WayXG
Results
1
comments of
WayXG
I think the llama3 with padding is obtained by adding a pad token to the original llama model. This can be done by calling the pair-pm/prepare_model.py script. I did so...