Top pick  ·  Efficient AI & Edge

D3D: Decoupled Diffusion Drafting for Asynchronous Speculative Decoding

D3D removes target-model hidden state dependence in diffusion drafting, enabling asynchronous speculative decoding with up to 7.48× speedup.

Why it matters: Builders can generate text much faster without sacrificing quality, saving compute and latency.

Read the breakdown →  Code

Trending  ·  LLMs & Language Models

SLA: Scaling Spatial-Language-Action Model via Continual Pre-training

SLA continually pre‑trains a vision‑language‑action model to align visual embeddings with 3D spatial representations, improving transferable spatial understanding for robots.

Why it matters: Better spatial grounding lets robots generalize manipulation across tasks without redesigning perception.

Read the breakdown →