A two-stage, value-guided decoding strategy with a margin-based reward adjustment is claimed to yield more faithful, detailed VLM captions at about a quarter of VisVM's inference cost.
Transfer q-star: Principled decoding for llm alignment.Advances in Neural Information Processing Systems, 37:101725–101761, 2024
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Dual-Stage Value-Guided Inference with Margin-Based Reward Adjustment for Fast and Faithful VLM Captioning
A two-stage, value-guided decoding strategy with a margin-based reward adjustment is claimed to yield more faithful, detailed VLM captions at about a quarter of VisVM's inference cost.