Bad performance on other dataset.

[qhdqhd]

Thanks for your great work!
I test your method on the blender dataset (lego), and I get good results.
However, when I test it on "fox" data from instant-ngp, the result is very bad.

In addition, the results are also bad on the nerfstudio dataset like, "poster"("https://drive.google.com/file/d/1dmjWGXlJnUxwosN6MVooCDQe970PkD-1/view?usp=sharing"),

I wonder why? Thank you!

[Giodiro]

Hi!
This should be a problem of adjusting some of the model hyperparameters. In particular I think the near/far planes will need to be tuned in order to get the scenes to converge properly.

[hasan1292]

@Giodiro I am facing the same problem on this dataset: [(https://drive.google.com/drive/folders/1v4MLNoSwxvSlWb26xvjxeoHpgjhi_s-s)] specifically on Face dataset.

This is my output

Can you tell me which parameters (near/far) needs to be changed like you said above? Where are they

[qhdqhd]

@Giodiro I am facing the same problem on this dataset: [(https://drive.google.com/drive/folders/1v4MLNoSwxvSlWb26xvjxeoHpgjhi_s-s)] specifically on Face dataset.

This is my output

Can you tell me which parameters (near/far) needs to be changed like you said above? Where are they

Basically setting near to 0 and far to 1000 can solve most problems.