Zongqi HE comments

Results 10 comments of


                                            Zongqi HE

help

same problem, have you solved it?

Question about reproduce SparseNeRF using MipNeRF360 Dataset

same issue

Issue with installing diff-gaussian-raterization-modified. Everything else works fine and get installed. But this package is creating problem. I have Torch with cu118 installed

Hi, I have run it in cpp10 and cu121, it works on my machine

Dataset GoogleDrive expired

> Hi! It seems like the google drive link for the dataset is expired or removed. May I know how I can get the training data now? Here is nerf_llff_data...

Dataset GoogleDrive expired

> > > Hi! It seems like the google drive link for the dataset is expired or removed. May I know how I can get the training data now? >...

How to apply the calculated pose and point cloud for downstream tasks?

> > [@LaFeuilleMorte](https://github.com/LaFeuilleMorte) Did you figure something out till today? Sounds super interesting to me and would love to know if you found a way to use mast3r for sfm...

The reproduced evaluated performance on tank dataset is much higher than the original paper reported.

I also achieved much better results on MipNeRF360 Dataset, with `--eval` flag, and without adding `--antialising`, `--d` and `exposure-compensation`

Slow speed than original 3DGS

It works, meanwhile PSNR decreased about 0.2？

关于新代码为何效果比原论文好疑问？

但是使用 `-i` 指令而非 `-r` 也会遇到新指标变高问题

How do you convert weight_l2 in your code, and what is distlosses? comparing with mipnerf

Ldist from MipNeRF-360, how above model.variables weight ?