- FantasyTalking: Realistic Talking Portrait Generation via Coherent Motion Synthesis
- FantasyID: Face Knowledge Enhanced ID-Preserving Video Generation
- Apr 28, 2025: 👋 We release the inference code and model weights of FantasyTalking.
- Apr 24, 2025: 👋 We release the inference code and model weights of FantasyID.
- Apr 14, 2025: 👨💻 We’re in the final stages of preparing FantasyID and FantasyTalking for the release.
- Giving Back to the Community: In our daily work, we benefit immensely from the resources, expertise, and support of the open source community, and we aim to give back by making our own projects open source.
- Attracting More Contributors: By open sourcing our code, we invite developers worldwide to collaborate—making our models smarter, our engineering more robust, and extending benefits to even more users.
- Building an Open Ecosystem: We believe that open source brings together diverse expertise to create a collaborative innovation platform—driving technological progress, industry growth, and broader societal impact.
