Yuxuan Cai

TL;DR: YOLObile as mentioned in this paper proposes a block-punched pruning scheme for any kernel size to improve computational efficiency on mobile devices, a GPU-CPU collaborative scheme is adopted along with advanced compiler-assisted optimizations.

TL;DR: A general category of fine-grained structured pruning applicable to various DNN layers is proposed, and a comprehensive, compiler automatic code generation framework supporting different DNNs and different pruning schemes are proposed, which bridge the gap of model compression and NAS.

TL;DR: In this article, the authors proposed a weight pruning framework for ReRAM-based mixed-signal DNN accelerators, which effectively reduces the required bits for ADC resolution and hence the overall area and power consumption of the accelerator without introducing any computational inaccuracy.

TL;DR: In this article, a fine-grained block-based pruning scheme is proposed to achieve real-time performance of large-scale neural network inference on mobile devices by integrating hardware-friendly, structured model compression with mobile-targeted compiler optimizations.