This work presents Whisper-Flash, a two-layer drafter built on a property of speech recognition: the words still to be written have already been spoken, which reads Whisper's encoded audio and accepted decoder states and proposes eight tokens in a single forward pass.
Hua-Peng Zhou, Hua-Yu Wang, Junkai Wu et al.· 0 citations
Experiments on video temporal grounding and highlight detection benchmarks show that TEMPURA substantially improves strong base VLMs across model families and scales, confirming that combining event-level reasoning with fine-grained temporal segmentation is an effective recipe for video temporal understanding.
TreeSpark is introduced, which reads a parent-conditioned distribution from the drafter's existing Markov head at negligible cost, calibrates it into an edge-acceptance estimate, and lets path survival govern everything else: best-first expansion, per-round stopping, and a load-adaptive serving policy.