Uno uses discrete diffusion for up to 3x lossless LLM inference speedup, talk Nov 13
burny_tech · x · 2026-11-12
IFM researcher Subham Sahoo will give an alphaXiv talk on Nov 13 introducing Uno, a discrete diffusion method that parallelizes token generation with lightweight diffusion adapters while the base model verifies them — no draft model needed, preserving the output distribution for up to 3x lossless speedup, claimed to beat top speculative decoding. New results to be presented include a version beating DFlash2, scaling to 400B, and faster RL post-training.
More from Infra
- Gratis launches: a local Mac AI assistant with no internet or API fees, built on Bittensor SN65 — markjeffrey · 2026-11-12
- ComfyUI update doubles Qwen LLM token rates on consumer GPUs, up to +251% on RTX 5090 — Lexius2129 · 2026-11-12
- HipKittens turns one: breaking AMD's 'CUDA moat' for a multi-silicon future — simran_s_arora · 2026-11-12
- 'Dear MacKenzie Scott, please send $8M for GPUs' becomes an AI meme — AIFlow_ML · 2026-11-12
- 768 tiny Pollen robots walk a simulated hospital ward on one GPU — MaziyarPanahi · 2026-11-12
- celld self-hosts Workers, Durable Objects and queues using just an object store bucket — irvinebroque · 2026-11-12