News for hackers. Stuff that matters.
Re:Welcome to Waifu Labs v2: How Do AIs Create? by spywaregorilla on Tuesday January 11, 2022 @10:59PM (#29902114)
Is there any way to approximate embeddings for a novel image?

novel meaning user provided, not generated by the model or in the training set.

Re:Welcome to Waifu Labs v2: How Do AIs Create? by liuru on Tuesday January 11, 2022 @11:24PM (#29902284)
Not at the moment in our tool, though this is an area of great curiosity and research for us!
Re:Welcome to Waifu Labs v2: How Do AIs Create? by spywaregorilla on Wednesday January 12, 2022 @08:59AM (#29906353)
Does the discriminator model translate the images into an embedding space of it's own? Could such a space be used to generate images themselves?