Skip to content

Latest commit

 

History

History
43 lines (39 loc) · 1.3 KB

INSTALL.md

File metadata and controls

43 lines (39 loc) · 1.3 KB
Raw

Installation

Requirements

  • Ubuntu18.04 with python 3.8
  • CUDA 11.3
  • Pytorch 1.12.1
  • Scripts for data downloading require rsync and aria2

Install dependencies and build solvent

  • prepare conda environment with installing all python dependencies (it takes a few minutes)

    conda env create --name=solvent -f environment.yml
conda activate solvent

  • install xformers and triton

    pip install ninja
pip install -v -U git+https://github.com/facebookresearch/xformers.git@b31f4a1#egg=xformers
pip install -U --index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/Triton-Nightly/pypi/simple/ triton-nightly

  • installation via apt-get

    sudo apt-get update
sudo apt-get install aria2 rsync

  • Download folding resources. Downloaded at solvent/resources

    bash datasets/download_chemical_props.sh

  • build solvent with compiling Openfold's CUDA kernel

    python setup.py install

Install evaluation program

  • Download and build TM-score program and locate at TMscore/ 
    # expected program structure
TMscore/
    TMscore
    TMscore.cpp