Jie Yang
Jie Yang
This PR adds `state_dict` and `load_state_dict` for current LR Schedulers. Please check it!
Platform: two different cards, 1080Ti, 2080Ti, run on 2080Ti will report it, 1080Ti is well.   Do you have any idea? Thanks!
Very wonderful work on the decompostion of NeRF. I want to try you demo. But, I have encounterd a problem when running the training script. `ModuleNotFoundError: No module named 'cosine_triline_feature_cuda'`...