- Home
- Agent skills
- huggingface-llm-trainer
huggingface-llm-trainer
Fine-tune language and vision models with TRL or Unsloth on Hugging Face Jobs
What it is
A Hugging Face skill that guides an agent through training or fine-tuning models on managed cloud GPUs with TRL or Unsloth. It covers SFT, DPO, GRPO and reward modelling, saving results to the Hub, and GGUF conversion for local use.
Good for
- Run an SFT fine-tune without a local GPU
- Train a DPO model from preference data
- Convert a trained model to GGUF
Install or run
/plugin marketplace add huggingface/skillsCommand as given in the project README. Read the repository before you run anything on your machine.
What you need
Hugging Face account with Jobs access and token
Sources
- https://github.com/huggingface/skills/tree/main/skills/huggingface-llm-trainer (project README, read 13 Nov 2026)
- https://api.github.com/repos/huggingface/skills (GitHub API, read 13 Nov 2026)
The summary is ours, written from the README. Spotted an error? See how we work.