This codebase is tested with torch==1.11.0
and torchvision==0.12.0
, with CUDA 11.3
and gcc 7.3.0
. In order to successfully reproduce the results reported in our paper, we recommend you to follow the exact same versions. However, similar versions that came out lately should be good as well.
conda create -n robodepth python=3.10
conda activate robodepth
# CUDA 10.2
conda install pytorch==1.11.0 torchvision==0.12.0 cudatoolkit=10.2 -c pytorch
# CUDA 11.3
conda install pytorch==1.11.0 torchvision==0.12.0 cudatoolkit=11.3 -c pytorch
# CPU Only
conda install pytorch==1.11.0 torchvision==0.12.0 cpuonly -c pytorch
Please install the following packages into the environment:
pip3 install opencv-python==4.6.0.66 timm==0.6.7 tensorboardX scikit-image==0.19.2 matplotlib yacs
pip install dotmap wandb einops tqdm
To create common corruptions, install the following packages into the environment:
pip install imagecorruptions
For the Monocular Depth Estimation Toolbox, install the following into the environment:
pip install mmcv-full==1.7.0 -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.11/index.html
cd zoo/Monocular-Depth-Estimation-Toolbox/
pip3 install -e .