Setting the scene for some future videos where I'll explore ways to improve diffusion models through various tricks. Here we learn why evaluating diffusion models is hard, that user preference is the gold standard, and that preference models like PickScore give us an approximation we can work with.
Code on github: github.com/johnowhitaker/dm_fun
PickScore on GitHub: github.com/yuvalkirstain/Pick... and paper: arxiv.org/pdf/2305.01569.pdf
Dall-e-3 announcement (openai.com/dall-e-3) and paper (cdn.openai.com/papers/dall-e-...)
EMU paper: scontent.fhio2-1.fna.fbcdn.ne...
Wurstchen: arxiv.org/pdf/2306.00637.pdf
SDXL paper: arxiv.org/pdf/2307.01952.pdf
22 окт 2023