multinerf
multinerf copied to clipboard
Ref-NeRF real capture dataset (GardenSphere) performance
Thank you for the great work. Could you please specify at what resolution the results from the real capture dataset in the Ref-NeRF paper were measured? I trained using the config mentioned in issue #53, and obtained an average PSNR of only 21.8002, compared to the 23.46 reported in the paper. Thanks a lot.