Mijeong Kim

I used the entropy_ray_zvals_lambda = 0.01 for the LLFF dataset. (The LLFF experiment in the supplementary document did not use the information gain reduction loss.) However, I recommend using information...

For the synthetic dataset, the information gain reduction loss is not necessary because the sampled images usually have wide-baseline. For the role of KL loss in the appendix, we used...

Please check the updated checkpoints on LLFF and DTU [here](https://drive.google.com/drive/folders/1sw1uAmSVw1K3Mw76yRCk1jD7muQH9yVS).

I found the answer at the closed issue about Thingi-32 Could I know the mesh models that are downloaded from TurboSquid?