MNIST数据集解压的正确方式：ValueError: cannot reshape array of size 9912406 into shape (60000,28,28,1)

win7 环境报错

1. File "C:/DCGAN/main.py", line 63, in main
2. sample_dir=FLAGS.sample_dir)
3. File "C:\DCGAN\model.py", line 74, in __init__
4. self.data_X, self.data_y = self.load_mnist()
5. File "C:\DCGAN\model.py", line 461, in load_mnist
6. trX = loaded[16:].reshape((60000,28,28,1)).astype(np.float)
7. ValueError: cannot reshape array of size 9912406 into shape (60000,28,28,1)

linux（Ubuntu）环境报错

1. File "main.py", line 63, in main
2. sample_dir=FLAGS.sample_dir)
3. File "/root/DCGAN-tensorflow-master/model.py", line 74, in __init__
4. self.data_X, self.data_y = self.load_mnist()
5. File "/root/DCGAN-tensorflow-master/model.py", line 459, in load_mnist
6. trX = loaded[16:].reshape((60000,28,28,1)).astype(np.float)
7. ValueError: cannot reshape array of size 9912406 into shape (60000,28,28,1)

原因

数据集解压方式错误，MNIST文件下载后是.gz格式，需要在linux环境下解压（win7环境用rar解压，报同样错误）

t10k-images-idx3-ubyte.gz  

t10k-labels-idx1-ubyte.gz 

train-images-idx3-ubyte.gz  

train-labels-idx1-ubyte.gz

解压后的文件列表：

t10k-images-idx3-ubyte

t10k-labels-idx1-ubyte

train-images-idx3-ubyte

train-labels-idx1-ubyte

linux（Ubuntu）环境gz解压指令：

gunzip –c train-labels-idx1-ubyte.gz > train-labels-idx1-ubyte
gunzip –c train-images-idx3-ubyte.gz   > train-images-idx3-ubyte
gunzip –c t10k-images-idx3-ubyte.gz  > t10k-images-idx3-ubyte

gunzip –c t10k-labels-idx1-ubyte.gz   > t10k-labels-idx1-ubyte