AutoTrain: No-code training for state-of-the-art models
This tool addresses the problem of complex model training for developers and researchers in industrial or open-source applications, though it is incremental as it builds on existing open-source models and best practices.
The authors tackled the lack of a unified tool for training models across various modalities and tasks by introducing AutoTrain, an open-source, no-code library that simplifies training for tasks like LLM finetuning, text classification, and image classification, supporting tens of thousands of models from Hugging Face Hub.
With the advancements in open-source models, training (or finetuning) models on custom datasets has become a crucial part of developing solutions which are tailored to specific industrial or open-source applications. Yet, there is no single tool which simplifies the process of training across different types of modalities or tasks. We introduce AutoTrain (aka AutoTrain Advanced) -- an open-source, no code tool/library which can be used to train (or finetune) models for different kinds of tasks such as: large language model (LLM) finetuning, text classification/regression, token classification, sequence-to-sequence task, finetuning of sentence transformers, visual language model (VLM) finetuning, image classification/regression and even classification and regression tasks on tabular data. AutoTrain Advanced is an open-source library providing best practices for training models on custom datasets. The library is available at https://github.com/huggingface/autotrain-advanced. AutoTrain can be used in fully local mode or on cloud machines and works with tens of thousands of models shared on Hugging Face Hub and their variations.