I work at an organization and we’ve been using Midjourney for a while now to generate portraits and people. However, the organization being in an Arab country, the results we are getting from Midjourney are most of the time not accurate and do not portray what we have in mind. We have a massive library (more then 500k pics) with faces of local people from all ages and genders. I would like to use a diffusion model and train it in order to be able to generate people based on the library we have.
1. Is this doable?
2. How much time will it take to launch the first prototype?
3. Hardware requirements
Job is closed
This job is already closed and no longer accepting applicants, sorry.
