AutoTrain

AutoTrain Advanced is Hugging Face's no-code tool for training machine learning models through a web UI or command-line interface. It handles LLM fine-tuning with SFT, DPO, ORPO, reward modeling, and generic trainers, plus text classification and regression, token classification, sequence-to-sequence tasks, extractive question answering, and image classification and scoring. Jobs run on Hugging Face Spaces, in Google Colab, or locally after installing the Python package, which requires Python 3.10 or newer; the software itself is free, with charges applying only to compute. The project is released under the Apache 2.0 license. An important caveat: the maintainers have declared the project no longer maintained, with no new features or bug fixes planned, and they recommend alternatives such as Axolotl, TRL, or the transformers Trainer for new work. It remains relevant to practitioners who want push-button fine-tuning on familiar Hugging Face infrastructure and to those maintaining existing AutoTrain pipelines.