DL之VGGNet:VGGNet算法的简介(论文介绍)、架构详解、案例应用等配图集合之详细攻略
目录
1、关于imagenet-vgg-verydeep-19.mat模型简介
相关文章
DL之VGGNet:VGGNet算法的简介(论文介绍)、架构详解、案例应用等配图集合之详细攻
DL之VGGNet:VGGNet算法的架构详解、损失函数、网络训练和测试之详细攻略
VGGNet 是2014 年ILSVRC竞赛分类任务的第二名(第一名是GoogLeNet)和定位任务的第一名,来自牛津大学VGG group 。
Abstract
In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16–19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively.
摘要
本文研究了卷积网络深度对大规模图像识别中卷积网络精度的影响。我们的主要贡献是对深度增加的网络进行彻底评估,这表明,通过将深度推至16-19重量层,可以显著改善现有技术配置。这些发现是我们提交的ImageNet挑战赛2014的基础,我们的团队在localisation和classification tracks中分别获得第一和第二名。
Conclusion
In this work we evaluated very deep convolutional networks (up to 19 weight layers) for largescale image classification. It was demonstrated that the representation depth is beneficial for the classification accuracy, and that the state-of-the-art performance on the ImageNet challenge dataset can be achieved using a conventional ConvNet architecture [10, 11] with substantially increased depth. Namely, our object localisation system won the ILSVRC-2014 localisation challenge, while our classification system took the second place in the classification challenge. Our results yet again confirm the importance of depth in visual representations.
结论
在这项工作中,我们评估了非常深的卷积网络(高达19个权重层)的大规模图像分类。结果表明,表示深度有利于分类的准确性,并且可以使用传统的ConvNet结构[10,11]在深度上显著增加,从而在ImageNet挑战赛数据集上实现最先进的性能。也就是说,我们的目标定位系统赢得了ILSVRC-2014定位挑战,而我们的分类系统在分类挑战中排名第二。我们的结果再次证实了深度在视觉表现中的重要性。
论文:Karen Simonyan and Andrew Zisserman(2014): Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv:1409.1556 [cs] (September 2014).
K. Simonyan and A. Zisserman. Very Deep Convolutional Networks for Large-Scale Image Recognition. ICLR, 2015.
Visual Geometry Group, Department of Engineering Science, University of Oxford
https://arxiv.org/abs/1409.1556v1
DL之VGGNet:VGGNet算法的架构详解、损失函数、网络训练和学习之详细攻略
1、VGG一般的网络结构及其参数
VGG11~VGG19
(1)、训练VGG系列模型所需要的显存:模型D(VGG16)性能效果最好
相关文章
DL之CNN优化技术:学习卷积神经网络CNN的优化、实践经验(练习调参)、从代码深刻认知CNN架构之练习技巧
Stanford CS20SI公开课里的一个神经网络style_transfer小模型的时候,用到了vgg pretrained模型,VGG-Net的结构图,来自论文《VERY DEEP CONVOLUTIONAL NETWORK SFORLARGE-SCALE IMAGE RECOGNITION》,发表于ICLR 2015上,比较起ALEXNET,VGG对图片有更精确的估值以及更省空间。包含了卷积层、池化层、全连接层、soft输出层等。更多详细内容见参考博客
网站声明:如果转载,请联系本站管理员。否则一切后果自行承担。
加入交流群
请使用微信扫一扫!