Looped-DiT: the first Looped Transformer for text-to-image generation
liuziwei7 · x · 2026-10-04
Inspired by rumors that Mythos and Astra may be powered by Looped Transformers, the team behind Looped-DiT introduces the first Looped Transformer architecture applied to text-to-image generation.
The author walks through the work in a 7-tweet thread with a code link, exploring how recurrence performs inside diffusion transformers (DiT).
Related event: Looped-DiT: Recycled Transformer Layers Beat Models 6.5x Larger(4 posts)→
More from Multimodal
- User Shares Coral Reef Creation Labeled 'Qwen 4 Max (xhigh)', an Unannounced Model Name — intellectronica · 2026-11-16
- Hand-editing 0.2% of weights plus 100 CPU-minutes turns an English TTS model into an 18MB Sanskrit one — kastnerkyle · 2026-11-16
- Zyphra open-sources textnorm tool, lifting TTS structured-data accuracy by up to 22 points — kastnerkyle · 2026-11-16
- AI music video tells a racing driver's comeback story — VictoryVisualsStudio · 2026-11-16
- MiniMax H3 official samples look stunning, but ComfyUI reproductions fall far short — cptdurr · 2026-11-16
- Testing Qwen Image 2.1 and Minimax H3 on anime-to-photo conversion and film photo restoration — Ant_6431 · 2026-11-16