1 paper · 1 filter
Hui Tang, Yao Lu, Qi Xuan
Despite the popularization of deep neural networks (DNNs) in many fields, it is still challenging to deploy state-of-the-art models to resource-constrained devices due to high comp…