Binarized Neural Networks on the ImageNet Classification Task

04/11/2016
by   Xundong Wu, et al.
0

We trained Binarized Neural Networks (BNNs) on the high resolution ImageNet ILSVRC-2102 dataset classification task and achieved a good performance. With a moderate size network of 13 layers, we obtained top-5 classification accuracy rate of 84.1 previous published results of 73.2 GoogleNET. We expect networks of better performance can be obtained by following our current strategies. We provide a detailed discussion and preliminary analysis on strategies used in the network training.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/16/2016

XNOR-Net: ImageNet Classification Using Binary Convolutional Neural Networks

We propose two efficient approximations to standard convolutional neural...
research
09/29/2021

The MAMe Dataset: On the relevance of High Resolution and Variable Shape image properties

In the image classification task, the most common approach is to resize ...
research
07/31/2016

Deep FisherNet for Object Classification

Despite the great success of convolutional neural networks (CNN) for the...
research
09/28/2022

Increasing the Accuracy of a Neural Network Using Frequency Selective Mesh-to-Grid Resampling

Neural networks are widely used for almost any task of recognizing image...
research
03/13/2017

What You Expect is NOT What You Get! Questioning Reconstruction/Classification Correlation of Stacked Convolutional Auto-Encoder Features

In this paper, we thoroughly investigate the quality of features produce...
research
01/20/2022

What can we learn from misclassified ImageNet images?

Understanding the patterns of misclassified ImageNet images is particula...
research
03/10/2022

Robustness Analysis of Classification Using Recurrent Neural Networks with Perturbed Sequential Input

For a given stable recurrent neural network (RNN) that is trained to per...

Please sign up or login with your details

Forgot password? Click here to reset