Picture for James Garland

James Garland

Low precision logarithmic number systems: Beyond base-2

Add code
Feb 12, 2021
Figure 1 for Low precision logarithmic number systems: Beyond base-2
Figure 2 for Low precision logarithmic number systems: Beyond base-2
Figure 3 for Low precision logarithmic number systems: Beyond base-2
Figure 4 for Low precision logarithmic number systems: Beyond base-2
Viaarxiv icon

Metric Embedding Sub-discrimination Study

Add code
Feb 05, 2021
Figure 1 for Metric Embedding Sub-discrimination Study
Figure 2 for Metric Embedding Sub-discrimination Study
Figure 3 for Metric Embedding Sub-discrimination Study
Figure 4 for Metric Embedding Sub-discrimination Study
Viaarxiv icon

HOBFLOPS CNNs: Hardware Optimized Bitsliced Floating-Point Operations Convolutional Neural Networks

Add code
Jul 11, 2020
Figure 1 for HOBFLOPS CNNs: Hardware Optimized Bitsliced Floating-Point Operations Convolutional Neural Networks
Figure 2 for HOBFLOPS CNNs: Hardware Optimized Bitsliced Floating-Point Operations Convolutional Neural Networks
Figure 3 for HOBFLOPS CNNs: Hardware Optimized Bitsliced Floating-Point Operations Convolutional Neural Networks
Figure 4 for HOBFLOPS CNNs: Hardware Optimized Bitsliced Floating-Point Operations Convolutional Neural Networks
Viaarxiv icon

Low Complexity Multiply Accumulate Unit for Weight-Sharing Convolutional Neural Networks

Add code
Jan 19, 2017
Figure 1 for Low Complexity Multiply Accumulate Unit for Weight-Sharing Convolutional Neural Networks
Figure 2 for Low Complexity Multiply Accumulate Unit for Weight-Sharing Convolutional Neural Networks
Figure 3 for Low Complexity Multiply Accumulate Unit for Weight-Sharing Convolutional Neural Networks
Figure 4 for Low Complexity Multiply Accumulate Unit for Weight-Sharing Convolutional Neural Networks
Viaarxiv icon