new version
Browse files- .ipynb_checkpoints/trainer_state-checkpoint.json +1458 -0
- config.json +3 -3
- generation_config.json +2 -2
- model.safetensors +1 -1
- runs/Dec17_21-04-23_notebook-7f74bd20-8f92-45b0-96c4-687ced4ee944/events.out.tfevents.1734440689.notebook-7f74bd20-8f92-45b0-96c4-687ced4ee944.50109.0 +3 -0
- tokenizer.json +2 -2
- tokenizer_config.json +1 -1
- trainer_state.json +0 -0
- training_args.bin +2 -2
.ipynb_checkpoints/trainer_state-checkpoint.json
ADDED
|
@@ -0,0 +1,1458 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9994767137624281,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 955,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.010465724751439037,
|
| 13 |
+
"grad_norm": 7.894938945770264,
|
| 14 |
+
"learning_rate": 2.0833333333333333e-07,
|
| 15 |
+
"logits/chosen": -0.3671875,
|
| 16 |
+
"logits/rejected": -0.2177734375,
|
| 17 |
+
"logps/chosen": -264.0,
|
| 18 |
+
"logps/rejected": -280.0,
|
| 19 |
+
"loss": 0.6932,
|
| 20 |
+
"rewards/accuracies": 0.21875,
|
| 21 |
+
"rewards/chosen": 0.000873565673828125,
|
| 22 |
+
"rewards/margins": 0.003082275390625,
|
| 23 |
+
"rewards/rejected": -0.0022125244140625,
|
| 24 |
+
"step": 10
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.020931449502878074,
|
| 28 |
+
"grad_norm": 8.288711547851562,
|
| 29 |
+
"learning_rate": 4.1666666666666667e-07,
|
| 30 |
+
"logits/chosen": -0.373046875,
|
| 31 |
+
"logits/rejected": -0.4296875,
|
| 32 |
+
"logps/chosen": -358.0,
|
| 33 |
+
"logps/rejected": -296.0,
|
| 34 |
+
"loss": 0.6947,
|
| 35 |
+
"rewards/accuracies": 0.2874999940395355,
|
| 36 |
+
"rewards/chosen": -0.0015106201171875,
|
| 37 |
+
"rewards/margins": -0.0001773834228515625,
|
| 38 |
+
"rewards/rejected": -0.00133514404296875,
|
| 39 |
+
"step": 20
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03139717425431711,
|
| 43 |
+
"grad_norm": 8.539895057678223,
|
| 44 |
+
"learning_rate": 6.249999999999999e-07,
|
| 45 |
+
"logits/chosen": -0.38671875,
|
| 46 |
+
"logits/rejected": -0.45703125,
|
| 47 |
+
"logps/chosen": -286.0,
|
| 48 |
+
"logps/rejected": -249.0,
|
| 49 |
+
"loss": 0.6912,
|
| 50 |
+
"rewards/accuracies": 0.3125,
|
| 51 |
+
"rewards/chosen": -0.000537872314453125,
|
| 52 |
+
"rewards/margins": -0.00101470947265625,
|
| 53 |
+
"rewards/rejected": 0.000484466552734375,
|
| 54 |
+
"step": 30
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"epoch": 0.04186289900575615,
|
| 58 |
+
"grad_norm": 7.561408996582031,
|
| 59 |
+
"learning_rate": 8.333333333333333e-07,
|
| 60 |
+
"logits/chosen": -0.30859375,
|
| 61 |
+
"logits/rejected": -0.2080078125,
|
| 62 |
+
"logps/chosen": -312.0,
|
| 63 |
+
"logps/rejected": -306.0,
|
| 64 |
+
"loss": 0.689,
|
| 65 |
+
"rewards/accuracies": 0.4625000059604645,
|
| 66 |
+
"rewards/chosen": 0.01068115234375,
|
| 67 |
+
"rewards/margins": 0.0240478515625,
|
| 68 |
+
"rewards/rejected": -0.01336669921875,
|
| 69 |
+
"step": 40
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.052328623757195186,
|
| 73 |
+
"grad_norm": 7.783715724945068,
|
| 74 |
+
"learning_rate": 9.999880027023293e-07,
|
| 75 |
+
"logits/chosen": -0.3671875,
|
| 76 |
+
"logits/rejected": -0.38671875,
|
| 77 |
+
"logps/chosen": -302.0,
|
| 78 |
+
"logps/rejected": -268.0,
|
| 79 |
+
"loss": 0.6772,
|
| 80 |
+
"rewards/accuracies": 0.625,
|
| 81 |
+
"rewards/chosen": 0.0257568359375,
|
| 82 |
+
"rewards/margins": 0.048828125,
|
| 83 |
+
"rewards/rejected": -0.02294921875,
|
| 84 |
+
"step": 50
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 0.06279434850863422,
|
| 88 |
+
"grad_norm": 6.941147327423096,
|
| 89 |
+
"learning_rate": 9.995681577335256e-07,
|
| 90 |
+
"logits/chosen": -0.34375,
|
| 91 |
+
"logits/rejected": -0.220703125,
|
| 92 |
+
"logps/chosen": -296.0,
|
| 93 |
+
"logps/rejected": -284.0,
|
| 94 |
+
"loss": 0.6664,
|
| 95 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 96 |
+
"rewards/chosen": 0.032470703125,
|
| 97 |
+
"rewards/margins": 0.059326171875,
|
| 98 |
+
"rewards/rejected": -0.0267333984375,
|
| 99 |
+
"step": 60
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.07326007326007326,
|
| 103 |
+
"grad_norm": 7.589351654052734,
|
| 104 |
+
"learning_rate": 9.985490234976131e-07,
|
| 105 |
+
"logits/chosen": -0.40625,
|
| 106 |
+
"logits/rejected": -0.478515625,
|
| 107 |
+
"logps/chosen": -316.0,
|
| 108 |
+
"logps/rejected": -250.0,
|
| 109 |
+
"loss": 0.6589,
|
| 110 |
+
"rewards/accuracies": 0.606249988079071,
|
| 111 |
+
"rewards/chosen": 0.056884765625,
|
| 112 |
+
"rewards/margins": 0.10205078125,
|
| 113 |
+
"rewards/rejected": -0.044921875,
|
| 114 |
+
"step": 70
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.0837257980115123,
|
| 118 |
+
"grad_norm": 7.692485809326172,
|
| 119 |
+
"learning_rate": 9.969318225629239e-07,
|
| 120 |
+
"logits/chosen": -0.4375,
|
| 121 |
+
"logits/rejected": -0.47265625,
|
| 122 |
+
"logps/chosen": -362.0,
|
| 123 |
+
"logps/rejected": -322.0,
|
| 124 |
+
"loss": 0.651,
|
| 125 |
+
"rewards/accuracies": 0.65625,
|
| 126 |
+
"rewards/chosen": 0.05908203125,
|
| 127 |
+
"rewards/margins": 0.15234375,
|
| 128 |
+
"rewards/rejected": -0.0927734375,
|
| 129 |
+
"step": 80
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.09419152276295134,
|
| 133 |
+
"grad_norm": 7.012711524963379,
|
| 134 |
+
"learning_rate": 9.947184949473476e-07,
|
| 135 |
+
"logits/chosen": -0.365234375,
|
| 136 |
+
"logits/rejected": -0.3203125,
|
| 137 |
+
"logps/chosen": -334.0,
|
| 138 |
+
"logps/rejected": -284.0,
|
| 139 |
+
"loss": 0.6364,
|
| 140 |
+
"rewards/accuracies": 0.637499988079071,
|
| 141 |
+
"rewards/chosen": 0.05908203125,
|
| 142 |
+
"rewards/margins": 0.169921875,
|
| 143 |
+
"rewards/rejected": -0.1103515625,
|
| 144 |
+
"step": 90
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.10465724751439037,
|
| 148 |
+
"grad_norm": 7.267467498779297,
|
| 149 |
+
"learning_rate": 9.919116957910565e-07,
|
| 150 |
+
"logits/chosen": -0.546875,
|
| 151 |
+
"logits/rejected": -0.50390625,
|
| 152 |
+
"logps/chosen": -330.0,
|
| 153 |
+
"logps/rejected": -280.0,
|
| 154 |
+
"loss": 0.6289,
|
| 155 |
+
"rewards/accuracies": 0.675000011920929,
|
| 156 |
+
"rewards/chosen": 0.07275390625,
|
| 157 |
+
"rewards/margins": 0.26953125,
|
| 158 |
+
"rewards/rejected": -0.1962890625,
|
| 159 |
+
"step": 100
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.1151229722658294,
|
| 163 |
+
"grad_norm": 7.708714485168457,
|
| 164 |
+
"learning_rate": 9.88514792171362e-07,
|
| 165 |
+
"logits/chosen": -0.453125,
|
| 166 |
+
"logits/rejected": -0.390625,
|
| 167 |
+
"logps/chosen": -306.0,
|
| 168 |
+
"logps/rejected": -298.0,
|
| 169 |
+
"loss": 0.6172,
|
| 170 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 171 |
+
"rewards/chosen": 0.0302734375,
|
| 172 |
+
"rewards/margins": 0.1875,
|
| 173 |
+
"rewards/rejected": -0.1572265625,
|
| 174 |
+
"step": 110
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.12558869701726844,
|
| 178 |
+
"grad_norm": 7.422197341918945,
|
| 179 |
+
"learning_rate": 9.845318590635185e-07,
|
| 180 |
+
"logits/chosen": -0.4765625,
|
| 181 |
+
"logits/rejected": -0.44140625,
|
| 182 |
+
"logps/chosen": -314.0,
|
| 183 |
+
"logps/rejected": -256.0,
|
| 184 |
+
"loss": 0.6065,
|
| 185 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 186 |
+
"rewards/chosen": 0.03759765625,
|
| 187 |
+
"rewards/margins": 0.349609375,
|
| 188 |
+
"rewards/rejected": -0.3125,
|
| 189 |
+
"step": 120
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.1360544217687075,
|
| 193 |
+
"grad_norm": 7.663195610046387,
|
| 194 |
+
"learning_rate": 9.799676744523238e-07,
|
| 195 |
+
"logits/chosen": -0.578125,
|
| 196 |
+
"logits/rejected": -0.51953125,
|
| 197 |
+
"logps/chosen": -328.0,
|
| 198 |
+
"logps/rejected": -278.0,
|
| 199 |
+
"loss": 0.6167,
|
| 200 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 201 |
+
"rewards/chosen": -0.00689697265625,
|
| 202 |
+
"rewards/margins": 0.32421875,
|
| 203 |
+
"rewards/rejected": -0.33203125,
|
| 204 |
+
"step": 130
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 0.14652014652014653,
|
| 208 |
+
"grad_norm": 7.429771900177002,
|
| 209 |
+
"learning_rate": 9.748277136003789e-07,
|
| 210 |
+
"logits/chosen": -0.5,
|
| 211 |
+
"logits/rejected": -0.296875,
|
| 212 |
+
"logps/chosen": -306.0,
|
| 213 |
+
"logps/rejected": -296.0,
|
| 214 |
+
"loss": 0.6256,
|
| 215 |
+
"rewards/accuracies": 0.637499988079071,
|
| 216 |
+
"rewards/chosen": -0.07666015625,
|
| 217 |
+
"rewards/margins": 0.1962890625,
|
| 218 |
+
"rewards/rejected": -0.2734375,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.15698587127158556,
|
| 223 |
+
"grad_norm": 7.726008415222168,
|
| 224 |
+
"learning_rate": 9.691181424798824e-07,
|
| 225 |
+
"logits/chosen": -0.56640625,
|
| 226 |
+
"logits/rejected": -0.5546875,
|
| 227 |
+
"logps/chosen": -276.0,
|
| 228 |
+
"logps/rejected": -264.0,
|
| 229 |
+
"loss": 0.6199,
|
| 230 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 231 |
+
"rewards/chosen": -0.07568359375,
|
| 232 |
+
"rewards/margins": 0.1953125,
|
| 233 |
+
"rewards/rejected": -0.271484375,
|
| 234 |
+
"step": 150
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.1674515960230246,
|
| 238 |
+
"grad_norm": 7.5920844078063965,
|
| 239 |
+
"learning_rate": 9.628458103758402e-07,
|
| 240 |
+
"logits/chosen": -0.609375,
|
| 241 |
+
"logits/rejected": -0.52734375,
|
| 242 |
+
"logps/chosen": -330.0,
|
| 243 |
+
"logps/rejected": -312.0,
|
| 244 |
+
"loss": 0.6082,
|
| 245 |
+
"rewards/accuracies": 0.612500011920929,
|
| 246 |
+
"rewards/chosen": -0.06787109375,
|
| 247 |
+
"rewards/margins": 0.294921875,
|
| 248 |
+
"rewards/rejected": -0.36328125,
|
| 249 |
+
"step": 160
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.17791732077446362,
|
| 253 |
+
"grad_norm": 6.563006401062012,
|
| 254 |
+
"learning_rate": 9.560182416695637e-07,
|
| 255 |
+
"logits/chosen": -0.5625,
|
| 256 |
+
"logits/rejected": -0.59375,
|
| 257 |
+
"logps/chosen": -310.0,
|
| 258 |
+
"logps/rejected": -292.0,
|
| 259 |
+
"loss": 0.6156,
|
| 260 |
+
"rewards/accuracies": 0.65625,
|
| 261 |
+
"rewards/chosen": -0.10205078125,
|
| 262 |
+
"rewards/margins": 0.25,
|
| 263 |
+
"rewards/rejected": -0.3515625,
|
| 264 |
+
"step": 170
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 0.18838304552590268,
|
| 268 |
+
"grad_norm": 7.841839790344238,
|
| 269 |
+
"learning_rate": 9.486436268123111e-07,
|
| 270 |
+
"logits/chosen": -0.5859375,
|
| 271 |
+
"logits/rejected": -0.55859375,
|
| 272 |
+
"logps/chosen": -330.0,
|
| 273 |
+
"logps/rejected": -308.0,
|
| 274 |
+
"loss": 0.6261,
|
| 275 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 276 |
+
"rewards/chosen": -0.1220703125,
|
| 277 |
+
"rewards/margins": 0.283203125,
|
| 278 |
+
"rewards/rejected": -0.40625,
|
| 279 |
+
"step": 180
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.1988487702773417,
|
| 283 |
+
"grad_norm": 6.8911027908325195,
|
| 284 |
+
"learning_rate": 9.40730812499903e-07,
|
| 285 |
+
"logits/chosen": -0.578125,
|
| 286 |
+
"logits/rejected": -0.5625,
|
| 287 |
+
"logps/chosen": -302.0,
|
| 288 |
+
"logps/rejected": -272.0,
|
| 289 |
+
"loss": 0.6092,
|
| 290 |
+
"rewards/accuracies": 0.6875,
|
| 291 |
+
"rewards/chosen": -0.08984375,
|
| 292 |
+
"rewards/margins": 0.322265625,
|
| 293 |
+
"rewards/rejected": -0.412109375,
|
| 294 |
+
"step": 190
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"epoch": 0.20931449502878074,
|
| 298 |
+
"grad_norm": 8.3724946975708,
|
| 299 |
+
"learning_rate": 9.322892910600958e-07,
|
| 300 |
+
"logits/chosen": -0.6171875,
|
| 301 |
+
"logits/rejected": -0.451171875,
|
| 302 |
+
"logps/chosen": -298.0,
|
| 303 |
+
"logps/rejected": -278.0,
|
| 304 |
+
"loss": 0.6,
|
| 305 |
+
"rewards/accuracies": 0.65625,
|
| 306 |
+
"rewards/chosen": -0.0247802734375,
|
| 307 |
+
"rewards/margins": 0.357421875,
|
| 308 |
+
"rewards/rejected": -0.380859375,
|
| 309 |
+
"step": 200
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.21978021978021978,
|
| 313 |
+
"grad_norm": 7.580628871917725,
|
| 314 |
+
"learning_rate": 9.233291890654476e-07,
|
| 315 |
+
"logits/chosen": -0.6953125,
|
| 316 |
+
"logits/rejected": -0.66796875,
|
| 317 |
+
"logps/chosen": -318.0,
|
| 318 |
+
"logps/rejected": -284.0,
|
| 319 |
+
"loss": 0.5919,
|
| 320 |
+
"rewards/accuracies": 0.699999988079071,
|
| 321 |
+
"rewards/chosen": -0.012451171875,
|
| 322 |
+
"rewards/margins": 0.419921875,
|
| 323 |
+
"rewards/rejected": -0.431640625,
|
| 324 |
+
"step": 210
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.2302459445316588,
|
| 328 |
+
"grad_norm": 7.269970893859863,
|
| 329 |
+
"learning_rate": 9.138612551853332e-07,
|
| 330 |
+
"logits/chosen": -0.72265625,
|
| 331 |
+
"logits/rejected": -0.68359375,
|
| 332 |
+
"logps/chosen": -286.0,
|
| 333 |
+
"logps/rejected": -302.0,
|
| 334 |
+
"loss": 0.5974,
|
| 335 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 336 |
+
"rewards/chosen": -0.05224609375,
|
| 337 |
+
"rewards/margins": 0.37890625,
|
| 338 |
+
"rewards/rejected": -0.431640625,
|
| 339 |
+
"step": 220
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.24071166928309787,
|
| 343 |
+
"grad_norm": 8.078968048095703,
|
| 344 |
+
"learning_rate": 9.03896847291683e-07,
|
| 345 |
+
"logits/chosen": -0.6328125,
|
| 346 |
+
"logits/rejected": -0.578125,
|
| 347 |
+
"logps/chosen": -348.0,
|
| 348 |
+
"logps/rejected": -308.0,
|
| 349 |
+
"loss": 0.5879,
|
| 350 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 351 |
+
"rewards/chosen": 0.003662109375,
|
| 352 |
+
"rewards/margins": 0.462890625,
|
| 353 |
+
"rewards/rejected": -0.458984375,
|
| 354 |
+
"step": 230
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 0.25117739403453687,
|
| 358 |
+
"grad_norm": 7.896542549133301,
|
| 359 |
+
"learning_rate": 8.934479188339137e-07,
|
| 360 |
+
"logits/chosen": -0.62890625,
|
| 361 |
+
"logits/rejected": -0.625,
|
| 362 |
+
"logps/chosen": -336.0,
|
| 363 |
+
"logps/rejected": -286.0,
|
| 364 |
+
"loss": 0.5862,
|
| 365 |
+
"rewards/accuracies": 0.668749988079071,
|
| 366 |
+
"rewards/chosen": -0.06787109375,
|
| 367 |
+
"rewards/margins": 0.36328125,
|
| 368 |
+
"rewards/rejected": -0.431640625,
|
| 369 |
+
"step": 240
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.2616431187859759,
|
| 373 |
+
"grad_norm": 7.647547245025635,
|
| 374 |
+
"learning_rate": 8.825270044993962e-07,
|
| 375 |
+
"logits/chosen": -0.6328125,
|
| 376 |
+
"logits/rejected": -0.443359375,
|
| 377 |
+
"logps/chosen": -372.0,
|
| 378 |
+
"logps/rejected": -336.0,
|
| 379 |
+
"loss": 0.5811,
|
| 380 |
+
"rewards/accuracies": 0.793749988079071,
|
| 381 |
+
"rewards/chosen": 0.034912109375,
|
| 382 |
+
"rewards/margins": 0.57421875,
|
| 383 |
+
"rewards/rejected": -0.5390625,
|
| 384 |
+
"step": 250
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 0.272108843537415,
|
| 388 |
+
"grad_norm": 7.487204551696777,
|
| 389 |
+
"learning_rate": 8.711472051766605e-07,
|
| 390 |
+
"logits/chosen": -0.73046875,
|
| 391 |
+
"logits/rejected": -0.66015625,
|
| 392 |
+
"logps/chosen": -296.0,
|
| 393 |
+
"logps/rejected": -312.0,
|
| 394 |
+
"loss": 0.5871,
|
| 395 |
+
"rewards/accuracies": 0.737500011920929,
|
| 396 |
+
"rewards/chosen": -0.02685546875,
|
| 397 |
+
"rewards/margins": 0.421875,
|
| 398 |
+
"rewards/rejected": -0.44921875,
|
| 399 |
+
"step": 260
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.282574568288854,
|
| 403 |
+
"grad_norm": 7.606942653656006,
|
| 404 |
+
"learning_rate": 8.593221722393789e-07,
|
| 405 |
+
"logits/chosen": -0.6875,
|
| 406 |
+
"logits/rejected": -0.62109375,
|
| 407 |
+
"logps/chosen": -298.0,
|
| 408 |
+
"logps/rejected": -280.0,
|
| 409 |
+
"loss": 0.5879,
|
| 410 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 411 |
+
"rewards/chosen": -0.034912109375,
|
| 412 |
+
"rewards/margins": 0.56640625,
|
| 413 |
+
"rewards/rejected": -0.6015625,
|
| 414 |
+
"step": 270
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 0.29304029304029305,
|
| 418 |
+
"grad_norm": 6.631251335144043,
|
| 419 |
+
"learning_rate": 8.470660911699782e-07,
|
| 420 |
+
"logits/chosen": -0.6171875,
|
| 421 |
+
"logits/rejected": -0.609375,
|
| 422 |
+
"logps/chosen": -308.0,
|
| 423 |
+
"logps/rejected": -274.0,
|
| 424 |
+
"loss": 0.5836,
|
| 425 |
+
"rewards/accuracies": 0.71875,
|
| 426 |
+
"rewards/chosen": -0.1591796875,
|
| 427 |
+
"rewards/margins": 0.486328125,
|
| 428 |
+
"rewards/rejected": -0.64453125,
|
| 429 |
+
"step": 280
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.3035060177917321,
|
| 433 |
+
"grad_norm": 5.671150207519531,
|
| 434 |
+
"learning_rate": 8.343936645425276e-07,
|
| 435 |
+
"logits/chosen": -0.5390625,
|
| 436 |
+
"logits/rejected": -0.55859375,
|
| 437 |
+
"logps/chosen": -318.0,
|
| 438 |
+
"logps/rejected": -296.0,
|
| 439 |
+
"loss": 0.5744,
|
| 440 |
+
"rewards/accuracies": 0.731249988079071,
|
| 441 |
+
"rewards/chosen": -0.119140625,
|
| 442 |
+
"rewards/margins": 0.50390625,
|
| 443 |
+
"rewards/rejected": -0.62109375,
|
| 444 |
+
"step": 290
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.3139717425431711,
|
| 448 |
+
"grad_norm": 7.8931989669799805,
|
| 449 |
+
"learning_rate": 8.213200943853158e-07,
|
| 450 |
+
"logits/chosen": -0.8125,
|
| 451 |
+
"logits/rejected": -0.5546875,
|
| 452 |
+
"logps/chosen": -288.0,
|
| 453 |
+
"logps/rejected": -292.0,
|
| 454 |
+
"loss": 0.6098,
|
| 455 |
+
"rewards/accuracies": 0.59375,
|
| 456 |
+
"rewards/chosen": -0.1357421875,
|
| 457 |
+
"rewards/margins": 0.28515625,
|
| 458 |
+
"rewards/rejected": -0.421875,
|
| 459 |
+
"step": 300
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.32443746729461015,
|
| 463 |
+
"grad_norm": 7.260922908782959,
|
| 464 |
+
"learning_rate": 8.07861063944276e-07,
|
| 465 |
+
"logits/chosen": -0.63671875,
|
| 466 |
+
"logits/rejected": -0.76171875,
|
| 467 |
+
"logps/chosen": -368.0,
|
| 468 |
+
"logps/rejected": -274.0,
|
| 469 |
+
"loss": 0.5672,
|
| 470 |
+
"rewards/accuracies": 0.731249988079071,
|
| 471 |
+
"rewards/chosen": 0.0260009765625,
|
| 472 |
+
"rewards/margins": 0.6015625,
|
| 473 |
+
"rewards/rejected": -0.578125,
|
| 474 |
+
"step": 310
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 0.3349031920460492,
|
| 478 |
+
"grad_norm": 6.909915924072266,
|
| 479 |
+
"learning_rate": 7.940327188691341e-07,
|
| 480 |
+
"logits/chosen": -0.72265625,
|
| 481 |
+
"logits/rejected": -0.7265625,
|
| 482 |
+
"logps/chosen": -306.0,
|
| 483 |
+
"logps/rejected": -282.0,
|
| 484 |
+
"loss": 0.5732,
|
| 485 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 486 |
+
"rewards/chosen": -0.050537109375,
|
| 487 |
+
"rewards/margins": 0.55078125,
|
| 488 |
+
"rewards/rejected": -0.6015625,
|
| 489 |
+
"step": 320
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.3453689167974882,
|
| 493 |
+
"grad_norm": 7.135429382324219,
|
| 494 |
+
"learning_rate": 7.798516478448514e-07,
|
| 495 |
+
"logits/chosen": -0.75390625,
|
| 496 |
+
"logits/rejected": -0.7265625,
|
| 497 |
+
"logps/chosen": -334.0,
|
| 498 |
+
"logps/rejected": -294.0,
|
| 499 |
+
"loss": 0.5739,
|
| 500 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 501 |
+
"rewards/chosen": 0.031494140625,
|
| 502 |
+
"rewards/margins": 0.486328125,
|
| 503 |
+
"rewards/rejected": -0.455078125,
|
| 504 |
+
"step": 330
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 0.35583464154892724,
|
| 508 |
+
"grad_norm": 6.795555591583252,
|
| 509 |
+
"learning_rate": 7.653348626915957e-07,
|
| 510 |
+
"logits/chosen": -0.73828125,
|
| 511 |
+
"logits/rejected": -0.734375,
|
| 512 |
+
"logps/chosen": -296.0,
|
| 513 |
+
"logps/rejected": -290.0,
|
| 514 |
+
"loss": 0.5676,
|
| 515 |
+
"rewards/accuracies": 0.699999988079071,
|
| 516 |
+
"rewards/chosen": 0.00927734375,
|
| 517 |
+
"rewards/margins": 0.490234375,
|
| 518 |
+
"rewards/rejected": -0.482421875,
|
| 519 |
+
"step": 340
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.3663003663003663,
|
| 523 |
+
"grad_norm": 6.563620567321777,
|
| 524 |
+
"learning_rate": 7.504997779571132e-07,
|
| 525 |
+
"logits/chosen": -0.63671875,
|
| 526 |
+
"logits/rejected": -0.62890625,
|
| 527 |
+
"logps/chosen": -330.0,
|
| 528 |
+
"logps/rejected": -288.0,
|
| 529 |
+
"loss": 0.5654,
|
| 530 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 531 |
+
"rewards/chosen": -0.09326171875,
|
| 532 |
+
"rewards/margins": 0.31640625,
|
| 533 |
+
"rewards/rejected": -0.408203125,
|
| 534 |
+
"step": 350
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.37676609105180536,
|
| 538 |
+
"grad_norm": 8.264630317687988,
|
| 539 |
+
"learning_rate": 7.353641900259823e-07,
|
| 540 |
+
"logits/chosen": -0.75,
|
| 541 |
+
"logits/rejected": -0.83203125,
|
| 542 |
+
"logps/chosen": -340.0,
|
| 543 |
+
"logps/rejected": -288.0,
|
| 544 |
+
"loss": 0.5644,
|
| 545 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 546 |
+
"rewards/chosen": -0.0208740234375,
|
| 547 |
+
"rewards/margins": 0.5078125,
|
| 548 |
+
"rewards/rejected": -0.52734375,
|
| 549 |
+
"step": 360
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.3872318158032444,
|
| 553 |
+
"grad_norm": 8.003825187683105,
|
| 554 |
+
"learning_rate": 7.199462557708097e-07,
|
| 555 |
+
"logits/chosen": -0.76171875,
|
| 556 |
+
"logits/rejected": -0.63671875,
|
| 557 |
+
"logps/chosen": -312.0,
|
| 558 |
+
"logps/rejected": -328.0,
|
| 559 |
+
"loss": 0.5807,
|
| 560 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 561 |
+
"rewards/chosen": -0.10888671875,
|
| 562 |
+
"rewards/margins": 0.494140625,
|
| 563 |
+
"rewards/rejected": -0.6015625,
|
| 564 |
+
"step": 370
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 0.3976975405546834,
|
| 568 |
+
"grad_norm": 6.812396049499512,
|
| 569 |
+
"learning_rate": 7.042644707709815e-07,
|
| 570 |
+
"logits/chosen": -0.66015625,
|
| 571 |
+
"logits/rejected": -0.6015625,
|
| 572 |
+
"logps/chosen": -290.0,
|
| 573 |
+
"logps/rejected": -245.0,
|
| 574 |
+
"loss": 0.5879,
|
| 575 |
+
"rewards/accuracies": 0.699999988079071,
|
| 576 |
+
"rewards/chosen": -0.03369140625,
|
| 577 |
+
"rewards/margins": 0.51171875,
|
| 578 |
+
"rewards/rejected": -0.546875,
|
| 579 |
+
"step": 380
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.40816326530612246,
|
| 583 |
+
"grad_norm": 7.9978485107421875,
|
| 584 |
+
"learning_rate": 6.883376471250955e-07,
|
| 585 |
+
"logits/chosen": -0.69921875,
|
| 586 |
+
"logits/rejected": -0.734375,
|
| 587 |
+
"logps/chosen": -302.0,
|
| 588 |
+
"logps/rejected": -284.0,
|
| 589 |
+
"loss": 0.5847,
|
| 590 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 591 |
+
"rewards/chosen": -0.035888671875,
|
| 592 |
+
"rewards/margins": 0.484375,
|
| 593 |
+
"rewards/rejected": -0.51953125,
|
| 594 |
+
"step": 390
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 0.4186289900575615,
|
| 598 |
+
"grad_norm": 7.784450531005859,
|
| 599 |
+
"learning_rate": 6.72184890883692e-07,
|
| 600 |
+
"logits/chosen": -0.734375,
|
| 601 |
+
"logits/rejected": -0.703125,
|
| 602 |
+
"logps/chosen": -286.0,
|
| 603 |
+
"logps/rejected": -253.0,
|
| 604 |
+
"loss": 0.5448,
|
| 605 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 606 |
+
"rewards/chosen": -0.034912109375,
|
| 607 |
+
"rewards/margins": 0.55078125,
|
| 608 |
+
"rewards/rejected": -0.5859375,
|
| 609 |
+
"step": 400
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.4290947148090005,
|
| 613 |
+
"grad_norm": 7.214269161224365,
|
| 614 |
+
"learning_rate": 6.558255791293571e-07,
|
| 615 |
+
"logits/chosen": -0.640625,
|
| 616 |
+
"logits/rejected": -0.73046875,
|
| 617 |
+
"logps/chosen": -302.0,
|
| 618 |
+
"logps/rejected": -276.0,
|
| 619 |
+
"loss": 0.5476,
|
| 620 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 621 |
+
"rewards/chosen": 0.00787353515625,
|
| 622 |
+
"rewards/margins": 0.61328125,
|
| 623 |
+
"rewards/rejected": -0.60546875,
|
| 624 |
+
"step": 410
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 0.43956043956043955,
|
| 628 |
+
"grad_norm": 8.752714157104492,
|
| 629 |
+
"learning_rate": 6.392793367316904e-07,
|
| 630 |
+
"logits/chosen": -0.73828125,
|
| 631 |
+
"logits/rejected": -0.7109375,
|
| 632 |
+
"logps/chosen": -334.0,
|
| 633 |
+
"logps/rejected": -306.0,
|
| 634 |
+
"loss": 0.5496,
|
| 635 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 636 |
+
"rewards/chosen": 0.0311279296875,
|
| 637 |
+
"rewards/margins": 0.5625,
|
| 638 |
+
"rewards/rejected": -0.52734375,
|
| 639 |
+
"step": 420
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.4500261643118786,
|
| 643 |
+
"grad_norm": 9.271011352539062,
|
| 644 |
+
"learning_rate": 6.225660128050247e-07,
|
| 645 |
+
"logits/chosen": -0.7734375,
|
| 646 |
+
"logits/rejected": -0.73828125,
|
| 647 |
+
"logps/chosen": -284.0,
|
| 648 |
+
"logps/rejected": -272.0,
|
| 649 |
+
"loss": 0.567,
|
| 650 |
+
"rewards/accuracies": 0.6875,
|
| 651 |
+
"rewards/chosen": -0.05224609375,
|
| 652 |
+
"rewards/margins": 0.53515625,
|
| 653 |
+
"rewards/rejected": -0.58984375,
|
| 654 |
+
"step": 430
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.4604918890633176,
|
| 658 |
+
"grad_norm": 8.128939628601074,
|
| 659 |
+
"learning_rate": 6.057056568971383e-07,
|
| 660 |
+
"logits/chosen": -0.65625,
|
| 661 |
+
"logits/rejected": -0.6953125,
|
| 662 |
+
"logps/chosen": -372.0,
|
| 663 |
+
"logps/rejected": -302.0,
|
| 664 |
+
"loss": 0.5309,
|
| 665 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 666 |
+
"rewards/chosen": 0.0830078125,
|
| 667 |
+
"rewards/margins": 0.6953125,
|
| 668 |
+
"rewards/rejected": -0.609375,
|
| 669 |
+
"step": 440
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.47095761381475665,
|
| 673 |
+
"grad_norm": 6.812074661254883,
|
| 674 |
+
"learning_rate": 5.887184949375242e-07,
|
| 675 |
+
"logits/chosen": -0.6484375,
|
| 676 |
+
"logits/rejected": -0.7265625,
|
| 677 |
+
"logps/chosen": -318.0,
|
| 678 |
+
"logps/rejected": -296.0,
|
| 679 |
+
"loss": 0.5451,
|
| 680 |
+
"rewards/accuracies": 0.78125,
|
| 681 |
+
"rewards/chosen": -0.00384521484375,
|
| 682 |
+
"rewards/margins": 0.64453125,
|
| 683 |
+
"rewards/rejected": -0.6484375,
|
| 684 |
+
"step": 450
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 0.48142333856619574,
|
| 688 |
+
"grad_norm": 7.323412895202637,
|
| 689 |
+
"learning_rate": 5.716249049740689e-07,
|
| 690 |
+
"logits/chosen": -0.66015625,
|
| 691 |
+
"logits/rejected": -0.6015625,
|
| 692 |
+
"logps/chosen": -340.0,
|
| 693 |
+
"logps/rejected": -328.0,
|
| 694 |
+
"loss": 0.5538,
|
| 695 |
+
"rewards/accuracies": 0.7562500238418579,
|
| 696 |
+
"rewards/chosen": 0.01373291015625,
|
| 697 |
+
"rewards/margins": 0.72265625,
|
| 698 |
+
"rewards/rejected": -0.7109375,
|
| 699 |
+
"step": 460
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.49188906331763477,
|
| 703 |
+
"grad_norm": 6.901383876800537,
|
| 704 |
+
"learning_rate": 5.544453927272492e-07,
|
| 705 |
+
"logits/chosen": -0.69140625,
|
| 706 |
+
"logits/rejected": -0.76171875,
|
| 707 |
+
"logps/chosen": -326.0,
|
| 708 |
+
"logps/rejected": -316.0,
|
| 709 |
+
"loss": 0.5626,
|
| 710 |
+
"rewards/accuracies": 0.731249988079071,
|
| 711 |
+
"rewards/chosen": -0.1630859375,
|
| 712 |
+
"rewards/margins": 0.55078125,
|
| 713 |
+
"rewards/rejected": -0.71484375,
|
| 714 |
+
"step": 470
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"epoch": 0.5023547880690737,
|
| 718 |
+
"grad_norm": 10.153675079345703,
|
| 719 |
+
"learning_rate": 5.372005669911693e-07,
|
| 720 |
+
"logits/chosen": -0.64453125,
|
| 721 |
+
"logits/rejected": -0.60546875,
|
| 722 |
+
"logps/chosen": -326.0,
|
| 723 |
+
"logps/rejected": -274.0,
|
| 724 |
+
"loss": 0.586,
|
| 725 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 726 |
+
"rewards/chosen": -0.1884765625,
|
| 727 |
+
"rewards/margins": 0.486328125,
|
| 728 |
+
"rewards/rejected": -0.67578125,
|
| 729 |
+
"step": 480
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.5128205128205128,
|
| 733 |
+
"grad_norm": 6.661205768585205,
|
| 734 |
+
"learning_rate": 5.199111149109497e-07,
|
| 735 |
+
"logits/chosen": -0.78125,
|
| 736 |
+
"logits/rejected": -0.6328125,
|
| 737 |
+
"logps/chosen": -304.0,
|
| 738 |
+
"logps/rejected": -278.0,
|
| 739 |
+
"loss": 0.5948,
|
| 740 |
+
"rewards/accuracies": 0.75,
|
| 741 |
+
"rewards/chosen": 0.003631591796875,
|
| 742 |
+
"rewards/margins": 0.65625,
|
| 743 |
+
"rewards/rejected": -0.65234375,
|
| 744 |
+
"step": 490
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.5232862375719518,
|
| 748 |
+
"grad_norm": 7.033756256103516,
|
| 749 |
+
"learning_rate": 5.025977771661266e-07,
|
| 750 |
+
"logits/chosen": -0.68359375,
|
| 751 |
+
"logits/rejected": -0.69921875,
|
| 752 |
+
"logps/chosen": -292.0,
|
| 753 |
+
"logps/rejected": -304.0,
|
| 754 |
+
"loss": 0.5819,
|
| 755 |
+
"rewards/accuracies": 0.606249988079071,
|
| 756 |
+
"rewards/chosen": -0.185546875,
|
| 757 |
+
"rewards/margins": 0.37890625,
|
| 758 |
+
"rewards/rejected": -0.5625,
|
| 759 |
+
"step": 500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.533751962323391,
|
| 763 |
+
"grad_norm": 7.984467506408691,
|
| 764 |
+
"learning_rate": 4.852813230898279e-07,
|
| 765 |
+
"logits/chosen": -0.63671875,
|
| 766 |
+
"logits/rejected": -0.640625,
|
| 767 |
+
"logps/chosen": -350.0,
|
| 768 |
+
"logps/rejected": -302.0,
|
| 769 |
+
"loss": 0.6104,
|
| 770 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 771 |
+
"rewards/chosen": 0.01904296875,
|
| 772 |
+
"rewards/margins": 0.62109375,
|
| 773 |
+
"rewards/rejected": -0.6015625,
|
| 774 |
+
"step": 510
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 0.54421768707483,
|
| 778 |
+
"grad_norm": 6.375154495239258,
|
| 779 |
+
"learning_rate": 4.679825257535794e-07,
|
| 780 |
+
"logits/chosen": -0.68359375,
|
| 781 |
+
"logits/rejected": -0.74609375,
|
| 782 |
+
"logps/chosen": -338.0,
|
| 783 |
+
"logps/rejected": -322.0,
|
| 784 |
+
"loss": 0.5167,
|
| 785 |
+
"rewards/accuracies": 0.737500011920929,
|
| 786 |
+
"rewards/chosen": 0.039794921875,
|
| 787 |
+
"rewards/margins": 0.69140625,
|
| 788 |
+
"rewards/rejected": -0.65234375,
|
| 789 |
+
"step": 520
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.554683411826269,
|
| 793 |
+
"grad_norm": 6.771985054016113,
|
| 794 |
+
"learning_rate": 4.507221370476223e-07,
|
| 795 |
+
"logits/chosen": -0.7421875,
|
| 796 |
+
"logits/rejected": -0.703125,
|
| 797 |
+
"logps/chosen": -312.0,
|
| 798 |
+
"logps/rejected": -320.0,
|
| 799 |
+
"loss": 0.547,
|
| 800 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 801 |
+
"rewards/chosen": 0.064453125,
|
| 802 |
+
"rewards/margins": 0.6796875,
|
| 803 |
+
"rewards/rejected": -0.61328125,
|
| 804 |
+
"step": 530
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 0.565149136577708,
|
| 808 |
+
"grad_norm": 7.244710922241211,
|
| 809 |
+
"learning_rate": 4.3352086278664377e-07,
|
| 810 |
+
"logits/chosen": -0.76171875,
|
| 811 |
+
"logits/rejected": -0.75,
|
| 812 |
+
"logps/chosen": -302.0,
|
| 813 |
+
"logps/rejected": -266.0,
|
| 814 |
+
"loss": 0.5402,
|
| 815 |
+
"rewards/accuracies": 0.6875,
|
| 816 |
+
"rewards/chosen": -0.07177734375,
|
| 817 |
+
"rewards/margins": 0.625,
|
| 818 |
+
"rewards/rejected": -0.6953125,
|
| 819 |
+
"step": 540
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.5756148613291471,
|
| 823 |
+
"grad_norm": 8.408710479736328,
|
| 824 |
+
"learning_rate": 4.1639933787077854e-07,
|
| 825 |
+
"logits/chosen": -0.6796875,
|
| 826 |
+
"logits/rejected": -0.58984375,
|
| 827 |
+
"logps/chosen": -306.0,
|
| 828 |
+
"logps/rejected": -294.0,
|
| 829 |
+
"loss": 0.552,
|
| 830 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 831 |
+
"rewards/chosen": 0.060302734375,
|
| 832 |
+
"rewards/margins": 0.6015625,
|
| 833 |
+
"rewards/rejected": -0.54296875,
|
| 834 |
+
"step": 550
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 0.5860805860805861,
|
| 838 |
+
"grad_norm": 7.500967979431152,
|
| 839 |
+
"learning_rate": 3.9937810153168016e-07,
|
| 840 |
+
"logits/chosen": -0.6171875,
|
| 841 |
+
"logits/rejected": -0.71484375,
|
| 842 |
+
"logps/chosen": -352.0,
|
| 843 |
+
"logps/rejected": -320.0,
|
| 844 |
+
"loss": 0.5782,
|
| 845 |
+
"rewards/accuracies": 0.65625,
|
| 846 |
+
"rewards/chosen": -0.08154296875,
|
| 847 |
+
"rewards/margins": 0.466796875,
|
| 848 |
+
"rewards/rejected": -0.546875,
|
| 849 |
+
"step": 560
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.5965463108320251,
|
| 853 |
+
"grad_norm": 6.775027751922607,
|
| 854 |
+
"learning_rate": 3.8247757269335957e-07,
|
| 855 |
+
"logits/chosen": -0.75390625,
|
| 856 |
+
"logits/rejected": -0.72265625,
|
| 857 |
+
"logps/chosen": -314.0,
|
| 858 |
+
"logps/rejected": -292.0,
|
| 859 |
+
"loss": 0.5366,
|
| 860 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 861 |
+
"rewards/chosen": 0.0106201171875,
|
| 862 |
+
"rewards/margins": 0.6328125,
|
| 863 |
+
"rewards/rejected": -0.625,
|
| 864 |
+
"step": 570
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.6070120355834642,
|
| 868 |
+
"grad_norm": 7.576064586639404,
|
| 869 |
+
"learning_rate": 3.657180254773445e-07,
|
| 870 |
+
"logits/chosen": -0.6640625,
|
| 871 |
+
"logits/rejected": -0.58203125,
|
| 872 |
+
"logps/chosen": -310.0,
|
| 873 |
+
"logps/rejected": -264.0,
|
| 874 |
+
"loss": 0.5501,
|
| 875 |
+
"rewards/accuracies": 0.762499988079071,
|
| 876 |
+
"rewards/chosen": -0.0281982421875,
|
| 877 |
+
"rewards/margins": 0.70703125,
|
| 878 |
+
"rewards/rejected": -0.734375,
|
| 879 |
+
"step": 580
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.6174777603349032,
|
| 883 |
+
"grad_norm": 7.297247409820557,
|
| 884 |
+
"learning_rate": 3.4911956488154694e-07,
|
| 885 |
+
"logits/chosen": -0.70703125,
|
| 886 |
+
"logits/rejected": -0.625,
|
| 887 |
+
"logps/chosen": -334.0,
|
| 888 |
+
"logps/rejected": -298.0,
|
| 889 |
+
"loss": 0.5798,
|
| 890 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 891 |
+
"rewards/chosen": -0.038330078125,
|
| 892 |
+
"rewards/margins": 0.5546875,
|
| 893 |
+
"rewards/rejected": -0.59375,
|
| 894 |
+
"step": 590
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"epoch": 0.6279434850863422,
|
| 898 |
+
"grad_norm": 9.57577133178711,
|
| 899 |
+
"learning_rate": 3.327021026620137e-07,
|
| 900 |
+
"logits/chosen": -0.640625,
|
| 901 |
+
"logits/rejected": -0.625,
|
| 902 |
+
"logps/chosen": -328.0,
|
| 903 |
+
"logps/rejected": -314.0,
|
| 904 |
+
"loss": 0.5525,
|
| 905 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 906 |
+
"rewards/chosen": -0.03515625,
|
| 907 |
+
"rewards/margins": 0.62890625,
|
| 908 |
+
"rewards/rejected": -0.6640625,
|
| 909 |
+
"step": 600
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.6384092098377813,
|
| 913 |
+
"grad_norm": 8.11950397491455,
|
| 914 |
+
"learning_rate": 3.16485333446493e-07,
|
| 915 |
+
"logits/chosen": -0.65625,
|
| 916 |
+
"logits/rejected": -0.5390625,
|
| 917 |
+
"logps/chosen": -300.0,
|
| 918 |
+
"logps/rejected": -356.0,
|
| 919 |
+
"loss": 0.5299,
|
| 920 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 921 |
+
"rewards/chosen": -0.0296630859375,
|
| 922 |
+
"rewards/margins": 0.66015625,
|
| 923 |
+
"rewards/rejected": -0.69140625,
|
| 924 |
+
"step": 610
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"epoch": 0.6488749345892203,
|
| 928 |
+
"grad_norm": 8.621895790100098,
|
| 929 |
+
"learning_rate": 3.004887111084704e-07,
|
| 930 |
+
"logits/chosen": -0.7421875,
|
| 931 |
+
"logits/rejected": -0.8046875,
|
| 932 |
+
"logps/chosen": -324.0,
|
| 933 |
+
"logps/rejected": -296.0,
|
| 934 |
+
"loss": 0.554,
|
| 935 |
+
"rewards/accuracies": 0.699999988079071,
|
| 936 |
+
"rewards/chosen": 0.06640625,
|
| 937 |
+
"rewards/margins": 0.640625,
|
| 938 |
+
"rewards/rejected": -0.57421875,
|
| 939 |
+
"step": 620
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.6593406593406593,
|
| 943 |
+
"grad_norm": 7.23579740524292,
|
| 944 |
+
"learning_rate": 2.8473142543001816e-07,
|
| 945 |
+
"logits/chosen": -0.73828125,
|
| 946 |
+
"logits/rejected": -0.71484375,
|
| 947 |
+
"logps/chosen": -304.0,
|
| 948 |
+
"logps/rejected": -300.0,
|
| 949 |
+
"loss": 0.554,
|
| 950 |
+
"rewards/accuracies": 0.7124999761581421,
|
| 951 |
+
"rewards/chosen": -0.0615234375,
|
| 952 |
+
"rewards/margins": 0.65625,
|
| 953 |
+
"rewards/rejected": -0.71875,
|
| 954 |
+
"step": 630
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.6698063840920984,
|
| 958 |
+
"grad_norm": 7.622311115264893,
|
| 959 |
+
"learning_rate": 2.6923237908145226e-07,
|
| 960 |
+
"logits/chosen": -0.765625,
|
| 961 |
+
"logits/rejected": -0.66796875,
|
| 962 |
+
"logps/chosen": -314.0,
|
| 963 |
+
"logps/rejected": -312.0,
|
| 964 |
+
"loss": 0.5305,
|
| 965 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 966 |
+
"rewards/chosen": 0.0216064453125,
|
| 967 |
+
"rewards/margins": 0.609375,
|
| 968 |
+
"rewards/rejected": -0.5859375,
|
| 969 |
+
"step": 640
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.6802721088435374,
|
| 973 |
+
"grad_norm": 7.722968578338623,
|
| 974 |
+
"learning_rate": 2.540101649454119e-07,
|
| 975 |
+
"logits/chosen": -0.57421875,
|
| 976 |
+
"logits/rejected": -0.625,
|
| 977 |
+
"logps/chosen": -318.0,
|
| 978 |
+
"logps/rejected": -304.0,
|
| 979 |
+
"loss": 0.5505,
|
| 980 |
+
"rewards/accuracies": 0.762499988079071,
|
| 981 |
+
"rewards/chosen": 0.0615234375,
|
| 982 |
+
"rewards/margins": 0.72265625,
|
| 983 |
+
"rewards/rejected": -0.66015625,
|
| 984 |
+
"step": 650
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 0.6907378335949764,
|
| 988 |
+
"grad_norm": 9.817713737487793,
|
| 989 |
+
"learning_rate": 2.3908304381256603e-07,
|
| 990 |
+
"logits/chosen": -0.7578125,
|
| 991 |
+
"logits/rejected": -0.58984375,
|
| 992 |
+
"logps/chosen": -300.0,
|
| 993 |
+
"logps/rejected": -290.0,
|
| 994 |
+
"loss": 0.5692,
|
| 995 |
+
"rewards/accuracies": 0.737500011920929,
|
| 996 |
+
"rewards/chosen": -0.1904296875,
|
| 997 |
+
"rewards/margins": 0.60546875,
|
| 998 |
+
"rewards/rejected": -0.796875,
|
| 999 |
+
"step": 660
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.7012035583464155,
|
| 1003 |
+
"grad_norm": 8.339869499206543,
|
| 1004 |
+
"learning_rate": 2.2446892247570255e-07,
|
| 1005 |
+
"logits/chosen": -0.70703125,
|
| 1006 |
+
"logits/rejected": -0.72265625,
|
| 1007 |
+
"logps/chosen": -306.0,
|
| 1008 |
+
"logps/rejected": -300.0,
|
| 1009 |
+
"loss": 0.5817,
|
| 1010 |
+
"rewards/accuracies": 0.75,
|
| 1011 |
+
"rewards/chosen": 0.0255126953125,
|
| 1012 |
+
"rewards/margins": 0.765625,
|
| 1013 |
+
"rewards/rejected": -0.7421875,
|
| 1014 |
+
"step": 670
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"epoch": 0.7116692830978545,
|
| 1018 |
+
"grad_norm": 8.644556045532227,
|
| 1019 |
+
"learning_rate": 2.1018533224847633e-07,
|
| 1020 |
+
"logits/chosen": -0.640625,
|
| 1021 |
+
"logits/rejected": -0.6015625,
|
| 1022 |
+
"logps/chosen": -302.0,
|
| 1023 |
+
"logps/rejected": -276.0,
|
| 1024 |
+
"loss": 0.5732,
|
| 1025 |
+
"rewards/accuracies": 0.706250011920929,
|
| 1026 |
+
"rewards/chosen": 0.10009765625,
|
| 1027 |
+
"rewards/margins": 0.69140625,
|
| 1028 |
+
"rewards/rejected": -0.59375,
|
| 1029 |
+
"step": 680
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.7221350078492935,
|
| 1033 |
+
"grad_norm": 6.904910087585449,
|
| 1034 |
+
"learning_rate": 1.9624940793459055e-07,
|
| 1035 |
+
"logits/chosen": -0.78515625,
|
| 1036 |
+
"logits/rejected": -0.67578125,
|
| 1037 |
+
"logps/chosen": -332.0,
|
| 1038 |
+
"logps/rejected": -306.0,
|
| 1039 |
+
"loss": 0.5436,
|
| 1040 |
+
"rewards/accuracies": 0.6875,
|
| 1041 |
+
"rewards/chosen": 0.07958984375,
|
| 1042 |
+
"rewards/margins": 0.625,
|
| 1043 |
+
"rewards/rejected": -0.546875,
|
| 1044 |
+
"step": 690
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"epoch": 0.7326007326007326,
|
| 1048 |
+
"grad_norm": 8.480828285217285,
|
| 1049 |
+
"learning_rate": 1.8267786727263424e-07,
|
| 1050 |
+
"logits/chosen": -0.64453125,
|
| 1051 |
+
"logits/rejected": -0.52734375,
|
| 1052 |
+
"logps/chosen": -272.0,
|
| 1053 |
+
"logps/rejected": -270.0,
|
| 1054 |
+
"loss": 0.5594,
|
| 1055 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1056 |
+
"rewards/chosen": -0.06591796875,
|
| 1057 |
+
"rewards/margins": 0.333984375,
|
| 1058 |
+
"rewards/rejected": -0.400390625,
|
| 1059 |
+
"step": 700
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7430664573521716,
|
| 1063 |
+
"grad_norm": 9.295394897460938,
|
| 1064 |
+
"learning_rate": 1.694869908812399e-07,
|
| 1065 |
+
"logits/chosen": -0.625,
|
| 1066 |
+
"logits/rejected": -0.73046875,
|
| 1067 |
+
"logps/chosen": -298.0,
|
| 1068 |
+
"logps/rejected": -278.0,
|
| 1069 |
+
"loss": 0.5747,
|
| 1070 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1071 |
+
"rewards/chosen": -0.1005859375,
|
| 1072 |
+
"rewards/margins": 0.478515625,
|
| 1073 |
+
"rewards/rejected": -0.578125,
|
| 1074 |
+
"step": 710
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 0.7535321821036107,
|
| 1078 |
+
"grad_norm": 7.032623767852783,
|
| 1079 |
+
"learning_rate": 1.5669260272861422e-07,
|
| 1080 |
+
"logits/chosen": -0.7109375,
|
| 1081 |
+
"logits/rejected": -0.7578125,
|
| 1082 |
+
"logps/chosen": -280.0,
|
| 1083 |
+
"logps/rejected": -282.0,
|
| 1084 |
+
"loss": 0.546,
|
| 1085 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 1086 |
+
"rewards/chosen": 0.0296630859375,
|
| 1087 |
+
"rewards/margins": 0.63671875,
|
| 1088 |
+
"rewards/rejected": -0.60546875,
|
| 1089 |
+
"step": 720
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.7639979068550498,
|
| 1093 |
+
"grad_norm": 7.640635967254639,
|
| 1094 |
+
"learning_rate": 1.4431005114987483e-07,
|
| 1095 |
+
"logits/chosen": -0.57421875,
|
| 1096 |
+
"logits/rejected": -0.68359375,
|
| 1097 |
+
"logps/chosen": -336.0,
|
| 1098 |
+
"logps/rejected": -300.0,
|
| 1099 |
+
"loss": 0.5721,
|
| 1100 |
+
"rewards/accuracies": 0.6875,
|
| 1101 |
+
"rewards/chosen": -0.0283203125,
|
| 1102 |
+
"rewards/margins": 0.56640625,
|
| 1103 |
+
"rewards/rejected": -0.59765625,
|
| 1104 |
+
"step": 730
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"epoch": 0.7744636316064888,
|
| 1108 |
+
"grad_norm": 8.33640193939209,
|
| 1109 |
+
"learning_rate": 1.323541904349636e-07,
|
| 1110 |
+
"logits/chosen": -0.58984375,
|
| 1111 |
+
"logits/rejected": -0.59765625,
|
| 1112 |
+
"logps/chosen": -338.0,
|
| 1113 |
+
"logps/rejected": -280.0,
|
| 1114 |
+
"loss": 0.5804,
|
| 1115 |
+
"rewards/accuracies": 0.6875,
|
| 1116 |
+
"rewards/chosen": 0.005828857421875,
|
| 1117 |
+
"rewards/margins": 0.62109375,
|
| 1118 |
+
"rewards/rejected": -0.6171875,
|
| 1119 |
+
"step": 740
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.7849293563579278,
|
| 1123 |
+
"grad_norm": 8.349899291992188,
|
| 1124 |
+
"learning_rate": 1.2083936300922237e-07,
|
| 1125 |
+
"logits/chosen": -0.55859375,
|
| 1126 |
+
"logits/rejected": -0.58203125,
|
| 1127 |
+
"logps/chosen": -332.0,
|
| 1128 |
+
"logps/rejected": -334.0,
|
| 1129 |
+
"loss": 0.573,
|
| 1130 |
+
"rewards/accuracies": 0.6875,
|
| 1131 |
+
"rewards/chosen": 0.004302978515625,
|
| 1132 |
+
"rewards/margins": 0.58984375,
|
| 1133 |
+
"rewards/rejected": -0.5859375,
|
| 1134 |
+
"step": 750
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 0.7953950811093669,
|
| 1138 |
+
"grad_norm": 7.29964542388916,
|
| 1139 |
+
"learning_rate": 1.0977938222801004e-07,
|
| 1140 |
+
"logits/chosen": -0.64453125,
|
| 1141 |
+
"logits/rejected": -0.71484375,
|
| 1142 |
+
"logps/chosen": -306.0,
|
| 1143 |
+
"logps/rejected": -286.0,
|
| 1144 |
+
"loss": 0.5763,
|
| 1145 |
+
"rewards/accuracies": 0.6875,
|
| 1146 |
+
"rewards/chosen": 0.053955078125,
|
| 1147 |
+
"rewards/margins": 0.625,
|
| 1148 |
+
"rewards/rejected": -0.5703125,
|
| 1149 |
+
"step": 760
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.8058608058608059,
|
| 1153 |
+
"grad_norm": 6.583461761474609,
|
| 1154 |
+
"learning_rate": 9.918751580599999e-08,
|
| 1155 |
+
"logits/chosen": -0.59375,
|
| 1156 |
+
"logits/rejected": -0.60546875,
|
| 1157 |
+
"logps/chosen": -340.0,
|
| 1158 |
+
"logps/rejected": -296.0,
|
| 1159 |
+
"loss": 0.5757,
|
| 1160 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1161 |
+
"rewards/chosen": 0.02490234375,
|
| 1162 |
+
"rewards/margins": 0.578125,
|
| 1163 |
+
"rewards/rejected": -0.5546875,
|
| 1164 |
+
"step": 770
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.8163265306122449,
|
| 1168 |
+
"grad_norm": 7.065498352050781,
|
| 1169 |
+
"learning_rate": 8.907646990103495e-08,
|
| 1170 |
+
"logits/chosen": -0.5703125,
|
| 1171 |
+
"logits/rejected": -0.59765625,
|
| 1172 |
+
"logps/chosen": -282.0,
|
| 1173 |
+
"logps/rejected": -270.0,
|
| 1174 |
+
"loss": 0.5388,
|
| 1175 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1176 |
+
"rewards/chosen": -0.0341796875,
|
| 1177 |
+
"rewards/margins": 0.61328125,
|
| 1178 |
+
"rewards/rejected": -0.6484375,
|
| 1179 |
+
"step": 780
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.826792255363684,
|
| 1183 |
+
"grad_norm": 6.218137741088867,
|
| 1184 |
+
"learning_rate": 7.945837387163424e-08,
|
| 1185 |
+
"logits/chosen": -0.65625,
|
| 1186 |
+
"logits/rejected": -0.68359375,
|
| 1187 |
+
"logps/chosen": -330.0,
|
| 1188 |
+
"logps/rejected": -340.0,
|
| 1189 |
+
"loss": 0.5702,
|
| 1190 |
+
"rewards/accuracies": 0.7437499761581421,
|
| 1191 |
+
"rewards/chosen": 0.048828125,
|
| 1192 |
+
"rewards/margins": 0.71875,
|
| 1193 |
+
"rewards/rejected": -0.671875,
|
| 1194 |
+
"step": 790
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 0.837257980115123,
|
| 1198 |
+
"grad_norm": 8.78946304321289,
|
| 1199 |
+
"learning_rate": 7.034476572643854e-08,
|
| 1200 |
+
"logits/chosen": -0.5703125,
|
| 1201 |
+
"logits/rejected": -0.578125,
|
| 1202 |
+
"logps/chosen": -300.0,
|
| 1203 |
+
"logps/rejected": -306.0,
|
| 1204 |
+
"loss": 0.5605,
|
| 1205 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1206 |
+
"rewards/chosen": -0.0103759765625,
|
| 1207 |
+
"rewards/margins": 0.53515625,
|
| 1208 |
+
"rewards/rejected": -0.546875,
|
| 1209 |
+
"step": 800
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.847723704866562,
|
| 1213 |
+
"grad_norm": 7.07961893081665,
|
| 1214 |
+
"learning_rate": 6.174657828304541e-08,
|
| 1215 |
+
"logits/chosen": -0.66015625,
|
| 1216 |
+
"logits/rejected": -0.54296875,
|
| 1217 |
+
"logps/chosen": -300.0,
|
| 1218 |
+
"logps/rejected": -296.0,
|
| 1219 |
+
"loss": 0.5812,
|
| 1220 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1221 |
+
"rewards/chosen": -0.01318359375,
|
| 1222 |
+
"rewards/margins": 0.515625,
|
| 1223 |
+
"rewards/rejected": -0.52734375,
|
| 1224 |
+
"step": 810
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"epoch": 0.858189429618001,
|
| 1228 |
+
"grad_norm": 6.761130332946777,
|
| 1229 |
+
"learning_rate": 5.36741260528415e-08,
|
| 1230 |
+
"logits/chosen": -0.640625,
|
| 1231 |
+
"logits/rejected": -0.71484375,
|
| 1232 |
+
"logps/chosen": -336.0,
|
| 1233 |
+
"logps/rejected": -286.0,
|
| 1234 |
+
"loss": 0.5308,
|
| 1235 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1236 |
+
"rewards/chosen": 0.12109375,
|
| 1237 |
+
"rewards/margins": 0.74609375,
|
| 1238 |
+
"rewards/rejected": -0.625,
|
| 1239 |
+
"step": 820
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 0.8686551543694401,
|
| 1243 |
+
"grad_norm": 6.658000469207764,
|
| 1244 |
+
"learning_rate": 4.613709286756412e-08,
|
| 1245 |
+
"logits/chosen": -0.625,
|
| 1246 |
+
"logits/rejected": -0.61328125,
|
| 1247 |
+
"logps/chosen": -302.0,
|
| 1248 |
+
"logps/rejected": -276.0,
|
| 1249 |
+
"loss": 0.5411,
|
| 1250 |
+
"rewards/accuracies": 0.762499988079071,
|
| 1251 |
+
"rewards/chosen": 0.053466796875,
|
| 1252 |
+
"rewards/margins": 0.7421875,
|
| 1253 |
+
"rewards/rejected": -0.6875,
|
| 1254 |
+
"step": 830
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"epoch": 0.8791208791208791,
|
| 1258 |
+
"grad_norm": 7.955780506134033,
|
| 1259 |
+
"learning_rate": 3.914452026243509e-08,
|
| 1260 |
+
"logits/chosen": -0.6796875,
|
| 1261 |
+
"logits/rejected": -0.65625,
|
| 1262 |
+
"logps/chosen": -332.0,
|
| 1263 |
+
"logps/rejected": -308.0,
|
| 1264 |
+
"loss": 0.549,
|
| 1265 |
+
"rewards/accuracies": 0.65625,
|
| 1266 |
+
"rewards/chosen": 0.018310546875,
|
| 1267 |
+
"rewards/margins": 0.4921875,
|
| 1268 |
+
"rewards/rejected": -0.474609375,
|
| 1269 |
+
"step": 840
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.8895866038723181,
|
| 1273 |
+
"grad_norm": 8.435918807983398,
|
| 1274 |
+
"learning_rate": 3.270479662980247e-08,
|
| 1275 |
+
"logits/chosen": -0.5546875,
|
| 1276 |
+
"logits/rejected": -0.73828125,
|
| 1277 |
+
"logps/chosen": -306.0,
|
| 1278 |
+
"logps/rejected": -272.0,
|
| 1279 |
+
"loss": 0.5817,
|
| 1280 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1281 |
+
"rewards/chosen": 0.068359375,
|
| 1282 |
+
"rewards/margins": 0.54296875,
|
| 1283 |
+
"rewards/rejected": -0.47265625,
|
| 1284 |
+
"step": 850
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 0.9000523286237572,
|
| 1288 |
+
"grad_norm": 8.153657913208008,
|
| 1289 |
+
"learning_rate": 2.6825647156302865e-08,
|
| 1290 |
+
"logits/chosen": -0.6015625,
|
| 1291 |
+
"logits/rejected": -0.5078125,
|
| 1292 |
+
"logps/chosen": -330.0,
|
| 1293 |
+
"logps/rejected": -348.0,
|
| 1294 |
+
"loss": 0.5387,
|
| 1295 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1296 |
+
"rewards/chosen": -0.0135498046875,
|
| 1297 |
+
"rewards/margins": 0.62890625,
|
| 1298 |
+
"rewards/rejected": -0.64453125,
|
| 1299 |
+
"step": 860
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 0.9105180533751962,
|
| 1303 |
+
"grad_norm": 8.690088272094727,
|
| 1304 |
+
"learning_rate": 2.151412455561441e-08,
|
| 1305 |
+
"logits/chosen": -0.62109375,
|
| 1306 |
+
"logits/rejected": -0.65625,
|
| 1307 |
+
"logps/chosen": -320.0,
|
| 1308 |
+
"logps/rejected": -286.0,
|
| 1309 |
+
"loss": 0.5855,
|
| 1310 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1311 |
+
"rewards/chosen": -0.018310546875,
|
| 1312 |
+
"rewards/margins": 0.5625,
|
| 1313 |
+
"rewards/rejected": -0.578125,
|
| 1314 |
+
"step": 870
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"epoch": 0.9209837781266352,
|
| 1318 |
+
"grad_norm": 7.3918070793151855,
|
| 1319 |
+
"learning_rate": 1.6776600607918356e-08,
|
| 1320 |
+
"logits/chosen": -0.66015625,
|
| 1321 |
+
"logits/rejected": -0.71484375,
|
| 1322 |
+
"logps/chosen": -304.0,
|
| 1323 |
+
"logps/rejected": -266.0,
|
| 1324 |
+
"loss": 0.5473,
|
| 1325 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1326 |
+
"rewards/chosen": 0.00152587890625,
|
| 1327 |
+
"rewards/margins": 0.5859375,
|
| 1328 |
+
"rewards/rejected": -0.5859375,
|
| 1329 |
+
"step": 880
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 0.9314495028780743,
|
| 1333 |
+
"grad_norm": 7.804713726043701,
|
| 1334 |
+
"learning_rate": 1.2618758516218186e-08,
|
| 1335 |
+
"logits/chosen": -0.6484375,
|
| 1336 |
+
"logits/rejected": -0.62890625,
|
| 1337 |
+
"logps/chosen": -322.0,
|
| 1338 |
+
"logps/rejected": -300.0,
|
| 1339 |
+
"loss": 0.5591,
|
| 1340 |
+
"rewards/accuracies": 0.78125,
|
| 1341 |
+
"rewards/chosen": 0.12109375,
|
| 1342 |
+
"rewards/margins": 0.6796875,
|
| 1343 |
+
"rewards/rejected": -0.55859375,
|
| 1344 |
+
"step": 890
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 0.9419152276295133,
|
| 1348 |
+
"grad_norm": 7.053552627563477,
|
| 1349 |
+
"learning_rate": 9.045586088686496e-09,
|
| 1350 |
+
"logits/chosen": -0.63671875,
|
| 1351 |
+
"logits/rejected": -0.67578125,
|
| 1352 |
+
"logps/chosen": -346.0,
|
| 1353 |
+
"logps/rejected": -292.0,
|
| 1354 |
+
"loss": 0.5594,
|
| 1355 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1356 |
+
"rewards/chosen": 0.1943359375,
|
| 1357 |
+
"rewards/margins": 0.61328125,
|
| 1358 |
+
"rewards/rejected": -0.41796875,
|
| 1359 |
+
"step": 900
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 0.9523809523809523,
|
| 1363 |
+
"grad_norm": 7.403259754180908,
|
| 1364 |
+
"learning_rate": 6.06136975521715e-09,
|
| 1365 |
+
"logits/chosen": -0.73828125,
|
| 1366 |
+
"logits/rejected": -0.69921875,
|
| 1367 |
+
"logps/chosen": -278.0,
|
| 1368 |
+
"logps/rejected": -264.0,
|
| 1369 |
+
"loss": 0.5563,
|
| 1370 |
+
"rewards/accuracies": 0.7875000238418579,
|
| 1371 |
+
"rewards/chosen": 0.035888671875,
|
| 1372 |
+
"rewards/margins": 0.6640625,
|
| 1373 |
+
"rewards/rejected": -0.62890625,
|
| 1374 |
+
"step": 910
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.9628466771323915,
|
| 1378 |
+
"grad_norm": 7.359611511230469,
|
| 1379 |
+
"learning_rate": 3.6696894253614442e-09,
|
| 1380 |
+
"logits/chosen": -0.64453125,
|
| 1381 |
+
"logits/rejected": -0.69140625,
|
| 1382 |
+
"logps/chosen": -322.0,
|
| 1383 |
+
"logps/rejected": -302.0,
|
| 1384 |
+
"loss": 0.5624,
|
| 1385 |
+
"rewards/accuracies": 0.731249988079071,
|
| 1386 |
+
"rewards/chosen": 0.05712890625,
|
| 1387 |
+
"rewards/margins": 0.671875,
|
| 1388 |
+
"rewards/rejected": -0.61328125,
|
| 1389 |
+
"step": 920
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 0.9733124018838305,
|
| 1393 |
+
"grad_norm": 7.091738700866699,
|
| 1394 |
+
"learning_rate": 1.8734141938160918e-09,
|
| 1395 |
+
"logits/chosen": -0.59375,
|
| 1396 |
+
"logits/rejected": -0.55078125,
|
| 1397 |
+
"logps/chosen": -328.0,
|
| 1398 |
+
"logps/rejected": -310.0,
|
| 1399 |
+
"loss": 0.5677,
|
| 1400 |
+
"rewards/accuracies": 0.6812499761581421,
|
| 1401 |
+
"rewards/chosen": 0.038330078125,
|
| 1402 |
+
"rewards/margins": 0.62890625,
|
| 1403 |
+
"rewards/rejected": -0.59375,
|
| 1404 |
+
"step": 930
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"epoch": 0.9837781266352695,
|
| 1408 |
+
"grad_norm": 6.281693935394287,
|
| 1409 |
+
"learning_rate": 6.746988986155999e-10,
|
| 1410 |
+
"logits/chosen": -0.7109375,
|
| 1411 |
+
"logits/rejected": -0.625,
|
| 1412 |
+
"logps/chosen": -330.0,
|
| 1413 |
+
"logps/rejected": -310.0,
|
| 1414 |
+
"loss": 0.5594,
|
| 1415 |
+
"rewards/accuracies": 0.7250000238418579,
|
| 1416 |
+
"rewards/chosen": 0.00750732421875,
|
| 1417 |
+
"rewards/margins": 0.6171875,
|
| 1418 |
+
"rewards/rejected": -0.609375,
|
| 1419 |
+
"step": 940
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 0.9942438513867086,
|
| 1423 |
+
"grad_norm": 7.982404708862305,
|
| 1424 |
+
"learning_rate": 7.498153615653758e-11,
|
| 1425 |
+
"logits/chosen": -0.671875,
|
| 1426 |
+
"logits/rejected": -0.68359375,
|
| 1427 |
+
"logps/chosen": -312.0,
|
| 1428 |
+
"logps/rejected": -314.0,
|
| 1429 |
+
"loss": 0.5419,
|
| 1430 |
+
"rewards/accuracies": 0.6937500238418579,
|
| 1431 |
+
"rewards/chosen": -0.027587890625,
|
| 1432 |
+
"rewards/margins": 0.59375,
|
| 1433 |
+
"rewards/rejected": -0.62109375,
|
| 1434 |
+
"step": 950
|
| 1435 |
+
}
|
| 1436 |
+
],
|
| 1437 |
+
"logging_steps": 10,
|
| 1438 |
+
"max_steps": 955,
|
| 1439 |
+
"num_input_tokens_seen": 0,
|
| 1440 |
+
"num_train_epochs": 1,
|
| 1441 |
+
"save_steps": 500,
|
| 1442 |
+
"stateful_callbacks": {
|
| 1443 |
+
"TrainerControl": {
|
| 1444 |
+
"args": {
|
| 1445 |
+
"should_epoch_stop": false,
|
| 1446 |
+
"should_evaluate": false,
|
| 1447 |
+
"should_log": false,
|
| 1448 |
+
"should_save": true,
|
| 1449 |
+
"should_training_stop": true
|
| 1450 |
+
},
|
| 1451 |
+
"attributes": {}
|
| 1452 |
+
}
|
| 1453 |
+
},
|
| 1454 |
+
"total_flos": 0.0,
|
| 1455 |
+
"train_batch_size": 8,
|
| 1456 |
+
"trial_name": null,
|
| 1457 |
+
"trial_params": null
|
| 1458 |
+
}
|
config.json
CHANGED
|
@@ -1,11 +1,11 @@
|
|
| 1 |
{
|
| 2 |
-
"_name_or_path": "
|
| 3 |
"architectures": [
|
| 4 |
"Qwen2ForCausalLM"
|
| 5 |
],
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
-
"eos_token_id":
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 1536,
|
| 11 |
"initializer_range": 0.02,
|
|
@@ -22,7 +22,7 @@
|
|
| 22 |
"sliding_window": null,
|
| 23 |
"tie_word_embeddings": true,
|
| 24 |
"torch_dtype": "bfloat16",
|
| 25 |
-
"transformers_version": "4.46.
|
| 26 |
"use_cache": false,
|
| 27 |
"use_mrope": false,
|
| 28 |
"use_sliding_window": false,
|
|
|
|
| 1 |
{
|
| 2 |
+
"_name_or_path": "models/Qwen2.5-1.5B-ultrachat200k",
|
| 3 |
"architectures": [
|
| 4 |
"Qwen2ForCausalLM"
|
| 5 |
],
|
| 6 |
"attention_dropout": 0.0,
|
| 7 |
"bos_token_id": 151643,
|
| 8 |
+
"eos_token_id": 151645,
|
| 9 |
"hidden_act": "silu",
|
| 10 |
"hidden_size": 1536,
|
| 11 |
"initializer_range": 0.02,
|
|
|
|
| 22 |
"sliding_window": null,
|
| 23 |
"tie_word_embeddings": true,
|
| 24 |
"torch_dtype": "bfloat16",
|
| 25 |
+
"transformers_version": "4.46.3",
|
| 26 |
"use_cache": false,
|
| 27 |
"use_mrope": false,
|
| 28 |
"use_sliding_window": false,
|
generation_config.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"bos_token_id": 151643,
|
| 3 |
-
"eos_token_id":
|
| 4 |
"max_new_tokens": 2048,
|
| 5 |
-
"transformers_version": "4.46.
|
| 6 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151645,
|
| 4 |
"max_new_tokens": 2048,
|
| 5 |
+
"transformers_version": "4.46.3"
|
| 6 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3087467144
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6c378882770842e588c5181a45492e6a710070ef0881fea319f3d959489b92e8
|
| 3 |
size 3087467144
|
runs/Dec17_21-04-23_notebook-7f74bd20-8f92-45b0-96c4-687ced4ee944/events.out.tfevents.1734440689.notebook-7f74bd20-8f92-45b0-96c4-687ced4ee944.50109.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:780a967f2211bee39e7d093e3292b6c91b2aa1baf92c9831dc73b66e851b3c94
|
| 3 |
+
size 71689
|
tokenizer.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9c5ae00e602b8860cbd784ba82a8aa14e8feecec692e7076590d014d7b7fdafa
|
| 3 |
+
size 11421896
|
tokenizer_config.json
CHANGED
|
@@ -195,7 +195,7 @@
|
|
| 195 |
"<|video_pad|>"
|
| 196 |
],
|
| 197 |
"bos_token": null,
|
| 198 |
-
"chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are
|
| 199 |
"clean_up_tokenization_spaces": false,
|
| 200 |
"eos_token": "<|im_end|>",
|
| 201 |
"errors": "replace",
|
|
|
|
| 195 |
"<|video_pad|>"
|
| 196 |
],
|
| 197 |
"bos_token": null,
|
| 198 |
+
"chat_template": "{%- if tools %}\n {{- '<|im_start|>system\\n' }}\n {%- if messages[0]['role'] == 'system' %}\n {{- messages[0]['content'] }}\n {%- else %}\n {{- 'You are a helpful assistant.' }}\n {%- endif %}\n {{- \"\\n\\n# Tools\\n\\nYou may call one or more functions to assist with the user query.\\n\\nYou are provided with function signatures within <tools></tools> XML tags:\\n<tools>\" }}\n {%- for tool in tools %}\n {{- \"\\n\" }}\n {{- tool | tojson }}\n {%- endfor %}\n {{- \"\\n</tools>\\n\\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\\n<tool_call>\\n{\\\"name\\\": <function-name>, \\\"arguments\\\": <args-json-object>}\\n</tool_call><|im_end|>\\n\" }}\n{%- else %}\n {%- if messages[0]['role'] == 'system' %}\n {{- '<|im_start|>system\\n' + messages[0]['content'] + '<|im_end|>\\n' }}\n {%- else %}\n {{- '<|im_start|>system\\nYou are a helpful assistant.<|im_end|>\\n' }}\n {%- endif %}\n{%- endif %}\n{%- for message in messages %}\n {%- if (message.role == \"user\") or (message.role == \"system\" and not loop.first) or (message.role == \"assistant\" and not message.tool_calls) %}\n {{- '<|im_start|>' + message.role + '\\n' + message.content + '<|im_end|>' + '\\n' }}\n {%- elif message.role == \"assistant\" %}\n {{- '<|im_start|>' + message.role }}\n {%- if message.content %}\n {{- '\\n' + message.content }}\n {%- endif %}\n {%- for tool_call in message.tool_calls %}\n {%- if tool_call.function is defined %}\n {%- set tool_call = tool_call.function %}\n {%- endif %}\n {{- '\\n<tool_call>\\n{\"name\": \"' }}\n {{- tool_call.name }}\n {{- '\", \"arguments\": ' }}\n {{- tool_call.arguments | tojson }}\n {{- '}\\n</tool_call>' }}\n {%- endfor %}\n {{- '<|im_end|>\\n' }}\n {%- elif message.role == \"tool\" %}\n {%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != \"tool\") %}\n {{- '<|im_start|>user' }}\n {%- endif %}\n {{- '\\n<tool_response>\\n' }}\n {{- message.content }}\n {{- '\\n</tool_response>' }}\n {%- if loop.last or (messages[loop.index0 + 1].role != \"tool\") %}\n {{- '<|im_end|>\\n' }}\n {%- endif %}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|im_start|>assistant\\n' }}\n{%- endif %}\n",
|
| 199 |
"clean_up_tokenization_spaces": false,
|
| 200 |
"eos_token": "<|im_end|>",
|
| 201 |
"errors": "replace",
|
trainer_state.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ac281589eee6ee1b6619ff2dd37a5f8ca4fa27790f53025aa0778b59e86249c
|
| 3 |
+
size 7608
|