update modeling_baichuan.py for torchscript mode with past_kv
#5
by
changwangss
- opened
to enable model inference with use_cache and return_dict from model.config.
GradientGuru
changed pull request status to
merged