Skip to content

Docker & Deployment

Optimus-DL provides a comprehensive set of Docker integration scripts and Dockerfiles designed for high-performance distributed training. This guide explains how to build, run, and deploy your training environment using Docker.

🏗 Build Your Environment

The docker/ directory contains everything needed to build a containerized training environment.

  • docker/Dockerfile: The main Dockerfile, which can be customized.
  • docker/build.sh: A helper script for building Docker images.
  • docker/docker-bake.hcl: A configuration file for Docker Buildx Bake, used for building multi-platform or multi-target images.

Quick Build

To build the default training image:

bash docker/build.sh

This will create an image named optimus-dl:latest (by default) with all the necessary dependencies installed, including FlashAttention, NCCL, and GPU-optimized libraries.

🚀 Running Your Container

When running the container, ensure you provide access to your GPUs and mount your project directory.

docker run --gpus all -it \
  -v $(pwd):/workspace \
  -w /workspace \
  optimus-dl:latest /bin/bash

Inside the container, you can run training jobs just like in your local environment.