资源分类
》 资源 》算法》ResNet-DenseNet-Notebook

ResNet-DenseNet-Notebook

2020-03-30 | |  45 |   0 |   0

ResNet-DenseNet-Notebook

This Jupyter notebook is using keras to build the residual or dense block, and compare with four models in cifar100 classification as following.

  • 18 layers residual network with ReLU and batch normalization

  • 18 layers plain network with ReLU and batch normalization

  • 18 layers residual network with SELU

  • 18 layers plain network with SELU

All model was trained with 100 epochs and Adam(learning rate:$10^{-5}$)

TL;DR

Residual Block Plot

Loss

Loss

Accuracy

Accuracy

Whole Model plot

Residual Network with ReLU and batch normalization

Residual Network with SELU


上一篇： depth_estimation_with_densenet-unet_hybrid

下一篇：quiz-w7-2-densenet

用户评价
全部评价

热门资源

  • seetafaceJNI

    项目介绍 基于中科院seetaface2进行封装的JAVA...

  • spark-corenlp

    This package wraps Stanford CoreNLP annotators ...

  • Keras-ResNeXt

    Keras ResNeXt Implementation of ResNeXt models...

  • capsnet-with-caps...

    CapsNet with capsule-wise convolution Project ...

  • shih-styletransfer

    shih-styletransfer Code from Style Transfer ...