VVS accelerates visual AR image generation by partially skipping verifications in speculative decoding, achieving 2.8x fewer target forward passes while preserving competitive quality.
Eagle-2: Faster Inference of Language Models with Dy- namic Draft Trees
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
VVS: Accelerating Speculative Decoding for Visual Autoregressive Generation via Partial Verification Skipping
VVS accelerates visual AR image generation by partially skipping verifications in speculative decoding, achieving 2.8x fewer target forward passes while preserving competitive quality.