paper-with-me

홈 › Papers

One Weight Bitwidth to Rule Them All

2020-08-22 · Ting-Wu Chin, Pierce I-Jen Chuang, Vikas Chandra, Diana Marculescu

Weight quantization for deep ConvNets has shown promising results for applications such as image classification and semantic segmentation and is especially important for applications where memory storage is limited. However, when aiming for quantization without accuracy degradation, different tasks may end up with different bitwidths. This creates complexity for software and hardware support and the complexity accumulates when one considers mixed-precision quantization, in which case each layer's weights use a different bitwidth. Our key insight is that optimizing for the least bitwidth subject to no accuracy degradation is not necessarily an optimal strategy. This is because one cannot decide optimality between two bitwidths if one has a smaller model size while the other has better accuracy. In this work, we take the first step to understand if some weight bitwidth is better than others by aligning all to the same model size using a width-multiplier. Under this setting, somewhat surprisingly, we show that using a single bitwidth for the whole network can achieve better accuracy compared to mixed-precision quantization targeting zero accuracy degradation when both have the same model size. In particular, our results suggest that when the number of channels becomes a target hyperparameter, a single weight bitwidth throughout the network shows superior results for model compression.

📄 PDF Abstract BibTeX arXiv:2008.09916

Code (0)

등록된 구현이 없습니다.

Tasks

Allimage-classificationImage ClassificationModel CompressionQuantizationSemantic Segmentation

Similar Papers 제목 키워드 기반

Bitwidth Heterogeneous Federated Learning with Progressive Weight Dequantization

2022-02-23 · Jaehong Yoon, Geon Park, Wonyong Jeong, Sung Ju Hwang

In practical federated learning scenarios, the participating devices may have different bitwidths for computation and memory storage by design. However, despite the progress made in device-heterogeneous federated learnin…

Federated Learning

DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients

2016-06-20 · Shuchang Zhou, Yuxin Wu, Zekun Ni, Xinyu Zhou 외

We propose DoReFa-Net, a method to train convolutional neural networks that have low bitwidth weights and activations using low bitwidth parameter gradients. In particular, during backward pass, parameter gradients are s…

CPUGPUQuantization

Efficient Bitwidth Search for Practical Mixed Precision Neural Network

2020-03-17 · Yuhang Li, Wei Wang, Haoli Bai, Ruihao Gong 외

Network quantization has rapidly become one of the most widely used methods to compress and accelerate deep neural networks. Recent efforts propose to quantize weights and activations from different layers with different…

Quantization

FAT: Learning Low-Bitwidth Parametric Representation via Frequency-Aware Transformation

2021-02-15 · Chaofan Tao, Rui Lin, Quan Chen, Zhaoyang Zhang 외

Learning convolutional neural networks (CNNs) with low bitwidth is challenging because performance may drop significantly after quantization. Prior arts often discretize the network weights by carefully tuning hyper-para…

Model CompressionNeural Network CompressionQuantization

WAVEQ: GRADIENT-BASED DEEP QUANTIZATION OF NEURAL NETWORKS THROUGH SINUSOIDAL REGULARIZATION

2021-01-01 · Ahmed T. Elthakeb, Prannoy Pilligundla, Tarek Elgindi, FatemehSadat Mireshghallah 외

Deep quantization of neural networks below eight bits can lead to superlinear benefits in storage and compute efficiency. However, homogeneously quantizing all the layers to the same level does not account for the distin…

Quantization