DeepFold 2

Features

• Acceptable(?) replica of AlphaFold model.
• Distributed inference (over multiple GPUS).

Setup

git clone [email protected]:DeepFoldProtein/DeepFold.git   # Clone the repository
cd DeepFold                                             # Change directory
conda env create -f environment.yml                     # Construct Conda environment
conda activate deepfold2-dev
pip install .                                           # Build and install the package

Inference

AlphaFold parameter (JAX parameter) is needed to run AlphaFold model of DeepFold framework.

conda activate deepfold2-dev

# Example
INPUT_FEATURES_PKL="_output/T1104/features.pkl"
CONFIG="conf/model/alphafold/model_1.yaml"
OUTPUT_BASE_DIR="_output"
JAX_PARAMS_DIR="_data/params"

python3 \
    "scripts/predict_from_pkl.py" \
    -f "$(INPUT_FEATURES_PKL)" \    # Input features pickle file
    -c "$(CONFIG_PATH)" \           # Configuration YAML
    -o "$(OUTPUT_BASE_DIR)" \       # Output directory base path
    -p "$(JAX_PARAMS_DIR)" \        # JAX parameter directory
    -nt 2                           # Two GPUs

• If you want to enable deterministic mode (for validation) add --deterministic flag.
• You can fix feature processing random seed with -data_random_seed option.
• You can determine how many GPUs to use with -nt flags and NVIDIA_VISIBLE_DEVICES environmental variable.

About NCCL

• Multi-GPU inference mode use NCCL (Nvidia Collective Communication Library).
• If the framework stuck on communication, set NCCL_P2P_DISABLE=1.
• Turn off ACS(Access Control Services) on BIOS.
• Turn off IOMMU(Input/Output Memory Management Unit) on BIOS to use RDMA/GPUDirect (if your system supports).
• You can disable ACS temporarily by run scripts/disable_acs.sh with root permission.

Environmental variabes

• Set DEBUG=1 to show debug messages.

Override configurations

You can override configurations with dot-list with --options <DOT_LIST_1> <DOT_LIST_2> ... argument.

Training

TBA

Copyright

Copyright 2023 DeepFold Protein Research Team