helixml / helix

Introduction

HelixML is a private GenAI platform. It allows users to deploy and manage large language models (LLMs), image models, and other open-source AI models in their own data center or VPC. This approach ensures complete data security and control, as opposed to relying on third-party cloud services.

HelixML is built on top of open source technologies and offers features like fine-tuning models, which can be done easily through drag-and-drop interfaces. The platform prioritizes user experience, providing a scalable and ergonomic platform while optimizing the tradeoff between GPU memory and latency.

Code Examples

Deployment:

# Clone the HelixML repository
          git clone https://github.com/helixml/helix.git
          cd helix
          
          # Create an environment file based on the example
          cp .env.example-prod .env
          
          # Ensure Keycloak realm settings are up to date with your .env file
          ./update-realm-settings.sh
          
          # Start the services
          docker-compose up -d

The dashboard will be available on http://localhost.

GPU Runner Attachment:

Documentation on attaching GPU runners can be found in the HelixML documentation.

HelixML Modules:

  • github.com/helixml/helix: The core Golang module for HelixML.

Building with Go:

go mod edit -module=github.com/helixml/helix
          go build -o helix

This example demonstrates how to build the HelixML binary using the correct module path.

Project Structure: