The generative modeling landscape has grown rapidly, encompassing a wide array of architectures—VAEs, GANs, diffusion models, flow-based approaches, and transformers. This proliferation has made it challenging to maintain a coherent view of the field’s structure and evolution. We introduce GeMOVA (Generative Models Visual Atlas), an interactive, open-source tool designed to map the ecosystem of generative models through a rich, graph-based interface. GeMOVA goes beyond categorization by capturing fine-grained semantic relationships between models—such as variation, evolution, component usage, inspiration, and methodological similarity. Users can explore models chronologically, thematically, or by influence, with direct access to key papers, summaries and Implementations. GeMOVA serves both as an educational reference and a research companion, and is designed to evolve with community contributions.
git clone github/gemovaWe actively welcome contributions from the research community! GeMOVA's value comes from comprehensive, up-to-date coverage.
- Non-technical: Suggest a model via issue - we'll add it for you
- Technical: See CONTRIBUTING.md for detailed guide
- ✅ Peer-reviewed generative models (NeurIPS, ICML, ICLR, CVPR, etc.)
- ✅ Influential preprints (>500 citations)
- ✅ Production systems (Stable Diffusion, DALL-E, etc.)
- ✅ Corrections to existing entries
- ✅ New relationships between models
- Fork → Add to
assets/data/nodes.json+links.json→ Test → PR - Review typically within 2-5 days
- All contributors credited in repository
This project is licensed under the MIT License.
If you use GeMOVA in academic work, please cite our paper (preprint coming soon).
Live Demo GeMOVA Paper (coming soon)
GeMOVA was designed and developed by:
- Mohamed El Baha & Fouad Oubari, please feel free to reach out via LinkedIn or email.