最近想使用DenseNet做特征提取,但是不知道DenseNet具体结构,所以做了一下DenseNet结构可视化。
# -*- coding: utf-8 -*- """ Created on Tue Feb 19 13:35:11 2019 @author: 13260 """ from keras.applications.densenet import DenseNet201,preprocess_input from keras.models import Model,load_model import numpy as np from keras.layers import Dense, GlobalAveragePooling2D from keras.preprocessing import image #base_model = DenseNet(weights='imagenet', include_top=False) base_model = DenseNet201(weights='imagenet', include_top=False) #base_model = load_model("F:/python/python-GenerRec/src/1019Resnet_gener_model_weights.h5") #base_model.get_layer() model = Model(inputs=base_model.input, outputs=base_model.output) model.summary() print('the number of layers in this model:'+str(len(model.layers))) 代码运行结果如图:
本站文章如无特殊说明,均为本站原创,如若转载,请注明出处:keras查看网络结构 - Python技术站