keras导入weights方式

ゝ一纸荒年。 2023-02-18 09:53 14阅读 0赞

keras导入weights方式

更多python视频教程请到菜鸟教程https://www.piaodoo.com/

keras源码engine中toplogy.py定义了加载权重的函数：

> load\_weights(self, filepath, by\_name=False)

其中默认by\_name为False,这时候加载权重按照网络拓扑结构加载,适合直接使用keras中自带的网络模型，如VGG16

VGG19/resnet50等，源码描述如下:

> If \`by\_name\` is False (default) weights are loaded  
> based on the network's topology, meaning the architecture  
> should be the same as when the weights were saved.  
> Note that layers that don't have weights are not taken  
> into account in the topological ordering, so adding or  
> removing layers is fine as long as they don't have weights.

若将by\_name改为True则加载权重按照layer的name进行,layer的name相同时加载权重,适合用于改变了

模型的相关结构或增加了节点但利用了原网络的主体结构情况下使用,源码描述如下：

> If \`by\_name\` is True, weights are loaded into layers  
> only if they share the same name. This is useful  
> for fine-tuning or transfer-learning models where  
> some of the layers have changed.

在进行边缘检测时,利用VGG网络的主体结构,网络中增加反卷积层,这时加载权重应该使用

> model.load\_weights(filepath,by\_name=True)

**补充知识：****Keras下实现mnist手写数字**

之前一直在用tensorflow，被同学推荐来用keras了，把之前文档中的mnist手写数字数据集拿来练手，

代码如下。

import struct
    import numpy as np
    import os

import keras  
from keras.models import Sequential  
from keras.layers import Dense  
from keras.optimizers import SGD

def load\_mnist(path, kind):  
labels\_path = os.path.join(path, ‘%s-labels.idx1-ubyte’ % kind)  
images\_path = os.path.join(path, ‘%s-images.idx3-ubyte’ % kind)  
with open(labels\_path, ‘rb’) as lbpath:  
magic, n = struct.unpack(’>II’, lbpath.read(8))  
labels = np.fromfile(lbpath, dtype=np.uint8)  
with open(images\_path, ‘rb’) as imgpath:  
magic, num, rows, cols = struct.unpack(">IIII", imgpath.read(16))  
images = np.fromfile(imgpath, dtype=np.uint8).reshape(len(labels), 784) \#28\*28=784  
return images, labels

\#loading train and test data  
X\_train, Y\_train = load\_mnist(’.\\data’, kind=‘train’)  
X\_test, Y\_test = load\_mnist(’.\\data’, kind=‘t10k’)

\#turn labels to one\_hot code  
Y\_train\_ohe = keras.utils.to\_categorical(Y\_train, num\_classes=10)

\#define models  
model = Sequential()

model.add(Dense(input\_dim=X\_train.shape\[1\],output\_dim=50,init=‘uniform’,activation=‘tanh’))  
model.add(Dense(input\_dim=50,output\_dim=50,init=‘uniform’,activation=‘tanh’))  
model.add(Dense(input\_dim=50,output\_dim=Y\_train\_ohe.shape\[1\],init=‘uniform’,activation=‘softmax’))

sgd = SGD(lr=0.001, decay=1e-7, momentum=0.9, nesterov=True)  
model.compile(loss=‘categorical\_crossentropy’, optimizer=sgd, metrics=\[“accuracy”\])

\#start training  
model.fit(X\_train,Y\_train\_ohe,epochs=50,batch\_size=300,shuffle=True,verbose=1,validation\_split=0.3)

\#count accuracy  
y\_train\_pred = model.predict\_classes(X\_train, verbose=0)

train\_acc = np.sum(Y\_train == y\_train\_pred, axis=0) / X\_train.shape\[0\]  
print(‘Training accuracy: %.2f%%’ % (train\_acc \* 100))

y\_test\_pred = model.predict\_classes(X\_test, verbose=0)  
test\_acc = np.sum(Y\_test == y\_test\_pred, axis=0) / X\_test.shape\[0\]  
print(‘Test accuracy: %.2f%%’ % (test\_acc \* 100))

训练结果如下：

Epoch 45/50
    42000/42000 [==============================] - 1s 17us/step - loss: 0.2174 - acc: 0.9380 - val_loss: 0.2341 - val_acc: 0.9323
    Epoch 46/50
    42000/42000 [==============================] - 1s 17us/step - loss: 0.2061 - acc: 0.9404 - val_loss: 0.2244 - val_acc: 0.9358
    Epoch 47/50
    42000/42000 [==============================] - 1s 17us/step - loss: 0.1994 - acc: 0.9413 - val_loss: 0.2295 - val_acc: 0.9347
    Epoch 48/50
    42000/42000 [==============================] - 1s 17us/step - loss: 0.2003 - acc: 0.9413 - val_loss: 0.2224 - val_acc: 0.9350
    Epoch 49/50
    42000/42000 [==============================] - 1s 18us/step - loss: 0.2013 - acc: 0.9417 - val_loss: 0.2248 - val_acc: 0.9359
    Epoch 50/50
    42000/42000 [==============================] - 1s 17us/step - loss: 0.1960 - acc: 0.9433 - val_loss: 0.2300 - val_acc: 0.9346
    Training accuracy: 94.11%
    Test accuracy: 93.61%

以上这篇keras导入weights方式就是小编分享给大家的全部内容了，希望能给大家一个参考，也希望大家多多

茂名论坛https://www.hnthzk.com/

化州橘红http://www.sfkyty.com/

茂名论坛http://www.nrso.net/

源码搜藏网http://www.cntkd.net/