Upload VideoRFT_cold_start_ckpt-268/trainer_state.json with huggingface_hub
Browse files
VideoRFT_cold_start_ckpt-268/trainer_state.json
ADDED
|
@@ -0,0 +1,1910 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 1.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 268,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.0037313432835820895,
|
| 14 |
+
"grad_norm": 12.533227920532227,
|
| 15 |
+
"learning_rate": 0.0,
|
| 16 |
+
"loss": 1.5383,
|
| 17 |
+
"step": 1
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.007462686567164179,
|
| 21 |
+
"grad_norm": 12.856209754943848,
|
| 22 |
+
"learning_rate": 1.111111111111111e-07,
|
| 23 |
+
"loss": 1.5393,
|
| 24 |
+
"step": 2
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.011194029850746268,
|
| 28 |
+
"grad_norm": 12.803821563720703,
|
| 29 |
+
"learning_rate": 2.222222222222222e-07,
|
| 30 |
+
"loss": 1.5559,
|
| 31 |
+
"step": 3
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.014925373134328358,
|
| 35 |
+
"grad_norm": 11.888638496398926,
|
| 36 |
+
"learning_rate": 3.333333333333333e-07,
|
| 37 |
+
"loss": 1.5721,
|
| 38 |
+
"step": 4
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.018656716417910446,
|
| 42 |
+
"grad_norm": 12.17424488067627,
|
| 43 |
+
"learning_rate": 4.444444444444444e-07,
|
| 44 |
+
"loss": 1.5967,
|
| 45 |
+
"step": 5
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.022388059701492536,
|
| 49 |
+
"grad_norm": 11.701398849487305,
|
| 50 |
+
"learning_rate": 5.555555555555555e-07,
|
| 51 |
+
"loss": 1.5471,
|
| 52 |
+
"step": 6
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.026119402985074626,
|
| 56 |
+
"grad_norm": 12.413783073425293,
|
| 57 |
+
"learning_rate": 6.666666666666666e-07,
|
| 58 |
+
"loss": 1.5562,
|
| 59 |
+
"step": 7
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.029850746268656716,
|
| 63 |
+
"grad_norm": 12.070161819458008,
|
| 64 |
+
"learning_rate": 7.777777777777778e-07,
|
| 65 |
+
"loss": 1.4952,
|
| 66 |
+
"step": 8
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.033582089552238806,
|
| 70 |
+
"grad_norm": 11.40228271484375,
|
| 71 |
+
"learning_rate": 8.888888888888888e-07,
|
| 72 |
+
"loss": 1.4731,
|
| 73 |
+
"step": 9
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.03731343283582089,
|
| 77 |
+
"grad_norm": 8.310558319091797,
|
| 78 |
+
"learning_rate": 1e-06,
|
| 79 |
+
"loss": 1.4464,
|
| 80 |
+
"step": 10
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.041044776119402986,
|
| 84 |
+
"grad_norm": 7.320647716522217,
|
| 85 |
+
"learning_rate": 9.999632180371775e-07,
|
| 86 |
+
"loss": 1.4187,
|
| 87 |
+
"step": 11
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.04477611940298507,
|
| 91 |
+
"grad_norm": 7.323790550231934,
|
| 92 |
+
"learning_rate": 9.99852877560361e-07,
|
| 93 |
+
"loss": 1.3817,
|
| 94 |
+
"step": 12
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.048507462686567165,
|
| 98 |
+
"grad_norm": 7.182246208190918,
|
| 99 |
+
"learning_rate": 9.99668994803708e-07,
|
| 100 |
+
"loss": 1.4159,
|
| 101 |
+
"step": 13
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.05223880597014925,
|
| 105 |
+
"grad_norm": 4.682595252990723,
|
| 106 |
+
"learning_rate": 9.99411596821493e-07,
|
| 107 |
+
"loss": 1.3502,
|
| 108 |
+
"step": 14
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.055970149253731345,
|
| 112 |
+
"grad_norm": 4.347743988037109,
|
| 113 |
+
"learning_rate": 9.990807214841286e-07,
|
| 114 |
+
"loss": 1.2414,
|
| 115 |
+
"step": 15
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.05970149253731343,
|
| 119 |
+
"grad_norm": 4.4329729080200195,
|
| 120 |
+
"learning_rate": 9.986764174725918e-07,
|
| 121 |
+
"loss": 1.3029,
|
| 122 |
+
"step": 16
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.06343283582089553,
|
| 126 |
+
"grad_norm": 4.8209710121154785,
|
| 127 |
+
"learning_rate": 9.98198744271263e-07,
|
| 128 |
+
"loss": 1.3404,
|
| 129 |
+
"step": 17
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.06716417910447761,
|
| 133 |
+
"grad_norm": 4.369340419769287,
|
| 134 |
+
"learning_rate": 9.976477721591745e-07,
|
| 135 |
+
"loss": 1.277,
|
| 136 |
+
"step": 18
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.0708955223880597,
|
| 140 |
+
"grad_norm": 4.031601428985596,
|
| 141 |
+
"learning_rate": 9.970235821996689e-07,
|
| 142 |
+
"loss": 1.3244,
|
| 143 |
+
"step": 19
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.07462686567164178,
|
| 147 |
+
"grad_norm": 3.9767301082611084,
|
| 148 |
+
"learning_rate": 9.963262662284735e-07,
|
| 149 |
+
"loss": 1.2718,
|
| 150 |
+
"step": 20
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.07835820895522388,
|
| 154 |
+
"grad_norm": 3.8867998123168945,
|
| 155 |
+
"learning_rate": 9.955559268401892e-07,
|
| 156 |
+
"loss": 1.307,
|
| 157 |
+
"step": 21
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.08208955223880597,
|
| 161 |
+
"grad_norm": 3.440000057220459,
|
| 162 |
+
"learning_rate": 9.947126773731948e-07,
|
| 163 |
+
"loss": 1.3288,
|
| 164 |
+
"step": 22
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.08582089552238806,
|
| 168 |
+
"grad_norm": 3.3316867351531982,
|
| 169 |
+
"learning_rate": 9.937966418929723e-07,
|
| 170 |
+
"loss": 1.2795,
|
| 171 |
+
"step": 23
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.08955223880597014,
|
| 175 |
+
"grad_norm": 3.2227070331573486,
|
| 176 |
+
"learning_rate": 9.928079551738541e-07,
|
| 177 |
+
"loss": 1.2795,
|
| 178 |
+
"step": 24
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.09328358208955224,
|
| 182 |
+
"grad_norm": 3.427421808242798,
|
| 183 |
+
"learning_rate": 9.917467626791925e-07,
|
| 184 |
+
"loss": 1.3093,
|
| 185 |
+
"step": 25
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.09701492537313433,
|
| 189 |
+
"grad_norm": 3.5163064002990723,
|
| 190 |
+
"learning_rate": 9.90613220539959e-07,
|
| 191 |
+
"loss": 1.2345,
|
| 192 |
+
"step": 26
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.10074626865671642,
|
| 196 |
+
"grad_norm": 3.148080587387085,
|
| 197 |
+
"learning_rate": 9.894074955317728e-07,
|
| 198 |
+
"loss": 1.252,
|
| 199 |
+
"step": 27
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.1044776119402985,
|
| 203 |
+
"grad_norm": 3.532881259918213,
|
| 204 |
+
"learning_rate": 9.88129765050364e-07,
|
| 205 |
+
"loss": 1.2532,
|
| 206 |
+
"step": 28
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.10820895522388059,
|
| 210 |
+
"grad_norm": 2.948366403579712,
|
| 211 |
+
"learning_rate": 9.867802170854723e-07,
|
| 212 |
+
"loss": 1.2385,
|
| 213 |
+
"step": 29
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.11194029850746269,
|
| 217 |
+
"grad_norm": 2.9284613132476807,
|
| 218 |
+
"learning_rate": 9.853590501931902e-07,
|
| 219 |
+
"loss": 1.2721,
|
| 220 |
+
"step": 30
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.11567164179104478,
|
| 224 |
+
"grad_norm": 3.193962812423706,
|
| 225 |
+
"learning_rate": 9.838664734667495e-07,
|
| 226 |
+
"loss": 1.1908,
|
| 227 |
+
"step": 31
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.11940298507462686,
|
| 231 |
+
"grad_norm": 3.3310251235961914,
|
| 232 |
+
"learning_rate": 9.82302706505756e-07,
|
| 233 |
+
"loss": 1.244,
|
| 234 |
+
"step": 32
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.12313432835820895,
|
| 238 |
+
"grad_norm": 2.9521822929382324,
|
| 239 |
+
"learning_rate": 9.806679793838827e-07,
|
| 240 |
+
"loss": 1.2565,
|
| 241 |
+
"step": 33
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.12686567164179105,
|
| 245 |
+
"grad_norm": 2.8528637886047363,
|
| 246 |
+
"learning_rate": 9.789625326150189e-07,
|
| 247 |
+
"loss": 1.2801,
|
| 248 |
+
"step": 34
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.13059701492537312,
|
| 252 |
+
"grad_norm": 2.9831106662750244,
|
| 253 |
+
"learning_rate": 9.77186617117883e-07,
|
| 254 |
+
"loss": 1.2023,
|
| 255 |
+
"step": 35
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.13432835820895522,
|
| 259 |
+
"grad_norm": 2.9576382637023926,
|
| 260 |
+
"learning_rate": 9.753404941791061e-07,
|
| 261 |
+
"loss": 1.1699,
|
| 262 |
+
"step": 36
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.13805970149253732,
|
| 266 |
+
"grad_norm": 2.873734474182129,
|
| 267 |
+
"learning_rate": 9.734244354147894e-07,
|
| 268 |
+
"loss": 1.2187,
|
| 269 |
+
"step": 37
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.1417910447761194,
|
| 273 |
+
"grad_norm": 2.7039084434509277,
|
| 274 |
+
"learning_rate": 9.71438722730542e-07,
|
| 275 |
+
"loss": 1.1593,
|
| 276 |
+
"step": 38
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.1455223880597015,
|
| 280 |
+
"grad_norm": 2.669891834259033,
|
| 281 |
+
"learning_rate": 9.693836482800043e-07,
|
| 282 |
+
"loss": 1.2048,
|
| 283 |
+
"step": 39
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.14925373134328357,
|
| 287 |
+
"grad_norm": 2.7127296924591064,
|
| 288 |
+
"learning_rate": 9.672595144218645e-07,
|
| 289 |
+
"loss": 1.2855,
|
| 290 |
+
"step": 40
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.15298507462686567,
|
| 294 |
+
"grad_norm": 2.5859594345092773,
|
| 295 |
+
"learning_rate": 9.650666336753727e-07,
|
| 296 |
+
"loss": 1.2285,
|
| 297 |
+
"step": 41
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.15671641791044777,
|
| 301 |
+
"grad_norm": 2.5001649856567383,
|
| 302 |
+
"learning_rate": 9.628053286743619e-07,
|
| 303 |
+
"loss": 1.2606,
|
| 304 |
+
"step": 42
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.16044776119402984,
|
| 308 |
+
"grad_norm": 2.5785489082336426,
|
| 309 |
+
"learning_rate": 9.604759321197773e-07,
|
| 310 |
+
"loss": 1.2644,
|
| 311 |
+
"step": 43
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.16417910447761194,
|
| 315 |
+
"grad_norm": 2.764547824859619,
|
| 316 |
+
"learning_rate": 9.58078786730729e-07,
|
| 317 |
+
"loss": 1.2096,
|
| 318 |
+
"step": 44
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.16791044776119404,
|
| 322 |
+
"grad_norm": 2.5018906593322754,
|
| 323 |
+
"learning_rate": 9.556142451940679e-07,
|
| 324 |
+
"loss": 1.2021,
|
| 325 |
+
"step": 45
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.17164179104477612,
|
| 329 |
+
"grad_norm": 2.342252254486084,
|
| 330 |
+
"learning_rate": 9.530826701124939e-07,
|
| 331 |
+
"loss": 1.2124,
|
| 332 |
+
"step": 46
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.17537313432835822,
|
| 336 |
+
"grad_norm": 2.6198010444641113,
|
| 337 |
+
"learning_rate": 9.504844339512094e-07,
|
| 338 |
+
"loss": 1.1813,
|
| 339 |
+
"step": 47
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.1791044776119403,
|
| 343 |
+
"grad_norm": 2.470140218734741,
|
| 344 |
+
"learning_rate": 9.478199189831182e-07,
|
| 345 |
+
"loss": 1.1665,
|
| 346 |
+
"step": 48
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.1828358208955224,
|
| 350 |
+
"grad_norm": 2.449758291244507,
|
| 351 |
+
"learning_rate": 9.450895172325822e-07,
|
| 352 |
+
"loss": 1.2287,
|
| 353 |
+
"step": 49
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.1865671641791045,
|
| 357 |
+
"grad_norm": 2.544707775115967,
|
| 358 |
+
"learning_rate": 9.422936304177438e-07,
|
| 359 |
+
"loss": 1.176,
|
| 360 |
+
"step": 50
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.19029850746268656,
|
| 364 |
+
"grad_norm": 2.420337200164795,
|
| 365 |
+
"learning_rate": 9.394326698914228e-07,
|
| 366 |
+
"loss": 1.1846,
|
| 367 |
+
"step": 51
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.19402985074626866,
|
| 371 |
+
"grad_norm": 2.5138134956359863,
|
| 372 |
+
"learning_rate": 9.36507056580594e-07,
|
| 373 |
+
"loss": 1.2089,
|
| 374 |
+
"step": 52
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.19776119402985073,
|
| 378 |
+
"grad_norm": 2.4552905559539795,
|
| 379 |
+
"learning_rate": 9.335172209244575e-07,
|
| 380 |
+
"loss": 1.1484,
|
| 381 |
+
"step": 53
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.20149253731343283,
|
| 385 |
+
"grad_norm": 2.3464629650115967,
|
| 386 |
+
"learning_rate": 9.304636028111092e-07,
|
| 387 |
+
"loss": 1.2083,
|
| 388 |
+
"step": 54
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.20522388059701493,
|
| 392 |
+
"grad_norm": 2.4915719032287598,
|
| 393 |
+
"learning_rate": 9.273466515128208e-07,
|
| 394 |
+
"loss": 1.2525,
|
| 395 |
+
"step": 55
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.208955223880597,
|
| 399 |
+
"grad_norm": 2.3456897735595703,
|
| 400 |
+
"learning_rate": 9.241668256199391e-07,
|
| 401 |
+
"loss": 1.2089,
|
| 402 |
+
"step": 56
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.2126865671641791,
|
| 406 |
+
"grad_norm": 2.378307819366455,
|
| 407 |
+
"learning_rate": 9.209245929734156e-07,
|
| 408 |
+
"loss": 1.2086,
|
| 409 |
+
"step": 57
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.21641791044776118,
|
| 413 |
+
"grad_norm": 2.4048099517822266,
|
| 414 |
+
"learning_rate": 9.176204305959725e-07,
|
| 415 |
+
"loss": 1.2151,
|
| 416 |
+
"step": 58
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.22014925373134328,
|
| 420 |
+
"grad_norm": 2.317758560180664,
|
| 421 |
+
"learning_rate": 9.14254824621921e-07,
|
| 422 |
+
"loss": 1.1891,
|
| 423 |
+
"step": 59
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.22388059701492538,
|
| 427 |
+
"grad_norm": 2.2456841468811035,
|
| 428 |
+
"learning_rate": 9.108282702256364e-07,
|
| 429 |
+
"loss": 1.1712,
|
| 430 |
+
"step": 60
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.22761194029850745,
|
| 434 |
+
"grad_norm": 2.3463008403778076,
|
| 435 |
+
"learning_rate": 9.073412715487043e-07,
|
| 436 |
+
"loss": 1.1491,
|
| 437 |
+
"step": 61
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.23134328358208955,
|
| 441 |
+
"grad_norm": 2.3797104358673096,
|
| 442 |
+
"learning_rate": 9.037943416257474e-07,
|
| 443 |
+
"loss": 1.176,
|
| 444 |
+
"step": 62
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.23507462686567165,
|
| 448 |
+
"grad_norm": 2.4389562606811523,
|
| 449 |
+
"learning_rate": 9.001880023089441e-07,
|
| 450 |
+
"loss": 1.2306,
|
| 451 |
+
"step": 63
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.23880597014925373,
|
| 455 |
+
"grad_norm": 2.3434224128723145,
|
| 456 |
+
"learning_rate": 8.965227841912488e-07,
|
| 457 |
+
"loss": 1.2211,
|
| 458 |
+
"step": 64
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.24253731343283583,
|
| 462 |
+
"grad_norm": 2.433840751647949,
|
| 463 |
+
"learning_rate": 8.92799226528328e-07,
|
| 464 |
+
"loss": 1.1713,
|
| 465 |
+
"step": 65
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.2462686567164179,
|
| 469 |
+
"grad_norm": 2.4185891151428223,
|
| 470 |
+
"learning_rate": 8.890178771592197e-07,
|
| 471 |
+
"loss": 1.2018,
|
| 472 |
+
"step": 66
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.25,
|
| 476 |
+
"grad_norm": 2.412198305130005,
|
| 477 |
+
"learning_rate": 8.851792924257315e-07,
|
| 478 |
+
"loss": 1.1599,
|
| 479 |
+
"step": 67
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.2537313432835821,
|
| 483 |
+
"grad_norm": 2.3101119995117188,
|
| 484 |
+
"learning_rate": 8.812840370905872e-07,
|
| 485 |
+
"loss": 1.1489,
|
| 486 |
+
"step": 68
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.2574626865671642,
|
| 490 |
+
"grad_norm": 2.3638556003570557,
|
| 491 |
+
"learning_rate": 8.773326842543347e-07,
|
| 492 |
+
"loss": 1.195,
|
| 493 |
+
"step": 69
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.26119402985074625,
|
| 497 |
+
"grad_norm": 2.266350746154785,
|
| 498 |
+
"learning_rate": 8.733258152710261e-07,
|
| 499 |
+
"loss": 1.2228,
|
| 500 |
+
"step": 70
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.26492537313432835,
|
| 504 |
+
"grad_norm": 2.3603343963623047,
|
| 505 |
+
"learning_rate": 8.692640196626857e-07,
|
| 506 |
+
"loss": 1.1813,
|
| 507 |
+
"step": 71
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.26865671641791045,
|
| 511 |
+
"grad_norm": 2.4589076042175293,
|
| 512 |
+
"learning_rate": 8.651478950325737e-07,
|
| 513 |
+
"loss": 1.1605,
|
| 514 |
+
"step": 72
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.27238805970149255,
|
| 518 |
+
"grad_norm": 2.421117067337036,
|
| 519 |
+
"learning_rate": 8.609780469772621e-07,
|
| 520 |
+
"loss": 1.1935,
|
| 521 |
+
"step": 73
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.27611940298507465,
|
| 525 |
+
"grad_norm": 2.3559322357177734,
|
| 526 |
+
"learning_rate": 8.567550889975361e-07,
|
| 527 |
+
"loss": 1.1674,
|
| 528 |
+
"step": 74
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.2798507462686567,
|
| 532 |
+
"grad_norm": 2.3851828575134277,
|
| 533 |
+
"learning_rate": 8.524796424081291e-07,
|
| 534 |
+
"loss": 1.1976,
|
| 535 |
+
"step": 75
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.2835820895522388,
|
| 539 |
+
"grad_norm": 2.437889575958252,
|
| 540 |
+
"learning_rate": 8.48152336246311e-07,
|
| 541 |
+
"loss": 1.1456,
|
| 542 |
+
"step": 76
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.2873134328358209,
|
| 546 |
+
"grad_norm": 2.3987743854522705,
|
| 547 |
+
"learning_rate": 8.437738071793394e-07,
|
| 548 |
+
"loss": 1.1622,
|
| 549 |
+
"step": 77
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.291044776119403,
|
| 553 |
+
"grad_norm": 2.2897541522979736,
|
| 554 |
+
"learning_rate": 8.393446994107876e-07,
|
| 555 |
+
"loss": 1.1765,
|
| 556 |
+
"step": 78
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.2947761194029851,
|
| 560 |
+
"grad_norm": 2.3303351402282715,
|
| 561 |
+
"learning_rate": 8.348656645857648e-07,
|
| 562 |
+
"loss": 1.1645,
|
| 563 |
+
"step": 79
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.29850746268656714,
|
| 567 |
+
"grad_norm": 2.247779607772827,
|
| 568 |
+
"learning_rate": 8.303373616950406e-07,
|
| 569 |
+
"loss": 1.1443,
|
| 570 |
+
"step": 80
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.30223880597014924,
|
| 574 |
+
"grad_norm": 2.2308011054992676,
|
| 575 |
+
"learning_rate": 8.257604569780896e-07,
|
| 576 |
+
"loss": 1.1206,
|
| 577 |
+
"step": 81
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.30597014925373134,
|
| 581 |
+
"grad_norm": 2.3626465797424316,
|
| 582 |
+
"learning_rate": 8.211356238250679e-07,
|
| 583 |
+
"loss": 1.1578,
|
| 584 |
+
"step": 82
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.30970149253731344,
|
| 588 |
+
"grad_norm": 2.356580972671509,
|
| 589 |
+
"learning_rate": 8.164635426777403e-07,
|
| 590 |
+
"loss": 1.165,
|
| 591 |
+
"step": 83
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.31343283582089554,
|
| 595 |
+
"grad_norm": 2.278188467025757,
|
| 596 |
+
"learning_rate": 8.117449009293668e-07,
|
| 597 |
+
"loss": 1.1798,
|
| 598 |
+
"step": 84
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.31716417910447764,
|
| 602 |
+
"grad_norm": 2.406306743621826,
|
| 603 |
+
"learning_rate": 8.069803928235687e-07,
|
| 604 |
+
"loss": 1.104,
|
| 605 |
+
"step": 85
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.3208955223880597,
|
| 609 |
+
"grad_norm": 2.3110954761505127,
|
| 610 |
+
"learning_rate": 8.021707193521865e-07,
|
| 611 |
+
"loss": 1.2179,
|
| 612 |
+
"step": 86
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.3246268656716418,
|
| 616 |
+
"grad_norm": 2.272700548171997,
|
| 617 |
+
"learning_rate": 7.973165881521433e-07,
|
| 618 |
+
"loss": 1.2139,
|
| 619 |
+
"step": 87
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.3283582089552239,
|
| 623 |
+
"grad_norm": 2.4564621448516846,
|
| 624 |
+
"learning_rate": 7.924187134013323e-07,
|
| 625 |
+
"loss": 1.1517,
|
| 626 |
+
"step": 88
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.332089552238806,
|
| 630 |
+
"grad_norm": 2.270390748977661,
|
| 631 |
+
"learning_rate": 7.874778157135415e-07,
|
| 632 |
+
"loss": 1.1607,
|
| 633 |
+
"step": 89
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.3358208955223881,
|
| 637 |
+
"grad_norm": 2.495291233062744,
|
| 638 |
+
"learning_rate": 7.824946220324311e-07,
|
| 639 |
+
"loss": 1.1512,
|
| 640 |
+
"step": 90
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.33955223880597013,
|
| 644 |
+
"grad_norm": 2.217815637588501,
|
| 645 |
+
"learning_rate": 7.774698655245801e-07,
|
| 646 |
+
"loss": 1.1866,
|
| 647 |
+
"step": 91
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.34328358208955223,
|
| 651 |
+
"grad_norm": 2.2916743755340576,
|
| 652 |
+
"learning_rate": 7.724042854716168e-07,
|
| 653 |
+
"loss": 1.168,
|
| 654 |
+
"step": 92
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.34701492537313433,
|
| 658 |
+
"grad_norm": 2.2644033432006836,
|
| 659 |
+
"learning_rate": 7.672986271614499e-07,
|
| 660 |
+
"loss": 1.157,
|
| 661 |
+
"step": 93
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.35074626865671643,
|
| 665 |
+
"grad_norm": 2.3119850158691406,
|
| 666 |
+
"learning_rate": 7.621536417786158e-07,
|
| 667 |
+
"loss": 1.1629,
|
| 668 |
+
"step": 94
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.35447761194029853,
|
| 672 |
+
"grad_norm": 2.192704916000366,
|
| 673 |
+
"learning_rate": 7.569700862937589e-07,
|
| 674 |
+
"loss": 1.1331,
|
| 675 |
+
"step": 95
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.3582089552238806,
|
| 679 |
+
"grad_norm": 2.263261318206787,
|
| 680 |
+
"learning_rate": 7.517487233522599e-07,
|
| 681 |
+
"loss": 1.1307,
|
| 682 |
+
"step": 96
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.3619402985074627,
|
| 686 |
+
"grad_norm": 2.248213768005371,
|
| 687 |
+
"learning_rate": 7.46490321162029e-07,
|
| 688 |
+
"loss": 1.1816,
|
| 689 |
+
"step": 97
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.3656716417910448,
|
| 693 |
+
"grad_norm": 2.3285648822784424,
|
| 694 |
+
"learning_rate": 7.411956533804818e-07,
|
| 695 |
+
"loss": 1.173,
|
| 696 |
+
"step": 98
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.3694029850746269,
|
| 700 |
+
"grad_norm": 2.1573896408081055,
|
| 701 |
+
"learning_rate": 7.358654990007122e-07,
|
| 702 |
+
"loss": 1.135,
|
| 703 |
+
"step": 99
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.373134328358209,
|
| 707 |
+
"grad_norm": 2.5231122970581055,
|
| 708 |
+
"learning_rate": 7.305006422368811e-07,
|
| 709 |
+
"loss": 1.1905,
|
| 710 |
+
"step": 100
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.376865671641791,
|
| 714 |
+
"grad_norm": 2.312589168548584,
|
| 715 |
+
"learning_rate": 7.251018724088366e-07,
|
| 716 |
+
"loss": 1.1561,
|
| 717 |
+
"step": 101
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.3805970149253731,
|
| 721 |
+
"grad_norm": 2.2655773162841797,
|
| 722 |
+
"learning_rate": 7.196699838259834e-07,
|
| 723 |
+
"loss": 1.1654,
|
| 724 |
+
"step": 102
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.3843283582089552,
|
| 728 |
+
"grad_norm": 2.186748504638672,
|
| 729 |
+
"learning_rate": 7.142057756704167e-07,
|
| 730 |
+
"loss": 1.1503,
|
| 731 |
+
"step": 103
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.3880597014925373,
|
| 735 |
+
"grad_norm": 2.224595546722412,
|
| 736 |
+
"learning_rate": 7.08710051879342e-07,
|
| 737 |
+
"loss": 1.1292,
|
| 738 |
+
"step": 104
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.3917910447761194,
|
| 742 |
+
"grad_norm": 2.307724952697754,
|
| 743 |
+
"learning_rate": 7.031836210267914e-07,
|
| 744 |
+
"loss": 1.119,
|
| 745 |
+
"step": 105
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.39552238805970147,
|
| 749 |
+
"grad_norm": 2.4213106632232666,
|
| 750 |
+
"learning_rate": 6.976272962046619e-07,
|
| 751 |
+
"loss": 1.2232,
|
| 752 |
+
"step": 106
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.39925373134328357,
|
| 756 |
+
"grad_norm": 2.2337331771850586,
|
| 757 |
+
"learning_rate": 6.920418949030855e-07,
|
| 758 |
+
"loss": 1.1748,
|
| 759 |
+
"step": 107
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.40298507462686567,
|
| 763 |
+
"grad_norm": 2.185559034347534,
|
| 764 |
+
"learning_rate": 6.864282388901543e-07,
|
| 765 |
+
"loss": 1.123,
|
| 766 |
+
"step": 108
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.40671641791044777,
|
| 770 |
+
"grad_norm": 2.2927253246307373,
|
| 771 |
+
"learning_rate": 6.807871540910154e-07,
|
| 772 |
+
"loss": 1.1138,
|
| 773 |
+
"step": 109
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.41044776119402987,
|
| 777 |
+
"grad_norm": 2.22214937210083,
|
| 778 |
+
"learning_rate": 6.751194704663543e-07,
|
| 779 |
+
"loss": 1.1275,
|
| 780 |
+
"step": 110
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.4141791044776119,
|
| 784 |
+
"grad_norm": 2.3897905349731445,
|
| 785 |
+
"learning_rate": 6.694260218902844e-07,
|
| 786 |
+
"loss": 1.1811,
|
| 787 |
+
"step": 111
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.417910447761194,
|
| 791 |
+
"grad_norm": 2.2676897048950195,
|
| 792 |
+
"learning_rate": 6.637076460276611e-07,
|
| 793 |
+
"loss": 1.1712,
|
| 794 |
+
"step": 112
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.4216417910447761,
|
| 798 |
+
"grad_norm": 2.2889180183410645,
|
| 799 |
+
"learning_rate": 6.57965184210838e-07,
|
| 800 |
+
"loss": 1.127,
|
| 801 |
+
"step": 113
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.4253731343283582,
|
| 805 |
+
"grad_norm": 2.279909610748291,
|
| 806 |
+
"learning_rate": 6.521994813158834e-07,
|
| 807 |
+
"loss": 1.1843,
|
| 808 |
+
"step": 114
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.4291044776119403,
|
| 812 |
+
"grad_norm": 2.211841106414795,
|
| 813 |
+
"learning_rate": 6.464113856382751e-07,
|
| 814 |
+
"loss": 1.1708,
|
| 815 |
+
"step": 115
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.43283582089552236,
|
| 819 |
+
"grad_norm": 2.2442774772644043,
|
| 820 |
+
"learning_rate": 6.406017487680937e-07,
|
| 821 |
+
"loss": 1.1481,
|
| 822 |
+
"step": 116
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.43656716417910446,
|
| 826 |
+
"grad_norm": 2.257807731628418,
|
| 827 |
+
"learning_rate": 6.347714254647283e-07,
|
| 828 |
+
"loss": 1.1584,
|
| 829 |
+
"step": 117
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.44029850746268656,
|
| 833 |
+
"grad_norm": 2.2966856956481934,
|
| 834 |
+
"learning_rate": 6.289212735311189e-07,
|
| 835 |
+
"loss": 1.1605,
|
| 836 |
+
"step": 118
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.44402985074626866,
|
| 840 |
+
"grad_norm": 2.299578905105591,
|
| 841 |
+
"learning_rate": 6.230521536875493e-07,
|
| 842 |
+
"loss": 1.1474,
|
| 843 |
+
"step": 119
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.44776119402985076,
|
| 847 |
+
"grad_norm": 2.2165260314941406,
|
| 848 |
+
"learning_rate": 6.171649294450112e-07,
|
| 849 |
+
"loss": 1.0987,
|
| 850 |
+
"step": 120
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.45149253731343286,
|
| 854 |
+
"grad_norm": 2.283634662628174,
|
| 855 |
+
"learning_rate": 6.112604669781572e-07,
|
| 856 |
+
"loss": 1.1926,
|
| 857 |
+
"step": 121
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.4552238805970149,
|
| 861 |
+
"grad_norm": 2.2140541076660156,
|
| 862 |
+
"learning_rate": 6.053396349978631e-07,
|
| 863 |
+
"loss": 1.1477,
|
| 864 |
+
"step": 122
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.458955223880597,
|
| 868 |
+
"grad_norm": 2.245781421661377,
|
| 869 |
+
"learning_rate": 5.994033046234162e-07,
|
| 870 |
+
"loss": 1.1804,
|
| 871 |
+
"step": 123
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.4626865671641791,
|
| 875 |
+
"grad_norm": 2.357011079788208,
|
| 876 |
+
"learning_rate": 5.934523492543488e-07,
|
| 877 |
+
"loss": 1.1231,
|
| 878 |
+
"step": 124
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.4664179104477612,
|
| 882 |
+
"grad_norm": 2.3178062438964844,
|
| 883 |
+
"learning_rate": 5.874876444419377e-07,
|
| 884 |
+
"loss": 1.2086,
|
| 885 |
+
"step": 125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.4701492537313433,
|
| 889 |
+
"grad_norm": 2.1921329498291016,
|
| 890 |
+
"learning_rate": 5.815100677603853e-07,
|
| 891 |
+
"loss": 1.1265,
|
| 892 |
+
"step": 126
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.47388059701492535,
|
| 896 |
+
"grad_norm": 2.1457953453063965,
|
| 897 |
+
"learning_rate": 5.755204986777049e-07,
|
| 898 |
+
"loss": 1.1093,
|
| 899 |
+
"step": 127
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.47761194029850745,
|
| 903 |
+
"grad_norm": 2.2648348808288574,
|
| 904 |
+
"learning_rate": 5.695198184263258e-07,
|
| 905 |
+
"loss": 1.1776,
|
| 906 |
+
"step": 128
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.48134328358208955,
|
| 910 |
+
"grad_norm": 2.249004602432251,
|
| 911 |
+
"learning_rate": 5.635089098734393e-07,
|
| 912 |
+
"loss": 1.1445,
|
| 913 |
+
"step": 129
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.48507462686567165,
|
| 917 |
+
"grad_norm": 2.2214391231536865,
|
| 918 |
+
"learning_rate": 5.574886573911055e-07,
|
| 919 |
+
"loss": 1.1263,
|
| 920 |
+
"step": 130
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.48880597014925375,
|
| 924 |
+
"grad_norm": 2.2356677055358887,
|
| 925 |
+
"learning_rate": 5.514599467261362e-07,
|
| 926 |
+
"loss": 1.1324,
|
| 927 |
+
"step": 131
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.4925373134328358,
|
| 931 |
+
"grad_norm": 2.17228627204895,
|
| 932 |
+
"learning_rate": 5.454236648697776e-07,
|
| 933 |
+
"loss": 1.1869,
|
| 934 |
+
"step": 132
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.4962686567164179,
|
| 938 |
+
"grad_norm": 2.2196078300476074,
|
| 939 |
+
"learning_rate": 5.393806999272089e-07,
|
| 940 |
+
"loss": 1.0961,
|
| 941 |
+
"step": 133
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.5,
|
| 945 |
+
"grad_norm": 2.144622325897217,
|
| 946 |
+
"learning_rate": 5.333319409868776e-07,
|
| 947 |
+
"loss": 1.137,
|
| 948 |
+
"step": 134
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 0.503731343283582,
|
| 952 |
+
"grad_norm": 2.4410345554351807,
|
| 953 |
+
"learning_rate": 5.272782779896897e-07,
|
| 954 |
+
"loss": 1.1667,
|
| 955 |
+
"step": 135
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 0.5074626865671642,
|
| 959 |
+
"grad_norm": 2.2752339839935303,
|
| 960 |
+
"learning_rate": 5.212206015980741e-07,
|
| 961 |
+
"loss": 1.1675,
|
| 962 |
+
"step": 136
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.5111940298507462,
|
| 966 |
+
"grad_norm": 2.130159854888916,
|
| 967 |
+
"learning_rate": 5.151598030649425e-07,
|
| 968 |
+
"loss": 1.1051,
|
| 969 |
+
"step": 137
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.5149253731343284,
|
| 973 |
+
"grad_norm": 2.1883411407470703,
|
| 974 |
+
"learning_rate": 5.090967741025598e-07,
|
| 975 |
+
"loss": 1.1488,
|
| 976 |
+
"step": 138
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 0.5186567164179104,
|
| 980 |
+
"grad_norm": 2.3985397815704346,
|
| 981 |
+
"learning_rate": 5.030324067513498e-07,
|
| 982 |
+
"loss": 1.1065,
|
| 983 |
+
"step": 139
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 0.5223880597014925,
|
| 987 |
+
"grad_norm": 2.1282200813293457,
|
| 988 |
+
"learning_rate": 4.969675932486503e-07,
|
| 989 |
+
"loss": 1.1379,
|
| 990 |
+
"step": 140
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 0.5261194029850746,
|
| 994 |
+
"grad_norm": 2.305114507675171,
|
| 995 |
+
"learning_rate": 4.909032258974402e-07,
|
| 996 |
+
"loss": 1.1841,
|
| 997 |
+
"step": 141
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 0.5298507462686567,
|
| 1001 |
+
"grad_norm": 2.107046604156494,
|
| 1002 |
+
"learning_rate": 4.848401969350576e-07,
|
| 1003 |
+
"loss": 1.1454,
|
| 1004 |
+
"step": 142
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.5335820895522388,
|
| 1008 |
+
"grad_norm": 2.1780588626861572,
|
| 1009 |
+
"learning_rate": 4.787793984019259e-07,
|
| 1010 |
+
"loss": 1.148,
|
| 1011 |
+
"step": 143
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.5373134328358209,
|
| 1015 |
+
"grad_norm": 2.2580158710479736,
|
| 1016 |
+
"learning_rate": 4.7272172201031047e-07,
|
| 1017 |
+
"loss": 1.1044,
|
| 1018 |
+
"step": 144
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 0.5410447761194029,
|
| 1022 |
+
"grad_norm": 2.2039361000061035,
|
| 1023 |
+
"learning_rate": 4.6666805901312243e-07,
|
| 1024 |
+
"loss": 1.1436,
|
| 1025 |
+
"step": 145
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 0.5447761194029851,
|
| 1029 |
+
"grad_norm": 2.189901351928711,
|
| 1030 |
+
"learning_rate": 4.6061930007279127e-07,
|
| 1031 |
+
"loss": 1.1706,
|
| 1032 |
+
"step": 146
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 0.5485074626865671,
|
| 1036 |
+
"grad_norm": 2.388174533843994,
|
| 1037 |
+
"learning_rate": 4.5457633513022235e-07,
|
| 1038 |
+
"loss": 1.1644,
|
| 1039 |
+
"step": 147
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 0.5522388059701493,
|
| 1043 |
+
"grad_norm": 2.34720778465271,
|
| 1044 |
+
"learning_rate": 4.4854005327386375e-07,
|
| 1045 |
+
"loss": 1.1303,
|
| 1046 |
+
"step": 148
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.5559701492537313,
|
| 1050 |
+
"grad_norm": 2.378136396408081,
|
| 1051 |
+
"learning_rate": 4.4251134260889446e-07,
|
| 1052 |
+
"loss": 1.1729,
|
| 1053 |
+
"step": 149
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.5597014925373134,
|
| 1057 |
+
"grad_norm": 2.287811040878296,
|
| 1058 |
+
"learning_rate": 4.364910901265606e-07,
|
| 1059 |
+
"loss": 1.136,
|
| 1060 |
+
"step": 150
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 0.5634328358208955,
|
| 1064 |
+
"grad_norm": 2.322864294052124,
|
| 1065 |
+
"learning_rate": 4.304801815736743e-07,
|
| 1066 |
+
"loss": 1.1419,
|
| 1067 |
+
"step": 151
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 0.5671641791044776,
|
| 1071 |
+
"grad_norm": 2.352736711502075,
|
| 1072 |
+
"learning_rate": 4.24479501322295e-07,
|
| 1073 |
+
"loss": 1.1571,
|
| 1074 |
+
"step": 152
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.5708955223880597,
|
| 1078 |
+
"grad_norm": 2.304757833480835,
|
| 1079 |
+
"learning_rate": 4.1848993223961466e-07,
|
| 1080 |
+
"loss": 1.1235,
|
| 1081 |
+
"step": 153
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 0.5746268656716418,
|
| 1085 |
+
"grad_norm": 2.2815308570861816,
|
| 1086 |
+
"learning_rate": 4.1251235555806236e-07,
|
| 1087 |
+
"loss": 1.1522,
|
| 1088 |
+
"step": 154
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.5783582089552238,
|
| 1092 |
+
"grad_norm": 2.3194005489349365,
|
| 1093 |
+
"learning_rate": 4.0654765074565125e-07,
|
| 1094 |
+
"loss": 1.185,
|
| 1095 |
+
"step": 155
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.582089552238806,
|
| 1099 |
+
"grad_norm": 2.444715738296509,
|
| 1100 |
+
"learning_rate": 4.005966953765839e-07,
|
| 1101 |
+
"loss": 1.1841,
|
| 1102 |
+
"step": 156
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 0.585820895522388,
|
| 1106 |
+
"grad_norm": 2.1941444873809814,
|
| 1107 |
+
"learning_rate": 3.94660365002137e-07,
|
| 1108 |
+
"loss": 1.1752,
|
| 1109 |
+
"step": 157
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 0.5895522388059702,
|
| 1113 |
+
"grad_norm": 2.2771668434143066,
|
| 1114 |
+
"learning_rate": 3.8873953302184283e-07,
|
| 1115 |
+
"loss": 1.1601,
|
| 1116 |
+
"step": 158
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 0.5932835820895522,
|
| 1120 |
+
"grad_norm": 2.2500874996185303,
|
| 1121 |
+
"learning_rate": 3.828350705549888e-07,
|
| 1122 |
+
"loss": 1.1226,
|
| 1123 |
+
"step": 159
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 0.5970149253731343,
|
| 1127 |
+
"grad_norm": 2.1883161067962646,
|
| 1128 |
+
"learning_rate": 3.769478463124506e-07,
|
| 1129 |
+
"loss": 1.171,
|
| 1130 |
+
"step": 160
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.6007462686567164,
|
| 1134 |
+
"grad_norm": 2.2116432189941406,
|
| 1135 |
+
"learning_rate": 3.710787264688811e-07,
|
| 1136 |
+
"loss": 1.1565,
|
| 1137 |
+
"step": 161
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.6044776119402985,
|
| 1141 |
+
"grad_norm": 2.2853686809539795,
|
| 1142 |
+
"learning_rate": 3.652285745352717e-07,
|
| 1143 |
+
"loss": 1.0897,
|
| 1144 |
+
"step": 162
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 0.6082089552238806,
|
| 1148 |
+
"grad_norm": 2.1353137493133545,
|
| 1149 |
+
"learning_rate": 3.593982512319063e-07,
|
| 1150 |
+
"loss": 1.1186,
|
| 1151 |
+
"step": 163
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 0.6119402985074627,
|
| 1155 |
+
"grad_norm": 2.3534247875213623,
|
| 1156 |
+
"learning_rate": 3.535886143617248e-07,
|
| 1157 |
+
"loss": 1.0777,
|
| 1158 |
+
"step": 164
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 0.6156716417910447,
|
| 1162 |
+
"grad_norm": 2.3094656467437744,
|
| 1163 |
+
"learning_rate": 3.478005186841167e-07,
|
| 1164 |
+
"loss": 1.1503,
|
| 1165 |
+
"step": 165
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 0.6194029850746269,
|
| 1169 |
+
"grad_norm": 2.3270974159240723,
|
| 1170 |
+
"learning_rate": 3.4203481578916196e-07,
|
| 1171 |
+
"loss": 1.1342,
|
| 1172 |
+
"step": 166
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.6231343283582089,
|
| 1176 |
+
"grad_norm": 2.2708523273468018,
|
| 1177 |
+
"learning_rate": 3.362923539723389e-07,
|
| 1178 |
+
"loss": 1.1387,
|
| 1179 |
+
"step": 167
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.6268656716417911,
|
| 1183 |
+
"grad_norm": 2.1796786785125732,
|
| 1184 |
+
"learning_rate": 3.3057397810971565e-07,
|
| 1185 |
+
"loss": 1.142,
|
| 1186 |
+
"step": 168
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 0.6305970149253731,
|
| 1190 |
+
"grad_norm": 2.2308785915374756,
|
| 1191 |
+
"learning_rate": 3.2488052953364573e-07,
|
| 1192 |
+
"loss": 1.1702,
|
| 1193 |
+
"step": 169
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 0.6343283582089553,
|
| 1197 |
+
"grad_norm": 2.259995698928833,
|
| 1198 |
+
"learning_rate": 3.192128459089846e-07,
|
| 1199 |
+
"loss": 1.1362,
|
| 1200 |
+
"step": 170
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 0.6380597014925373,
|
| 1204 |
+
"grad_norm": 2.3103954792022705,
|
| 1205 |
+
"learning_rate": 3.135717611098457e-07,
|
| 1206 |
+
"loss": 1.1067,
|
| 1207 |
+
"step": 171
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 0.6417910447761194,
|
| 1211 |
+
"grad_norm": 2.234175205230713,
|
| 1212 |
+
"learning_rate": 3.0795810509691457e-07,
|
| 1213 |
+
"loss": 1.2034,
|
| 1214 |
+
"step": 172
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.6455223880597015,
|
| 1218 |
+
"grad_norm": 2.2961885929107666,
|
| 1219 |
+
"learning_rate": 3.023727037953382e-07,
|
| 1220 |
+
"loss": 1.1075,
|
| 1221 |
+
"step": 173
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.6492537313432836,
|
| 1225 |
+
"grad_norm": 2.269788980484009,
|
| 1226 |
+
"learning_rate": 2.9681637897320864e-07,
|
| 1227 |
+
"loss": 1.178,
|
| 1228 |
+
"step": 174
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 0.6529850746268657,
|
| 1232 |
+
"grad_norm": 2.201759099960327,
|
| 1233 |
+
"learning_rate": 2.912899481206582e-07,
|
| 1234 |
+
"loss": 1.1494,
|
| 1235 |
+
"step": 175
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 0.6567164179104478,
|
| 1239 |
+
"grad_norm": 2.200390338897705,
|
| 1240 |
+
"learning_rate": 2.857942243295831e-07,
|
| 1241 |
+
"loss": 1.1496,
|
| 1242 |
+
"step": 176
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 0.6604477611940298,
|
| 1246 |
+
"grad_norm": 2.2247750759124756,
|
| 1247 |
+
"learning_rate": 2.803300161740166e-07,
|
| 1248 |
+
"loss": 1.1213,
|
| 1249 |
+
"step": 177
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 0.664179104477612,
|
| 1253 |
+
"grad_norm": 2.380087375640869,
|
| 1254 |
+
"learning_rate": 2.748981275911633e-07,
|
| 1255 |
+
"loss": 1.1356,
|
| 1256 |
+
"step": 178
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.667910447761194,
|
| 1260 |
+
"grad_norm": 2.4810879230499268,
|
| 1261 |
+
"learning_rate": 2.6949935776311895e-07,
|
| 1262 |
+
"loss": 1.0748,
|
| 1263 |
+
"step": 179
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 0.6716417910447762,
|
| 1267 |
+
"grad_norm": 2.3208250999450684,
|
| 1268 |
+
"learning_rate": 2.641345009992878e-07,
|
| 1269 |
+
"loss": 1.1345,
|
| 1270 |
+
"step": 180
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 0.6753731343283582,
|
| 1274 |
+
"grad_norm": 2.3277993202209473,
|
| 1275 |
+
"learning_rate": 2.588043466195182e-07,
|
| 1276 |
+
"loss": 1.1355,
|
| 1277 |
+
"step": 181
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 0.6791044776119403,
|
| 1281 |
+
"grad_norm": 2.25797438621521,
|
| 1282 |
+
"learning_rate": 2.535096788379709e-07,
|
| 1283 |
+
"loss": 1.1447,
|
| 1284 |
+
"step": 182
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.6828358208955224,
|
| 1288 |
+
"grad_norm": 2.347884178161621,
|
| 1289 |
+
"learning_rate": 2.4825127664774006e-07,
|
| 1290 |
+
"loss": 1.1625,
|
| 1291 |
+
"step": 183
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 0.6865671641791045,
|
| 1295 |
+
"grad_norm": 2.4898929595947266,
|
| 1296 |
+
"learning_rate": 2.4302991370624106e-07,
|
| 1297 |
+
"loss": 1.1616,
|
| 1298 |
+
"step": 184
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.6902985074626866,
|
| 1302 |
+
"grad_norm": 2.4398884773254395,
|
| 1303 |
+
"learning_rate": 2.378463582213842e-07,
|
| 1304 |
+
"loss": 1.1143,
|
| 1305 |
+
"step": 185
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 0.6940298507462687,
|
| 1309 |
+
"grad_norm": 2.2198681831359863,
|
| 1310 |
+
"learning_rate": 2.327013728385502e-07,
|
| 1311 |
+
"loss": 1.1477,
|
| 1312 |
+
"step": 186
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 0.6977611940298507,
|
| 1316 |
+
"grad_norm": 2.267277479171753,
|
| 1317 |
+
"learning_rate": 2.2759571452838322e-07,
|
| 1318 |
+
"loss": 1.0987,
|
| 1319 |
+
"step": 187
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 0.7014925373134329,
|
| 1323 |
+
"grad_norm": 2.1941375732421875,
|
| 1324 |
+
"learning_rate": 2.225301344754199e-07,
|
| 1325 |
+
"loss": 1.1439,
|
| 1326 |
+
"step": 188
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 0.7052238805970149,
|
| 1330 |
+
"grad_norm": 2.2704875469207764,
|
| 1331 |
+
"learning_rate": 2.1750537796756897e-07,
|
| 1332 |
+
"loss": 1.0848,
|
| 1333 |
+
"step": 189
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 0.7089552238805971,
|
| 1337 |
+
"grad_norm": 2.16701078414917,
|
| 1338 |
+
"learning_rate": 2.1252218428645845e-07,
|
| 1339 |
+
"loss": 1.144,
|
| 1340 |
+
"step": 190
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.7126865671641791,
|
| 1344 |
+
"grad_norm": 2.198181629180908,
|
| 1345 |
+
"learning_rate": 2.0758128659866768e-07,
|
| 1346 |
+
"loss": 1.1051,
|
| 1347 |
+
"step": 191
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 0.7164179104477612,
|
| 1351 |
+
"grad_norm": 2.232703924179077,
|
| 1352 |
+
"learning_rate": 2.026834118478567e-07,
|
| 1353 |
+
"loss": 1.2096,
|
| 1354 |
+
"step": 192
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 0.7201492537313433,
|
| 1358 |
+
"grad_norm": 2.25785231590271,
|
| 1359 |
+
"learning_rate": 1.978292806478134e-07,
|
| 1360 |
+
"loss": 1.1266,
|
| 1361 |
+
"step": 193
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 0.7238805970149254,
|
| 1365 |
+
"grad_norm": 2.3111367225646973,
|
| 1366 |
+
"learning_rate": 1.930196071764312e-07,
|
| 1367 |
+
"loss": 1.0998,
|
| 1368 |
+
"step": 194
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 0.7276119402985075,
|
| 1372 |
+
"grad_norm": 2.1398303508758545,
|
| 1373 |
+
"learning_rate": 1.8825509907063326e-07,
|
| 1374 |
+
"loss": 1.1229,
|
| 1375 |
+
"step": 195
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 0.7313432835820896,
|
| 1379 |
+
"grad_norm": 2.1584665775299072,
|
| 1380 |
+
"learning_rate": 1.8353645732225975e-07,
|
| 1381 |
+
"loss": 1.1356,
|
| 1382 |
+
"step": 196
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.7350746268656716,
|
| 1386 |
+
"grad_norm": 2.3748433589935303,
|
| 1387 |
+
"learning_rate": 1.7886437617493204e-07,
|
| 1388 |
+
"loss": 1.1666,
|
| 1389 |
+
"step": 197
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.7388059701492538,
|
| 1393 |
+
"grad_norm": 2.160381317138672,
|
| 1394 |
+
"learning_rate": 1.7423954302191045e-07,
|
| 1395 |
+
"loss": 1.1376,
|
| 1396 |
+
"step": 198
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 0.7425373134328358,
|
| 1400 |
+
"grad_norm": 2.382323741912842,
|
| 1401 |
+
"learning_rate": 1.6966263830495935e-07,
|
| 1402 |
+
"loss": 1.194,
|
| 1403 |
+
"step": 199
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 0.746268656716418,
|
| 1407 |
+
"grad_norm": 2.287235975265503,
|
| 1408 |
+
"learning_rate": 1.6513433541423527e-07,
|
| 1409 |
+
"loss": 1.1851,
|
| 1410 |
+
"step": 200
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 0.75,
|
| 1414 |
+
"grad_norm": 2.2451207637786865,
|
| 1415 |
+
"learning_rate": 1.606553005892125e-07,
|
| 1416 |
+
"loss": 1.1282,
|
| 1417 |
+
"step": 201
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 0.753731343283582,
|
| 1421 |
+
"grad_norm": 2.1960198879241943,
|
| 1422 |
+
"learning_rate": 1.562261928206608e-07,
|
| 1423 |
+
"loss": 1.1387,
|
| 1424 |
+
"step": 202
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.7574626865671642,
|
| 1428 |
+
"grad_norm": 2.217029333114624,
|
| 1429 |
+
"learning_rate": 1.5184766375368912e-07,
|
| 1430 |
+
"loss": 1.1082,
|
| 1431 |
+
"step": 203
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 0.7611940298507462,
|
| 1435 |
+
"grad_norm": 2.3279242515563965,
|
| 1436 |
+
"learning_rate": 1.4752035759187103e-07,
|
| 1437 |
+
"loss": 1.1234,
|
| 1438 |
+
"step": 204
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 0.7649253731343284,
|
| 1442 |
+
"grad_norm": 2.2879576683044434,
|
| 1443 |
+
"learning_rate": 1.4324491100246384e-07,
|
| 1444 |
+
"loss": 1.1566,
|
| 1445 |
+
"step": 205
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 0.7686567164179104,
|
| 1449 |
+
"grad_norm": 2.181703805923462,
|
| 1450 |
+
"learning_rate": 1.3902195302273778e-07,
|
| 1451 |
+
"loss": 1.1188,
|
| 1452 |
+
"step": 206
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 0.7723880597014925,
|
| 1456 |
+
"grad_norm": 2.1797292232513428,
|
| 1457 |
+
"learning_rate": 1.3485210496742637e-07,
|
| 1458 |
+
"loss": 1.131,
|
| 1459 |
+
"step": 207
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 0.7761194029850746,
|
| 1463 |
+
"grad_norm": 2.2674946784973145,
|
| 1464 |
+
"learning_rate": 1.3073598033731427e-07,
|
| 1465 |
+
"loss": 1.064,
|
| 1466 |
+
"step": 208
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.7798507462686567,
|
| 1470 |
+
"grad_norm": 2.479985237121582,
|
| 1471 |
+
"learning_rate": 1.2667418472897383e-07,
|
| 1472 |
+
"loss": 1.1558,
|
| 1473 |
+
"step": 209
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 0.7835820895522388,
|
| 1477 |
+
"grad_norm": 2.2408385276794434,
|
| 1478 |
+
"learning_rate": 1.2266731574566536e-07,
|
| 1479 |
+
"loss": 1.1508,
|
| 1480 |
+
"step": 210
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 0.7873134328358209,
|
| 1484 |
+
"grad_norm": 2.1557300090789795,
|
| 1485 |
+
"learning_rate": 1.1871596290941276e-07,
|
| 1486 |
+
"loss": 1.1424,
|
| 1487 |
+
"step": 211
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 0.7910447761194029,
|
| 1491 |
+
"grad_norm": 2.078749179840088,
|
| 1492 |
+
"learning_rate": 1.1482070757426854e-07,
|
| 1493 |
+
"loss": 1.1176,
|
| 1494 |
+
"step": 212
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 0.7947761194029851,
|
| 1498 |
+
"grad_norm": 2.2646126747131348,
|
| 1499 |
+
"learning_rate": 1.1098212284078035e-07,
|
| 1500 |
+
"loss": 1.115,
|
| 1501 |
+
"step": 213
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 0.7985074626865671,
|
| 1505 |
+
"grad_norm": 2.2255332469940186,
|
| 1506 |
+
"learning_rate": 1.07200773471672e-07,
|
| 1507 |
+
"loss": 1.1115,
|
| 1508 |
+
"step": 214
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.8022388059701493,
|
| 1512 |
+
"grad_norm": 2.2112481594085693,
|
| 1513 |
+
"learning_rate": 1.0347721580875123e-07,
|
| 1514 |
+
"loss": 1.1206,
|
| 1515 |
+
"step": 215
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 0.8059701492537313,
|
| 1519 |
+
"grad_norm": 2.2681076526641846,
|
| 1520 |
+
"learning_rate": 9.981199769105602e-08,
|
| 1521 |
+
"loss": 1.1265,
|
| 1522 |
+
"step": 216
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 0.8097014925373134,
|
| 1526 |
+
"grad_norm": 2.082890033721924,
|
| 1527 |
+
"learning_rate": 9.62056583742527e-08,
|
| 1528 |
+
"loss": 1.1319,
|
| 1529 |
+
"step": 217
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 0.8134328358208955,
|
| 1533 |
+
"grad_norm": 2.279475212097168,
|
| 1534 |
+
"learning_rate": 9.265872845129569e-08,
|
| 1535 |
+
"loss": 1.1473,
|
| 1536 |
+
"step": 218
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 0.8171641791044776,
|
| 1540 |
+
"grad_norm": 2.2021682262420654,
|
| 1541 |
+
"learning_rate": 8.917172977436355e-08,
|
| 1542 |
+
"loss": 1.0919,
|
| 1543 |
+
"step": 219
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 0.8208955223880597,
|
| 1547 |
+
"grad_norm": 2.2297630310058594,
|
| 1548 |
+
"learning_rate": 8.574517537807896e-08,
|
| 1549 |
+
"loss": 1.1787,
|
| 1550 |
+
"step": 220
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.8246268656716418,
|
| 1554 |
+
"grad_norm": 2.4313836097717285,
|
| 1555 |
+
"learning_rate": 8.237956940402757e-08,
|
| 1556 |
+
"loss": 1.2248,
|
| 1557 |
+
"step": 221
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 0.8283582089552238,
|
| 1561 |
+
"grad_norm": 2.3627495765686035,
|
| 1562 |
+
"learning_rate": 7.907540702658455e-08,
|
| 1563 |
+
"loss": 1.1571,
|
| 1564 |
+
"step": 222
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 0.832089552238806,
|
| 1568 |
+
"grad_norm": 2.1874005794525146,
|
| 1569 |
+
"learning_rate": 7.583317438006093e-08,
|
| 1570 |
+
"loss": 1.1259,
|
| 1571 |
+
"step": 223
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 0.835820895522388,
|
| 1575 |
+
"grad_norm": 2.172623872756958,
|
| 1576 |
+
"learning_rate": 7.26533484871793e-08,
|
| 1577 |
+
"loss": 1.1047,
|
| 1578 |
+
"step": 224
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 0.8395522388059702,
|
| 1582 |
+
"grad_norm": 2.259795904159546,
|
| 1583 |
+
"learning_rate": 6.953639718889076e-08,
|
| 1584 |
+
"loss": 1.1339,
|
| 1585 |
+
"step": 225
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 0.8432835820895522,
|
| 1589 |
+
"grad_norm": 2.313694953918457,
|
| 1590 |
+
"learning_rate": 6.648277907554234e-08,
|
| 1591 |
+
"loss": 1.1695,
|
| 1592 |
+
"step": 226
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.8470149253731343,
|
| 1596 |
+
"grad_norm": 2.4433200359344482,
|
| 1597 |
+
"learning_rate": 6.349294341940592e-08,
|
| 1598 |
+
"loss": 1.1657,
|
| 1599 |
+
"step": 227
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 0.8507462686567164,
|
| 1603 |
+
"grad_norm": 2.412324905395508,
|
| 1604 |
+
"learning_rate": 6.056733010857713e-08,
|
| 1605 |
+
"loss": 1.1327,
|
| 1606 |
+
"step": 228
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 0.8544776119402985,
|
| 1610 |
+
"grad_norm": 2.2885019779205322,
|
| 1611 |
+
"learning_rate": 5.770636958225617e-08,
|
| 1612 |
+
"loss": 1.1136,
|
| 1613 |
+
"step": 229
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 0.8582089552238806,
|
| 1617 |
+
"grad_norm": 2.278291940689087,
|
| 1618 |
+
"learning_rate": 5.4910482767417834e-08,
|
| 1619 |
+
"loss": 1.1669,
|
| 1620 |
+
"step": 230
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 0.8619402985074627,
|
| 1624 |
+
"grad_norm": 2.267599582672119,
|
| 1625 |
+
"learning_rate": 5.218008101688171e-08,
|
| 1626 |
+
"loss": 1.0824,
|
| 1627 |
+
"step": 231
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 0.8656716417910447,
|
| 1631 |
+
"grad_norm": 2.1955244541168213,
|
| 1632 |
+
"learning_rate": 4.951556604879048e-08,
|
| 1633 |
+
"loss": 1.1282,
|
| 1634 |
+
"step": 232
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.8694029850746269,
|
| 1638 |
+
"grad_norm": 2.2082724571228027,
|
| 1639 |
+
"learning_rate": 4.691732988750613e-08,
|
| 1640 |
+
"loss": 1.1488,
|
| 1641 |
+
"step": 233
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 0.8731343283582089,
|
| 1645 |
+
"grad_norm": 2.3228390216827393,
|
| 1646 |
+
"learning_rate": 4.4385754805932095e-08,
|
| 1647 |
+
"loss": 1.1906,
|
| 1648 |
+
"step": 234
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 0.8768656716417911,
|
| 1652 |
+
"grad_norm": 2.336118698120117,
|
| 1653 |
+
"learning_rate": 4.192121326927073e-08,
|
| 1654 |
+
"loss": 1.1737,
|
| 1655 |
+
"step": 235
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 0.8805970149253731,
|
| 1659 |
+
"grad_norm": 2.16275954246521,
|
| 1660 |
+
"learning_rate": 3.9524067880222665e-08,
|
| 1661 |
+
"loss": 1.0883,
|
| 1662 |
+
"step": 236
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 0.8843283582089553,
|
| 1666 |
+
"grad_norm": 2.21974515914917,
|
| 1667 |
+
"learning_rate": 3.719467132563819e-08,
|
| 1668 |
+
"loss": 1.1684,
|
| 1669 |
+
"step": 237
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 0.8880597014925373,
|
| 1673 |
+
"grad_norm": 2.315598249435425,
|
| 1674 |
+
"learning_rate": 3.493336632462718e-08,
|
| 1675 |
+
"loss": 1.14,
|
| 1676 |
+
"step": 238
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.8917910447761194,
|
| 1680 |
+
"grad_norm": 2.2459030151367188,
|
| 1681 |
+
"learning_rate": 3.274048557813552e-08,
|
| 1682 |
+
"loss": 1.159,
|
| 1683 |
+
"step": 239
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 0.8955223880597015,
|
| 1687 |
+
"grad_norm": 2.26613450050354,
|
| 1688 |
+
"learning_rate": 3.061635171999566e-08,
|
| 1689 |
+
"loss": 1.1201,
|
| 1690 |
+
"step": 240
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 0.8992537313432836,
|
| 1694 |
+
"grad_norm": 2.2701683044433594,
|
| 1695 |
+
"learning_rate": 2.8561277269457895e-08,
|
| 1696 |
+
"loss": 1.1172,
|
| 1697 |
+
"step": 241
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 0.9029850746268657,
|
| 1701 |
+
"grad_norm": 2.3200957775115967,
|
| 1702 |
+
"learning_rate": 2.6575564585210486e-08,
|
| 1703 |
+
"loss": 1.1658,
|
| 1704 |
+
"step": 242
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 0.9067164179104478,
|
| 1708 |
+
"grad_norm": 2.3541018962860107,
|
| 1709 |
+
"learning_rate": 2.4659505820893823e-08,
|
| 1710 |
+
"loss": 1.1541,
|
| 1711 |
+
"step": 243
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 0.9104477611940298,
|
| 1715 |
+
"grad_norm": 2.2937629222869873,
|
| 1716 |
+
"learning_rate": 2.2813382882116982e-08,
|
| 1717 |
+
"loss": 1.1193,
|
| 1718 |
+
"step": 244
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.914179104477612,
|
| 1722 |
+
"grad_norm": 2.193416118621826,
|
| 1723 |
+
"learning_rate": 2.1037467384981022e-08,
|
| 1724 |
+
"loss": 1.1164,
|
| 1725 |
+
"step": 245
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 0.917910447761194,
|
| 1729 |
+
"grad_norm": 2.16845965385437,
|
| 1730 |
+
"learning_rate": 1.9332020616117218e-08,
|
| 1731 |
+
"loss": 1.1283,
|
| 1732 |
+
"step": 246
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 0.9216417910447762,
|
| 1736 |
+
"grad_norm": 2.2416491508483887,
|
| 1737 |
+
"learning_rate": 1.769729349424415e-08,
|
| 1738 |
+
"loss": 1.1192,
|
| 1739 |
+
"step": 247
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 0.9253731343283582,
|
| 1743 |
+
"grad_norm": 2.2076077461242676,
|
| 1744 |
+
"learning_rate": 1.6133526533250563e-08,
|
| 1745 |
+
"loss": 1.1723,
|
| 1746 |
+
"step": 248
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 0.9291044776119403,
|
| 1750 |
+
"grad_norm": 2.1859934329986572,
|
| 1751 |
+
"learning_rate": 1.4640949806809522e-08,
|
| 1752 |
+
"loss": 1.1843,
|
| 1753 |
+
"step": 249
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 0.9328358208955224,
|
| 1757 |
+
"grad_norm": 2.211667060852051,
|
| 1758 |
+
"learning_rate": 1.3219782914527633e-08,
|
| 1759 |
+
"loss": 1.1576,
|
| 1760 |
+
"step": 250
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.9365671641791045,
|
| 1764 |
+
"grad_norm": 2.188645124435425,
|
| 1765 |
+
"learning_rate": 1.1870234949636071e-08,
|
| 1766 |
+
"loss": 1.1575,
|
| 1767 |
+
"step": 251
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 0.9402985074626866,
|
| 1771 |
+
"grad_norm": 2.152869701385498,
|
| 1772 |
+
"learning_rate": 1.0592504468227126e-08,
|
| 1773 |
+
"loss": 1.1052,
|
| 1774 |
+
"step": 252
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 0.9440298507462687,
|
| 1778 |
+
"grad_norm": 2.4283602237701416,
|
| 1779 |
+
"learning_rate": 9.386779460041016e-09,
|
| 1780 |
+
"loss": 1.1159,
|
| 1781 |
+
"step": 253
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 0.9477611940298507,
|
| 1785 |
+
"grad_norm": 2.2025809288024902,
|
| 1786 |
+
"learning_rate": 8.25323732080746e-09,
|
| 1787 |
+
"loss": 1.1317,
|
| 1788 |
+
"step": 254
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 0.9514925373134329,
|
| 1792 |
+
"grad_norm": 2.24324369430542,
|
| 1793 |
+
"learning_rate": 7.1920448261457715e-09,
|
| 1794 |
+
"loss": 1.1619,
|
| 1795 |
+
"step": 255
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 0.9552238805970149,
|
| 1799 |
+
"grad_norm": 2.3615405559539795,
|
| 1800 |
+
"learning_rate": 6.20335810702749e-09,
|
| 1801 |
+
"loss": 1.0635,
|
| 1802 |
+
"step": 256
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.9589552238805971,
|
| 1806 |
+
"grad_norm": 2.3955984115600586,
|
| 1807 |
+
"learning_rate": 5.287322626805202e-09,
|
| 1808 |
+
"loss": 1.1217,
|
| 1809 |
+
"step": 257
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 0.9626865671641791,
|
| 1813 |
+
"grad_norm": 2.260822296142578,
|
| 1814 |
+
"learning_rate": 4.444073159810768e-09,
|
| 1815 |
+
"loss": 1.144,
|
| 1816 |
+
"step": 258
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 0.9664179104477612,
|
| 1820 |
+
"grad_norm": 2.247169256210327,
|
| 1821 |
+
"learning_rate": 3.6737337715264662e-09,
|
| 1822 |
+
"loss": 1.1194,
|
| 1823 |
+
"step": 259
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 0.9701492537313433,
|
| 1827 |
+
"grad_norm": 2.3124468326568604,
|
| 1828 |
+
"learning_rate": 2.9764178003311436e-09,
|
| 1829 |
+
"loss": 1.1558,
|
| 1830 |
+
"step": 260
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 0.9738805970149254,
|
| 1834 |
+
"grad_norm": 2.3551888465881348,
|
| 1835 |
+
"learning_rate": 2.3522278408253936e-09,
|
| 1836 |
+
"loss": 1.1263,
|
| 1837 |
+
"step": 261
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 0.9776119402985075,
|
| 1841 |
+
"grad_norm": 2.2351226806640625,
|
| 1842 |
+
"learning_rate": 1.8012557287367391e-09,
|
| 1843 |
+
"loss": 1.1721,
|
| 1844 |
+
"step": 262
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 0.9813432835820896,
|
| 1848 |
+
"grad_norm": 2.252028226852417,
|
| 1849 |
+
"learning_rate": 1.3235825274081625e-09,
|
| 1850 |
+
"loss": 1.2214,
|
| 1851 |
+
"step": 263
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 0.9850746268656716,
|
| 1855 |
+
"grad_norm": 2.247479200363159,
|
| 1856 |
+
"learning_rate": 9.19278515871369e-10,
|
| 1857 |
+
"loss": 1.1489,
|
| 1858 |
+
"step": 264
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 0.9888059701492538,
|
| 1862 |
+
"grad_norm": 2.2732558250427246,
|
| 1863 |
+
"learning_rate": 5.884031785068356e-10,
|
| 1864 |
+
"loss": 1.1366,
|
| 1865 |
+
"step": 265
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 0.9925373134328358,
|
| 1869 |
+
"grad_norm": 2.1621742248535156,
|
| 1870 |
+
"learning_rate": 3.310051962920335e-10,
|
| 1871 |
+
"loss": 1.0937,
|
| 1872 |
+
"step": 266
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 0.996268656716418,
|
| 1876 |
+
"grad_norm": 2.086367130279541,
|
| 1877 |
+
"learning_rate": 1.4712243963893589e-10,
|
| 1878 |
+
"loss": 1.1372,
|
| 1879 |
+
"step": 267
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 1.0,
|
| 1883 |
+
"grad_norm": 2.332995891571045,
|
| 1884 |
+
"learning_rate": 3.6781962822529657e-11,
|
| 1885 |
+
"loss": 1.14,
|
| 1886 |
+
"step": 268
|
| 1887 |
+
}
|
| 1888 |
+
],
|
| 1889 |
+
"logging_steps": 1.0,
|
| 1890 |
+
"max_steps": 268,
|
| 1891 |
+
"num_input_tokens_seen": 0,
|
| 1892 |
+
"num_train_epochs": 1,
|
| 1893 |
+
"save_steps": 81,
|
| 1894 |
+
"stateful_callbacks": {
|
| 1895 |
+
"TrainerControl": {
|
| 1896 |
+
"args": {
|
| 1897 |
+
"should_epoch_stop": false,
|
| 1898 |
+
"should_evaluate": false,
|
| 1899 |
+
"should_log": false,
|
| 1900 |
+
"should_save": true,
|
| 1901 |
+
"should_training_stop": true
|
| 1902 |
+
},
|
| 1903 |
+
"attributes": {}
|
| 1904 |
+
}
|
| 1905 |
+
},
|
| 1906 |
+
"total_flos": 1.0169113530733691e+18,
|
| 1907 |
+
"train_batch_size": 4,
|
| 1908 |
+
"trial_name": null,
|
| 1909 |
+
"trial_params": null
|
| 1910 |
+
}
|