hackater

Can you go into a bit more detail? What architecture did you use? Is the month training time really just training with mini batches with a constant learning rate? Or are these many failed attempts until you trained a successful model for a few days in the end?

I particularly interested in the image generation part (the DDPM/SGM)

How do you learn about this? Does anyone have any good recommendation on books, where you learn how to do these kind of very low level optimizations?