代表性论文
[1]Mengshu Sun, Kaidi Xu, Xue Lin, Yongli Hu, and Baocai Yin.Hardware- Friendly3DCNNAccelerationWithBalancedKernelGroupSparsity.InIEEETransactions on Computer-Aided Design of Integrated Circuits and Systems (TCAD), volume 43, issue 10, pages 3027–3040, 2024.
[2]MengshuSun,PuZhao,MehmetGungor,MassoudPedram,MiriamLeeser,and Xue Lin.3d cnn acceleration on fpga using hardware-aware pruning.InACM/IEEE Design Automation Conference (DAC), pages 1–6, 2020.
[3]MengshuSun,ZhengangLi,AlecLu,HaoyuMa,GengYuan,YanyueXie,Hao Tang,YanyuLi,MiriamLeeser,ZhangyangWang,XueLin,andZhenmanFang.Fpga-aware automaticaccelerationframeworkforvisiontransformerwithmixed-schemequantization:late breaking results.InProceedings of the ACM/IEEE Design Automation Conference (DAC), pages 1394–1395, 2022.
[4]MengshuSun, Zhengang Li, Alec Lu, Yanyu Li, Sung-En Chang, Xiaolong Ma, Xue Lin, and Zhenman Fang.Film-qnn: Efficient fpga acceleration of deep neural networks with intra-layer, mixed-precision quantization.InProceedings of the ACM/SIGDA International Symposium on Field-Programmable Gate Arrays (FPGA), pages 134–145, 2022.
[5]Wei Niu,Mengshu Sun, Zhengang Li, Jou-An Chen, Jiexiong Guan, Xipeng Shen,YanzhiWang,SijiaLiu,XueLin,andBinRen.Rt3d: Achievingreal-timeexecutionof 3d convolutional neural networks on mobile devices.InProceedings of the AAAI Conference on Artificial Intelligence (AAAI), volume 35, pages 9179–9187, 2021.
[6]PeiyanDong,MengshuSun,AlecLu,YanyueXie,KennethLiu,ZhenglunKong, Xin Meng, Zhengang Li, Xue Lin, Zhenman Fang, and Yanzhi Wang.Heatvit: Hardware- efficient adaptive token pruning for vision transformers.InIEEE International Symposium on High-Performance Computer Architecture (HPCA), pages 442-455, 2023.
[7]Sung-En Chang, Yanyu Li,Mengshu Sun, Runbin Shi, Hayden K-H So, Xuehai Qian, YanzhiWang, andXueLin.Mixandmatch: Anovelfpga-centricdeepneuralnetwork quantization framework.InIEEE International Symposium on High-Performance Computer Architecture (HPCA), pages 208–220, 2021.
[8]Sung-EnChang,YanyuLi,MengshuSun,WeiwenJiang,SijiaLiu,YanzhiWang,and Xue Lin.Rmsmp: A novel deep neural network quantization framework with row-wise mixed schemes and multiple precisions.InProceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 5251–5260, 2021.