Inquiry for sparse_view_eval scripts

Hi, thank you for sharing this great work! I noticed that there are inference scripts for both single-view and sparse-view settings. However, it seems that the evaluation script only supports the single_view_eval mode. Could you please advise on how to perform quantitative evaluation for multiple or sparse reference view settings? Thanks in advance!