PSNR anomaly #1236
Description
when I used the latest 3dgs to reconstruct the MipNerf dataset, I found that the measured indicators were much higher than those given in the paper, and many scenarios hit more than 30, is it normal? Or am I having a problem running it