bowang-lab/scGPT
A foundation model applying GPT-style transformer architecture to single-cell multi-omics data for cell embedding and biological analysis.
Not currently ranked — collecting fresh signals.
star history
scGPT is a foundation model for single-cell multi-omics built using generative AI techniques. It applies transformer-based architectures similar to GPT to biological single-cell data, enabling tasks such as cell embedding, zero-shot integration, and continual pretraining. The model comes with pretrained checkpoints and integrates with HuggingFace for model loading across PyTorch CPU, GPU, and flash-attention backends.
Frequently asked
- What is bowang-lab/scGPT?
- A foundation model applying GPT-style transformer architecture to single-cell multi-omics data for cell embedding and biological analysis.
- Is scGPT open source?
- Yes — bowang-lab/scGPT is open source, released under the MIT license.
- What language is scGPT written in?
- bowang-lab/scGPT is primarily written in Jupyter Notebook.
- How popular is scGPT?
- bowang-lab/scGPT has 1.6k stars on GitHub.
- Where can I find scGPT?
- bowang-lab/scGPT is on GitHub at https://github.com/bowang-lab/scGPT.