Ivan Puhachov
Ivan Puhachov
First of all, does your pytorch see cuda? Try: ```python import torch print (torch.cuda.is_available) ``` If it does see, then set up the CUDA_HOME path as its ask you to...
This is what was done in its successor - [CenterNet ("Objects as Points")](https://arxiv.org/pdf/1904.07850.pdf)
Had similar warning, fixed with: ``` pip3 install -U xformers==0.0.23 --index-url https://download.pytorch.org/whl/cu118 ``` See xformers repo (https://github.com/facebookresearch/xformers)
see https://github.com/nerfstudio-project/gsplat/issues/159
Adding `mp.jupyter()` solves the issue. 
I believe you can download the model manually - https://drive.usercontent.google.com/download?id=1rp8kdyLPvErw2dTmqtjISRVvQLj6Yzje&export=download (414 MB)
I fixed it on Ubuntu 22.04 as follows: 1. remove `ffmpeg` installed with conda: `conda remove --force ffmpeg` 2. `sudo apt install libx264-dev` 3. `sudo apt-get install ffmpeg` 4. verify...
maybe try changing engine to cycles? I found that eevee makes some rounding on depthmaps
check the params that are passed with `project_gaussians`. In recent versions redundant projection matrix was dropped (see #97). Just removing this line should work
Had the same problem. Solved it with this (assuming you installed a separate conda environment with requirements) Locate `op_kernel.h` in anaconda: ``` find /home/share/anaconda3/ | grep "op_kernel" ``` And find...