TY  - RPRT
TI  - Dual-Stage Value-Guided Inference with Margin-Based Reward Adjustment for Fast and Faithful VLM Captioning
AU  - Ankan Deria
AU  - Adinath Madhavrao Dukre
AU  - Feilong Tang
AU  - Sara Atito
AU  - Sudipta Roy
AU  - Muhammad Awais
AU  - Muhammad Haris Khan
AU  - Imran Razzak
PY  - 2025
UR  - https://arxiv.org/abs/2506.15649
ID  - 2506.15649
ER  -