Back to Resources
🦎

Axolotl

Training & Fine-tuning

Streamlined fine-tuning of LLMs with PEFT and full-param support.

8kstars900forksPython
View on GitHub

About

Axolotl provides a config-driven interface for fine-tuning LLMs using LoRA, QLoRA, and full fine-tuning, with support for multiple dataset formats and FSDP/DeepSpeed.

Key Features

  • LoRA
  • QLoRA
  • FSDP
  • DeepSpeed
  • Multi-dataset

Tags

Fine-tuningLoRALLMPEFT

Related Resources

🎛️

PEFT

Parameter-efficient fine-tuning methods by Hugging Face.

🎖️

TRL

Transformer Reinforcement Learning for RLHF and alignment.

🚀

DeepSpeed

Deep learning optimization library for large-scale training.