Open SainingZhang opened 1 week ago
Thank you very much! Wonderful work! I wonder how many iterations you trained for outdoor scenes for both stage 1 and 2?
Thank you for your interest. We trained the model for 300,000 iterations for each stage.
Thank you. Do you use tinycudann 1.7?
Thank you. Do you use tinycudann 1.7?
Yes, pip install git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch
will install tinycudann==1.7.
Hi! We don't modify package versions when working with torch 1.12.1 and cuda 11.3. We have tested the environment installation commands provided in the README.md on an RTX3090 and RTX4090.
If you intend to use torch 1.13.1, the crucial step is to verify that the cuda version aligns with the torch version. Also, make sure that tinycudann works correctly in your environment.