- Linux
- Python 3.9+
- PyTorch 1.10+
- CUDA 11.0+
- NCCL 2+
- GCC 5+ Requirements are not in principle hard requirements, since we are using basic pytorch/torchvision operations, but there might be some differences (not tested).
a. Create a conda virtual environment and activate it.
conda create -n idisc python=3.9 -y
conda activate idisc
b. Install PyTorch and torchvision,
conda install pytorch torchvision -c pytorch
Note: Make sure that your compilation CUDA version and runtime CUDA version match. You can check the supported CUDA version for precompiled packages on the PyTorch website.
c. Install IDisc addiitonal requirements and export to PYTHONPATH
(add to .bashrc in order not to export in every new shell)
pip install -r requirements.txt
export PYTHONPATH="$PWD:$PYTHONPATH"
d. Install Deformable Attention
export IDISC-REPO=${PWD}
cd ${IDISC-REPO}/idisc/models/ops/
bash ./make.sh