Axolotl Fine-Tuning

What You Get

YAML-driven LLM fine-tuning with Axolotl. Configure LoRA, QLoRA, DPO, and GRPO training runs without writing boilerplate code.

Features

• One-YAML configuration for full training runs
• LoRA/QLoRA with automatic rank selection
• DPO and GRPO reward model training
• Multi-GPU distributed training support
• Weights & Biases integration for experiment tracking
• Automatic dataset formatting (Alpaca, ShareGPT, chat)

Requirements

• Hermes Agent or OpenClaw v2.0+
• NVIDIA GPU with 16GB+ VRAM
• CUDA 11.8+

Installation

hermes skills install axolotl-fine-tuning