First screen, captured .
Looks similar
Training Text-to-Image Models Without a VAE(linum.ai)
49 points by schopra909 | 14 comments
Linum research blog post introducing Pyramid-JiT, a multiresolution prediction method to speed convergence of VAE-free pixel-space text-to-image models.
Training a 4B model to produce 81% faster query plans than Postgres(rohanbansal.com)
647 points by polyphilz | 130 comments
Blog post by Rohan Bansal on training a 4B Qwen model with agentic reinforcement learning to generate query plans faster than Postgres's optimizer.
The Training Example Lie Bracket(pbement.com)
25 points by pb1729 | 12 comments
Diffusion models explained simply(seangoedecke.com)
123 points by onnnon | 25 comments