hugging-face-model-trainer

Submit TRL training jobs to Hugging Face Jobs with inline scripts.

6|1|Updated Jan 17, 2026
One-click install
npx skills add https://github.com/issdandavis/SCBE-AETHERMOORE --skill hugging-face-model-trainer-issdandavis
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hugging-face-model-trainer
Source: https://github.com/issdandavis/SCBE-AETHERMOORE/tree/main/external/codex-skills-live/hugging-face-model-trainer
Command: npx skills add https://github.com/issdandavis/SCBE-AETHERMOORE --skill hugging-face-model-trainer-issdandavis

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires trl, peft, transformers, accelerate, trackio, datasets, sentencepiece, protobuf, numpy, gguf, huggingface_hub, tensorboard, unsloth, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This skill enables cloud-based Transformer Reinforcement Learning (TRL) training on Hugging Face Jobs, removing the need for local GPUs and simplifying production-grade experimentation.

Core Features & Use Cases

  • Supports TRL training methods including SFT, DPO, GRPO, and Reward Modeling, plus GGUF conversion for local deployment.
  • Facilitates Hub pushes, Trackio monitoring, dataset validation, and cost/time estimation within production pipelines.
  • Provides templates and guidance for inline UV/TRL scripts and TRL-maintained examples to accelerate workflows.

Quick Start

Submit a training job via hf_jobs with an inline Python script or a reference to a production script to begin end-to-end TRL training and deployment.

Frequently Asked Questions about hugging-face-model-trainer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run TRL fine-tuning on Hugging Face Jobs without a local GPU?

You can run TRL fine-tuning on Hugging Face Jobs by submitting an inline Python script or referencing a production script via hf_jobs, enabling cloud-based training without local GPU hardware. It relies on TRL, PEFT, and Hugging Face tooling for execution.

What training methods does TRL support for LLM fine-tuning on Hugging Face?

TRL supports SFT, DPO, GRPO, and Reward Modeling for LLM fine-tuning. This skill facilitates these methods on Hugging Face Jobs, providing templates and guidance for inline scripts to accelerate production-scale training workflows.

Can I convert my fine-tuned Hugging Face model to GGUF for local deployment?

Yes, you can convert fine-tuned models to GGUF format for local deployment. The skill enables end-to-end workflows including SFT, DPO, GRPO, and GGUF conversion, allowing you to push models to the Hub after training.

Do I need to set up an HF_TOKEN secret to use Hugging Face Jobs for training?

Yes, you need an HF_TOKEN configured in secrets to use Hugging Face Jobs for training. This token enables hub pushes, Trackio monitoring, dataset validation, and end-to-end execution through inline script submission.

How does Trackio monitor TRL training experiments on Hugging Face Jobs?

Trackio monitors TRL training experiments by integrating with Hugging Face Jobs pipelines. The skill facilitates Trackio monitoring alongside hub pushes, dataset validation, and cost estimation for production-scale training experiments.

Why use PEFT and Hugging Face Jobs for production-scale LLM training instead of local GPUs?

Using PEFT and Hugging Face Jobs removes the need for local GPUs and simplifies production-grade experimentation. It supports cloud-based TRL training with dataset validation, cost estimation, and Trackio monitoring for scalable workflows.