The provided exported anoconda environment file is mainly tested on NVIDIA GeForce RTX 3090
and Tesla V100
with cuda 11.4. If it does not work on your machine, please consider installing the environment manually by following the instructions below.
- initialize a python 3.6 environment
conda create -n sem2nerf python=3.6.7
conda activate sem2nerf
- install pytorch 1.7.0. Choose a correct CUDA version from here. Below we provide the commands for cuda11
conda install pytorch==1.7.0 torchvision==0.8.0 torchaudio==0.7.0 cudatoolkit=11.0 -c pytorch
- install related python dependencies
pip install matplotlib==3.2.1 pyyaml==5.3.1 opencv-python==4.2.0.34 timm==0.5.4 tensorboard==2.2.1 tqdm==4.46.0 scikit-video==1.1.11
The scripts/inference3d.py
leverages ffmpeg
to generate video files. If you do not have ffmpeg installed on your machine, please install it via conda, e.g.,
conda install -c conda-forge ffmpeg