TinyAlign: Boosting Lightweight Vision-Language Models by Mitigating Modal Alignment Bottlenecks

v1v2 (latest)

TinyAlign: Boosting Lightweight Vision-Language Models by Mitigating Modal Alignment Bottlenecks

19 May 2025

ArXiv (abs)PDF HTML

Papers citing "TinyAlign: Boosting Lightweight Vision-Language Models by Mitigating Modal Alignment Bottlenecks"

1 / 1 papers shown

Title
$Δ$ -AttnMask: Attention-Guided Masked Hidden States for Efficient Data Selection and Augmentation Jucheng Hu Suorong Yang Dongzhan Zhou 112 0 0 08 Aug 2025