hugging-face-jobs

What it does

This skill enables users to execute workloads within the Hugging Face Jobs environment. It supports running User Volume (UV) scripts and containerized applications using Docker. The skill facilitates selecting appropriate hardware resources for tasks and provides cost estimations based on resource usage, while also handling authentication via tokens.

When to use it

• You need to run a custom data processing script that requires specific libraries not readily available in standard environments.
• You want to deploy a machine learning model training job using Docker containers.
• You require an environment with dedicated GPU resources for computationally intensive tasks.
• You’re looking to estimate the cost of running a particular workload on Hugging Face infrastructure before committing resources.

Key capabilities

• UV script execution
• Docker-based job support
• Hardware resource selection
• Cost estimation
• Token-based authentication

Example prompts

• "Run this UV script to preprocess the dataset: [script content]"
• "Deploy my Docker container for model training, requesting 2 GPUs."
• "Estimate the cost of running a job with 8GB RAM and a single GPU for 1 hour."

Tips & gotchas

• Ensure you have a valid Hugging Face token configured for authentication.
• Docker images must be built and pushed to a container registry accessible by Hugging Face Jobs.