tensorflow2.0入门-手写数字识别

tensorflow2.0入门-手写数字识别

代码实现：

import os

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import tensorflow as tf
from tensorflow import keras

# 载入数据
fashion_mnist = keras.datasets.fashion_mnist
(x_train_all, y_train_all), (x_test, y_test) = fashion_mnist.load_data()
x_valid, x_train = x_train_all[:5000], x_train_all[5000:]
y_valid, y_train = y_train_all[:5000], y_train_all[5000:]

# [9 0 0 ... 5 8 3]
# print(y_valid)

# 数据归一化
from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
# x_train: [None, 28, 28] -> [None,784]
x_train_scaled = scaler.fit_transform(x_train.astype(np.float32).reshape(-1, 1)).reshape(-1, 28, 28)
x_valid_scaled = scaler.transform(x_valid.astype(np.float32).reshape(-1, 1)).reshape(-1, 28, 28)
x_test_scaled = scaler.transform(x_test.astype(np.float32).reshape(-1, 1)).reshape(-1, 28, 28)

model = tf.keras.models.Sequential()
model.add(tf.keras.layers.Flatten(input_shape=[28, 28]))
model.add(tf.keras.layers.Dense(300, activation='relu'))
model.add(tf.keras.layers.Dense(100, activation='relu'))
model.add(tf.keras.layers.Dense(10, activation='softmax'))

model.compile(loss='sparse_categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])
# 使用了三个callback：Tensorboard, earlystopping, ModelCheckpoint
# logdir = './callbacks'
logdir = os.path.join("callbacks")
if not os.path.exists(logdir):
    os.mkdir(logdir)
output_model_file = os.path.join(logdir, "fashion_mnist_model.h5")
print("out:", output_model_file)
callbacks = [
    tf.keras.callbacks.TensorBoard(logdir),
    tf.keras.callbacks.ModelCheckpoint(output_model_file, save_best_only=True),  # 默认值：monitor='val_loss'
    tf.keras.callbacks.EarlyStopping(min_delta=1e-3, patience=5)  # 默认值：monitor='val_loss',min_delta=0,patience=0
]
history = model.fit(x_train_scaled, y_train, epochs=10, validation_data=(x_valid_scaled, y_valid), callbacks=callbacks)


def plot_learning_curves(history):
    pd.DataFrame(history.history).plot(figsize=(8, 5))
    plt.grid(True)
    plt.gca().set_ylim(0, 1)
    plt.show()


plot_learning_curves(history)