使用神经网络工具TensorFlow训练FashionMNIST

数据集FashionMNIST

FashionMNIST 是一个替代 MNIST 手写数字集的图像数据集。其涵盖了来自 10 种类别的共 7 万个不同商品的正面图片。
FashionMNIST 的大小、格式和训练集/测试集划分与原始的 MNIST 完全一致。60000/10000 的训练测试数据划分，28x28 的灰度图片。

加载数据集

网络加载数据集

fashion_mnist = keras.datasets.fashion_mnist
(train_images, train_labels),(test_images, test_labels) = fashion_mnist.load_data()

本地加载数据集
定义加载数据的函数，./datasets/fashion/为保存gz数据的文件夹，该文件夹下有4个文件
'train-labels-idx1-ubyte.gz',
'train-images-idx3-ubyte.gz',
't10k-labels-idx1-ubyte.gz',
't10k-images-idx3-ubyte.gz'

def load_data(data_folder):
 files = [
     'train-labels-idx1-ubyte.gz', 'train-images-idx3-ubyte.gz',
     't10k-labels-idx1-ubyte.gz', 't10k-images-idx3-ubyte.gz'
 ]

 paths = []
 for fname in files:
     paths.append(os.path.join(data_folder, fname))

 with gzip.open(paths[0], 'rb') as lbpath:
     y_train = np.frombuffer(lbpath.read(), np.uint8, offset=8)

 with gzip.open(paths[1], 'rb') as imgpath:
     x_train = np.frombuffer(
         imgpath.read(), np.uint8, offset=16).reshape(len(y_train), 28, 28)

 with gzip.open(paths[2], 'rb') as lbpath:
     y_test = np.frombuffer(lbpath.read(), np.uint8, offset=8)

 with gzip.open(paths[3], 'rb') as imgpath:
     x_test = np.frombuffer(
         imgpath.read(), np.uint8, offset=16).reshape(len(y_test), 28, 28)

 return (x_train, y_train), (x_test, y_test)
(train_images, train_labels), (test_images, test_labels) = load_data('./datasets/fashion/')

查看训练集的个数
```
print(train_images.shape)
```
展示训练集内的某个图像
```
plt.imshow(train_images[0])
plt.show()
```

构建神经网络

简单神经网络添加flatten层、全连接层
卷积神经网络添加卷积层、池化层、flatten层和全连接层
全连接层最后输出10，分别对应FashionMNIST的10个类别
构建完成后可以使用 model.summary() 预览神经网络的结构

构建简单的神经网络

model = keras.Sequential([
 keras.layers.Flatten(input_shape=(28, 28)),
 keras.layers.Dense(128, activation=tf.nn.relu),
 keras.layers.Dense(10, activation=tf.nn.softmax)
])

神经网络结构预览
(784+1)128=100480 (128+1)10=1290

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
flatten (Flatten)            (None, 784)               0         
_________________________________________________________________
dense (Dense)                (None, 128)               100480    
_________________________________________________________________
dense_1 (Dense)              (None, 10)                1290      
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________

构建卷积神经网络

model = keras.Sequential([
 keras.layers.Conv2D(64, (3, 3), activation='relu', input_shape=(28, 28, 1)),
 keras.layers.MaxPool2D((2, 2)),
 keras.layers.Conv2D(64, (3, 3), activation='relu'),
 keras.layers.MaxPool2D((2, 2)),
 keras.layers.Flatten(input_shape=(28, 28)),
 keras.layers.Dense(128, activation=tf.nn.relu),
 keras.layers.Dense(10, activation=tf.nn.softmax)
])

卷积神经网络结构预览

Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d (Conv2D)              (None, 26, 26, 64)        640       
_________________________________________________________________
max_pooling2d (MaxPooling2D) (None, 13, 13, 64)        0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 11, 11, 64)        36928     
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 5, 5, 64)          0         
_________________________________________________________________
flatten (Flatten)            (None, 1600)              0         
_________________________________________________________________
dense (Dense)                (None, 128)               204928    
_________________________________________________________________
dense_1 (Dense)              (None, 10)                1290      
=================================================================
Total params: 243,786
Trainable params: 243,786
Non-trainable params: 0
_________________________________________________________________

训练神经网络

训练模型,epoch设置训练次数,设置epoch为10，训练10组训练集
多分类问题loss函数使用交叉熵更为合适，优化策略使用Adam

train_images = train_images/255
model.compile(optimizer=tf.optimizers.Adam(),
              loss=tf.losses.sparse_categorical_crossentropy,
              metrics=['accuracy'])
model.fit(train_images.reshape(-1, 28, 28, 1), train_labels, epochs=10)

测试模型

使用训练集测试模型的准确度
```
test_images_scaled = test_images/255
model.evaluate(test_images_scaled.reshape(-1, 28, 28, 1), test_labels)
```
测试结果

#简单的神经网络
313/313 [==============================] - 0s 500us/step - loss: 0.3448 - accuracy: 0.8847
#卷积神经网络
313/313 [==============================] - 1s 3ms/step - loss: 0.2859 - accuracy: 0.9111
使用模型预测测试集的标签
预测测试集第一个图像的标签
```
print(np.argmax(model.predict((test_images[0]/255).reshape(-1, 28, 28, 1))))
```
测试结果：9

模型训练结果对比

简单的神经网络和卷积神经网络，损失函数loss和准确度对比

Epoch	简单的神经网络(loss/accuracy)	卷积神经网络(loss/accuracy)
1	0.4960/0.8269	0.4513/0.8354
2	0.3750/0.8658	0.3036/0.8893
3	0.3376/0.8771	0.2563/0.9050
4	0.3154/0.8840	0.2221/0.9174
5	0.2960/0.8920	0.1951/0.9274
6	0.2802/0.8958	0.1745/0.9354
7	0.2695/0.9002	0.1540/0.9430
8	0.2602/0.9028	0.1367/0.9492
9	0.2494/0.9062	0.1193/0.9543
10	0.2406/0.9097	0.1077/0.9593

由表格可以得出结论：

两个神经网络的训练结果都不错，loss函数的数值都在不断降低
卷积神经网络比简单的神经网络效果更好，证明一定情况下神经网络的宽度和深度更大的神经网络效果更好，但是所需的性能要求更高，训练的时间更长。

使用神经网络工具TensorFlow训练FashionMNIST Github完整代码

附录

代码使用的库文件

import numpy as np
import os
import gzip
import matplotlib.pyplot as plt
from tensorflow import keras
from tensorflow.keras import models
import tensorflow as tf