scFTAT: a Single-Cell Annotation Method Integrating FFT and Improved Transformer

• ScFTAT is a cell-type recognition approach based on FFT and an improved Transformer. It aims to provide researchers with a deep learning tool for single-cell classifying tasks.
• Current version: 1.0

• Setting up scFTAT requires an environment with Python 3.7 or newer, Tensorflow, and other necessary machine-learning libraries.
• Install all dependencies using pip: pip install -r requirements.txt.
• scFTAT does not require database configuration.
• After setting up the above files, execute the code in the train-test.py file in order. Note that the pkl file needs to be converted to CSV format before running.
• Since scFTAT is primarily used for research and development, it does not have a specific deployment guide. Please integrate it into your project or workflow as needed.

• Please write appropriate tests for your contributions and ensure all tests pass.
• Submit pull requests for any changes. Project maintainers will review according to the project's coding standards.

• Have any questions or issues related to the repository, please contact Dr. Binhua Tang (bh.tang@hhu.edu.cn).