VGG 中草药识别

一些基础

关于VGG

使用小尺寸卷积核和池化层组成的基础模块 一共包含13层卷积和3层全连接层。VGG网络使用3×3的卷积层和池化层组成的基础模块来提取特征，三层全连接层放在网络的最后组成分类器，最后一层全连接层的输出即为分类的预测。多用于图像分类。

参考链接

中药草识别

模型优化：

可以改变batch_size的大小。在网络训练过程中，batch_size过大或者过小都会影响训练的性能和速度，batch_size过小，花费时间多，同时梯度震荡严重，不利于收敛；batch_size过大，不同batch的梯度方向没有任何变化，容易陷入局部极小值。
在合理范围内，增大batch_size会提高显存的利用率，提高大矩阵乘法的并行化效率，减少每个epoch需要训练的迭代次数。在一定范围内，batch size越大，其确定的下降方向越准，引起训练时准确率震荡越小。

paddle.nn.Dropout

Dropout是一种正则化手段，该算子根据给定的丢弃概率 p ，在训练过程中随机将一些神经元输出设置为0，通过阻止神经元节点间的相关性来减少过拟合。默认: 0.5。

数据准备

引入包

import os
import zipfile
import random
import json
import paddle
import sys
import numpy as np
from PIL import Image
import matplotlib.pyplot as plt
from paddle.io import Dataset
random.seed(200)

解压数据集

def unzip_data(src_path,target_path):
    if(not os.path.isdir(target_path + "Chinese Medicine")):     
        z = zipfile.ZipFile(src_path, 'r')
        z.extractall(path=target_path)
        z.close()

生成数据列表

def get_data_list(target_path,train_list_path,eval_list_path):
    '''
    生成数据列表
    '''
    #存放所有类别的信息
    class_detail = []
    #获取所有类别保存的文件夹名称
    data_list_path=target_path+"Chinese Medicine/"
    class_dirs = os.listdir(data_list_path)  
    #总的图像数量
    all_class_images = 0
    #存放类别标签
    class_label=0
    #存放类别数目
    class_dim = 0
    #存储要写进eval.txt和train.txt中的内容
    trainer_list=[]
    eval_list=[]
    #读取每个类别，['baihe', 'gouqi','jinyinhua','huaihua','dangshen']
    for class_dir in class_dirs:
        if class_dir != ".DS_Store":
            class_dim += 1
            #每个类别的信息
            class_detail_list = {}
            eval_sum = 0
            trainer_sum = 0
            #统计每个类别有多少张图片
            class_sum = 0
            #获取类别路径 
            path = data_list_path  + class_dir
            # 获取所有图片
            img_paths = os.listdir(path)
            for img_path in img_paths:                                  # 遍历文件夹下的每个图片
                name_path = path + '/' + img_path                       # 每张图片的路径
                if class_sum % 8 == 0:                                  # 每8张图片取一个做验证数据
                    eval_sum += 1                                       # test_sum为测试数据的数目
                    eval_list.append(name_path + "\t%d" % class_label + "\n")
                else:
                    trainer_sum += 1 
                    trainer_list.append(name_path + "\t%d" % class_label + "\n")#trainer_sum测试数据的数目
                class_sum += 1                                          #每类图片的数目
                all_class_images += 1                                   #所有类图片的数目
             
            # 说明的json文件的class_detail数据
            class_detail_list['class_name'] = class_dir             #类别名称
            class_detail_list['class_label'] = class_label          #类别标签
            class_detail_list['class_eval_images'] = eval_sum       #该类数据的测试集数目
            class_detail_list['class_trainer_images'] = trainer_sum #该类数据的训练集数目
            class_detail.append(class_detail_list)  
            #初始化标签列表
            train_parameters['label_dict'][str(class_label)] = class_dir
            class_label += 1 
            
    #初始化分类数
    train_parameters['class_dim'] = class_dim
    
    #乱序  
    random.shuffle(eval_list)
    with open(eval_list_path, 'a') as f:
        for eval_image in eval_list:
            f.write(eval_image) 
            
    random.shuffle(trainer_list)
    with open(train_list_path, 'a') as f2:
        for train_image in trainer_list:
            f2.write(train_image) 

    # 说明的json文件信息
    readjson = {}
    readjson['all_class_name'] = data_list_path                  #文件父目录
    readjson['all_class_images'] = all_class_images
    readjson['class_detail'] = class_detail
    jsons = json.dumps(readjson, sort_keys=True, indent=4, separators=(',', ': '))
    with open(train_parameters['readme_path'],'w') as f:
        f.write(jsons)
    print ('生成数据列表完成！')
# 当然，我们目前只是定义了以上两个函数，要想完成上述两个步骤，我们需要初始化数据集存取存放路径，调用上述解压数据集函数与生成数据列表函数完成读入数据之前的准备工作


train_parameters = {
    "src_path":"/home/aistudio/data/data105575/Chinese Medicine.zip",    #原始数据集路径
    "target_path":"/home/aistudio/data/",                     #要解压的路径
    "train_list_path": "/home/aistudio/data/train.txt",       #train.txt路径
    "eval_list_path": "/home/aistudio/data/eval.txt",         #eval.txt路径
    "label_dict":{},                                          #标签字典
    "readme_path": "/home/aistudio/data/readme.json",         #readme.json路径
    "class_dim": -1,                                          #分类数
}
src_path=train_parameters['src_path']
target_path=train_parameters['target_path']
train_list_path=train_parameters['train_list_path']
eval_list_path=train_parameters['eval_list_path']

# 调用解压函数解压数据集
unzip_data(src_path,target_path)


# 划分训练集与验证集，乱序，生成数据列表
#每次生成数据列表前，首先清空train.txt和eval.txt
with open(train_list_path, 'w') as f: 
    f.seek(0)
    f.truncate() 
with open(eval_list_path, 'w') as f: 
    f.seek(0)
    f.truncate() 
#生成数据列表   
get_data_list(target_path,train_list_path,eval_list_path)

效果：包含图片文件的路径和标签，分成了训练集和测试集。

生成的json文件：

定义数据读取器

class dataset(Dataset):
    def __init__(self, data_path, mode='train'):
        """
        数据读取器
        :param data_path: 数据集所在路径
        :param mode: train or eval
        """
        super().__init__()
        self.data_path = data_path
        self.img_paths = []
        self.labels = []

        if mode == 'train':
            with open(os.path.join(self.data_path, "train.txt"), "r", encoding="utf-8") as f:
                self.info = f.readlines()
            for img_info in self.info:
                img_path, label = img_info.strip().split('\t')
                self.img_paths.append(img_path)
                self.labels.append(int(label))

        else:
            with open(os.path.join(self.data_path, "eval.txt"), "r", encoding="utf-8") as f:
                self.info = f.readlines()
            for img_info in self.info:
                img_path, label = img_info.strip().split('\t')
                self.img_paths.append(img_path)
                self.labels.append(int(label))


    def __getitem__(self, index):
        """
        获取一组数据
        :param index: 文件索引号
        :return:
        """
        # 第一步打开图像文件并获取label值
        img_path = self.img_paths[index]
        img = Image.open(img_path)
        if img.mode != 'RGB':
            img = img.convert('RGB') 
        img = img.resize((224, 224), Image.BILINEAR)
        #img = rand_flip_image(img)
        img = np.array(img).astype('float32')
        img = img.transpose((2, 0, 1)) / 255
        label = self.labels[index]
        label = np.array([label], dtype="int64")
        return img, label

    def print_sample(self, index: int = 0):
        print("文件名", self.img_paths[index], "\t标签值", self.labels[index])

    def __len__(self):
        return len(self.img_paths)

数据加载

这里我们使用paddle.io.DataLoader模块实现数据加载，并且指定训练集批大小batch_size为32，乱序读入；验证集批大小为8，不打乱顺序。

train_dataset = dataset('/home/aistudio/data',mode='train')
train_loader = paddle.io.DataLoader(train_dataset, batch_size=32, shuffle=True)  # 传入dataset类作为参数，该参数必须是paddle.io.Dataset的一个实例化
#评估数据加载
eval_dataset = dataset('/home/aistudio/data',mode='eval')
eval_loader = paddle.io.DataLoader(eval_dataset, batch_size = 8, shuffle=False)  # shuffle代表是否打乱顺序，返回一个迭代器，根据batch_sampler 给定的顺序迭代一次给定的 dataset

模型搭建

定义卷积池化网络

# 定义卷积池化网络
class ConvPool(paddle.nn.Layer):
    '''卷积+池化'''
    def __init__(self,
                 num_channels,
                 num_filters, 
                 filter_size,
                 pool_size,
                 pool_stride,
                 groups,
                 conv_stride=1, 
                 conv_padding=1,
                 ):
        super(ConvPool, self).__init__()  

        # groups代表卷积层的数量
        for i in range(groups):
            self.add_sublayer(   #添加子层实例
                'bb_%d' % i,
                paddle.nn.Conv2D(         # layer
                in_channels=num_channels, #通道数
                out_channels=num_filters,   #卷积核个数
                kernel_size=filter_size,   #卷积核大小
                stride=conv_stride,        #步长
                padding = conv_padding,    #padding
                )
            )
            self.add_sublayer(
                'relu%d' % i,
                paddle.nn.ReLU()
            )
            num_channels = num_filters
            

        self.add_sublayer(
            'Maxpool',
            paddle.nn.MaxPool2D(
            kernel_size=pool_size,           #池化核大小
            stride=pool_stride               #池化步长
            )
        )

    def forward(self, inputs):
        x = inputs
        for prefix, sub_layer in self.named_children():
            # print(prefix,sub_layer)
            x = sub_layer(x)
        return x

VGG网络

class VGGNet(paddle.nn.Layer):
    def __init__(self):
        super(VGGNet, self).__init__()       
        # 5个卷积池化操作
        self.convpool01 = ConvPool(
            3, 64, 3, 2, 2, 2)  #3:通道数，64：卷积核个数，3:卷积核大小，2:池化核大小，2:池化步长，2:连续卷积个数
        self.convpool02 = ConvPool(
            64, 128, 3, 2, 2, 2)
        self.convpool03 = ConvPool(
            128, 256, 3, 2, 2, 3) 
        self.convpool04 = ConvPool(
            256, 512, 3, 2, 2, 3)
        self.convpool05 = ConvPool(
            512, 512, 3, 2, 2, 3)       
        self.pool_5_shape = 512 * 7* 7
        # 三个全连接层
        self.fc01 = paddle.nn.Linear(self.pool_5_shape, 4096)
        self.drop1 = paddle.nn.Dropout(p=0.5)
        self.fc02 = paddle.nn.Linear(4096, 4096)
        self.drop2 = paddle.nn.Dropout(p=0.5)
        self.fc03 = paddle.nn.Linear(4096, train_parameters['class_dim'])

    def forward(self, inputs, label=None):
        # print('input_shape:', inputs.shape) #[8, 3, 224, 224]
        """前向计算"""
        out = self.convpool01(inputs)
        # print('convpool01_shape:', out.shape)           #[8, 64, 112, 112]
        out = self.convpool02(out)
        # print('convpool02_shape:', out.shape)           #[8, 128, 56, 56]
        out = self.convpool03(out)
        # print('convpool03_shape:', out.shape)           #[8, 256, 28, 28]
        out = self.convpool04(out)
        # print('convpool04_shape:', out.shape)           #[8, 512, 14, 14]
        out = self.convpool05(out)
        # print('convpool05_shape:', out.shape)           #[8, 512, 7, 7]         

        out = paddle.reshape(out, shape=[-1, 512*7*7])
        out = self.fc01(out)
        out = self.drop1(out)
        out = self.fc02(out)
        out = self.drop2(out)
        out = self.fc03(out)
        
        if label is not None:
            acc = paddle.metric.accuracy(input=out, label=label)
            return out, acc
        else:
            return out

参数配置

# 参数配置，要保留之前数据集准备阶段配置的参数，所以使用update更新字典
train_parameters.update({
    "input_size": [3, 224, 224],                              #输入图片的shape
    "num_epochs": 35,                                         #训练轮数
    "skip_steps": 10,                                         #训练时输出日志的间隔
    "save_steps": 100,                                         #训练时保存模型参数的间隔
    "learning_strategy": {                                    #优化函数相关的配置
        "lr": 0.0001                                          #超参数学习率
    },
    "checkpoints": "/home/aistudio/work/checkpoints"          #保存的路径
})

模型训练

model = VGGNet()
model.train()
# 配置loss函数
cross_entropy = paddle.nn.CrossEntropyLoss()
# 配置参数优化器
optimizer = paddle.optimizer.Adam(learning_rate=train_parameters['learning_strategy']['lr'],
                                  parameters=model.parameters()) 

steps = 0
Iters, total_loss, total_acc = [], [], []

for epo in range(train_parameters['num_epochs']):
    for _, data in enumerate(train_loader()):
        steps += 1
        x_data = data[0]
        y_data = data[1]
        predicts, acc = model(x_data, y_data)
        loss = cross_entropy(predicts, y_data)
        loss.backward()
        optimizer.step()
        optimizer.clear_grad()
        if steps % train_parameters["skip_steps"] == 0:
            Iters.append(steps)
            total_loss.append(loss.numpy()[0])
            total_acc.append(acc.numpy()[0])
            #打印中间过程
            print('epo: {}, step: {}, loss is: {}, acc is: {}'\
                  .format(epo, steps, loss.numpy(), acc.numpy()))
        #保存模型参数
        if steps % train_parameters["save_steps"] == 0:
            save_path = train_parameters["checkpoints"]+"/"+"save_dir_" + str(steps) + '.pdparams'
            print('save model to: ' + save_path)
            paddle.save(model.state_dict(),save_path)
paddle.save(model.state_dict(),train_parameters["checkpoints"]+"/"+"save_dir_final.pdparams")
draw_process("trainning loss","red",Iters,total_loss,"trainning loss")
draw_process("trainning acc","green",Iters,total_acc,"trainning acc")

评估与预测

评估

model__state_dict = paddle.load('work/checkpoints/save_dir_final.pdparams')
model_eval = VGGNet()
model_eval.set_state_dict(model__state_dict) 
model_eval.eval()
accs = []
# 开始评估
for _, data in enumerate(eval_loader()):
    x_data = data[0]
    y_data = data[1]
    predicts = model_eval(x_data)
    acc = paddle.metric.accuracy(predicts, y_data)
    accs.append(acc.numpy()[0])
print('模型在验证集上的准确率为：',np.mean(accs))

模型推理

def load_image(img_path):
    '''
    预测图片预处理
    '''
    img = Image.open(img_path) 
    if img.mode != 'RGB': 
        img = img.convert('RGB') 
    img = img.resize((224, 224), Image.BILINEAR)
    img = np.array(img).astype('float32') 
    img = img.transpose((2, 0, 1)) / 255 # HWC to CHW 及归一化
    return img


label_dic = train_parameters['label_dict']

import time
# 加载训练过程保存的最后一个模型
model__state_dict = paddle.load('work/checkpoints/save_dir_final.pdparams')
model_predict = VGGNet()
model_predict.set_state_dict(model__state_dict) 
model_predict.eval()
infer_imgs_path = os.listdir("infer")
# print(infer_imgs_path)

# 预测所有图片
for infer_img_path in infer_imgs_path:
    infer_img = load_image("infer/"+infer_img_path)
    infer_img = infer_img[np.newaxis,:, : ,:]  #reshape(-1,3,224,224)
    infer_img = paddle.to_tensor(infer_img)
    result = model_predict(infer_img)
    lab = np.argmax(result.numpy())
    print("样本: {},被预测为:{}".format(infer_img_path,label_dic[str(lab)]))
    img = Image.open("infer/"+infer_img_path)
    plt.imshow(img)
    plt.axis('off')
    plt.show()
    sys.stdout.flush()
    time.sleep(0.5)