bytedance / bytedance/InfiniteYou
Support for Z-Image?
- Dominant language
- Python
- Stars
- 2.7k
- Forks
- 285
- PR merge metrics
- No merged PRs in 30d
Description
Thanks for open sourcing this incredible project! 🙏
Since there is [no training code being released](https://github.com/bytedance/InfiniteYou/issues/34#issuecomment-2819272769) is the InfiniteYou team considering training a version of the InfuseNet for Z-Image?
The InfiniteYou results are amazing and the Z-Image model has superior inference speed, prompt adherence, and general image quality compared to Flux.1-dev.
The architecture is different but they are both transformer models. It seems like the combo would amazing!
Contributor guide
No contributing guide indexed for this repository
Research direction
The issue does not identify any files, tests, or entry points for adding Z-Image support. Start by locating the existing InfuseNet training and inference paths and comparing their assumptions with the Z-Image architecture. Done would require a defined compatibility plan, training support, and validation that the resulting model preserves the project's identity-preserving results.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 20/100