TensorFlow 2.0 快速入门（专业人员） -- MNIST数据集

下载并安装 TensorFlow 2.0 软件包，将 Tensorflow 导入您的程序：

from __future__ import absolute_import, division, print_function, unicode_literals

import tensorflow as tf

from tensorflow.keras.layers import Dense, Flatten, Conv2D
from tensorflow.keras import Model

加载并准备 MNIST 数据集

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

# Add a channels dimension
x_train = x_train[..., tf.newaxis]
x_test = x_test[..., tf.newaxis]

使用 tf.data 来将数据集切分为 batch 以及混淆数据集：

train_ds = tf.data.Dataset.from_tensor_slices(
    (x_train, y_train)).shuffle(10000).batch(32)
test_ds = tf.data.Dataset.from_tensor_slices((x_test, y_test)).batch(32)

使用 Keras 模型子类化（model subclassing） API 构建 tf.keras 模型：

class MyModel(Model):
  def __init__(self):
    super(MyModel, self).__init__()
    self.conv1 = Conv2D(32, 3, activation='relu')
    self.flatten = Flatten()
    self.d1 = Dense(128, activation='relu')
    self.d2 = Dense(10, activation='softmax')

  def call(self, x):
    x = self.conv1(x)
    x = self.flatten(x)
    x = self.d1(x)
    return self.d2(x)

model = MyModel()

为训练选择优化器与损失函数：

loss_object = tf.keras.losses.SparseCategoricalCrossentropy()

optimizer = tf.keras.optimizers.Adam()

选择衡量指标来度量模型的损失值（loss）和准确率（accuracy）。这些指标在 epoch 上累积值，然后打印出整体结果。

train_loss = tf.keras.metrics.Mean(name='train_loss')
train_accuracy = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')

test_loss = tf.keras.metrics.Mean(name='test_loss')
test_accuracy = tf.keras.metrics.SparseCategoricalAccuracy(name='test_accuracy')

使用 tf.GradientTape 来训练模型：

@tf.function
def train_step(images, labels):
  with tf.GradientTape() as tape:
    predictions = model(images)
    loss = loss_object(labels, predictions)
  gradients = tape.gradient(loss, model.trainable_variables)
  optimizer.apply_gradients(zip(gradients, model.trainable_variables))

  train_loss(loss)
  train_accuracy(labels, predictions)

测试模型：

@tf.function
def test_step(images, labels):
  predictions = model(images)
  t_loss = loss_object(labels, predictions)

  test_loss(t_loss)
  test_accuracy(labels, predictions)

EPOCHS = 5

for epoch in range(EPOCHS):
  # 在下一个epoch开始时，重置评估指标
  train_loss.reset_states()
  train_accuracy.reset_states()
  test_loss.reset_states()
  test_accuracy.reset_states()

  for images, labels in train_ds:
    train_step(images, labels)

  for test_images, test_labels in test_ds:
    test_step(test_images, test_labels)

  template = 'Epoch {}, Loss: {}, Accuracy: {}, Test Loss: {}, Test Accuracy: {}'
  print (template.format(epoch+1,
                         train_loss.result(),
                         train_accuracy.result()*100,
                         test_loss.result(),
                         test_accuracy.result()*100))

WARNING:tensorflow:Layer my_model is casting an input tensor from dtype float64 to the layer's dtype of float32, which is new behavior in TensorFlow 2.  The layer has dtype float32 because it's dtype defaults to floatx.

If you intended to run this layer in float32, you can safely ignore this warning. If in doubt, this warning is likely only an issue if you are porting a TensorFlow 1.X model to TensorFlow 2.

To change all layers to have dtype float64 by default, call `tf.keras.backend.set_floatx('float64')`. To change just this layer, pass dtype='float64' to the layer constructor. If you are the author of this layer, you can disable autocasting by passing autocast=False to the base Layer constructor.

Epoch 1, Loss: 0.13130314648151398, Accuracy: 96.03833770751953, Test Loss: 0.06053972616791725, Test Accuracy: 97.91999816894531
Epoch 2, Loss: 0.042836885899305344, Accuracy: 98.63333129882812, Test Loss: 0.05354950577020645, Test Accuracy: 98.23999786376953
Epoch 3, Loss: 0.023272410035133362, Accuracy: 99.25, Test Loss: 0.0571180060505867, Test Accuracy: 98.29000091552734
Epoch 4, Loss: 0.013985390774905682, Accuracy: 99.51000213623047, Test Loss: 0.061239469796419144, Test Accuracy: 98.3499984741211
Epoch 5, Loss: 0.008612685836851597, Accuracy: 99.70166778564453, Test Loss: 0.060723815113306046, Test Accuracy: 98.44999694824219

该图片分类器现在在此数据集上训练得到了接近 98% 的准确率（accuracy）。