电脑环境：
语言环境：Python 3.8.0
编译器：Jupyter Notebook
深度学习环境：tensorflow 2.17.0
一、前期工作
1、ResNet-50总体结构
2、设置GPU
python
import tensorflow as tf
gpus = tf.config.list_physical_devices('GPU')
if gpus:
tf.config.experimental.set_memory_growth(gpus[0], True)
tf.config.set_visible_devices(gpus[0], 'GPU')3、导入数据
python
import matplotlib.pyplot as plt
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
import os, PIL, pathlib
import numpy as np
from tensorflow import keras
from keras import layers, models
data_dir = './bird_photos'
data_dir = pathlib.Path(data_dir)
image_count = len(list(data_dir.glob('*/*')))
image_count565
二、数据预处理
1、加载数据
python
batch_size = 8
img_height = 224
img_width = 224
train_ds = tf.keras.preprocessing.image_dataset_from_directory(
data_dir,
validation_split=0.2,
subset='training',
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)
val_ds = tf.keras.preprocessing.image_dataset_from_directory(
data_dir,
validation_split=0.2,
subset='validation',
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)我们可以通过class_names输出数据集的标签，按字母顺序对应于目录名称。
python
class_names = train_ds.class_names
class_names['Bananaquit', 'Black Skimmer', 'Black Throated Bushtiti', 'Cockatoo']
2、可视化数据
python
plt.figure(figsize=(10, 4))
for images, labels in train_ds.take(1):
for i in range(8):
ax = plt.subplot(2, 4, i+1)
plt.imshow(images[i].numpy().astype('uint8'))
plt.title(class_names[labels[i]])
plt.axis('off')
python
plt.imshow(images[0].numpy().astype('uint8'))
3、再次检查数据
python
for image_batch, labels_batch in train_ds:
print(image_batch.shape)
print(labels_batch.shape)
break(8, 224, 224, 3)
(8,)
4、配置数据集
python
AUTOTUNE = tf.data.experimental.AUTOTUNE
train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)三、构建ResNet-50模型
python
from keras import layers
from keras.layers import Input, Activation, BatchNormalization, Flatten
from keras.layers import Dense, Conv2D, MaxPooling2D, ZeroPadding2D, AveragePooling2D
from keras.models import Model
def identity_block(input_tensor, kernel_size, filters, stage, block):
filters1, filters2, filters3 = filters
name_base = str(stage) + block + '_identity_block_'
x = Conv2D(filters1, (1, 1), name=name_base + 'conv1')(input_tensor)
x = BatchNormalization(name=name_base + 'bn1')(x)
x = Activation('relu', name=name_base + 'relu1')(x)
x = Conv2D(filters2, kernel_size, padding='same', name=name_base + 'conv2')(x)
x = BatchNormalization(name=name_base + 'bn2')(x)
x = Activation('relu', name=name_base + 'relu2')(x)
x = Conv2D(filters3, (1, 1), name=name_base + 'conv3')(x)
x = BatchNormalization(name=name_base + 'bn3')(x)
x = layers.add([x, input_tensor], name=name_base + 'add')
x = Activation('relu', name=name_base + 'relu3')(x)
return x
def conv_block(input_tensor, kernel_size, filters, stage, block, strides=(2, 2)):
filters1, filters2, filters3 = filters
res_name_base = str(stage) + block + '_conv_block_res_'
name_base = str(stage) + block + '_conv_block_'
x = Conv2D(filters1, (1, 1), strides=strides, name=name_base + 'conv1')(input_tensor)
x = BatchNormalization(name=name_base + 'bn1')(x)
x = Activation('relu', name=name_base + 'relu1')(x)
x = Conv2D(filters2, kernel_size, padding='same', name=name_base + 'conv2')(x)
x = BatchNormalization(name=name_base + 'bn2')(x)
x = Activation('relu', name=name_base + 'relu2')(x)
x = Conv2D(filters3, (1, 1), name=name_base + 'conv3')(x)
x = BatchNormalization(name=name_base + 'bn3')(x)
shortcut = Conv2D(filters3, (1, 1), strides=strides, name=res_name_base + 'conv')(input_tensor)
shortcut = BatchNormalization(name=res_name_base + 'bn')(shortcut)
x = layers.add([x, shortcut], name=name_base + 'add')
x = Activation('relu', name=name_base + 'relu3')(x)
return x
def ResNet50(input_shape=(224, 224, 3), classes=1000):
img_input = Input(shape=input_shape)
x = ZeroPadding2D((3, 3))(img_input)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1')(x)
x = BatchNormalization(name='bn_conv1')(x)
x = Activation('relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2))(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='c')
x = identity_block(x, 3, [128, 128, 512], stage=3, block='d')
x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='c')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='d')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='e')
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='f')
x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
x = AveragePooling2D((7, 7), name='avg_pool')(x)
x = Flatten()(x)
x = Dense(classes, activation='softmax', name='fc1000')(x)
model = Model(img_input, x, name='resnet50')
# 加载预训练模型
model.load_weights('./resnet50_weights_tf_dim_ordering_tf_kernels.h5')
return model
model = ResNet50()
model.summary()
python
Model: "resnet50"
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Layer (type) ┃ Output Shape ┃ Param # ┃ Connected to ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━┩
│ input_layer (InputLayer) │ (None, 224, 224, 3) │ 0 │ - │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ zero_padding2d │ (None, 230, 230, 3) │ 0 │ input_layer[0][0] │
│ (ZeroPadding2D) │ │ │ │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ conv1 (Conv2D) │ (None, 112, 112, 64) │ 9,472 │ zero_padding2d[0][0] │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ bn_conv1 │ (None, 112, 112, 64) │ 256 │ conv1[0][0] │
│ (BatchNormalization) │ │ │ │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ activation (Activation) │ (None, 112, 112, 64) │ 0 │ bn_conv1[0][0] │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ max_pooling2d │ (None, 55, 55, 64) │ 0 │ activation[0][0] │
│ (MaxPooling2D) │ │ │ │
..............................................................
..............................................................
..............................................................
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ avg_pool │ (None, 1, 1, 2048) │ 0 │ 5c_identity_block_rel... │
│ (AveragePooling2D) │ │ │ │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ flatten (Flatten) │ (None, 2048) │ 0 │ avg_pool[0][0] │
├───────────────────────────┼────────────────────────┼────────────────┼────────────────────────┤
│ fc1000 (Dense) │ (None, 1000) │ 2,049,000 │ flatten[0][0] │
└───────────────────────────┴────────────────────────┴────────────────┴────────────────────────┘
Total params: 25,636,712 (97.80 MB)
Trainable params: 25,583,592 (97.59 MB)
Non-trainable params: 53,120 (207.50 KB)四、编译
python
opt = tf.keras.optimizers.Adam(learning_rate=1e-4)
model.compile(optimizer=opt,
loss='sparse_categorical_crossentropy',
metrics=['accuracy'])五、训练模型
python
epochs = 10
history = model.fit(
train_ds,
validation_data=val_ds,
epochs=epochs
)
python
Epoch 1/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 269s 1s/step - accuracy: 0.5021 - loss: 3.6748 - val_accuracy: 0.9646 - val_loss: 0.1640
Epoch 2/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 6s 99ms/step - accuracy: 0.9636 - loss: 0.2068 - val_accuracy: 0.9823 - val_loss: 0.0241
Epoch 3/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 6s 97ms/step - accuracy: 0.9800 - loss: 0.0443 - val_accuracy: 0.9912 - val_loss: 0.0115
Epoch 4/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 10s 98ms/step - accuracy: 0.9943 - loss: 0.0286 - val_accuracy: 0.9912 - val_loss: 0.0183
Epoch 5/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 11s 104ms/step - accuracy: 0.9945 - loss: 0.0377 - val_accuracy: 1.0000 - val_loss: 0.0108
Epoch 6/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 10s 100ms/step - accuracy: 0.9995 - loss: 0.0038 - val_accuracy: 0.9735 - val_loss: 0.0359
Epoch 7/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 10s 104ms/step - accuracy: 1.0000 - loss: 0.0024 - val_accuracy: 0.9912 - val_loss: 0.0196
Epoch 8/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 10s 98ms/step - accuracy: 1.0000 - loss: 6.2409e-04 - val_accuracy: 0.9912 - val_loss: 0.0139
Epoch 9/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 6s 106ms/step - accuracy: 1.0000 - loss: 5.9430e-04 - val_accuracy: 1.0000 - val_loss: 0.0103
Epoch 10/10
57/57 ━━━━━━━━━━━━━━━━━━━━ 6s 99ms/step - accuracy: 1.0000 - loss: 3.5871e-04 - val_accuracy: 1.0000 - val_loss: 0.0094六、模型评估
python
acc = history.history['accuracy']
val_acc = history.history['val_accuracy']
loss = history.history['loss']
val_loss = history.history['val_loss']
epochs_range = range(epochs)
plt.figure(figsize=(12, 4))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation 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.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()
七、预测
python
plt.figure(figsize=(10, 4))
for images, labels in train_ds.take(1):
for i in range(8):
ax = plt.subplot(2, 4, i+1)
plt.imshow(images[i].numpy().astype('uint8'))
img_array = tf.expand_dims(images[i], 0)
predictions = model.predict(img_array)
plt.title(class_names[np.argmax(predictions)])
plt.axis('off')
八、总结
在一般的卷积神经网络中，由于深度的增加，可能会带来梯度爆炸，梯度消失，ResNet的残差网络结构可以有效解决这些问题。