DiffusionBlocks pub.sakana.ai/diffusionblo... from @sakanaai.bsky.social is one of these results that unlock a whole new branch of research papers.
Implications of reframing as diffusion goes beyond memory efficiency.
Test-time inpainting, guidance, and more would find new interpretations and uses.
pub.sakana.ai
DiffusionBlocks: Training Neural Networks One Block at a Time
A principled framework that converts a residual network into independently trainable blocks via a diffusion interpretation, achieving B× memory reduction without sacrificing performance.
