no code implementations • 15 May 2024 • Honghui Chen, Yuhang Qiu, Jiabao Wang, Pingping Chen, Nam Ling
It reduces the training overhead of PLM while avoiding training fit oscillations.
no code implementations • 12 Apr 2024 • Yuhang Qiu, Honghui Chen, Xingbo Dong, Zheng Lin, Iman Yi Liao, Massimo Tistarelli, Zhe Jin
The first module, an interpretable dense registration module, establishes a Vision Transformer (ViT)-based Siamese Network to capture long-range dependencies and the global context in fingerprint pairs.