Closed XLY43 closed 10 months ago
You should mainly modify the gaussian_model.py. Here is the reference, https://github.com/graphdeco-inria/gaussian-splatting/blob/main/scene/gaussian_model.py
Note that the current torch implementation is computationally-expensive for real applications. In case you need to train your model in real application, you may replace the gaussian renderer with the official gaussian_renderer (CUDA implementation).
Thanks for pointing the direction - Yeah it seems computationally expensive with my tests but for now, I'm not too concerned with real application.
Awesome repo! Are there any plans for implementing the adaptive control of the gaussians? If I want to implement it myself, what files should I look at in particular? Thanks.