Axolotl

Open-source LLM fine-tuning framework supporting multiple training techniques

Code & DevelopmentFree

Axolotl is an open-source LLM fine-tuning framework that supports a broad range of training techniques—full fine-tuning, LoRA, QLoRA, FSDP, and multi-GPU distributed training—through a unified YAML configuration. Researchers and ML engineers use Axolotl to fine-tune open-source models on custom datasets without writing training loops from scratch. Its broad compatibility with model architectures and training strategies makes it a popular choice in the open-source LLM fine-tuning community.

Key Features

• ✓Multiple training techniques
• ✓LoRA and QLoRA
• ✓Multi-GPU support
• ✓YAML configuration
• ✓Broad model support
• ✓Open source