MIPI 2022 Challenge on Under-Display Camera Image Restoration: Methods and Results
Published in ArXiv, 2022
Our team participated in the Under-Display Camera (UDC) Image Restoration track on MIPI 2022 and got 5th place finally.
Publications
Towards Feature Distribution Alignment and Diversity Enhancement for Data-Free Quantization
Published in IEEE International Conference on Data Mining (ICDM), 2022
We propose a new and effective data-free quantization method termed ClusterQ, which utilizes the semantic feature distribution alignment for synthetic data generation.
Dictionary Pair-based Data-Free Fast Deep Neural Network Compression
Published in IEEE International Conference on Data Mining (ICDM), 2021
We propose a new deep model compression method, termed Dictionary Pair-based Data-Free Fast DNN Compression, which aims at reducing the memory consumption of DNNs without extra training and can greatly improve the compression efficiency.
Preprints & Working Papers
Zhao Zhang, Yangcheng Gao, et al. “SelectQ: Calibration Data Selection for Post-Training Quantization,” submitted to CVPR, 2023.
Yan Luo#, Yangcheng Gao#, et al. “Long-Range Zero-Shot Generative Deep Network Quantization,” submitted to CVPR, 2023.
Yangcheng Gao, Zhao Zhang, et al. “ClusterQ: Semantic Feature Distribution Alignment for Data-Free Quantization,” submitted to TNNLS, 2022.
Yangcheng Gao, Zhao Zhang, “Fast Data-free Deep Neural Network Compression with Dictionary Pair Learning,” submitted to KAIS, 2022.