tensorflow--Mnist数据集

Mnist数据集一共有7万张图片:

  • 6万张28*28像素点的0~9手写数字图片和标签,用于训练
  • 1万张28*28像素点的0~9手写数字图片和标签,用于测试
import tensorflow as tf
import matplotlib.pyplot as plt

# 导入数据
mnist = tf.keras.datasets.mnist
(x_train, y_train), (x_test, y_test) = mnist.load_data()

# 看看数据样式,有个大概认识
print(x_train.shape, y_train.shape, x_test.shape, y_test.shape)
print(x_train[0].shape)
print(x_train[0])

# 可视化数据(随便看一个)
plt.imshow(x_train[24], cmap="gray")
plt.show()

运行结果:

(60000, 28, 28) (60000,) (10000, 28, 28) (10000,)
(28, 28)
[[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   3  18  18  18 126 136
  175  26 166 255 247 127   0   0   0   0]
 [  0   0   0   0   0   0   0   0  30  36  94 154 170 253 253 253 253 253
  225 172 253 242 195  64   0   0   0   0]
 [  0   0   0   0   0   0   0  49 238 253 253 253 253 253 253 253 253 251
   93  82  82  56  39   0   0   0   0   0]
 [  0   0   0   0   0   0   0  18 219 253 253 253 253 253 198 182 247 241
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0  80 156 107 253 253 205  11   0  43 154
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0  14   1 154 253  90   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0 139 253 190   2   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0  11 190 253  70   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0  35 241 225 160 108   1
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0  81 240 253 253 119
   25   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0  45 186 253 253
  150  27   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0  16  93 252
  253 187   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0 249
  253 249  64   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0  46 130 183 253
  253 207   2   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0  39 148 229 253 253 253
  250 182   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0  24 114 221 253 253 253 253 201
   78   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0  23  66 213 253 253 253 253 198  81   2
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0  18 171 219 253 253 253 253 195  80   9   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0  55 172 226 253 253 253 253 244 133  11   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0 136 253 253 253 212 135 132  16   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]]

图片如下:


1.png

Sequential实现手写数字识别

import tensorflow as tf


mnist = tf.keras.datasets.mnist
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0  # 输入特征归一化,把0·255的数值变为0·1之间,更适合神经网络吸收

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(),  # 拉直为一维数组
    tf.keras.layers.Dense(128, activation="relu"),  # 第一层128个神经元,relu激活函数
    tf.keras.layers.Dense(10, activation="softmax")  # 第二层10个神经元,softmax使输出符合概率分布
])

# 训练方法:adam优化器,损失函数,评测指标
model.compile(optimizer="adam",
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics=["sparse_categorical_accuracy"])

model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test),
          validation_freq=1)
model.summary()

运行结果:

Epoch 5/5
   1/1875 [..............................] - ETA: 0s - loss: 0.1622 - sparse_categorical_accuracy: 0.9375
  50/1875 [..............................] - ETA: 2s - loss: 0.0566 - sparse_categorical_accuracy: 0.9812
  98/1875 [>.............................] - ETA: 2s - loss: 0.0531 - sparse_categorical_accuracy: 0.9831
 147/1875 [=>............................] - ETA: 1s - loss: 0.0478 - sparse_categorical_accuracy: 0.9853
 196/1875 [==>...........................] - ETA: 1s - loss: 0.0475 - sparse_categorical_accuracy: 0.9850
 235/1875 [==>...........................] - ETA: 1s - loss: 0.0444 - sparse_categorical_accuracy: 0.9859
 276/1875 [===>..........................] - ETA: 1s - loss: 0.0439 - sparse_categorical_accuracy: 0.9868
 321/1875 [====>.........................] - ETA: 1s - loss: 0.0422 - sparse_categorical_accuracy: 0.9876
 366/1875 [====>.........................] - ETA: 1s - loss: 0.0420 - sparse_categorical_accuracy: 0.9880
 407/1875 [=====>........................] - ETA: 1s - loss: 0.0425 - sparse_categorical_accuracy: 0.9878
 452/1875 [======>.......................] - ETA: 1s - loss: 0.0430 - sparse_categorical_accuracy: 0.9876
 502/1875 [=======>......................] - ETA: 1s - loss: 0.0425 - sparse_categorical_accuracy: 0.9877
 552/1875 [=======>......................] - ETA: 1s - loss: 0.0444 - sparse_categorical_accuracy: 0.9870
 596/1875 [========>.....................] - ETA: 1s - loss: 0.0450 - sparse_categorical_accuracy: 0.9869
 640/1875 [=========>....................] - ETA: 1s - loss: 0.0453 - sparse_categorical_accuracy: 0.9867
 689/1875 [==========>...................] - ETA: 1s - loss: 0.0453 - sparse_categorical_accuracy: 0.9868
 739/1875 [==========>...................] - ETA: 1s - loss: 0.0455 - sparse_categorical_accuracy: 0.9867
 782/1875 [===========>..................] - ETA: 1s - loss: 0.0458 - sparse_categorical_accuracy: 0.9867
 827/1875 [============>.................] - ETA: 1s - loss: 0.0452 - sparse_categorical_accuracy: 0.9870
 877/1875 [=============>................] - ETA: 1s - loss: 0.0455 - sparse_categorical_accuracy: 0.9868
 926/1875 [=============>................] - ETA: 1s - loss: 0.0458 - sparse_categorical_accuracy: 0.9867
 968/1875 [==============>...............] - ETA: 1s - loss: 0.0457 - sparse_categorical_accuracy: 0.9867
1003/1875 [===============>..............] - ETA: 1s - loss: 0.0458 - sparse_categorical_accuracy: 0.9866
1036/1875 [===============>..............] - ETA: 1s - loss: 0.0455 - sparse_categorical_accuracy: 0.9867
1085/1875 [================>.............] - ETA: 0s - loss: 0.0455 - sparse_categorical_accuracy: 0.9867
1130/1875 [=================>............] - ETA: 0s - loss: 0.0456 - sparse_categorical_accuracy: 0.9866
1175/1875 [=================>............] - ETA: 0s - loss: 0.0457 - sparse_categorical_accuracy: 0.9867
1224/1875 [==================>...........] - ETA: 0s - loss: 0.0460 - sparse_categorical_accuracy: 0.9865
1273/1875 [===================>..........] - ETA: 0s - loss: 0.0458 - sparse_categorical_accuracy: 0.9866
1319/1875 [====================>.........] - ETA: 0s - loss: 0.0454 - sparse_categorical_accuracy: 0.9866
1367/1875 [====================>.........] - ETA: 0s - loss: 0.0459 - sparse_categorical_accuracy: 0.9865
1415/1875 [=====================>........] - ETA: 0s - loss: 0.0458 - sparse_categorical_accuracy: 0.9864
1463/1875 [======================>.......] - ETA: 0s - loss: 0.0462 - sparse_categorical_accuracy: 0.9863
1508/1875 [=======================>......] - ETA: 0s - loss: 0.0465 - sparse_categorical_accuracy: 0.9862
1555/1875 [=======================>......] - ETA: 0s - loss: 0.0469 - sparse_categorical_accuracy: 0.9861
1603/1875 [========================>.....] - ETA: 0s - loss: 0.0474 - sparse_categorical_accuracy: 0.9859
1650/1875 [=========================>....] - ETA: 0s - loss: 0.0473 - sparse_categorical_accuracy: 0.9859
1696/1875 [==========================>...] - ETA: 0s - loss: 0.0468 - sparse_categorical_accuracy: 0.9861
1744/1875 [==========================>...] - ETA: 0s - loss: 0.0471 - sparse_categorical_accuracy: 0.9859
1795/1875 [===========================>..] - ETA: 0s - loss: 0.0471 - sparse_categorical_accuracy: 0.9858
1844/1875 [============================>.] - ETA: 0s - loss: 0.0468 - sparse_categorical_accuracy: 0.9858
1875/1875 [==============================] - 3s 1ms/step - loss: 0.0467 - sparse_categorical_accuracy: 0.9859 - val_loss: 0.0869 - val_sparse_categorical_accuracy: 0.9750
Model: "sequential"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
flatten (Flatten)            multiple                  0         
_________________________________________________________________
dense (Dense)                multiple                  100480    
_________________________________________________________________
dense_1 (Dense)              multiple                  1290      
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________
©著作权归作者所有,转载或内容合作请联系作者
平台声明:文章内容(如有图片或视频亦包括在内)由作者上传并发布,文章内容仅代表作者本人观点,简书系信息发布平台,仅提供信息存储服务。