Besides the obvious pedagogical challenges, I also faced technical challenges as I wanted to:
- only use models with open weights
- include all students and teachers into the generated images and videos within 1 hour
I evaluated several different models for that, but settled for Z-Image and WanAnimate in the end.
May I ask Why? What’s the rationale? I have a hard time understanding how this helps foundational education at such a young age…