⚽开发平台:jupyter lab
🎈运行环境:python3、tensorflow2.x
第2章 tensorflow和keras快速入门 | iris数据集(自定义全连接层)
1. iris数据集 - 程序部分
1.1 程序2.10(完整代码)
import tensorflow as tf
import numpy as np
##加载数据
from sklearn.datasets import load_iris
data = load_iris()
## 转换数据形式
iris_data = np.float32(data.data)
iris_target = data.target
iris_data.shape,iris_target.shape #((150, 4), (150,))
## 对目标函数进行热编码化,数目为3
set(iris_target) # {0, 1, 2}
iris_target = np.float32(tf.keras.utils.to_categorical(iris_target,num_classes = 3))
iris_target
## 训练集的划分
train_data = tf.data.Dataset.from_tensor_slices((iris_data,iris_target)).batch(128)
## 自定义的层--全连接层
class MyLayer(tf.keras.layers.Layer):
def __init__(self,output_dim):
self.output_dim = output_dim
super(MyLayer,self).__init__()
def build(self,input_shape):
self.weight = tf.Variable(tf.random.normal([input_shape[-1],self.output_dim]),name="dense_weight")
self.bias = tf.Variable(tf.random.normal([self.output_dim]),name="bias_weight")
super(MyLayer,self).build(input_shape)
def call(self,input_tensor):
out = tf.matmul(input_tensor,self.weight)+self.bias
out = tf.nn.relu(out)
out = tf.keras.layers.Dropout(0.1)(out)
return out
## 模型构建
input_xs = tf.keras.Input(shape = (4,),name = "input_xs") ## 输入层
out = tf.keras.layers.Dense(32,activation='relu',name = 'dense_1')(input_xs) ## 隐藏层
out = MyLayer(32)(out) ## 自定义层
out = MyLayer(48)(out) ## 自定义层
out = tf.keras.layers.Dense(64,activation = 'relu',name ='dense_2')(out) ## 隐藏层
logits = tf.keras.layers.Dense(3,activation="softmax",name = 'prediction')(out) ## 输出层
model = tf.keras.Model(inputs = input_xs,outputs = logits) ## 构建模型
opt = tf.optimizers.Adam(0.005) ## 设置优化器学习率
## 模型编译并进行训练
model.compile(optimizer=tf.optimizers.Adam(0.005),loss=tf.losses.categorical_crossentropy,metrics=['acc'])
hist = model.fit(train_data,epochs=300)
## 模型评估
loss,accuracy = model.evaluate(iris_data,iris_target)
print('test loss',loss)
print('accuracy',accuracy)
1.2 数据展示
import matplotlib.pyplot as plt
### 精确度变化
plt.plot(range(300),hist.history.get('acc'))
### 损失度变化
plt.plot(range(300),hist.history.get('loss'))
1.3 模型参数
### 模型参数获取
model.summary()
2.知识点
2.1 自定义的层–全连接层详解
## 自定义的层--全连接层
## 1.init是对参数做定义和声明;
## 2.build函数是对权重可变参数做声明
## 3.call中含有真正实现所定义类的作用
class MyLayer(tf.keras.layers.Layer): #显示继承自Layer层
def __init__(self,output_dim): #init中显示确定参数
self.output_dim = output_dim #参数载入进类中
super(MyLayer,self).__init__() #向父类注册
def build(self,input_shape): #自定义的参数一般要加上self
self.weight = tf.Variable(tf.random.normal([input_shape[-1],self.output_dim]),name="dense_weight")
self.bias = tf.Variable(tf.random.normal([self.output_dim]),name="bias_weight")
super(MyLayer,self).build(input_shape)
def call(self,input_tensor): #包含主要层的实现,这里声明Tensor
out = tf.matmul(input_tensor,self.weight) + self.bias # 计算
out = tf.nn.relu(out) #计算
out = tf.keras.layers.Dropout(0.1)(out) #计算
return out #返回结果