Looped Diffusion Transformer
Looped-DiT reuses shared Transformer blocks within each denoising step, letting a 260M-parameter text-to-image model beat one 6.5x larger at 4.9x lower compute.
1 stories, newest first · one card per story
Looped-DiT reuses shared Transformer blocks within each denoising step, letting a 260M-parameter text-to-image model beat one 6.5x larger at 4.9x lower compute.
Today 1.6K readers · 2.5K views · 779 reads · all time 189K views