python根据label.txt将未分类的图片数据集分类，一个文件夹为一个单独的类别

代码如下，改下train_data的路径即可

import numpy as np
import os
import shutil


def read_txt(full_path: str) -> dict:
    m = {}
    i = 0
    for line in open(full_path).readlines():
        i += 1
        tmp = line[:-1]
        # label = np.append(label, tmp)
        file_name = str(i) + '.jpg'
        m[file_name] = tmp
    return m


# 根据label.txt处理数据集，对每一个类别建立一个文件夹
def deal_train(path):
    file_names = np.array([])
    label_dict = {}
    for _, _, files in os.walk(path):
        for file in files:
            if os.path.splitext(file)[1] == '.txt':
                full_path = os.path.join(os.getcwd(), path[2:], file)
                label_dict = read_txt(full_path)
                break

        for file in files:
            if os.path.splitext(file)[1] == '.jpg':
                # file_names = np.append(file_names, file)
                category = label_dict[file]
                if category == '1' or category == '-1':
                    full_path = os.path.join(os.getcwd(), path[2:], category)
                    if not os.path.exists(full_path):
                        os.mkdir(full_path)
                    # 将当前文件拷贝到新建的目录下
                    if not os.path.exists(os.path.join(full_path, file)):  # 判断目标文件夹不存在该文件才进行拷贝
                        shutil.copy(os.path.join(os.getcwd(), path[2:], file), os.path.join(full_path, file))


train_path = './data/train'
validate_path = './data/validate'

if __name__ == '__main__':
    # deal_train(train_path)
    deal_train(validate_path)