NanoDiffuser: sub-400 MiB text-to-image model running in-browser via WebGPU
jason_mayes · x · 2026-09-06
NanoDiffuser is a text-to-image generator under 400 MiB that runs entirely locally in the browser via WebGPU, producing images in under 3 seconds on mid-tier laptops and under 6 seconds on mid-tier phones. It's a heavily modified Stable Diffusion, and the author built the conversion pipeline and compatibility patches enabling browser inference; full technical details are on the model page.
More from Infra
- Hot Chips 2026: Irrational Analysis publishes investment-driven recap — jwt0625 · 2026-09-06
- New method predicts transformer training divergence before the run starts — burkov · 2026-09-06
- Wall Street veteran: the world's biggest data center consumes water like just 3 golf courses — rohanpaul_ai · 2026-09-06
- Japan says $550B U.S. investment pact advancing, AI and semiconductors to play major role — Polymarket · 2026-09-06
- QuixiAI open-sources SlimServe, the inference stack behind the 8x3090 Qwen deployment — QuixiAI · 2026-09-06
- Serving Qwen3.8-Flash-Next at 262K context on 8 RTX 3090s hits 1,200 tok/s — QuixiAI · 2026-09-06