |
+
"loss": 0.0021,
|
| 2712 |
+
"step": 3860
|
| 2713 |
+
},
|
| 2714 |
+
{
|
| 2715 |
+
"epoch": 0.9212457845665543,
|
| 2716 |
+
"grad_norm": 0.380859375,
|
| 2717 |
+
"learning_rate": 8.944356748224152e-06,
|
| 2718 |
+
"loss": 0.0016,
|
| 2719 |
+
"step": 3870
|
| 2720 |
+
},
|
| 2721 |
+
{
|
| 2722 |
+
"epoch": 0.9236262646300337,
|
| 2723 |
+
"grad_norm": 0.486328125,
|
| 2724 |
+
"learning_rate": 8.924625098658247e-06,
|
| 2725 |
+
"loss": 0.0022,
|
| 2726 |
+
"step": 3880
|
| 2727 |
+
},
|
| 2728 |
+
{
|
| 2729 |
+
"epoch": 0.9260067446935132,
|
| 2730 |
+
"grad_norm": 0.52734375,
|
| 2731 |
+
"learning_rate": 8.904893449092345e-06,
|
| 2732 |
+
"loss": 0.0013,
|
| 2733 |
+
"step": 3890
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 0.9283872247569926,
|
| 2737 |
+
"grad_norm": 0.6953125,
|
| 2738 |
+
"learning_rate": 8.885161799526441e-06,
|
| 2739 |
+
"loss": 0.0025,
|
| 2740 |
+
"step": 3900
|
| 2741 |
+
},
|
| 2742 |
+
{
|
| 2743 |
+
"epoch": 0.9307677048204721,
|
| 2744 |
+
"grad_norm": 1.1640625,
|
| 2745 |
+
"learning_rate": 8.865430149960538e-06,
|
| 2746 |
+
"loss": 0.0026,
|
| 2747 |
+
"step": 3910
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 0.9331481848839516,
|
| 2751 |
+
"grad_norm": 0.86328125,
|
| 2752 |
+
"learning_rate": 8.845698500394633e-06,
|
| 2753 |
+
"loss": 0.002,
|
| 2754 |
+
"step": 3920
|
| 2755 |
+
},
|
| 2756 |
+
{
|
| 2757 |
+
"epoch": 0.935528664947431,
|
| 2758 |
+
"grad_norm": 1.046875,
|
| 2759 |
+
"learning_rate": 8.825966850828729e-06,
|
| 2760 |
+
"loss": 0.0034,
|
| 2761 |
+
"step": 3930
|
| 2762 |
+
},
|
| 2763 |
+
{
|
| 2764 |
+
"epoch": 0.9379091450109105,
|
| 2765 |
+
"grad_norm": 0.435546875,
|
| 2766 |
+
"learning_rate": 8.806235201262826e-06,
|
| 2767 |
+
"loss": 0.0019,
|
| 2768 |
+
"step": 3940
|
| 2769 |
+
},
|
| 2770 |
+
{
|
| 2771 |
+
"epoch": 0.94028962507439,
|
| 2772 |
+
"grad_norm": 1.28125,
|
| 2773 |
+
"learning_rate": 8.786503551696922e-06,
|
| 2774 |
+
"loss": 0.0018,
|
| 2775 |
+
"step": 3950
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 0.9426701051378694,
|
| 2779 |
+
"grad_norm": 0.50390625,
|
| 2780 |
+
"learning_rate": 8.766771902131017e-06,
|
| 2781 |
+
"loss": 0.0017,
|
| 2782 |
+
"step": 3960
|
| 2783 |
+
},
|
| 2784 |
+
{
|
| 2785 |
+
"epoch": 0.945050585201349,
|
| 2786 |
+
"grad_norm": 1.0234375,
|
| 2787 |
+
"learning_rate": 8.747040252565113e-06,
|
| 2788 |
+
"loss": 0.002,
|
| 2789 |
+
"step": 3970
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 0.9474310652648285,
|
| 2793 |
+
"grad_norm": 2.390625,
|
| 2794 |
+
"learning_rate": 8.727308602999212e-06,
|
| 2795 |
+
"loss": 0.0029,
|
| 2796 |
+
"step": 3980
|
| 2797 |
+
},
|
| 2798 |
+
{
|
| 2799 |
+
"epoch": 0.9498115453283079,
|
| 2800 |
+
"grad_norm": 0.796875,
|
| 2801 |
+
"learning_rate": 8.707576953433308e-06,
|
| 2802 |
+
"loss": 0.0017,
|
| 2803 |
+
"step": 3990
|
| 2804 |
+
},
|
| 2805 |
+
{
|
| 2806 |
+
"epoch": 0.9521920253917874,
|
| 2807 |
+
"grad_norm": 1.28125,
|
| 2808 |
+
"learning_rate": 8.687845303867403e-06,
|
| 2809 |
+
"loss": 0.0023,
|
| 2810 |
+
"step": 4000
|
| 2811 |
+
},
|
| 2812 |
+
{
|
| 2813 |
+
"epoch": 0.9545725054552668,
|
| 2814 |
+
"grad_norm": 0.94921875,
|
| 2815 |
+
"learning_rate": 8.6681136543015e-06,
|
| 2816 |
+
"loss": 0.0015,
|
| 2817 |
+
"step": 4010
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 0.9569529855187463,
|
| 2821 |
+
"grad_norm": 0.515625,
|
| 2822 |
+
"learning_rate": 8.648382004735596e-06,
|
| 2823 |
+
"loss": 0.0018,
|
| 2824 |
+
"step": 4020
|
| 2825 |
+
},
|
| 2826 |
+
{
|
| 2827 |
+
"epoch": 0.9593334655822258,
|
| 2828 |
+
"grad_norm": 1.1328125,
|
| 2829 |
+
"learning_rate": 8.628650355169693e-06,
|
| 2830 |
+
"loss": 0.0013,
|
| 2831 |
+
"step": 4030
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 0.9617139456457052,
|
| 2835 |
+
"grad_norm": 1.453125,
|
| 2836 |
+
"learning_rate": 8.608918705603789e-06,
|
| 2837 |
+
"loss": 0.0021,
|
| 2838 |
+
"step": 4040
|
| 2839 |
+
},
|
| 2840 |
+
{
|
| 2841 |
+
"epoch": 0.9640944257091847,
|
| 2842 |
+
"grad_norm": 0.41796875,
|
| 2843 |
+
"learning_rate": 8.589187056037884e-06,
|
| 2844 |
+
"loss": 0.0031,
|
| 2845 |
+
"step": 4050
|
| 2846 |
+
},
|
| 2847 |
+
{
|
| 2848 |
+
"epoch": 0.9664749057726642,
|
| 2849 |
+
"grad_norm": 2.046875,
|
| 2850 |
+
"learning_rate": 8.56945540647198e-06,
|
| 2851 |
+
"loss": 0.0028,
|
| 2852 |
+
"step": 4060
|
| 2853 |
+
},
|
| 2854 |
+
{
|
| 2855 |
+
"epoch": 0.9688553858361436,
|
| 2856 |
+
"grad_norm": 0.8359375,
|
| 2857 |
+
"learning_rate": 8.549723756906079e-06,
|
| 2858 |
+
"loss": 0.0018,
|
| 2859 |
+
"step": 4070
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 0.9712358658996231,
|
| 2863 |
+
"grad_norm": 0.6484375,
|
| 2864 |
+
"learning_rate": 8.529992107340175e-06,
|
| 2865 |
+
"loss": 0.0016,
|
| 2866 |
+
"step": 4080
|
| 2867 |
+
},
|
| 2868 |
+
{
|
| 2869 |
+
"epoch": 0.9736163459631025,
|
| 2870 |
+
"grad_norm": 1.078125,
|
| 2871 |
+
"learning_rate": 8.51026045777427e-06,
|
| 2872 |
+
"loss": 0.0018,
|
| 2873 |
+
"step": 4090
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 0.975996826026582,
|
| 2877 |
+
"grad_norm": 1.140625,
|
| 2878 |
+
"learning_rate": 8.490528808208366e-06,
|
| 2879 |
+
"loss": 0.0018,
|
| 2880 |
+
"step": 4100
|
| 2881 |
+
},
|
| 2882 |
+
{
|
| 2883 |
+
"epoch": 0.9783773060900615,
|
| 2884 |
+
"grad_norm": 1.34375,
|
| 2885 |
+
"learning_rate": 8.470797158642463e-06,
|
| 2886 |
+
"loss": 0.0019,
|
| 2887 |
+
"step": 4110
|
| 2888 |
+
},
|
| 2889 |
+
{
|
| 2890 |
+
"epoch": 0.9807577861535409,
|
| 2891 |
+
"grad_norm": 0.64453125,
|
| 2892 |
+
"learning_rate": 8.45106550907656e-06,
|
| 2893 |
+
"loss": 0.0022,
|
| 2894 |
+
"step": 4120
|
| 2895 |
+
},
|
| 2896 |
+
{
|
| 2897 |
+
"epoch": 0.9831382662170204,
|
| 2898 |
+
"grad_norm": 1.0078125,
|
| 2899 |
+
"learning_rate": 8.431333859510654e-06,
|
| 2900 |
+
"loss": 0.0021,
|
| 2901 |
+
"step": 4130
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 0.9855187462804998,
|
| 2905 |
+
"grad_norm": 0.91015625,
|
| 2906 |
+
"learning_rate": 8.41160220994475e-06,
|
| 2907 |
+
"loss": 0.0014,
|
| 2908 |
+
"step": 4140
|
| 2909 |
+
},
|
| 2910 |
+
{
|
| 2911 |
+
"epoch": 0.9878992263439794,
|
| 2912 |
+
"grad_norm": 1.9140625,
|
| 2913 |
+
"learning_rate": 8.391870560378849e-06,
|
| 2914 |
+
"loss": 0.0027,
|
| 2915 |
+
"step": 4150
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 0.9902797064074589,
|
| 2919 |
+
"grad_norm": 1.4453125,
|
| 2920 |
+
"learning_rate": 8.372138910812946e-06,
|
| 2921 |
+
"loss": 0.002,
|
| 2922 |
+
"step": 4160
|
| 2923 |
+
},
|
| 2924 |
+
{
|
| 2925 |
+
"epoch": 0.9926601864709383,
|
| 2926 |
+
"grad_norm": 1.7578125,
|
| 2927 |
+
"learning_rate": 8.35240726124704e-06,
|
| 2928 |
+
"loss": 0.0018,
|
| 2929 |
+
"step": 4170
|
| 2930 |
+
},
|
| 2931 |
+
{
|
| 2932 |
+
"epoch": 0.9950406665344178,
|
| 2933 |
+
"grad_norm": 1.0625,
|
| 2934 |
+
"learning_rate": 8.332675611681137e-06,
|
| 2935 |
+
"loss": 0.0015,
|
| 2936 |
+
"step": 4180
|
| 2937 |
+
},
|
| 2938 |
+
{
|
| 2939 |
+
"epoch": 0.9974211465978973,
|
| 2940 |
+
"grad_norm": 1.1953125,
|
| 2941 |
+
"learning_rate": 8.312943962115233e-06,
|
| 2942 |
+
"loss": 0.0015,
|
| 2943 |
+
"step": 4190
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 0.9998016266613767,
|
| 2947 |
+
"grad_norm": 0.89453125,
|
| 2948 |
+
"learning_rate": 8.29321231254933e-06,
|
| 2949 |
+
"loss": 0.0023,
|
| 2950 |
+
"step": 4200
|
| 2951 |
+
},
|
| 2952 |
+
{
|
| 2953 |
+
"epoch": 1.0021424320571315,
|
| 2954 |
+
"grad_norm": 0.50390625,
|
| 2955 |
+
"learning_rate": 8.273480662983425e-06,
|
| 2956 |
+
"loss": 0.0015,
|
| 2957 |
+
"step": 4210
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 1.004522912120611,
|
| 2961 |
+
"grad_norm": 1.421875,
|
| 2962 |
+
"learning_rate": 8.253749013417521e-06,
|
| 2963 |
+
"loss": 0.0015,
|
| 2964 |
+
"step": 4220
|
| 2965 |
+
},
|
| 2966 |
+
{
|
| 2967 |
+
"epoch": 1.0069033921840904,
|
| 2968 |
+
"grad_norm": 0.765625,
|
| 2969 |
+
"learning_rate": 8.234017363851618e-06,
|
| 2970 |
+
"loss": 0.0014,
|
| 2971 |
+
"step": 4230
|
| 2972 |
+
},
|
| 2973 |
+
{
|
| 2974 |
+
"epoch": 1.0092838722475699,
|
| 2975 |
+
"grad_norm": 0.5625,
|
| 2976 |
+
"learning_rate": 8.214285714285716e-06,
|
| 2977 |
+
"loss": 0.0014,
|
| 2978 |
+
"step": 4240
|
| 2979 |
+
},
|
| 2980 |
+
{
|
| 2981 |
+
"epoch": 1.0116643523110493,
|
| 2982 |
+
"grad_norm": 0.671875,
|
| 2983 |
+
"learning_rate": 8.19455406471981e-06,
|
| 2984 |
+
"loss": 0.0013,
|
| 2985 |
+
"step": 4250
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 1.0140448323745288,
|
| 2989 |
+
"grad_norm": 0.455078125,
|
| 2990 |
+
"learning_rate": 8.174822415153907e-06,
|
| 2991 |
+
"loss": 0.0013,
|
| 2992 |
+
"step": 4260
|
| 2993 |
+
},
|
| 2994 |
+
{
|
| 2995 |
+
"epoch": 1.0164253124380083,
|
| 2996 |
+
"grad_norm": 0.74609375,
|
| 2997 |
+
"learning_rate": 8.155090765588004e-06,
|
| 2998 |
+
"loss": 0.0009,
|
| 2999 |
+
"step": 4270
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 1.0188057925014877,
|
| 3003 |
+
"grad_norm": 2.265625,
|
| 3004 |
+
"learning_rate": 8.1353591160221e-06,
|
| 3005 |
+
"loss": 0.0018,
|
| 3006 |
+
"step": 4280
|
| 3007 |
+
},
|
| 3008 |
+
{
|
| 3009 |
+
"epoch": 1.0211862725649672,
|
| 3010 |
+
"grad_norm": 0.63671875,
|
| 3011 |
+
"learning_rate": 8.115627466456195e-06,
|
| 3012 |
+
"loss": 0.0014,
|
| 3013 |
+
"step": 4290
|
| 3014 |
+
},
|
| 3015 |
+
{
|
| 3016 |
+
"epoch": 1.0235667526284467,
|
| 3017 |
+
"grad_norm": 1.265625,
|
| 3018 |
+
"learning_rate": 8.095895816890292e-06,
|
| 3019 |
+
"loss": 0.0012,
|
| 3020 |
+
"step": 4300
|
| 3021 |
+
},
|
| 3022 |
+
{
|
| 3023 |
+
"epoch": 1.0259472326919261,
|
| 3024 |
+
"grad_norm": 1.375,
|
| 3025 |
+
"learning_rate": 8.076164167324388e-06,
|
| 3026 |
+
"loss": 0.0012,
|
| 3027 |
+
"step": 4310
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 1.0283277127554056,
|
| 3031 |
+
"grad_norm": 0.578125,
|
| 3032 |
+
"learning_rate": 8.056432517758485e-06,
|
| 3033 |
+
"loss": 0.0018,
|
| 3034 |
+
"step": 4320
|
| 3035 |
+
},
|
| 3036 |
+
{
|
| 3037 |
+
"epoch": 1.030708192818885,
|
| 3038 |
+
"grad_norm": 0.8046875,
|
| 3039 |
+
"learning_rate": 8.036700868192581e-06,
|
| 3040 |
+
"loss": 0.0013,
|
| 3041 |
+
"step": 4330
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 1.0330886728823647,
|
| 3045 |
+
"grad_norm": 0.58984375,
|
| 3046 |
+
"learning_rate": 8.016969218626678e-06,
|
| 3047 |
+
"loss": 0.0012,
|
| 3048 |
+
"step": 4340
|
| 3049 |
+
},
|
| 3050 |
+
{
|
| 3051 |
+
"epoch": 1.0354691529458442,
|
| 3052 |
+
"grad_norm": 0.5234375,
|
| 3053 |
+
"learning_rate": 7.997237569060774e-06,
|
| 3054 |
+
"loss": 0.0012,
|
| 3055 |
+
"step": 4350
|
| 3056 |
+
},
|
| 3057 |
+
{
|
| 3058 |
+
"epoch": 1.0378496330093236,
|
| 3059 |
+
"grad_norm": 1.03125,
|
| 3060 |
+
"learning_rate": 7.97750591949487e-06,
|
| 3061 |
+
"loss": 0.0014,
|
| 3062 |
+
"step": 4360
|
| 3063 |
+
},
|
| 3064 |
+
{
|
| 3065 |
+
"epoch": 1.040230113072803,
|
| 3066 |
+
"grad_norm": 0.99609375,
|
| 3067 |
+
"learning_rate": 7.957774269928966e-06,
|
| 3068 |
+
"loss": 0.0009,
|
| 3069 |
+
"step": 4370
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 1.0426105931362826,
|
| 3073 |
+
"grad_norm": 0.60546875,
|
| 3074 |
+
"learning_rate": 7.938042620363062e-06,
|
| 3075 |
+
"loss": 0.0016,
|
| 3076 |
+
"step": 4380
|
| 3077 |
+
},
|
| 3078 |
+
{
|
| 3079 |
+
"epoch": 1.044991073199762,
|
| 3080 |
+
"grad_norm": 1.3515625,
|
| 3081 |
+
"learning_rate": 7.918310970797159e-06,
|
| 3082 |
+
"loss": 0.0012,
|
| 3083 |
+
"step": 4390
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 1.0473715532632415,
|
| 3087 |
+
"grad_norm": 0.671875,
|
| 3088 |
+
"learning_rate": 7.898579321231255e-06,
|
| 3089 |
+
"loss": 0.0012,
|
| 3090 |
+
"step": 4400
|
| 3091 |
+
},
|
| 3092 |
+
{
|
| 3093 |
+
"epoch": 1.049752033326721,
|
| 3094 |
+
"grad_norm": 1.2109375,
|
| 3095 |
+
"learning_rate": 7.87884767166535e-06,
|
| 3096 |
+
"loss": 0.0018,
|
| 3097 |
+
"step": 4410
|
| 3098 |
+
},
|
| 3099 |
+
{
|
| 3100 |
+
"epoch": 1.0521325133902004,
|
| 3101 |
+
"grad_norm": 0.61328125,
|
| 3102 |
+
"learning_rate": 7.859116022099448e-06,
|
| 3103 |
+
"loss": 0.0091,
|
| 3104 |
+
"step": 4420
|
| 3105 |
+
},
|
| 3106 |
+
{
|
| 3107 |
+
"epoch": 1.0545129934536799,
|
| 3108 |
+
"grad_norm": 0.384765625,
|
| 3109 |
+
"learning_rate": 7.839384372533545e-06,
|
| 3110 |
+
"loss": 0.0018,
|
| 3111 |
+
"step": 4430
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 1.0568934735171593,
|
| 3115 |
+
"grad_norm": 0.56640625,
|
| 3116 |
+
"learning_rate": 7.819652722967641e-06,
|
| 3117 |
+
"loss": 0.0018,
|
| 3118 |
+
"step": 4440
|
| 3119 |
+
},
|
| 3120 |
+
{
|
| 3121 |
+
"epoch": 1.0592739535806388,
|
| 3122 |
+
"grad_norm": 0.224609375,
|
| 3123 |
+
"learning_rate": 7.799921073401736e-06,
|
| 3124 |
+
"loss": 0.0013,
|
| 3125 |
+
"step": 4450
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 1.0616544336441183,
|
| 3129 |
+
"grad_norm": 1.234375,
|
| 3130 |
+
"learning_rate": 7.780189423835833e-06,
|
| 3131 |
+
"loss": 0.0012,
|
| 3132 |
+
"step": 4460
|
| 3133 |
+
},
|
| 3134 |
+
{
|
| 3135 |
+
"epoch": 1.0640349137075977,
|
| 3136 |
+
"grad_norm": 0.291015625,
|
| 3137 |
+
"learning_rate": 7.760457774269929e-06,
|
| 3138 |
+
"loss": 0.0013,
|
| 3139 |
+
"step": 4470
|
| 3140 |
+
},
|
| 3141 |
+
{
|
| 3142 |
+
"epoch": 1.0664153937710772,
|
| 3143 |
+
"grad_norm": 1.125,
|
| 3144 |
+
"learning_rate": 7.740726124704026e-06,
|
| 3145 |
+
"loss": 0.007,
|
| 3146 |
+
"step": 4480
|
| 3147 |
+
},
|
| 3148 |
+
{
|
| 3149 |
+
"epoch": 1.0687958738345567,
|
| 3150 |
+
"grad_norm": 0.3359375,
|
| 3151 |
+
"learning_rate": 7.72099447513812e-06,
|
| 3152 |
+
"loss": 0.0016,
|
| 3153 |
+
"step": 4490
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 1.0711763538980361,
|
| 3157 |
+
"grad_norm": 1.3125,
|
| 3158 |
+
"learning_rate": 7.701262825572217e-06,
|
| 3159 |
+
"loss": 0.0015,
|
| 3160 |
+
"step": 4500
|
| 3161 |
+
},
|
| 3162 |
+
{
|
| 3163 |
+
"epoch": 1.0735568339615156,
|
| 3164 |
+
"grad_norm": 3.921875,
|
| 3165 |
+
"learning_rate": 7.681531176006315e-06,
|
| 3166 |
+
"loss": 0.0013,
|
| 3167 |
+
"step": 4510
|
| 3168 |
+
},
|
| 3169 |
+
{
|
| 3170 |
+
"epoch": 1.075937314024995,
|
| 3171 |
+
"grad_norm": 2.140625,
|
| 3172 |
+
"learning_rate": 7.661799526440412e-06,
|
| 3173 |
+
"loss": 0.0014,
|
| 3174 |
+
"step": 4520
|
| 3175 |
+
},
|
| 3176 |
+
{
|
| 3177 |
+
"epoch": 1.0783177940884745,
|
| 3178 |
+
"grad_norm": 1.234375,
|
| 3179 |
+
"learning_rate": 7.642067876874506e-06,
|
| 3180 |
+
"loss": 0.0013,
|
| 3181 |
+
"step": 4530
|
| 3182 |
+
},
|
| 3183 |
+
{
|
| 3184 |
+
"epoch": 1.080698274151954,
|
| 3185 |
+
"grad_norm": 1.8984375,
|
| 3186 |
+
"learning_rate": 7.622336227308603e-06,
|
| 3187 |
+
"loss": 0.001,
|
| 3188 |
+
"step": 4540
|
| 3189 |
+
},
|
| 3190 |
+
{
|
| 3191 |
+
"epoch": 1.0830787542154334,
|
| 3192 |
+
"grad_norm": 0.671875,
|
| 3193 |
+
"learning_rate": 7.6026045777426995e-06,
|
| 3194 |
+
"loss": 0.001,
|
| 3195 |
+
"step": 4550
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 1.0854592342789129,
|
| 3199 |
+
"grad_norm": 1.7109375,
|
| 3200 |
+
"learning_rate": 7.582872928176795e-06,
|
| 3201 |
+
"loss": 0.0011,
|
| 3202 |
+
"step": 4560
|
| 3203 |
+
},
|
| 3204 |
+
{
|
| 3205 |
+
"epoch": 1.0878397143423923,
|
| 3206 |
+
"grad_norm": 0.4375,
|
| 3207 |
+
"learning_rate": 7.563141278610892e-06,
|
| 3208 |
+
"loss": 0.0013,
|
| 3209 |
+
"step": 4570
|
| 3210 |
+
},
|
| 3211 |
+
{
|
| 3212 |
+
"epoch": 1.0902201944058718,
|
| 3213 |
+
"grad_norm": 0.953125,
|
| 3214 |
+
"learning_rate": 7.543409629044987e-06,
|
| 3215 |
+
"loss": 0.0016,
|
| 3216 |
+
"step": 4580
|
| 3217 |
+
},
|
| 3218 |
+
{
|
| 3219 |
+
"epoch": 1.0926006744693513,
|
| 3220 |
+
"grad_norm": 1.25,
|
| 3221 |
+
"learning_rate": 7.523677979479084e-06,
|
| 3222 |
+
"loss": 0.0013,
|
| 3223 |
+
"step": 4590
|
| 3224 |
+
},
|
| 3225 |
+
{
|
| 3226 |
+
"epoch": 1.0949811545328307,
|
| 3227 |
+
"grad_norm": 0.333984375,
|
| 3228 |
+
"learning_rate": 7.503946329913181e-06,
|
| 3229 |
+
"loss": 0.0008,
|
| 3230 |
+
"step": 4600
|
| 3231 |
+
},
|
| 3232 |
+
{
|
| 3233 |
+
"epoch": 1.0973616345963102,
|
| 3234 |
+
"grad_norm": 0.640625,
|
| 3235 |
+
"learning_rate": 7.484214680347277e-06,
|
| 3236 |
+
"loss": 0.0015,
|
| 3237 |
+
"step": 4610
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 1.0997421146597897,
|
| 3241 |
+
"grad_norm": 1.9453125,
|
| 3242 |
+
"learning_rate": 7.464483030781374e-06,
|
| 3243 |
+
"loss": 0.002,
|
| 3244 |
+
"step": 4620
|
| 3245 |
+
},
|
| 3246 |
+
{
|
| 3247 |
+
"epoch": 1.1021225947232691,
|
| 3248 |
+
"grad_norm": 0.69921875,
|
| 3249 |
+
"learning_rate": 7.44475138121547e-06,
|
| 3250 |
+
"loss": 0.0012,
|
| 3251 |
+
"step": 4630
|
| 3252 |
+
},
|
| 3253 |
+
{
|
| 3254 |
+
"epoch": 1.1045030747867486,
|
| 3255 |
+
"grad_norm": 0.62109375,
|
| 3256 |
+
"learning_rate": 7.4250197316495664e-06,
|
| 3257 |
+
"loss": 0.0013,
|
| 3258 |
+
"step": 4640
|
| 3259 |
+
},
|
| 3260 |
+
{
|
| 3261 |
+
"epoch": 1.106883554850228,
|
| 3262 |
+
"grad_norm": 0.95703125,
|
| 3263 |
+
"learning_rate": 7.405288082083662e-06,
|
| 3264 |
+
"loss": 0.0015,
|
| 3265 |
+
"step": 4650
|
| 3266 |
+
},
|
| 3267 |
+
{
|
| 3268 |
+
"epoch": 1.1092640349137075,
|
| 3269 |
+
"grad_norm": 0.76953125,
|
| 3270 |
+
"learning_rate": 7.3855564325177595e-06,
|
| 3271 |
+
"loss": 0.0015,
|
| 3272 |
+
"step": 4660
|
| 3273 |
+
},
|
| 3274 |
+
{
|
| 3275 |
+
"epoch": 1.111644514977187,
|
| 3276 |
+
"grad_norm": 1.59375,
|
| 3277 |
+
"learning_rate": 7.365824782951855e-06,
|
| 3278 |
+
"loss": 0.0015,
|
| 3279 |
+
"step": 4670
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 1.1140249950406664,
|
| 3283 |
+
"grad_norm": 1.3046875,
|
| 3284 |
+
"learning_rate": 7.346093133385952e-06,
|
| 3285 |
+
"loss": 0.0011,
|
| 3286 |
+
"step": 4680
|
| 3287 |
+
},
|
| 3288 |
+
{
|
| 3289 |
+
"epoch": 1.1164054751041461,
|
| 3290 |
+
"grad_norm": 0.84765625,
|
| 3291 |
+
"learning_rate": 7.326361483820047e-06,
|
| 3292 |
+
"loss": 0.0011,
|
| 3293 |
+
"step": 4690
|
| 3294 |
+
},
|
| 3295 |
+
{
|
| 3296 |
+
"epoch": 1.1187859551676254,
|
| 3297 |
+
"grad_norm": 1.703125,
|
| 3298 |
+
"learning_rate": 7.306629834254144e-06,
|
| 3299 |
+
"loss": 0.0019,
|
| 3300 |
+
"step": 4700
|
| 3301 |
+
},
|
| 3302 |
+
{
|
| 3303 |
+
"epoch": 1.121166435231105,
|
| 3304 |
+
"grad_norm": 0.48046875,
|
| 3305 |
+
"learning_rate": 7.28689818468824e-06,
|
| 3306 |
+
"loss": 0.0008,
|
| 3307 |
+
"step": 4710
|
| 3308 |
+
},
|
| 3309 |
+
{
|
| 3310 |
+
"epoch": 1.1235469152945845,
|
| 3311 |
+
"grad_norm": 0.3359375,
|
| 3312 |
+
"learning_rate": 7.267166535122337e-06,
|
| 3313 |
+
"loss": 0.001,
|
| 3314 |
+
"step": 4720
|
| 3315 |
+
},
|
| 3316 |
+
{
|
| 3317 |
+
"epoch": 1.125927395358064,
|
| 3318 |
+
"grad_norm": 0.58984375,
|
| 3319 |
+
"learning_rate": 7.2474348855564325e-06,
|
| 3320 |
+
"loss": 0.0012,
|
| 3321 |
+
"step": 4730
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 1.1283078754215434,
|
| 3325 |
+
"grad_norm": 0.76171875,
|
| 3326 |
+
"learning_rate": 7.227703235990529e-06,
|
| 3327 |
+
"loss": 0.0014,
|
| 3328 |
+
"step": 4740
|
| 3329 |
+
},
|
| 3330 |
+
{
|
| 3331 |
+
"epoch": 1.130688355485023,
|
| 3332 |
+
"grad_norm": 0.359375,
|
| 3333 |
+
"learning_rate": 7.2079715864246256e-06,
|
| 3334 |
+
"loss": 0.0011,
|
| 3335 |
+
"step": 4750
|
| 3336 |
+
},
|
| 3337 |
+
{
|
| 3338 |
+
"epoch": 1.1330688355485024,
|
| 3339 |
+
"grad_norm": 1.15625,
|
| 3340 |
+
"learning_rate": 7.188239936858722e-06,
|
| 3341 |
+
"loss": 0.0013,
|
| 3342 |
+
"step": 4760
|
| 3343 |
+
},
|
| 3344 |
+
{
|
| 3345 |
+
"epoch": 1.1354493156119818,
|
| 3346 |
+
"grad_norm": 0.91796875,
|
| 3347 |
+
"learning_rate": 7.168508287292818e-06,
|
| 3348 |
+
"loss": 0.0012,
|
| 3349 |
+
"step": 4770
|
| 3350 |
+
},
|
| 3351 |
+
{
|
| 3352 |
+
"epoch": 1.1378297956754613,
|
| 3353 |
+
"grad_norm": 0.8828125,
|
| 3354 |
+
"learning_rate": 7.148776637726914e-06,
|
| 3355 |
+
"loss": 0.0012,
|
| 3356 |
+
"step": 4780
|
| 3357 |
+
},
|
| 3358 |
+
{
|
| 3359 |
+
"epoch": 1.1402102757389407,
|
| 3360 |
+
"grad_norm": 0.69140625,
|
| 3361 |
+
"learning_rate": 7.12904498816101e-06,
|
| 3362 |
+
"loss": 0.0018,
|
| 3363 |
+
"step": 4790
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 1.1425907558024202,
|
| 3367 |
+
"grad_norm": 0.54296875,
|
| 3368 |
+
"learning_rate": 7.109313338595107e-06,
|
| 3369 |
+
"loss": 0.0009,
|
| 3370 |
+
"step": 4800
|
| 3371 |
+
},
|
| 3372 |
+
{
|
| 3373 |
+
"epoch": 1.1449712358658997,
|
| 3374 |
+
"grad_norm": 1.015625,
|
| 3375 |
+
"learning_rate": 7.089581689029203e-06,
|
| 3376 |
+
"loss": 0.0016,
|
| 3377 |
+
"step": 4810
|
| 3378 |
+
},
|
| 3379 |
+
{
|
| 3380 |
+
"epoch": 1.1473517159293791,
|
| 3381 |
+
"grad_norm": 1.015625,
|
| 3382 |
+
"learning_rate": 7.0698500394632995e-06,
|
| 3383 |
+
"loss": 0.0014,
|
| 3384 |
+
"step": 4820
|
| 3385 |
+
},
|
| 3386 |
+
{
|
| 3387 |
+
"epoch": 1.1497321959928586,
|
| 3388 |
+
"grad_norm": 0.79296875,
|
| 3389 |
+
"learning_rate": 7.050118389897395e-06,
|
| 3390 |
+
"loss": 0.0013,
|
| 3391 |
+
"step": 4830
|
| 3392 |
+
},
|
| 3393 |
+
{
|
| 3394 |
+
"epoch": 1.152112676056338,
|
| 3395 |
+
"grad_norm": 0.71875,
|
| 3396 |
+
"learning_rate": 7.0303867403314925e-06,
|
| 3397 |
+
"loss": 0.0014,
|
| 3398 |
+
"step": 4840
|
| 3399 |
+
},
|
| 3400 |
+
{
|
| 3401 |
+
"epoch": 1.1544931561198175,
|
| 3402 |
+
"grad_norm": 0.7734375,
|
| 3403 |
+
"learning_rate": 7.010655090765588e-06,
|
| 3404 |
+
"loss": 0.0017,
|
| 3405 |
+
"step": 4850
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 1.156873636183297,
|
| 3409 |
+
"grad_norm": 0.91015625,
|
| 3410 |
+
"learning_rate": 6.990923441199685e-06,
|
| 3411 |
+
"loss": 0.0014,
|
| 3412 |
+
"step": 4860
|
| 3413 |
+
},
|
| 3414 |
+
{
|
| 3415 |
+
"epoch": 1.1592541162467764,
|
| 3416 |
+
"grad_norm": 1.046875,
|
| 3417 |
+
"learning_rate": 6.97119179163378e-06,
|
| 3418 |
+
"loss": 0.0013,
|
| 3419 |
+
"step": 4870
|
| 3420 |
+
},
|
| 3421 |
+
{
|
| 3422 |
+
"epoch": 1.161634596310256,
|
| 3423 |
+
"grad_norm": 0.75,
|
| 3424 |
+
"learning_rate": 6.951460142067877e-06,
|
| 3425 |
+
"loss": 0.0014,
|
| 3426 |
+
"step": 4880
|
| 3427 |
+
},
|
| 3428 |
+
{
|
| 3429 |
+
"epoch": 1.1640150763737354,
|
| 3430 |
+
"grad_norm": 0.34375,
|
| 3431 |
+
"learning_rate": 6.931728492501973e-06,
|
| 3432 |
+
"loss": 0.0011,
|
| 3433 |
+
"step": 4890
|
| 3434 |
+
},
|
| 3435 |
+
{
|
| 3436 |
+
"epoch": 1.1663955564372148,
|
| 3437 |
+
"grad_norm": 0.78515625,
|
| 3438 |
+
"learning_rate": 6.91199684293607e-06,
|
| 3439 |
+
"loss": 0.0014,
|
| 3440 |
+
"step": 4900
|
| 3441 |
+
},
|
| 3442 |
+
{
|
| 3443 |
+
"epoch": 1.1687760365006943,
|
| 3444 |
+
"grad_norm": 1.078125,
|
| 3445 |
+
"learning_rate": 6.892265193370166e-06,
|
| 3446 |
+
"loss": 0.0011,
|
| 3447 |
+
"step": 4910
|
| 3448 |
+
},
|
| 3449 |
+
{
|
| 3450 |
+
"epoch": 1.1711565165641737,
|
| 3451 |
+
"grad_norm": 0.63671875,
|
| 3452 |
+
"learning_rate": 6.872533543804262e-06,
|
| 3453 |
+
"loss": 0.0011,
|
| 3454 |
+
"step": 4920
|
| 3455 |
+
},
|
| 3456 |
+
{
|
| 3457 |
+
"epoch": 1.1735369966276532,
|
| 3458 |
+
"grad_norm": 0.71875,
|
| 3459 |
+
"learning_rate": 6.852801894238359e-06,
|
| 3460 |
+
"loss": 0.0012,
|
| 3461 |
+
"step": 4930
|
| 3462 |
+
},
|
| 3463 |
+
{
|
| 3464 |
+
"epoch": 1.1759174766911327,
|
| 3465 |
+
"grad_norm": 0.671875,
|
| 3466 |
+
"learning_rate": 6.833070244672455e-06,
|
| 3467 |
+
"loss": 0.001,
|
| 3468 |
+
"step": 4940
|
| 3469 |
+
},
|
| 3470 |
+
{
|
| 3471 |
+
"epoch": 1.1782979567546121,
|
| 3472 |
+
"grad_norm": 0.8203125,
|
| 3473 |
+
"learning_rate": 6.813338595106551e-06,
|
| 3474 |
+
"loss": 0.0077,
|
| 3475 |
+
"step": 4950
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 1.1806784368180916,
|
| 3479 |
+
"grad_norm": 0.380859375,
|
| 3480 |
+
"learning_rate": 6.793606945540647e-06,
|
| 3481 |
+
"loss": 0.0013,
|
| 3482 |
+
"step": 4960
|
| 3483 |
+
},
|
| 3484 |
+
{
|
| 3485 |
+
"epoch": 1.183058916881571,
|
| 3486 |
+
"grad_norm": 1.1796875,
|
| 3487 |
+
"learning_rate": 6.773875295974743e-06,
|
| 3488 |
+
"loss": 0.0016,
|
| 3489 |
+
"step": 4970
|
| 3490 |
+
},
|
| 3491 |
+
{
|
| 3492 |
+
"epoch": 1.1854393969450505,
|
| 3493 |
+
"grad_norm": 0.6640625,
|
| 3494 |
+
"learning_rate": 6.75414364640884e-06,
|
| 3495 |
+
"loss": 0.0014,
|
| 3496 |
+
"step": 4980
|
| 3497 |
+
},
|
| 3498 |
+
{
|
| 3499 |
+
"epoch": 1.18781987700853,
|
| 3500 |
+
"grad_norm": 0.453125,
|
| 3501 |
+
"learning_rate": 6.734411996842936e-06,
|
| 3502 |
+
"loss": 0.0018,
|
| 3503 |
+
"step": 4990
|
| 3504 |
+
},
|
| 3505 |
+
{
|
| 3506 |
+
"epoch": 1.1902003570720094,
|
| 3507 |
+
"grad_norm": 0.98828125,
|
| 3508 |
+
"learning_rate": 6.7146803472770325e-06,
|
| 3509 |
+
"loss": 0.0015,
|
| 3510 |
+
"step": 5000
|
| 3511 |
+
}
|
| 3512 |
+
],
|
| 3513 |
+
"logging_steps": 10,
|
| 3514 |
+
"max_steps": 8402,
|
| 3515 |
+
"num_input_tokens_seen": 0,
|
| 3516 |
+
"num_train_epochs": 2,
|
| 3517 |
+
"save_steps": 200,
|
| 3518 |
+
"stateful_callbacks": {
|
| 3519 |
+
"TrainerControl": {
|
| 3520 |
+
"args": {
|
| 3521 |
+
"should_epoch_stop": false,
|
| 3522 |
+
"should_evaluate": false,
|
| 3523 |
+
"should_log": false,
|
| 3524 |
+
"should_save": true,
|
| 3525 |
+
"should_training_stop": false
|
| 3526 |
+
},
|
| 3527 |
+
"attributes": {}
|
| 3528 |
+
}
|
| 3529 |
+
},
|
| 3530 |
+
"total_flos": 1.1364638902363423e+18,
|
| 3531 |
+
"train_batch_size": 6,
|
| 3532 |
+
"trial_name": null,
|
| 3533 |
+
"trial_params": null
|
| 3534 |
+
}
|