spark-deep-learning icon indicating copy to clipboard operation
spark-deep-learning copied to clipboard
verified publisher icon databricks

unable to read images using readImages method on windows

Open arunnatva opened this issue 8 years ago • 1 comments

Hi, I have installed spark 2.1.0, python 3.5 on windows. Installed tensorflow-gpu, keras and other packages required for image processing. Copied images from linux box to windows box. However, reading images using "readImages" throws error as follows:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0x80 in position 0: invalid start byte

arunnatva avatar Nov 22 '17 18:11 arunnatva

when I run the python command on windows to get default encoding, it shows as "cp1252". I am suspecting that the code in readImages is assuming that the image is with utf-8 encoding, but it is infact encoded with cp1252. Could you please suggest a work around for this. ?

arunnatva avatar Nov 22 '17 18:11 arunnatva