Amber

“instancenorm” in encode stage

1

The instancenorm para about "track_running_states=True" and "affine=True". In the training stage, should I set "track_running_states=True" and "affine=True"?

Install submodules\diff-gaussian-rasterization failed

12

Problems occur when installing submodules. Windows11, cuda version:11.8, python 3.7, torch: 1.13.1 Processing d:\work\gaussian-splatting\submodules\diff-gaussian-rasterization Preparing metadata (setup.py) ... done Building wheels for collected packages: diff-gaussian-rasterization Building wheel for diff-gaussian-rasterization (setup.py)...

generate pointcloud

1

Hi, Tanks a lot for your job. How to generate a RGB pointcloud when finish training?

merge two gaussian splatting scene

2

Hi, I have question about how to merge two gaussian splatting scene? when I load two scene into super-splat, they are sperated to differnt coordinate(cannot merge), and cannot be export...

I already trained a model, when finetune the same model, always "unexpected key" & "missing key" notification.   and config like this: