Zobrazeno 1 - 10
of 131
pro vyhledávání: '"Li, Yunqiang"'
Autor:
Li, Yunqiang, van Gemert, Jan C., Hoefler, Torsten, Moons, Bert, Eleftheriou, Evangelos, Verhoef, Bram-Ernst
Deep learning algorithms are increasingly employed at the edge. However, edge devices are resource constrained and thus require efficient deployment of deep neural networks. Pruning methods are a key tool for edge deployment as they can improve stora
Externí odkaz:
http://arxiv.org/abs/2307.08483
Binary Neural Networks (BNNs) are compact and efficient by using binary weights instead of real-valued weights. Current BNNs use latent real-valued weights during training, where several training hyper-parameters are inherited from real-valued networ
Externí odkaz:
http://arxiv.org/abs/2303.02452
Autor:
Li, Yunqiang1,2 (AUTHOR) chenying@jimuyida.com, Huang, Shuowen1 (AUTHOR) huangshuowen@whu.edu.cn, Chen, Ying2 (AUTHOR) dingyong@jimuyida.com, Ding, Yong2 (AUTHOR), Zhao, Pengcheng1 (AUTHOR) pengcheng.zhao@whu.edu.cn, Hu, Qingwu1 (AUTHOR) ac_zxj@whu.edu.cn, Zhang, Xujie1,3 (AUTHOR)
Publikováno v:
Remote Sensing. Sep2024, Vol. 16 Issue 17, p3188. 18p.
Autor:
Cao, Yangyang, Wang, Zheng, Wan, Jieru, He, Yuzhu, Li, Yunqiang, Wang, Sheng, Wang, Yanli, Song, Dalei, Zhang, Tao
Publikováno v:
In Journal of Colloid And Interface Science September 2024 669:912-926
Binary networks are extremely efficient as they use only two symbols to define the network: $\{+1,-1\}$. One can make the prior distribution of these symbols a design choice. The recent IR-Net of Qin et al. argues that imposing a Bernoulli distributi
Externí odkaz:
http://arxiv.org/abs/2112.03406
Publikováno v:
In Colloids and Surfaces A: Physicochemical and Engineering Aspects 20 March 2024 685
Autor:
Li, Yunqiang, van Gemert, Jan
Unsupervised hashing is important for indexing huge image or video collections without having expensive annotations available. Hashing aims to learn short binary codes for compact storage and efficient semantic retrieval. We propose an unsupervised d
Externí odkaz:
http://arxiv.org/abs/2012.12334
Publikováno v:
In Journal of Environmental Chemical Engineering February 2024 12(1)
Current weakly supervised object localization and segmentation rely on class-discriminative visualization techniques to generate pseudo-labels for pixel-level training. Such visualization methods, including class activation mapping (CAM) and Grad-CAM
Externí odkaz:
http://arxiv.org/abs/2010.08644
Batch normalization (BN) allows training very deep networks by normalizing activations by mini-batch sample statistics which renders BN unstable for small batch sizes. Current small-batch solutions such as Instance Norm, Layer Norm, and Group Norm us
Externí odkaz:
http://arxiv.org/abs/2010.07160