It depends on your needs. Vulkan is more portable but CUDA has more community support behind it, especially in deep learning.

I guess most researchers didn't want to go back to the challenges of training GANs, so they stuck with the usual diffusion models instead of exploring this area further. But I agree, it definitely has some potential.

@@necrashter agree. but from my understand stable diffusion was trained on a VAE-GAN and used diffusion to construct the latent. So i guess they did use gans in a way. - I honestly would not be surprised if diffusion models went out of fashion again. in favor of some sort of one shot model.

​@@Champignon1000Yes, in the latent diffusion paper they mention that their image compression model can be interpreted as a VQGAN. I think diffusion still provides a strong framework but this is a fast-moving field and we never know what the future holds.

Projenizde başarılar. Sascha Willems'ın Vulkan örnekleri reposunda dökümantasyonun eksik olduğu alanlarla ilgili örnekler içerebiliyordu. Mesh shader için de bir örnek var, belki faydalı olur: github.com/SaschaWillems/Vulkan/blob/master/examples/meshshader/meshshader.cpp

This is the music I made for a game. Gameplay footage plays in the background.

Well. IDK what is it. But good luck with this stuff man@@necrashter

Thanks @@mojo-wizard. You can play the game on itch.io if you are interested. The link is in the description.

I'm a teaching assistant in this course. You might be missing some context if you're not a student in this course, but it could still be helpful if you want to learn more about buffer overflow vulnerabilities. However, I wouldn't consider it a beginner-friendly video. This is a demonstration video, so you're expected to have a basic understanding.

xD