上篇博客说了ResNet和DenseNet的原理，这次说说具体实现

ResNet

def basic_block(input, in_features, out_features, stride, is_training, keep_prob):"""Residual block"""if stride == 1:shortcut = inputelse:shortcut = tf.nn.avg_pool(input, [ 1, stride, stride, 1 ], [1, stride, stride, 1 ], 'VALID')shortcut = tf.pad(shortcut, [[0, 0], [0, 0], [0, 0],[(out_features-in_features)//2, (out_features-in_features)//2]])current = conv2d(input, in_features, out_features, 3, stride)current = tf.nn.dropout(current, keep_prob)current = tf.contrib.layers.batch_norm(current, scale=True, is_training=is_training, updates_collections=None)current = tf.nn.relu(current)current = conv2d(current, out_features, out_features, 3, 1)current = tf.nn.dropout(current, keep_prob)current = tf.contrib.layers.batch_norm(current, scale=True, is_training=is_training, updates_collections=None)return current + shortcutdef block_stack(input, in_features, out_features, stride, depth, is_training, keep_prob):"""Stack Residual block"""current = basic_block(input, in_features, out_features, stride, is_training, keep_prob)for _d in xrange(depth - 1):current = basic_block(current, out_features, out_features, 1, is_training, keep_prob)return current

DenseNet

def conv2d(input, in_features, out_features, kernel_size, with_bias=False):W = weight_variable([ kernel_size, kernel_size, in_features, out_features ])conv = tf.nn.conv2d(input, W, [ 1, 1, 1, 1 ], padding='SAME')if with_bias:return conv + bias_variable([ out_features ])return convdef batch_activ_conv(current, in_features, out_features, kernel_size, is_training, keep_prob):"""BatchNorm+Relu+conv+dropout"""current = tf.contrib.layers.batch_norm(current, scale=True, is_training=is_training, updates_collections=None)current = tf.nn.relu(current)current = conv2d(current, in_features, out_features, kernel_size)current = tf.nn.dropout(current, keep_prob)return currentdef block(input, layers, in_features, growth, is_training, keep_prob):"""Dense Block"""current = inputfeatures = in_featuresfor idx in xrange(layers):tmp = batch_activ_conv(current, features, growth, 3, is_training, keep_prob)current = tf.concat(3, (current, tmp))features += growthreturn current, featuresdef model():"""DenseNet on ImageNet"""current = tf.reshape(xs, [ -1, 32, 32, 3 ])  # Inputcurrent = conv2d(current, 3, 16, 3)current, features = block(current, layers, 16, 12, is_training, keep_prob)current = batch_activ_conv(current, features, features, 1, is_training, keep_prob)current = avg_pool(current, 2)current, features = block(current, layers, features, 12, is_training, keep_prob)current = batch_activ_conv(current, features, features, 1, is_training, keep_prob)current = avg_pool(current, 2)current, features = block(current, layers, features, 12, is_training, keep_prob)current = tf.contrib.layers.batch_norm(current, scale=True, is_training=is_training, updates_collections=None)current = tf.nn.relu(current)current = avg_pool(current, 8)final_dim = featurescurrent = tf.reshape(current, [ -1, final_dim ])Wfc = weight_variable([ final_dim, label_count ])bfc = bias_variable([ label_count ])ys_ = tf.nn.softmax( tf.matmul(current, Wfc) + bfc )

代码不是完整的，只是表达最navie的思想核心部分

ResNet DenseNet(实践篇)相关推荐

[深度学习-总结]Deep learning中8大模型介绍与比较(LeNet5,AlexNet,VGG,Inception,MobileNets,ResNet,DenseNet,Senet)
深度学习 9中模型介绍与比较 0. CNN 结构演化 1. LeNet5 2. AlexNet 3. VGG 为什么使用2个3x3卷积核可以来代替5*5卷积核 4. 1*1卷积 5. Inceptio ...
系统学习深度学习（二十）--ResNet,DenseNet,以及残差家族
转自:http://blog.csdn.net/cv_family_z/article/details/50328175 CVPR2016 https://github.com/KaimingHe/d ...
ResNet DenseNet(原理篇)
这篇博客讲现在很流行的两种网络模型,ResNet和DenseNet,其实可以把DenseNet看做是ResNet的特例文章地址: [1]Deep Residual Learning for Im ...
PyTorch - FashionMNIST + LeNet / AlexNet / VGG / GooLeNet / NiN / ResNet / DenseNet
文章目录项目说明数据集说明 - FashionMNIST 算法说明 - LeNet-5 LeNet-5 网络结构代码实现数据准备下载数据集查看数据定义网络训练设置参数训练方法验证 ...
极深网络（ResNet/DenseNet）: Skip Connection为何有效及其它
/* 版权声明:可以任意转载,转载时请标明文章原始出处和作者信息 .*/ Residual Network通过引入Skip Connection到CNN网络结构中,使得网络深度达到了千层的规模,并且其 ...
resnet论文_ResNet还是DenseNet？即插即用的DS涨点神器来了！
DSNet比ResNet取得了更好的结果,并且具有与DenseNet相当的性能,但需要的计算资源更少.其中改进的DS2Res2Net性能非常强大. 作者:ChaucerG Date:2020-10-2 ...
一点就分享系列（实践篇6——上篇）【迟到补发_详解yolov8】Yolo-high_level系列融入YOLOv8 旨在研究和兼容使用【3月份开始持续补更】
一点就分享系列(实践篇5-补更篇)[迟到补发]-Yolo系列算法开源项目融入V8旨在研究和兼容使用[持续更新] 题外话 [最近一直在研究开放多模态泛化模型的应用事情,所以这部分内容会更新慢一些,文章和 ...
densenet tensorflow 中文汉字手写识别
densenet 中文汉字手写识别,代码如下: import tensorflow as tf import os import random import math import tensorflo ...
CNN网络架构演进：从LeNet到DenseNet
原文来源:https://www.cnblogs.com/skyfsm/p/8451834.html 卷积神经网络可谓是现在深度学习领域中大红大紫的网络框架,尤其在计算机视觉领域更是一枝独秀.CNN从 ...

ResNet DenseNet(实践篇)

ResNet

DenseNet

ResNet DenseNet(实践篇)相关推荐

最新文章

热门文章