Speculative-Decoding

an archive of posts with this tag

Aug 25, 2026	Paper Breakdown: SpecVLM — Speeding Up Video LLMs with Verifier-Guided Token Pruning