FashionMNIST数据集保存为图片

Myth丶恋晨 2023-01-03 04:52 218阅读 0赞

这个保存就很简单，因为使用过pytorch里面加载FashionMNIST数据集，就会生成这样一个文件

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pob3VfNDM4_size_16_color_FFFFFF_t_70][]

里面有两个文件

![20210107161745101.png][]

于是加载就非常简单了

import torch
    import cv2
    import numpy as np
    
    train_data=torch.load("training.pt")[0].numpy()
    test_data=torch.load("test.pt")[0].numpy()
    print(train_data.shape)
    print(test_data.shape)
    #print(test_data[0])
    for i  in range(train_data.shape[0]):     
        cv2.imwrite("train\\"+str(i)+".jpg",train_data[i][:,:,np.newaxis])
    for i  in range(test_data.shape[0]):     
        cv2.imwrite("test\\"+str(i)+".jpg",test_data[i][:,:,np.newaxis])

运行之前自己在路径下新建train和test文件夹

运行结果

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pob3VfNDM4_size_16_color_FFFFFF_t_70 1][]

### 友情提示：如想要处理好的图片文件和源文件，可以去我的csdn资源下载 ###

![20210107164646987.png][]

[https://download.csdn.net/download/zhou\_438/14040224][https_download.csdn.net_download_zhou_438_14040224]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pob3VfNDM4_size_16_color_FFFFFF_t_70]: /images/20221119/c26ae041ca7a42e98739b0575f2b6242.png
[20210107161745101.png]: /images/20221119/171eea0d82ba4c2db3ca9658c7452f37.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3pob3VfNDM4_size_16_color_FFFFFF_t_70 1]: /images/20221119/d364ba4c95a04d7597700d54f124575e.png
[20210107164646987.png]: /images/20221119/1a6cd0eab7054a398b287868330a7262.png
[https_download.csdn.net_download_zhou_438_14040224]: https://download.csdn.net/download/zhou_438/14040224