- Neural Networks with Keras Cookbook
- V Kishore Ayyadevara
- 261字
- 2021-07-02 12:46:27
How to do it...
A deep neural network architecture is built by adding multiple hidden layers between input and output layers, as follows:
- Load the dataset and scale it:
(X_train, y_train), (X_test, y_test) = mnist.load_data()
num_pixels = X_train.shape[1] * X_train.shape[2]
X_train = X_train.reshape(X_train.shape[0], num_pixels).astype('float32')
X_test = X_test.reshape(X_test.shape[0], num_pixels).astype('float32')
X_train = X_train/255
X_test = X_test/255
y_train = np_utils.to_categorical(y_train)
y_test = np_utils.to_categorical(y_test)
num_classes = y_test.shape[1]
- Build a model with multiple hidden layers connecting the input and output layers:
model = Sequential()
model.add(Dense(1000, input_dim=784, activation='relu'))
model.add(Dense(1000,activation='relu'))
model.add(Dense(1000,activation='relu'))
model.add(Dense(10, activation='softmax'))
The preceding model architecture results in a model summary, as follows:

Note that the preceding model results in a higher number of parameters, as a result of deep architectures (as there are multiple hidden layers in the model).
- Now that the model is set up, let's compile and fit the model:
model.compile(loss='categorical_crossentropy', optimizer=adam, metrics=['accuracy'])
history = model.fit(X_train, y_train, validation_data=(X_test, y_test), epochs=250, batch_size=1024, verbose=1)
The preceding results in a model with an accuracy of 98.6%, which is slightly better than the accuracies we observed with the model architectures that we saw earlier. The training and test loss and accuracy are as follows (the code to generate the plots in the following diagram remains the same as the code we used in step 8 of the Training a vanilla neural network recipe):

Note that, in this scenario, there is a considerable gap between training and test loss, indicating that the deep feedforward neural network specialized on training data. Again, in the sections on overfitting, we will learn about ways to avoid overfitting on training data.
- ReSharper Essentials
- .NET 4.0面向對象編程漫談:基礎篇
- Practical Data Science Cookbook(Second Edition)
- 碼上行動:零基礎學會Python編程(ChatGPT版)
- FreeSWITCH 1.6 Cookbook
- 小程序開發原理與實戰
- Learning JavaScript Data Structures and Algorithms
- Kotlin從基礎到實戰
- PHP+MySQL+Dreamweaver動態網站開發從入門到精通(第3版)
- Python全棧數據工程師養成攻略(視頻講解版)
- 基于ARM Cortex-M4F內核的MSP432 MCU開發實踐
- HTML5開發精要與實例詳解
- Python大學實用教程
- 青少年學Python(第2冊)
- C#面向對象程序設計(第2版)