用Python将文件夹中的图片拆分为训练集、验证集、测试集

好的，可以使用Python中的os和shutil模块来实现将文件夹中的图片拆分为训练集、验证集、测试集的功能。可以按照以下步骤进行操作：

1. 导入必要的模块：

import os
import random
import shutil

2. 设置文件夹路径以及需要拆分的图片数量：

folder_path = "your_folder_path"
train_path = "your_train_folder_path"
valid_path = "your_valid_folder_path"
test_path = "your_test_folder_path"
split_ratio = [0.7, 0.2, 0.1]  # 分割比例，训练集占70%，验证集占20%，测试集占10%

3. 获取文件夹中的所有图片路径并打乱顺序：

file_names = os.listdir(folder_path)
random.shuffle(file_names)

4. 根据分割比例计算出各自需要的图片数量：

num_files = len(file_names)
train_split = int(split_ratio[0] * num_files)
valid_split = int(split_ratio[1] * num_files)
test_split = int(split_ratio[2] * num_files)

5. 将图片分别复制到训练集、验证集、测试集对应的文件夹中：

for i, file_name in enumerate(file_names):
    if i < train_split:
        shutil.copy(os.path.join(folder_path, file_name), train_path)
    elif i < train_split + valid_split:
        shutil.copy(os.path.join(folder_path, file_name), valid_path)
    else:
        shutil.copy(os.path.join(folder_path, file_name), test_path)

完整代码如下：

import os
import random
import shutil

folder_path = "your_folder_path"
train_path = "your_train_folder_path"
valid_path = "your_valid_folder_path"
test_path = "your_test_folder_path"
split_ratio = [0.7, 0.2, 0.1]

file_names = os.listdir(folder_path)
random.shuffle(file_names)

num_files = len(file_names)
train_split = int(split_ratio[0] * num_files)
valid_split = int(split_ratio[1] * num_files)
test_split = int(split_ratio[2] * num_files)

for i, file_name in enumerate(file_names):
    if i < train_split:
        shutil.copy(os.path.join(folder_path, file_name), train_path)
    elif i < train_split + valid_split:
        shutil.copy(os.path.join(folder_path, file_name), valid_path)
    else:
        shutil.copy(os.path.join(folder_path, file_name), test_path)

希望这个代码能够帮助到你！