AI智能
改变未来

一小时学会TensorFlow2之自定义层


目录
  • 概述
  • Sequential
  • Model & Layer
  • 案例
  • 数据集介绍
  • 完整代码

概述

通过自定义网络, 我们可以自己创建网络并和现有的网络串联起来, 从而实现各种各样的网络结构.

Sequential

Sequential 是 Keras 的一个网络容器. 可以帮助我们将多层网络封装在一起.

通过 Sequential 我们可以把现有的层已经我们自己的层实现结合, 一次前向传播就可以实现数据从第一层到最后一层的计算.

格式:

tf.keras.Sequential(layers=None, name=None)

例子:

# 5层网络模型model = tf.keras.Sequential([tf.keras.layers.Dense(256, activation=tf.nn.relu),tf.keras.layers.Dense(128, activation=tf.nn.relu),tf.keras.layers.Dense(64, activation=tf.nn.relu),tf.keras.layers.Dense(32, activation=tf.nn.relu),tf.keras.layers.Dense(10)])

Model & Layer

通过 Model 和 Layer 的__init__call()我们可以自定义层和模型.

Model:

class My_Model(tf.keras.Model):  # 继承Modeldef __init__(self):\"\"\"初始化\"\"\"super(My_Model, self).__init__()self.fc1 = My_Dense(784, 256)  # 第一层self.fc2 = My_Dense(256, 128)  # 第二层self.fc3 = My_Dense(128, 64)  # 第三层self.fc4 = My_Dense(64, 32)  # 第四层self.fc5 = My_Dense(32, 10)  # 第五层def call(self, inputs, training=None):\"\"\"在Model被调用的时候执行:param inputs: 输入:param training: 默认为None:return: 返回输出\"\"\"x = self.fc1(inputs)x = tf.nn.relu(x)x = self.fc2(x)x = tf.nn.relu(x)x = self.fc3(x)x = tf.nn.relu(x)x = self.fc4(x)x = tf.nn.relu(x)x = self.fc5(x)return x

Layer:

class My_Dense(tf.keras.layers.Layer):  # 继承Layerdef __init__(self, input_dim, output_dim):\"\"\"初始化:param input_dim::param output_dim:\"\"\"super(My_Dense, self).__init__()# 添加变量self.kernel = self.add_variable(\"w\", [input_dim, output_dim])  # 权重self.bias = self.add_variable(\"b\", [output_dim])  # 偏置def call(self, inputs, training=None):\"\"\"在Layer被调用的时候执行, 计算结果:param inputs: 输入:param training: 默认为None:return: 返回计算结果\"\"\"# y = w * x + bout = inputs @ self.kernel + self.biasreturn out

案例

数据集介绍

CIFAR-10 是由 10 类不同的物品组成的 6 万张彩色图片的数据集. 其中 5 万张为训练集, 1 万张为测试集.

完整代码

import tensorflow as tfdef pre_process(x, y):# 转换xx = 2 * tf.cast(x, dtype=tf.float32) / 255 - 1  # 转换为-1~1的形式x = tf.reshape(x, [-1, 32 * 32 * 3])  # 把x铺平# 转换yy = tf.convert_to_tensor(y)  # 转换为0~1的形式y = tf.one_hot(y, depth=10)  # 转成one_hot编码# 返回x, yreturn x, ydef get_data():\"\"\"获取数据:return:\"\"\"# 获取数据(X_train, y_train), (X_test, y_test) = tf.keras.datasets.cifar10.load_data()# 调试输出维度print(X_train.shape)  # (50000, 32, 32, 3)print(y_train.shape)  # (50000, 1)# squeezey_train = tf.squeeze(y_train)  # (50000, 1) => (50000,)y_test = tf.squeeze(y_test)  # (10000, 1) => (10000,)# 分割训练集train_db = tf.data.Dataset.from_tensor_slices((X_train, y_train)).shuffle(10000, seed=0)train_db = train_db.batch(batch_size).map(pre_process).repeat(iteration_num)  # 迭代20次# 分割测试集test_db = tf.data.Dataset.from_tensor_slices((X_test, y_test)).shuffle(10000, seed=0)test_db = test_db.batch(batch_size).map(pre_process)return train_db, test_dbclass My_Dense(tf.keras.layers.Layer):  # 继承Layerdef __init__(self, input_dim, output_dim):\"\"\"初始化:param input_dim::param output_dim:\"\"\"super(My_Dense, self).__init__()# 添加变量self.kernel = self.add_weight(\"w\", [input_dim, output_dim])  # 权重self.bias = self.add_weight(\"b\", [output_dim])  # 偏置def call(self, inputs, training=None):\"\"\"在Layer被调用的时候执行, 计算结果:param inputs: 输入:param training: 默认为None:return: 返回计算结果\"\"\"# y = w * x + bout = inputs @ self.kernel + self.biasreturn outclass My_Model(tf.keras.Model):  # 继承Modeldef __init__(self):\"\"\"初始化\"\"\"super(My_Model, self).__init__()self.fc1 = My_Dense(32 * 32 * 3, 256)  # 第一层self.fc2 = My_Dense(256, 128)  # 第二层self.fc3 = My_Dense(128, 64)  # 第三层self.fc4 = My_Dense(64, 32)  # 第四层self.fc5 = My_Dense(32, 10)  # 第五层def call(self, inputs, training=None):\"\"\"在Model被调用的时候执行:param inputs: 输入:param training: 默认为None:return: 返回输出\"\"\"x = self.fc1(inputs)x = tf.nn.relu(x)x = self.fc2(x)x = tf.nn.relu(x)x = self.fc3(x)x = tf.nn.relu(x)x = self.fc4(x)x = tf.nn.relu(x)x = self.fc5(x)return x# 定义超参数batch_size = 256  # 一次训练的样本数目learning_rate = 0.001  # 学习率iteration_num = 20  # 迭代次数optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate)  # 优化器loss = tf.losses.CategoricalCrossentropy(from_logits=True)  # 损失network = My_Model()  # 实例化网络# 调试输出summarynetwork.build(input_shape=[None, 32 * 32 * 3])print(network.summary())# 组合network.compile(optimizer=optimizer,loss=loss,metrics=[\"accuracy\"])if __name__ == \"__main__\":# 获取分割的数据集train_db, test_db = get_data()# 拟合network.fit(train_db, epochs=5, validation_data=test_db, validation_freq=1)

输出结果:

Model: \”my__model\”
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
my__dense (My_Dense) multiple 786688
_________________________________________________________________
my__dense_1 (My_Dense) multiple 32896
_________________________________________________________________
my__dense_2 (My_Dense) multiple 8256
_________________________________________________________________
my__dense_3 (My_Dense) multiple 2080
_________________________________________________________________
my__dense_4 (My_Dense) multiple 330
=================================================================
Total params: 830,250
Trainable params: 830,250
Non-trainable params: 0
_________________________________________________________________
None
(50000, 32, 32, 3)
(50000, 1)
2021-06-15 14:35:26.600766: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:176] None of the MLIR Optimization Passes are enabled (registered 2)
Epoch 1/5
3920/3920 [==============================] – 39s 10ms/step – loss: 0.9676 – accuracy: 0.6595 – val_loss: 1.8961 – val_accuracy: 0.5220
Epoch 2/5
3920/3920 [==============================] – 41s 10ms/step – loss: 0.3338 – accuracy: 0.8831 – val_loss: 3.3207 – val_accuracy: 0.5141
Epoch 3/5
3920/3920 [==============================] – 41s 10ms/step – loss: 0.1713 – accuracy: 0.9410 – val_loss: 4.2247 – val_accuracy: 0.5122
Epoch 4/5
3920/3920 [==============================] – 41s 10ms/step – loss: 0.1237 – accuracy: 0.9581 – val_loss: 4.9458 – val_accuracy: 0.5050
Epoch 5/5
3920/3920 [==============================] – 42s 11ms/step – loss: 0.1003 – accuracy: 0.9666 – val_loss: 5.2425 – val_accuracy: 0.5097

到此这篇关于一小时学会TensorFlow2之自定义层的文章就介绍到这了,更多相关TensorFlow2自定义层内容请搜索脚本之家以前的文章或继续浏览下面的相关文章希望大家以后多多支持脚本之家!

您可能感兴趣的文章:

  • 解决Pytorch自定义层出现多Variable共享内存错误问题
  • 解决Keras 自定义层时遇到版本的问题
  • keras Lambda自定义层实现数据的切片方式,Lambda传参数
  • 使用Keras加载含有自定义层或函数的模型操作
赞(0) 打赏
未经允许不得转载:爱站程序员基地 » 一小时学会TensorFlow2之自定义层