skills.rest
Supporter09
Supporter09COMMUNITY
Explore all skills by this author →
GitHub

llama-factory

Configure and troubleshoot LLaMA-Factory LLM fine-tuning workflows end to end.

Updated May 4, 2026
Search Tags:#multimodal#quantization#distributed training#lora#vllm#qlora#llm fine-tuning
One-click install
npx skills add https://github.com/Supporter09/Face_Anti_Spoofing_Biometric --skill llama-factory-supporter09
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llama-factory
Source: https://github.com/Supporter09/Face_Anti_Spoofing_Biometric/tree/main/.claude/skills/llama-factory
Command: npx skills add https://github.com/Supporter09/Face_Anti_Spoofing_Biometric --skill llama-factory-supporter09

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

LLaMA-Factory removes the complexity of building and operating modern LLM fine-tuning pipelines by centralizing training, inference, evaluation, export, and deployment guidance in one skill.

Core Features & Use Cases

  • Training workflows: Set up pretraining, supervised fine-tuning, reward modeling, PPO, DPO, KTO, and ORPO with LoRA, QLoRA, freeze, or full-parameter methods.
  • Production optimization: Apply quantization, acceleration, distributed training, monitoring, and adapter merging for efficient local or scalable deployment.
  • Use case: Use this skill when you need to fine-tune a base model on custom text or multimodal data, validate the result, and export a deployable checkpoint.

Quick Start

Use the llama-factory skill to design the best fine-tuning, evaluation, and deployment setup for my model, dataset, and hardware.

Frequently Asked Questions about llama-factory

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

FAQPage Schema
How do I fine-tune an LLM with LoRA or QLoRA on custom datasets?▼

LLM fine-tuning with LoRA or QLoRA is configured by preparing your custom dataset, selecting the adapter training method, and setting the quantization parameters to train efficiently on limited hardware.

What's the best way to set up distributed training for multimodal LLMs?▼

Distributed training for multimodal LLMs is set up by configuring the environment across local, cloud, or NPU platforms, enabling you to handle combined text and image data with scalable hardware allocation.

Can I use vllm for inference and evaluation after training an adapter?▼

Yes, vllm is supported for inference. After training your adapter, you can merge it and run evaluation to validate performance before exporting a deployable checkpoint.

Does this workflow support full parameter fine-tuning and DPO?▼

Full parameter fine-tuning and DPO are fully supported. You can configure pretraining, supervised fine-tuning, reward modeling, PPO, DPO, KTO, and ORPO with freeze or full-parameter methods.

How does quantization work when deploying large models locally?▼

Quantization reduces the memory footprint of large models for local deployment. By applying quantization and acceleration techniques, you optimize the model for efficient inference on local hardware.

What are the limitations of using QLoRA compared to full fine-tuning?▼

QLoRA trains adapters with quantized base weights to save memory, but full fine-tuning updates all parameters for maximum performance. QLoRA is constrained by adapter capacity and quantization precision loss.

Related Skills

Explore More Software Engineering Skills→
iani-kuliiani-kuli

unsloth

Optimize LLM fine-tuning with LoRA and QLoRA to minimize VRAM usage.

Community
Advanced
ScientiaCapitalScientiaCapital
1

model-deployment

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

Community
Advanced
Ced3-hanCed3-han

openrlhf-training

Run distributed RLHF training for large language models with Ray and vLLM.

Community
Advanced
gagan114662gagan114662

axolotl

Generates YAML configurations for fine-tuning LLMs with Axolotl frameworks and advanced techniques such as LoRA and QLoRA.

Community
Advanced