Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous Driving 事件

Fast-dDrive: Efficient Block-Diffusion VLM for Autonomous Driving arXiv:2605.23163v2 Announce Type: replace Abstract: End-to-end autonomous driving via Vision-Language-Action (VLA) models demands a precarious balance between high-fidelity trajectory planning and efficient inference. Existing paradigms typically fall short: autoregressive (AR) VLAs are memory-bandwidth-bound on edge hardware and prone to exposure-bias drift, while full-sequence diffusion models preclude KV-cache reuse and suffer