Ruilong Li(李瑞龙)

> In regards to the `Memory Allocation` plot, does the `nerfacto` model now use more GPU memory (the colors are greyed of I can't tell for sure)? If so, why?...

Hi these code does not have in-place operations. You would want to look for something like `+=`, `*=` in the entire pipeline. Inplace operation is in general forbidden in any...

Hi it is hard to tell what could be wrong from this among of information. One possible trigger of this might be zero shape tensor being sent to tiny-cuda-nn.

Not a problem at all. We recently update the code from `render_image_with_occgrid` to `render_image_with_occgrid_test` for faster rendering (see https://github.com/KAIR-BAIR/nerfacc/pull/217) but it is totally compatible to just use `render_image_with_occgrid`. High levelly,...

That's a fantastic question. You are right it is wasteful to query sigma twice. But it is also not trivial to avoid that. The reason is in sampling we use...

I see this error `fatal error: ATen/cuda/CUDAGeneratorImpl.h: No such file or directory`. For me this file lives in: /home/ruilongli/anaconda3/envs/nerfacc/lib/python3.9/site-packages/torch/include/ATen/cuda/CUDAGeneratorImpl.h You might want to check if your torch is installed in...

Yes

Could you dump the input of [this function here ](https://github.com/nerfstudio-project/nerfacc/blob/master/nerfacc/estimators/occ_grid.py#L164-L173)and share it so I can take a look?

I can't use these pasted outputs to examine the code.. Could you save them into a file (say .pth or .npz) and upload it here?

Hi I will check this issue after ECCV's ddl tmr!