List your company

CNN Inference Acceleration on Low-Bit Quantized CNNs

22.4Performance Density (GOPS/kLUT)Architecture [10]
3.71128.563113.41518.2669Apr 29, 2021

Evaluation results

MethodLinks
Architecture [10]Device=FPGA, Clock (MHz)=90, Bit Width (bits)=12021.0422.4
Proposed ArchitectureDevice=ASIC, Clock (MHz)=400, Bit Width (bits)=1-22021.047.72
Architecture [21]Device=FPGA, Clock (MHz)=143, Bit Width (bits)=1-22021.044.43