Python 时装设计师使用CNN，过度装修？

最后我得到了这个结果：

cnn_model = Sequential()

cnn_model.add(Conv2D(64,3, 3, input_shape = (28,28,1), activation='relu'))
cnn_model.add(MaxPooling2D(pool_size = (2, 2)))

cnn_model.add(Dropout(0.25))

cnn_model.add(Flatten())
cnn_model.add(Dense(output_dim = 32, activation = 'relu'))
cnn_model.add(Dense(output_dim = 10, activation = 'sigmoid'))

cnn_model.compile(loss ='sparse_categorical_crossentropy', optimizer=Adam(lr=0.001),metrics =['accuracy'])

epochs = 50

history = cnn_model.fit(X_train,
                        y_train,
                        batch_size = 512,
                        nb_epoch = epochs,
                        verbose = 1,
                        validation_data = (X_validate, y_validate))

我的问题如下:

模型是过拟合还是过拟合

我应该增加新纪元的数量吗

---

由于模型过于合适，您可以

cnn\u model.fit中使用
Shuffle=True
对数据进行Shuffle

处理

使用

提前停止

使用

正则化

完整的代码，使用与您相同的体系结构，减少过拟合，如下所示。损失有所增加，但我们可以通过添加更多的卷积层和池层来改善它

Epoch 50/50
48000/48000 [==============================] - 35s 728us/step - loss: 0.1265 - accuracy: 0.9537 - val_loss: 0.2425 - val_accuracy: 0.9167
training loss=0.125 ,validation loss=0.2425
training accuracy=95.3% ,validation accuracy=91.67

---

我认为情节本身就说明了问题。这太过合适了，不，更多的时代也无济于事。

# To support both python 2 and python 3
from __future__ import division, print_function, unicode_literals
from io import open

# Common imports
import numpy as np
import os
import tensorflow as tf
from tensorflow.keras.layers import Input, Conv2D, MaxPool2D, Dense, Dropout, Flatten
from tensorflow.keras.models import Sequential
from tensorflow.keras.optimizers import Adam
import matplotlib.pyplot as plt
from tensorflow.keras.regularizers import l2
from matplotlib.pyplot import axis as ax

tf.__version__

X = tf.placeholder(tf.float32, shape=[None, 784], name="X")
X_reshaped = tf.reshape(X, shape=[-1, 28, 28, 1])
y = tf.placeholder(tf.int32, shape=[None], name="y")

# instantiate regularizer
Regularizer = l2(0.001)

cnn_model = Sequential()

cnn_model.add(Conv2D(64,3, 3, input_shape = (28,28,1), activation='relu', data_format='channels_last', 
                    activity_regularizer=Regularizer, kernel_regularizer=Regularizer))

cnn_model.add(MaxPool2D(pool_size = (2, 2)))

cnn_model.add(Dropout(0.25))

cnn_model.add(Flatten())

cnn_model.add(Dense(units = 32, activation = 'relu', 
                    activity_regularizer=Regularizer, kernel_regularizer=Regularizer))

cnn_model.add(Dense(units = 10, activation = 'sigmoid', 
                    activity_regularizer=Regularizer, kernel_regularizer=Regularizer))

cnn_model.compile(loss ='sparse_categorical_crossentropy', optimizer=Adam(lr=0.001),metrics =['accuracy'])

epochs = 50

cnn_model.summary()

(X_train, y_train), (X_test, y_test) = tf.keras.datasets.fashion_mnist.load_data()
X_train = X_train.astype(np.float32).reshape(-1, 28*28) / 255.0
X_train_reshaped = tf.reshape(X_train, shape=[-1, 28, 28, 1])
X_test = X_test.astype(np.float32).reshape(-1, 28*28) / 255.0
X_test_reshaped = tf.reshape(X_test, shape=[-1, 28, 28, 1])

y_train = tf.cast(y_train, dtype = tf.int32)

y_test = tf.cast(y_test, dtype = tf.int32)

def shuffle_batch(X, y, batch_size):
    rnd_idx = np.random.permutation(len(X))
    n_batches = len(X) // batch_size
    for batch_idx in np.array_split(rnd_idx, n_batches):
        X_batch, y_batch = X[batch_idx], y[batch_idx]
        yield X_batch, y_batch

#steps_per_epoch = X_train_reshaped.shape[0]//512
steps_per_epoch = X_train_reshaped.shape[0].value//512

epochs = 50

callback = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=15)

history = cnn_model.fit(x = X_train_reshaped,
                        y = y_train,
                        batch_size = 512,
                        epochs = epochs, callbacks=[callback],
                        verbose = 1, validation_data = (X_test_reshaped, y_test),
                        validation_steps = 10, steps_per_epoch=steps_per_epoch, shuffle = True)

print(history.history.keys())

#  "Accuracy"
plt.plot(history.history['acc'])
plt.plot(history.history['val_acc'])
plt.title('model accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'validation'], loc='upper left')
plt.show()
# "Loss"
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('model loss')
axes = plt.axes()
axes.set_ylim([0.2, 1.5])
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'validation'], loc='upper left')
plt.show()