Hi,
Thank you so much for the great work! I was wondering if you have also tried the NeRF-synthetic dataset, where the Gaussians are initialized randomly (in the original paper). If so, how are the results for this type of synthetic bounded scenes? Thanks in advance!