Migrating Your GitHub CI to Hugging Face Jobs

2026-06-09 · Source: Hugging Face - Blog · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

The Trackio project successfully migrated its GitHub CI workflows to Hugging Face Jobs, demonstrating a viable alternative to GitHub-hosted runners. This migration resulted in a significant 30% reduction in CI time for CPU jobs, decreasing from 1m40s to 1m10s, and crucially enabled the execution of a new GPU test suite on actual CUDA hardware, completing in 45s on a `t4-small` instance for less than a cent. The setup leverages a custom `huggingface/jobs-actions` bridge, a dispatcher Space (e.g., `cpu-upgrade`), and a GitHub App to manage ephemeral self-hosted runners within Hugging Face's serverless infrastructure. This approach allows for precise hardware flavor selection, including `t4-small` or `h200` GPUs, and custom Docker images like `mcr.microsoft.com/playwright:v1.60.0-jammy` or `nvidia/cuda:12.4.0-runtime-ubuntu22.04`, offering enhanced flexibility and performance for machine learning projects.

Key takeaway

For MLOps Engineers or Machine Learning Engineers evaluating CI infrastructure, migrating GitHub Actions to Hugging Face Jobs offers significant advantages. You can achieve up to 30% faster CPU test execution and enable crucial GPU-accelerated testing for ML models, all while maintaining GitHub as the CI orchestrator. This approach provides granular control over hardware and Docker images, potentially reducing costs and improving debugging with CLI-accessible logs. Consider this migration to enhance your ML project's testing capabilities and efficiency.

Key insights

Migrating GitHub CI to Hugging Face Jobs cuts CPU test times by 30% and enables cost-effective GPU testing via custom runners and hardware.

Principles

• CI benefits from specific hardware and clean environments.
• Ephemeral runners can bridge CI platforms effectively.
• Custom Docker images optimize CI performance.

Method

Duplicate a dispatcher Space, create and install a GitHub App for webhook handling, configure Space secrets, then update GitHub Actions `runs-on` labels to `hf-jobs-*` to trigger HF Jobs.

In practice

• Use `hf-jobs-cpu-upgrade` for CPU CI.
• Use `hf-jobs-t4-small` for GPU CI.
• Optimize with specific Docker images like Playwright or CUDA.

Topics

• GitHub Actions
• Hugging Face Jobs
• CI/CD
• GPU Acceleration
• MLOps Infrastructure
• Docker Images

Code references

• huggingface/blog
• gradio-app/trackio
• huggingface/jobs-actions
• user-attachments/assets
• organizations/YOUR-GITHUB-ORG

Best for: MLOps Engineer, Machine Learning Engineer, Software Engineer

Related on AIssential

• Build an AI app with FastAPI and Docker - Coding Tutorial with Tips — Integrate Hugging Face models with FastAPI and Docker for efficient, production-ready ML applications.
• GitHub Actions Custom Runner Images Reach General Availability — GitHub has announced the General Availability of custom images for its hosted runners, a feature that exited public preview in October and is designed to streamline CI/CD workflows by pre-baking tooli…
• Versioning and Testing Data Solutions: Applying CI and Unit Tests on Interview-style Queries — Robust data solutions require versioning, unit testing, and automated CI to ensure long-term reliability.
• GitHub Actions: Providing Data Scientists With New Superpowers. — GitHub Actions provides powerful automation for MLOps, enabling CI/CD for machine learning models and data science workflows.
• Who needs VCs when you have friends like these? — Community-driven funding and product development can validate demand and scale infrastructure effectively for AI cloud services.
• The Pulse: is GitHub still best for AI-native development? — GitHub's declining reliability and strategic drift challenge its status as a leading AI-native development platform.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Hugging Face - Blog.