update modeling_baichuan.py for torchscript mode with past_kv

#5

to enable model inference with use_cache and return_dict from model.config.

GradientGuru changed pull request status to merged

Sign up or log in to comment