Tensorflow2（预课程）---7.6、cifar10分类-层方式-卷积神经网络-Inception10
Tensorflow2（预课程）---7.6、cifar10分类-层方式-卷积神经网络-Inception10
一、总结
一句话总结：
InceptionNet：一层内使用不同尺寸卷积核，提升感知力使用批标准化，缓解梯度消失
InceptionNet：1、1*1卷积；2、3*3卷积+1*1卷积；3、5*5卷积+1*1卷积；4、1*1卷积+3*3卷积；
class ConvBNRelu(Model): def __init__(self, ch, kernelsz=3, strides=1, padding='same'): super(ConvBNRelu, self).__init__() self.model = tf.keras.models.Sequential([ Conv2D(ch, kernelsz, strides=strides, padding=padding), BatchNormalization(), Activation('relu') ]) def call(self, x): x = self.model(x, training=False) #在training=False时，BN通过整个训练集计算均值、方差去做批归一化，training=True时，通过当前batch的均值、方差去做批归一化。推理时 training=False效果好 return x class InceptionBlk(Model): def __init__(self, ch, strides=1): super(InceptionBlk, self).__init__() self.ch = ch self.strides = strides self.c1 = ConvBNRelu(ch, kernelsz=1, strides=strides) self.c2_1 = ConvBNRelu(ch, kernelsz=1, strides=strides) self.c2_2 = ConvBNRelu(ch, kernelsz=3, strides=1) self.c3_1 = ConvBNRelu(ch, kernelsz=1, strides=strides) self.c3_2 = ConvBNRelu(ch, kernelsz=5, strides=1) self.p4_1 = MaxPool2D(3, strides=1, padding='same') self.c4_2 = ConvBNRelu(ch, kernelsz=1, strides=strides) def call(self, x): x1 = self.c1(x) x2_1 = self.c2_1(x) x2_2 = self.c2_2(x2_1) x3_1 = self.c3_1(x) x3_2 = self.c3_2(x3_1) x4_1 = self.p4_1(x) x4_2 = self.c4_2(x4_1) # concat along axis=channel x = tf.concat([x1, x2_2, x3_2, x4_2], axis=3) return x class Inception10(Model): def __init__(self, num_blocks, num_classes, init_ch=16, **kwargs): super(Inception10, self).__init__(**kwargs) self.in_channels = init_ch self.out_channels = init_ch self.num_blocks = num_blocks self.init_ch = init_ch self.c1 = ConvBNRelu(init_ch) self.blocks = tf.keras.models.Sequential() # 代码非常简单：4个layer，两个block，block的前一个layer都不一样 for block_id in range(num_blocks): for layer_id in range(2): # 第一个InceptionBlock if layer_id == 0: block = InceptionBlk(self.out_channels, strides=2) else: block = InceptionBlk(self.out_channels, strides=1) self.blocks.add(block) # enlarger out_channels per block # channel数每次增加2倍，resnet和vgg里面好像也是这么做的 self.out_channels *= 2 self.p1 = GlobalAveragePooling2D() self.f1 = Dense(num_classes, activation='softmax') def call(self, x): x = self.c1(x) x = self.blocks(x) x = self.p1(x) y = self.f1(x) return y model = Inception10(num_blocks=2, num_classes=10)
二、cifar10分类-层方式-卷积神经网络-Inception10
博客对应课程的视频位置：
步骤
1、读取数据集
2、拆分数据集（拆分成训练数据集和测试数据集）
3、构建模型
4、训练模型
5、检验模型需求
cifar10（物品分类）
该数据集共有60000张彩色图像，这些图像是32*32，分为10个类，每类6000张图。这里面有50000张用于训练，构成了5个训练批，每一批10000张图；另外10000用于测试，单独构成一批。测试批的数据里，取自10类中的每一类，每一类随机取1000张。抽剩下的就随机排列组成了训练批。注意一个训练批中的各类图像并不一定数量相同，总的来看训练批，每一类都有5000张图。
In [1]:
import pandas as pd
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
1、读取数据集
直接从tensorflow的dataset来读取数据集即可
In [2]:
(train_x, train_y), (test_x, test_y) = tf.keras.datasets.cifar10.load_data()
print(train_x.shape, train_y.shape)
这是32*32的彩色图，rgb三个通道如何处理呢
In [3]:
plt.imshow(train_x[0])
plt.show()
In [4]:
plt.figure()
plt.imshow(train_x[1])
plt.figure()
plt.imshow(train_x[2])
plt.show()
In [5]:
print(test_y)
In [6]:
# 像素值 RGB
np.max(train_x[0])
Out[6]:
2、拆分数据集（拆分成训练数据集和测试数据集）
上一步做了拆分数据集的工作
In [7]:
# 图片数据如何归一化
# 直接除255即可
train_x = train_x/255.0
test_x = test_x/255.0
In [8]:
# 像素值 RGB
np.max(train_x[0])
Out[8]:
In [9]:
train_y=train_y.flatten()
test_y=test_y.flatten()
train_y = tf.one_hot(train_y, depth=10)
test_y = tf.one_hot(test_y, depth=10)
print(test_y.shape)
3、构建模型
应该构建一个怎么样的模型：
In [11]:
from tensorflow.keras.layers import Conv2D, BatchNormalization, Activation, MaxPool2D, Dropout, Flatten, Dense, \
GlobalAveragePooling2D
from tensorflow.keras import Model
In [13]:
class ConvBNRelu(Model):
def __init__(self, ch, kernelsz=3, strides=1, padding='same'):
super(ConvBNRelu, self).__init__()
self.model = tf.keras.models.Sequential([
Conv2D(ch, kernelsz, strides=strides, padding=padding),
BatchNormalization(),
Activation('relu')
])
def call(self, x):
x = self.model(x, training=False) #在training=False时，BN通过整个训练集计算均值、方差去做批归一化，training=True时，通过当前batch的均值、方差去做批归一化。推理时 training=False效果好
return x
class InceptionBlk(Model):
def __init__(self, ch, strides=1):
super(InceptionBlk, self).__init__()
self.ch = ch
self.strides = strides
self.c1 = ConvBNRelu(ch, kernelsz=1, strides=strides)
self.c2_1 = ConvBNRelu(ch, kernelsz=1, strides=strides)
self.c2_2 = ConvBNRelu(ch, kernelsz=3, strides=1)
self.c3_1 = ConvBNRelu(ch, kernelsz=1, strides=strides)
self.c3_2 = ConvBNRelu(ch, kernelsz=5, strides=1)
self.p4_1 = MaxPool2D(3, strides=1, padding='same')
self.c4_2 = ConvBNRelu(ch, kernelsz=1, strides=strides)
def call(self