tensorflow中tf.keras模块的实现

from tensorflow.keras import Sequential
from tensorflow.keras.layers import Dense, Flatten, Conv2D

model = Sequential([
    Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),
    Flatten(),
    Dense(128, activation='relu'),
    Dense(10, activation='softmax')
])

from tensorflow.keras import Model, Input
from tensorflow.keras.layers import Dense, Concatenate

inputs = Input(shape=(784,))
x = Dense(64, activation='relu')(inputs)
x = Dense(32, activation='relu')(x)
outputs = Dense(10, activation='softmax')(x)
model = Model(inputs=inputs, outputs=outputs)

from tensorflow.keras import Model
from tensorflow.keras.layers import Dense

class MyModel(Model):
    def __init__(self):
        super(MyModel, self).__init__()
        self.dense1 = Dense(64, activation='relu')
        self.dense2 = Dense(10, activation='softmax')
    
    def call(self, inputs):
        x = self.dense1(inputs)
        return self.dense2(x)

from tensorflow.keras import layers

# 常用层类型
# - Dense: 全连接层
# - Conv2D/Conv1D/Conv3D: 卷积层
# - LSTM/GRU/SimpleRNN: 循环层
# - Dropout: 丢弃层
# - BatchNormalization: 批量归一化
# - Embedding: 嵌入层
# - MaxPooling2D/AveragePooling2D: 池化层
# - LayerNormalization: 层归一化

from tensorflow.keras import losses

# 常用损失函数
# - BinaryCrossentropy: 二分类交叉熵
# - CategoricalCrossentropy: 多分类交叉熵
# - MeanSquaredError: 均方误差
# - MeanAbsoluteError: 平均绝对误差
# - Huber: Huber损失（回归问题）
# - SparseCategoricalCrossentropy: 稀疏多分类交叉熵

from tensorflow.keras import optimizers

# 常用优化器
# - SGD: 随机梯度下降（可带动量）
# - Adam: 自适应矩估计
# - RMSprop: 均方根传播
# - Adagrad: 自适应梯度
# - Nadam: Nesterov Adam

from tensorflow.keras import metrics

# 常用指标
# - Accuracy: 准确率
# - Precision: 精确率
# - Recall: 召回率
# - AUC: ROC曲线下面积
# - MeanSquaredError: 均方误差
# - MeanAbsoluteError: 平均绝对误差

from tensorflow.keras import callbacks

# 常用回调
# - ModelCheckpoint: 模型保存
# - EarlyStopping: 早停
# - TensorBoard: TensorBoard可视化
# - ReduceLROnPlateau: 动态调整学习率
# - CSVLogger: 训练日志记录

from tensorflow.keras.preprocessing import image, text, sequence

# 图像预处理
# - ImageDataGenerator: 图像增强（TF 2.x 风格）
# - load_img, img_to_array: 图像加载转换

# 文本预处理
# - Tokenizer: 文本分词
# - pad_sequences: 序列填充

from tensorflow.keras.applications import (
    VGG16, ResNet50, MobileNet,
    InceptionV3, EfficientNetB0
)

# 加载预训练模型
base_model = ResNet50(weights='imagenet', include_top=False)

from tensorflow.keras import utils

# 常用工具
# - to_categorical: 类别编码
# - plot_model: 模型结构可视化
# - normalize: 数据标准化

import tensorflow as tf
from tensorflow.keras import layers, models

# 1. 数据准备
(x_train, y_train), (x_test, y_test) = tf.keras.datasets.mnist.load_data()
x_train = x_train.reshape(-1, 28, 28, 1).astype('float32') / 255.0
x_test = x_test.reshape(-1, 28, 28, 1).astype('float32') / 255.0

# 2. 构建模型
model = models.Sequential([
    layers.Conv2D(32, (3, 3), activation='relu', input_shape=(28, 28, 1)),
    layers.MaxPooling2D((2, 2)),
    layers.Conv2D(64, (3, 3), activation='relu'),
    layers.MaxPooling2D((2, 2)),
    layers.Flatten(),
    layers.Dense(128, activation='relu'),
    layers.Dropout(0.5),
    layers.Dense(10, activation='softmax')
])

# 3. 编译模型
model.compile(
    optimizer='adam',
    loss='sparse_categorical_crossentropy',
    metrics=['accuracy']
)

# 4. 训练模型
history = model.fit(
    x_train, y_train,
    epochs=10,
    batch_size=32,
    validation_split=0.2,
    callbacks=[
        tf.keras.callbacks.EarlyStopping(patience=3),
        tf.keras.callbacks.ModelCheckpoint('best_model.h5')
    ]
)

# 5. 评估模型
test_loss, test_acc = model.evaluate(x_test, y_test)

# 6. 使用模型预测
predictions = model.predict(x_test[:5])

from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences

# 1. 文本预处理
tokenizer = Tokenizer(num_words=10000)
tokenizer.fit_on_texts(texts)
sequences = tokenizer.texts_to_sequences(texts)
padded_sequences = pad_sequences(sequences, maxlen=200)

# 2. 构建文本分类模型
model = models.Sequential([
    layers.Embedding(10000, 128, input_length=200),
    layers.Bidirectional(layers.LSTM(64, return_sequences=True)),
    layers.GlobalMaxPooling1D(),
    layers.Dense(64, activation='relu'),
    layers.Dense(1, activation='sigmoid')  # 二分类
])

class CustomLayer(layers.Layer):
    def __init__(self, units=32):
        super(CustomLayer, self).__init__()
        self.units = units
    
    def build(self, input_shape):
        self.w = self.add_weight(
            shape=(input_shape[-1], self.units),
            initializer='random_normal',
            trainable=True
        )
        self.b = self.add_weight(
            shape=(self.units,),
            initializer='zeros',
            trainable=True
        )
    
    def call(self, inputs):
        return tf.matmul(inputs, self.w) + self.b

def custom_loss(y_true, y_pred):
    mse = tf.keras.losses.mean_squared_error(y_true, y_pred)
    penalty = tf.reduce_mean(tf.square(y_pred))
    return mse + 0.01 * penalty

# 多输入
input1 = Input(shape=(64,))
input2 = Input(shape=(128,))

# 多输出
output1 = Dense(1, name='regression')(merged)
output2 = Dense(5, activation='softmax', name='classification')(merged)

model = Model(inputs=[input1, input2], outputs=[output1, output2])

# 保存整个模型
model.save('my_model.h5')

# 保存为SavedModel格式（用于TF Serving）
model.save('my_model', save_format='tf')

# 转换为TensorFlow Lite（移动端）
converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_model = converter.convert()