2 ms·
I believe this could be achieved by providing a seed image to the diffusion model and generating memoji based on it. This way fine tuning isn't required.
by gokuldas011011 2y ago
I believe this could be achieved by providing a seed image to the diffusion model and generating memoji based on it. This way fine tuning isn't required.
- raverbashing 2y agoYup this is pretty much it, and DALLE and others can do this already