AutoTrain Advanced is a no-code machine learning model training and deployment solution developed by Hugging Face, with the goal of making training and deploying advanced machine learning models faster and easier. Users do not need to write code; they only need to upload data in the correct format to start training a model in a few clicks.
Main capabilities and uses: The project provides a no-code interface for training machine learning models. It emphasizes that data in the correct format must be uploaded to create a project, and help regarding data formats and pricing can be found in its documentation. The documentation address is https://hf.co/docs/autotrain/。
Usage:
- Run the UI on Colab or Hugging Face Spaces: You can deploy AutoTrain to Hugging Face Spaces, or run the AutoTrain UI on Colab via ngrok.
- Local installation: The AutoTrain-Advanced Python package can be installed via PIP and requires Python >= 3.10. git lfs needs to be installed, and torch, torchaudio, and torchvision also need to be installed. It is recommended to run autotrain in a conda environment; you can first create a new conda environment, then start the application.
- Command line and configuration: If you do not like using the UI, you can train via the command line using AutoTrain Configs, or use the AutoTrain CLI directly. When training with configuration files, the corresponding commands can be run. Example configuration files are located in the configs directory of the repository, including an example configuration for fine-tuning SmolLM2.
Pricing: AutoTrain is free. You only need to pay for the resources used if you decide to run AutoTrain on Hugging Face Spaces. When running locally, you only need to pay for the resources used on your own infrastructure.
Maintenance status: The project is no longer maintained; no new features will be added, and bugs will not be fixed. The official recommendation is to use Axolotl, TRL, or transformers.Trainer instead.
Limitations and notes: The materials do not provide a specific list of supported tasks, supported model types, interface screenshots, or detailed feature descriptions; information such as price, performance, and results is not yet provided.
