Axolotl
Visit ToolAxolotl is an open-source framework for fine-tuning large language models (LLMs). It supports various models and training methods, including LoRA, QLoRA, and DPO.
At a glance
Trending
Axolotl is an open-source framework for fine-tuning large language models (LLMs). It supports various models and training methods, including LoRA, QLoRA, and DPO.
Trending
About
Axolotl is a free and open-source tool designed to streamline post-training and fine-tuning for the latest large language models (LLMs). It offers extensive support for multiple models like GPT-OSS, LLaMA, Mistral, and Mixtral, as well as multimodal training for vision-language models. Users can leverage various training methods including full fine-tuning, LoRA, QLoRA, GPTQ, and preference tuning (DPO, IPO, KTO, ORPO). Axolotl provides easy configuration through YAML files, performance optimizations like Flash Attention and multi-GPU training, and flexible dataset handling from local, HuggingFace, and cloud sources. It is cloud-ready with Docker images and PyPI packages.
Capabilities
Pricing & Plans
Open Source
Free
FAQs
Trending