skills.rest
RobbieJ
RobbieJCOMMUNITY
Explore all skills by this author →
𝕏@robbiejGitHub

training-hub

Automates end-to-end LLM fine-tuning workflows with the traininghub library.

2|Updated Dec 9, 2025
Search Tags:#validation#lora#jsonl#vllm#sft#training-hub#osft
One-click install
npx skills add https://github.com/RobbieJ/training-hub-skills --skill training-hub
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: training-hub
Source: https://github.com/RobbieJ/training-hub-skills/tree/main
Command: npx skills add https://github.com/RobbieJ/training-hub-skills --skill training-hub

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires datasets, and includes scripts (resource) components.

What problem does it solve?

Fine-tune LLMs efficiently by using Red Hat's training-hub library, consolidating SFT, LoRA/QLoRA, and OSFT workflows into a single framework.

Core Features & Use Cases

  • End-to-end training: dataset prep, hardware configuration, scalable training across single to multi-node clusters, evaluation with lm-eval, and inference serving with vLLM.
  • Data validation and memory estimation: validate JSONL data, estimate GPU memory, and recommend hardware configurations.
  • Flexible backends and formats: supports SFT, LoRA, OSFT, and multiple model families with pre-tested configurations.

Quick Start

Run a full fine-tuning workflow on your dataset using training-hub.

Frequently Asked Questions about training-hub

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

FAQPage Schema
How do I streamline end-to-end LLM fine-tuning workflows for SFT and LoRA?▼

You can streamline LLM fine-tuning through a framework that automates dataset preparation, multi-node scaling, memory estimation, and vLLM deployment for SFT, LoRA, and OSFT workflows.

What is the best way to validate JSONL datasets for LLM training?▼

Validate JSONL datasets for LLM training by enforcing strict data format validation checks before execution, ensuring your inputs meet the structural requirements for SFT and OSFT processing.

Can I scale fine-tuning across multi-node clusters and deploy with vLLM?▼

Yes, you can scale fine-tuning across multi-node clusters and deploy the resulting models using vLLM, with automated hardware configuration and memory estimation recommending the optimal setup.

Does this workflow support model evaluation with lm-eval after training?▼

Yes, the workflow supports model evaluation with lm-eval, integrating the evaluation step directly into the end-to-end fine-tuning pipeline to verify model performance before deployment.

How do I estimate GPU memory requirements before starting a LoRA training run?▼

Estimate GPU memory requirements by using built-in memory estimation APIs that analyze your selected model family and training method to recommend appropriate hardware configurations.

When should I choose OSFT over standard SFT for model fine-tuning?▼

Choose OSFT over standard SFT when your use case requires the specific orthogonal fine-tuning approach, utilizing the pre-tested configurations provided for multiple model families within the framework.

Related Skills

Explore More Software Engineering Skills→
yuxuan-louyuxuan-lou
1

claw-finetune

Gather fine-tuning requirements, validate inputs, and prepare LoRA/SFT runs for Tinker or HPC-AI.

Community
Intermediate
ScientiaCapitalScientiaCapital
1

model-deployment

Export fine-tuned models to GGUF and deploy with vLLM or Ollama.

Community
Advanced
Z43LZ43L

distributed-llm-pretraining-torchtitan

Configure torchtitan TOML files for distributed LLM pretraining with 4D parallelism.

Community
Advanced
librefanglibrefang
10

llm-finetuning

Fine-tune large language models with LoRA and QLoRA for domain adaptation.

Official
Advanced