NVIDIA DLI 深度学习基础 答案 领取证书

最后一节作业是水果分类的任务，一共6类，使用之前学习的知识在代码段上进行填空。

加载ImageNet预训练的基础模型

from tensorflow import keras

base_model = keras.applications.VGG16(
    weights="imagenet",
    input_shape=(224, 224, 3),
    include_top=False)

冻结基础模型

# Freeze base model
base_model.trainable = False

向模型添加新层

# Create inputs with correct shape
inputs = keras.Input(shape=(224, 224, 3))

x = base_model(inputs, training=False)

# Add pooling layer or flatten layer
x = keras.layers.GlobalAveragePooling2D()(x)

# Add final dense layer
outputs = keras.layers.Dense(6, activation = 'softmax')(x)

# Combine inputs and outputs to create model
model = keras.Model(inputs, outputs)

model.summary()

编译模型

model.compile(loss='categorical_crossentropy', optimizer='adam',  metrics=['accuracy'])

扩充数据

from tensorflow.keras.preprocessing.image import ImageDataGenerator

datagen_train = ImageDataGenerator(featurewise_center=True,  # set input mean to 0 over the dataset
        samplewise_center=True,  # set each sample mean to 0
        rotation_range=10,  # randomly rotate images in the range (degrees, 0 to 180)
        zoom_range = 0.1, # Randomly zoom image 
        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)
        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)
        horizontal_flip=True,  # randomly flip images
        vertical_flip=False)
datagen_valid = ImageDataGenerator(featurewise_center=True,  # set input mean to 0 over the dataset
        samplewise_center=True,  # set each sample mean to 0
        rotation_range=10,  # randomly rotate images in the range (degrees, 0 to 180)
        zoom_range = 0.1, # Randomly zoom image 
        width_shift_range=0.1,  # randomly shift images horizontally (fraction of total width)
        height_shift_range=0.1,  # randomly shift images vertically (fraction of total height)
        horizontal_flip=True,  # randomly flip images
        vertical_flip=False)

加载数据集

# load and iterate training dataset
train_it = datagen_train.flow_from_directory(
    "data/fruits/train", 
     target_size=(224, 224), 
    color_mode="rgb",
    class_mode="categorical",
)

# load and iterate validation dataset
valid_it = datagen_valid.flow_from_directory(
    "data/fruits/valid",                                     
    target_size=(224, 224), 
    color_mode="rgb",
    class_mode="categorical",
)

训练模型
现在开始训练模型！将训练和测试数据集传递给fit函数，并设置所需的训练次数（epochs）

model.fit(train_it,
          validation_data=valid_it,
          steps_per_epoch=train_it.samples/train_it.batch_size,
          validation_steps=valid_it.samples/valid_it.batch_size,
          epochs=10)

其实到这里已经满足了评估需求，达到了92%以上的准确率
所以可以不进行微调的部分，直接运行后边的代码

就可以生成证书了