Basic Introduction
Welcome to the OmniGenBench documentation!
What is OmniGenBench?
OmniGenBench is a modular, extensible benchmarking platform designed to evaluate genomic foundation models (GFMs) across a wide range of genomics tasks. Whether you are working on RNA structure prediction, gene function classification, or multi-species generalization, OmniGenBench provides a unified and reproducible framework for assessing model performance.
Who is OmniGenBench for?
- Researchers developing or evaluating genomic models
- Bioinformaticians seeking standardized benchmarks
- Data scientists interested in large-scale genomics tasks
Key Features
- Modular Benchmarks: Easily plug in new tasks or datasets.
- Automatic Dataset Management: Datasets are downloaded and managed automatically.
- Reproducible Results: Built-in support for random seeds and result tracking.
- Flexible Model Integration: Benchmark any model compatible with the Hugging Face ecosystem.
- Comprehensive Reporting: Generate detailed performance reports and visualizations.
Where should I start?
- Visit the Getting Started Guide for a hands-on example.
- Check out the API Reference for detailed usage.
- Explore our Tutorials for advanced workflows.