paper-with-me

홈 › Papers

Towards the Limit of Network Quantization

2016-12-05 · Yoojin Choi, Mostafa El-Khamy, Jungwon Lee

Network quantization is one of network compression techniques to reduce the redundancy of deep neural networks. It reduces the number of distinct network parameter values by quantization in order to save the storage for them. In this paper, we design network quantization schemes that minimize the performance loss due to quantization given a compression ratio constraint. We analyze the quantitative relation of quantization errors to the neural network loss function and identify that the Hessian-weighted distortion measure is locally the right objective function for the optimization of network quantization. As a result, Hessian-weighted k-means clustering is proposed for clustering network parameters to quantize. When optimal variable-length binary codes, e.g., Huffman codes, are employed for further compression, we derive that the network quantization problem can be related to the entropy-constrained scalar quantization (ECSQ) problem in information theory and consequently propose two solutions of ECSQ for network quantization, i.e., uniform quantization and an iterative solution similar to Lloyd's algorithm. Finally, using the simple uniform quantization followed by Huffman coding, we show from our experiments that the compression ratios of 51.25, 22.17 and 40.65 are achievable for LeNet, 32-layer ResNet and AlexNet, respectively.

📄 PDF Abstract BibTeX arXiv:1612.01543

Code (0)

등록된 구현이 없습니다.

Tasks

ClusteringQuantization

Methods 이 논문이 사용한 방법론

Average Pooling 설명 없음
Local Response Normalization Local Response Normalization is a normalization layer that implements the idea of lateral inhibition. Lateral inhibition is a concept in neurobiology that refers to the…
Grouped Convolution A Grouped Convolution uses a group of convolutions - multiple kernels per layer - resulting in multiple channel outputs per layer. This leads to wider networks helping a…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…
LeNet LeNet is a classic convolutional neural network employing the use of convolutions, pooling and fully connected layers. It was used for the handwritten digit recognition task…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
How do I speak to a person at Expedia?-/+/ To make a reservation or communicate with Expedia, the quickest option is typically to call their customer service at +1-805-330-4056 or +1-805-330-4056. You can also use the live…

Similar Papers 제목 키워드 기반

Adaptive quantization with mixed-precision based on low-cost proxy

2024-02-27 · Junzhe Chen, Qiao Yang, Senmao Tian, Shunli Zhang

It is critical to deploy complicated neural network models on hardware with limited resources. This paper proposes a novel model quantization method, named the Low-Cost Proxy-Based Adaptive Mixed-Precision Model Quantiza…

Neural Architecture SearchQuantization

Hardware-friendly Deep Learning by Network Quantization and Binarization

2021-12-01 · Haotong Qin

Quantization is emerging as an efficient approach to promote hardware-friendly deep learning and run deep neural networks on resource-limited hardware. However, it still causes a significant decrease to the network in ac…

BinarizationDeep LearningQuantization

An Investigation on Different Underlying Quantization Schemes for Pre-trained Language Models

2020-10-14 · Zihan Zhao, Yuncong Liu, Lu Chen, Qi Liu 외

Recently, pre-trained language models like BERT have shown promising performance on multiple natural language processing tasks. However, the application of these models has been limited due to their huge size. To reduce …

ClusteringQuantization

Real-Time Distributed Model Predictive Control with Limited Communication Data Rates

2022-08-26 · Yujia Yang, Ye Wang, Chris Manzie, Ye Pu

The application of distributed model predictive controllers (DMPC) for multi-agent systems (MASs) necessitates communication between agents, yet the consequence of communication data rates is typically overlooked. This w…

Distributed OptimizationModel Predictive ControlQuantization

TopGQ: Fast GNN Post-Training Quantization Leveraging Topology Information

2026-08-31 · Dain Kwon, Kanghyun Choi, Hyeyoon Lee, Sunjong Park 외 arxiv

Existing GNN quantization methods suffer from considerable quantization overhead, which severely limits their practical usage in real-world scenarios. To this end, we present TopGQ, an accurate post-training GNN quantiza…