search

Arlen0615 / Convert-own-data-to-MNIST-format

This project is convert own data to MNIST format. It may help you to quickly test new DL model without modify too much code.
MIT License
40 stars 21 forks source link

list index out of range #2

Open boloslaw opened 6 years ago

boloslaw commented 6 years ago

What does it mean?

C:\PROGRAMOWANIE\Python_OCR\Convert-own-data-to-MNIST-format-master\Convert-own- data-to-MNIST-format-master>python convert_to_mnist_format.py C:\PROGRAMOWANIE\P ython\OCR_Math\training_images 10 Traceback (most recent call last): File "convert_to_mnist_format.py", line 215, in main(sys.argv) File "convert_to_mnist_format.py", line 199, in main labelsAndFiles, argv[2]) File "convert_to_mnist_format.py", line 115, in make_arrays imShape = imageio.imread(labelsAndFiles[0][1]).shape IndexError: list index out of range

SunnerLi commented 6 years ago

Hi @boloslaw , I help my friend to deal with this project, and also encounter this problem. Some experience can share to you. I just run the following command:

$ python3 convert_to_mnist_format.py dataset 0

But the same error occurs. After I trace the code, the variable labelsAndFiles is empty. The rough guess is that the image format is not PNG. In the make_array function, the program only consider .png format. Maybe you can just revise like this in line 87:

for file in os.listdir(dirname):
    if (file.endswith('.png')):
        fullname = os.path.join(dirname, file)

to become like this:

for file in os.listdir(dirname):
    if (file.endswith('.png') or file.endswith('.jpg')):
        fullname = os.path.join(dirname, file)

Moreover, just append other condition if your image format is neither PNG nor JPG. The only thing you need to do is waiting it complete!