Search Results for author: Xiangyu He

Found 19 papers, 9 papers with code

Towards Accurate Post-training Network Quantization via Bit-Split and Stitching

1 code implementation ICML 2020 Peisong Wang, Qiang Chen, Xiangyu He, Jian Cheng

Network quantization is essential for deploying deep models to IoT devices due to the high efficiency, no matter on special hardware like TPU or general hardware like CPU and GPU.

Image Classification Instance Segmentation +4

ProxyBNN: Learning Binarized Neural Networks via Proxy Matrices

no code implementations ECCV 2020 Xiangyu He, Zitao Mo, Ke Cheng, Weixiang Xu, Qinghao Hu, Peisong Wang, Qingshan Liu, Jian Cheng

The matrix composed of basis vectors is referred to as the proxy matrix, and auxiliary variables serve as the coefficients of this linear combination.

Binarization Quantization

$Z^*$: Zero-shot Style Transfer via Attention Rearrangement

no code implementations25 Nov 2023 Yingying Deng, Xiangyu He, Fan Tang, WeiMing Dong

Despite the remarkable progress in image style transfer, formulating style in the context of art is inherently subjective and challenging.

Style Transfer

Soft Threshold Ternary Networks

1 code implementation4 Apr 2022 Weixiang Xu, Xiangyu He, Tianli Zhao, Qinghao Hu, Peisong Wang, Jian Cheng

The latest STTN shows that ResNet-18 with ternary weights and ternary activations achieves up to 68. 2% Top-1 accuracy on ImageNet.


Revisiting L1 Loss in Super-Resolution: A Probabilistic View and Beyond

no code implementations25 Jan 2022 Xiangyu He, Jian Cheng

Super-resolution as an ill-posed problem has many high-resolution candidates for a low-resolution input.

Image Restoration Super-Resolution

APRIL: Finding the Achilles' Heel on Privacy for Vision Transformers

1 code implementation CVPR 2022 Jiahao Lu, Xi Sheryl Zhang, Tianli Zhao, Xiangyu He, Jian Cheng

Showing how vision Transformers are at the risk of privacy leakage via gradients, we urge the significance of designing privacy-safer Transformer models and defending schemes.

Federated Learning

Improving Binary Neural Networks through Fully Utilizing Latent Weights

no code implementations12 Oct 2021 Weixiang Xu, Qiang Chen, Xiangyu He, Peisong Wang, Jian Cheng

Binary Neural Networks (BNNs) rely on a real-valued auxiliary variable W to help binary training.

Architecture Aware Latency Constrained Sparse Neural Networks

no code implementations1 Sep 2021 Tianli Zhao, Qinghao Hu, Xiangyu He, Weixiang Xu, Jiaxing Wang, Cong Leng, Jian Cheng

Acceleration of deep neural networks to meet a specific latency constraint is essential for their deployment on mobile devices.

Generative Zero-shot Network Quantization

no code implementations21 Jan 2021 Xiangyu He, Qinghao Hu, Peisong Wang, Jian Cheng

Convolutional neural networks are able to learn realistic image priors from numerous training samples in low-level image generation and restoration.

Data Free Quantization Image Generation

Dynamic Dual Gating Neural Networks

1 code implementation ICCV 2021 Fanrong Li, Gang Li, Xiangyu He, Jian Cheng

In particular, dynamic dual gating can provide 59. 7% saving in computing of ResNet50 with 76. 41% top-1 accuracy on ImageNet, which has advanced the state-of-the-art.

Location-aware Upsampling for Semantic Segmentation

1 code implementation13 Nov 2019 Xiangyu He, Zitao Mo, Qiang Chen, Anda Cheng, Peisong Wang, Jian Cheng

Many successful learning targets such as minimizing dice loss and cross-entropy loss have enabled unprecedented breakthroughs in segmentation tasks.

Segmentation Semantic Segmentation

SpatialFlow: Bridging All Tasks for Panoptic Segmentation

1 code implementation19 Oct 2019 Qiang Chen, Anda Cheng, Xiangyu He, Peisong Wang, Jian Cheng

Object location is fundamental to panoptic segmentation as it is related to all things and stuff in the image scene.

Instance Segmentation Object +3

A System-Level Solution for Low-Power Object Detection

no code implementations24 Sep 2019 Fanrong Li, Zitao Mo, Peisong Wang, Zejian Liu, Jiayun Zhang, Gang Li, Qinghao Hu, Xiangyu He, Cong Leng, Yang Zhang, Jian Cheng

As a case study, we evaluate our object detection system on a real-world surveillance video with input size of 512x512, and it turns out that the system can achieve an inference speed of 18 fps at the cost of 6. 9W (with display) with an mAP of 66. 4 verified on the PASCAL VOC 2012 dataset.

Object object-detection +2

Compact Global Descriptor for Neural Networks

1 code implementation23 Jul 2019 Xiangyu He, Ke Cheng, Qiang Chen, Qinghao Hu, Peisong Wang, Jian Cheng

Long-range dependencies modeling, widely used in capturing spatiotemporal correlation, has shown to be effective in CNN dominated computer vision tasks.

Audio Classification Deep Attention +2

K-Nearest Neighbors Hashing

no code implementations CVPR 2019 Xiangyu He, Peisong Wang, Jian Cheng

Hashing based approximate nearest neighbor search embeds high dimensional data to compact binary codes, which enables efficient similarity search and storage.

ODE-Inspired Network Design for Single Image Super-Resolution

no code implementations CVPR 2019 Xiangyu He, Zitao Mo, Peisong Wang, Yang Liu, Mingyuan Yang, Jian Cheng

By casting the numerical schemes in ODE as blueprints, we derive two types of network structures: LF-block and RK-block, which correspond to the Leapfrog method and Runge-Kutta method in numerical ordinary differential equations.

Image Super-Resolution Structured Prediction

Learning Compression from Limited Unlabeled Data

no code implementations ECCV 2018 Xiangyu He, Jian Cheng

Through quantization or pruning, most methods may compress a large number of parameters but ignore the core role in performance degradation, which is the Gaussian conjugate prior induced by batch normalization.

Model Compression Quantization

Cannot find the paper you are looking for? You can Submit a new open access paper.