Axolotl

Config-driven fine-tuning: train without writing code

What it is

An open fine-tuning framework describing dataset, model and training parameters in one config file, removing the need to write training scripts. The community provides many reusable configurations.

Highlights

  • Full training flow described in a config file
  • No hand-written training scripts
  • Many reusable community configs
  • Supports multiple tuning methods and models

How to use

Prepare a dataset and config file, run one command, then start training and get weights.

License

Released under Apache-2.0. Read the terms before commercial use or redistribution, especially if you plan to offer it as a service.