docker_installation.md

Run with Docker

1. Install Docker (with GPU Support)

   Ensure that Docker is installed and configured with GPU support. Follow these steps:

   • Install Docker if not already installed.
   • Install the NVIDIA Container Toolkit to enable GPU support.
   • Verify the setup with (using a version close to our environment):

     docker run --rm --gpus all nvidia/cuda:12.6.3-base-ubuntu22.04 nvidia-smi

2. Pull the Docker image The image contains all necessary dependencies (PyTorch, HMMER, Kalign, CUTLASS, etc.), but does not include the Protenix source code by default.

   docker pull ai4s-share-public-cn-beijing.cr.volces.com/release/protenix:1.0.0.4

3. Clone this repository

   git clone https://github.com/bytedance/protenix.git 
   cd ./protenix

4. Run Docker with an interactive shell Mount the current directory to /app inside the container. If you have external data or weights (e.g., in /root/protenix), consider mounting them as well.

   docker run --gpus all -it \
       -v "$(pwd)":/app \
       -v /dev/shm:/dev/shm \
       ai4s-share-public-cn-beijing.cr.volces.com/release/protenix:1.0.0.4 \
       /bin/bash

5. Install Protenix and Verify Once inside the container, install Protenix in editable mode and verify the installation:

   cd /app
   pip install -e .
   
   # Verify the installation by checking the help message
   protenix --help

After completing these steps, you can proceed with inference or training. See Inference Guide for more details.