tensorflow的简单使用

参考文章：

https://www.tensorflow.org/tutorials/keras/classification

https://www.datacamp.com/tutorial/tensorflow-tutorial

一、引入模块

import tensorflow as tf
import tensorflow.keras as keras
import numpy as np
import matplotlib.pyplot as plt
import skimage
from skimage import transform
from skimage.color import rgb2gray

二、加载数据集

        1、使用minist数据集

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

        2、自定义数据集

def load_data(data_directory):
    directories = [d for d in os.listdir(data_directory)
                   if os.path.isdir(os.path.join(data_directory, d))]
    labels = []
    images = []
    all_image_paths = []
    for d in directories:
        label_directory = os.path.join(data_directory, d)
        file_names = [os.path.join(label_directory, f)
                      for f in os.listdir(label_directory)
                      # if f.endswith(".ppm")
                      ]
        for f in file_names:
            images.append(skimage.io.imread(f))
            labels.append(int(d))
            all_image_paths.append(f)
    return images, labels
 
ROOT_PATH = "E:/dataset/"
train_data_directory = os.path.join(ROOT_PATH, "test/training")
test_data_directory = os.path.join(ROOT_PATH, "test/testing")
 
train_images, train_labels = load_data(train_data_directory)
test_images, test_labels = load_data(test_data_directory)
 
# 训练数据
train_images = [transform.resize(image, (28, 28)) for image in train_images]
train_images = np.array(train_images)
train_images = rgb2gray(train_images)
train_labels = np.array(train_labels)
# 测试数据
test_images = [transform.resize(image, (28, 28)) for image in test_images]
test_images = np.array(test_images)
test_images = rgb2gray(test_images)
test_labels = np.array(test_labels)

三、构建模型

# 创建模型
model = keras.Sequential([
        keras.layers.Flatten(input_shape=(28, 28)),
        keras.layers.Dense(128, activation='relu'),  # 激活函数 relu
        keras.layers.Dense(128)
    ])
# 优化器
model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])
# 训练模型
model.fit(train_images, train_labels, epochs=100)  # epochs迭代次数
# 评估准确率
test_loss, test_acc = model.evaluate(test_images, test_labels, verbose=2)
 
print('\nTest test_loss:{0} accuracy:{1}'.format(test_loss, test_acc))

四、进行预测

probability_model = tf.keras.Sequential([model, tf.keras.layers.Softmax()])
predictions = probability_model.predict(test_images)
pre_value = np.argmax(predictions[2])  # 预测标签
truth_value = test_labels[2]  # 实际标签
 
print('\nTest pre_value:{0} truth_value:{1}'.format(pre_value, truth_value))
 
# 可视化 流程图
tf.keras.utils.plot_model(model, 'model1.png', show_shapes=True, show_dtype=True,show_layer_names=True)
 
# 查看图片
fig = plt.figure(figsize=(10, 10))
for i in range(10):
    truth = test_labels[i]
    prediction = np.argmax(predictions[i])
    plt.subplot(3, 4, 1 + i)  # 五行 两列
    plt.axis('off')  # 坐标轴 off关闭坐标轴
    color = 'green' if truth == prediction else 'red'
    plt.text(40, 10, "Truth:        {0}\nPrediction: {1}".format(truth, prediction),
             fontsize=12, color=color)
    plt.imshow(test_images[i], cmap="gray")  # gray显示灰度图 默认显示热图
# 显示图片
plt.show()