https://tf.wiki/

 https://github.com/snowkylin/tensorflow-handbook

https://blog.csdn.net/lzs781/article/details/104742043/

 

官网

https://tensorflow.google.cn/tutorials/images/classification

一、生成模型 , 为了增加训练的精确率，可以使 epochs 值变大

import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Conv2D, Flatten, Dropout, MaxPooling2D
import os
import matplotlib.pyplot as plt


#


# 1. 训练路径
PATH = r'C:\Users\wuhao\Desktop\cats_and_dogs_filtered\cats_and_dogs_filtered'
train_dir = os.path.join(PATH, 'train')
train_cats_dir = os.path.join(train_dir, 'cats')
train_dogs_dir = os.path.join(train_dir, 'dogs')


batch_size = 128
epochs = 5
IMG_HEIGHT = 150
IMG_WIDTH = 150

# 2.转化为生成器
train_image_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1./255)
train_data_gen = train_image_generator.flow_from_directory(batch_size=batch_size,
                                                           directory=train_dir,
                                                           shuffle=True,
                                                           target_size=(IMG_HEIGHT, IMG_WIDTH),
                                                           class_mode='binary')


sample_training_images, _ = next(train_data_gen)


# 3.展示图片（可有可无）
def plot_images(images_arr):
    fig, axes = plt.subplots(1, 5, figsize=(20, 20))
    axes = axes.flatten()
    for img, ax in zip(images_arr, axes):
        ax.imshow(img)
        ax.axis('off')
    plt.tight_layout()
    plt.show()


# 显示 5张 图片
plot_images(sample_training_images[:5])

# 4. 创建模型

model = Sequential([
    Conv2D(16, 3, padding='same', activation='relu', input_shape=(IMG_HEIGHT, IMG_WIDTH ,3)),
    MaxPooling2D(),
    Conv2D(32, 3, padding='same', activation='relu'),
    MaxPooling2D(),
    Conv2D(64, 3, padding='same', activation='relu'),
    MaxPooling2D(),
    Flatten(),
    Dense(512, activation='relu'),
    Dense(1)
])


# 5. 编译模型
model.compile(
    optimizer='adam',
    loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
    metrics=['accuracy']
)

model.summary()


# 6.训练模型
num_cats_tr = len(os.listdir(train_cats_dir))
num_dogs_tr = len(os.listdir(train_dogs_dir))
total_train = num_cats_tr + num_dogs_tr
history = model.fit_generator(
    train_data_gen,
    steps_per_epoch=total_train // batch_size,
    epochs=epochs,
)
# 7.训练结果可视化
acc = history.history['accuracy']
loss = history.history['loss']
epochs_range = range(epochs)
plt.figure(figsize=(8, 8))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

# 9. 保存训练模型
model.save('path_to_my_model.h5')

2、加载模型

import tensorflow as tf
import os

#

batch_size = 128
epochs = 5
IMG_HEIGHT = 150
IMG_WIDTH = 150
PATH = r'C:\Users\wuhao\Desktop\cats_and_dogs_filtered\cats_and_dogs_filtered'
validation_dir = os.path.join(PATH, 'validation')
# 1.加载模型
new_model = tf.keras.models.load_model('path_to_my_model.h5')

new_model.summary()

# 2.获取验证的生成器
validation_image_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1./255)
val_data_gen = validation_image_generator.flow_from_directory(batch_size=batch_size,
                                                              directory=validation_dir,
                                                              target_size=(IMG_HEIGHT, IMG_WIDTH),
                                                              class_mode='binary')
# 3.获取评估
res = new_model.evaluate(val_data_gen)
print(res)

 三、预测

import tensorflow as tf
import os
import numpy as np
from tensorflow.keras.preprocessing import image
import cv2 as cv

# 1.加载模型
new_model = tf.keras.models.load_model('path_to_my_model.h5')


# 3.评估
# res = new_model.evaluate(val_data_gen)
# print(res)

tmp_path = r"C:\Users\wuhao\Desktop\cats_and_dogs_filtered\cats_and_dogs_filtered\validation\cats\cat.2000.jpg"
tmp1_path = r"C:\Users\wuhao\Desktop\cats_and_dogs_filtered\cats_and_dogs_filtered\validation\dogs\dog.2001.jpg"
img = image.load_img(tmp1_path, target_size=(150, 150))

x = image.img_to_array(img)

abc = x.reshape((1, 150, 150, 3))

result = new_model.predict(abc)
print(result)

 补充上面

probability_model = tf.keras.Sequential([
    model,
    tf.keras.layers.Softmax()
])

predictions = probability_model.predict(test_images)
# 查看 index为 1 的图片
print(np.argmax(predictions[1]))

# 查看单张图片
img = test_images[1]
# 由(x, x) => (1, x, x)
img = (np.expand_dims(img, 0))

res = probability_model.predict(img)

print(np.argmax(res))


a = model.predict(img)
print(np.argmax(a))