Skip to content

Installation

The best way to install baktfold is using conda, so you can install Foldseek (the only non-Python dependency) along with the Python dependencies.

We would highly recommend installing conda via miniforge. For more information, please follow the installation instructions (https://github.com/conda-forge/miniforge). To install the latest version on your system (your operating system and architecture will automatically be detected):

curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
sh ./Miniforge3-$(uname)-$(uname -m).sh

We then detail installing Baktfold in four different scenarios

  1. Your machine does not have a GPU
  2. Your machine has a modern NVIDIA GPU installed
  3. Your machine is manufactured by Apple and has an Apple Silicon chip

  4. Your machine has a modern GPU installed from a different vendor (e.g. AMD)

  5. If you have no GPU, the default bioconda install (CPU-only) will suffice:

conda create -n baktfoldENV -c conda-forge -c bioconda baktfold
  1. If your machine has an NVIDIA GPU, you can create a conda environment with a CUDA-compatible PyTorch installation in one line automatically:
conda create -n baktfoldENV -c conda-forge -c bioconda baktfold pytorch=*=cuda*
  1. If your Apple machine has an Apple Silicon chip with integrated GPU cores (e.g. M-series), you can create a conda environment with a GPU compatible PyTorch version as follows by installing from the PyTorch channel:
conda create -n baktfoldENV 
conda activate baktfoldENV
conda install -c pytorch pytorch torchvision torchaudio
conda install -c conda-forge -c bioconda baktfold
  1. If you have a GPU from a different vendor (e.g. AMD) or have older hardware (e.g. CUDA drivers), we direct you to follow the instructions at https://pytorch.org and https://pytorch.org/get-started/previous-versions/ to install the appropriate version of PyTorch for your machine using pip . For example, to install baktfold on a machine with an AMD GPU and ROCM 7.2
conda create -n baktfoldENV pip
conda activate baktfoldENV
pip install torch torchvision --index-url https://download.pytorch.org/whl/rocm7.2
conda install -c conda-forge -c bioconda baktfold

Pip

You can also install baktfold using pip.

pip install baktfold

You will need to have Foldseek (ideally v10.941cd33) installed and available in the $PATH.

Source

You can install the latest version of baktfold with potentially untested and unreleased changes into a conda environment using conda as follows:

conda create -n baktfoldENV pip foldseek python=3.13
conda activate baktfoldENV
git clone https://github.com/gbouras13/baktfold.git
cd baktfold 
pip install -e .

Database Installation

To download and install the baktfold database:

baktfold install -t <threads>

If you would like to specify a particular location for the database (e.g. if you use baktfold on a shared server), please use -d

baktfold install -d <path/to/databse_dir> -t <threads>
  • Note: You will need at least 15GB of free space.

If you have an NVIDIA GPU available, you may wish to accelerate Foldseek using GPU. To do this, you will need to format the databases appropriately as follows

baktfold install -d <path/to/databse_dir> --foldseek-gpu -t <threads>

Usage: baktfold install [OPTIONS]

  Installs ProstT5 model and baktfold database

Options:
  -h, --help             Show this message and exit.
  -V, --version          Show the version and exit.
  -d, --database TEXT    Path to install Baktfold's database
  --foldseek-gpu         Enable Foldseek-GPU acceleration
  -t, --threads INTEGER  Number of threads  [default: 1]