Simultaneous or sequential integration of geometric-median filter pruning with 4-bit additive-power-of-two quantization compresses ResNet and VGG models on CIFAR-10 by about 15x with modest accuracy loss.
Block and subword-s caling floating-point (BSFP) : An efficient non-uniform quantizati on for low precision inference,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.NE 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Integrating Pruning with Quantization for Efficient Deep Neural Networks Compression
Simultaneous or sequential integration of geometric-median filter pruning with 4-bit additive-power-of-two quantization compresses ResNet and VGG models on CIFAR-10 by about 15x with modest accuracy loss.