Closed ggarimagoyal closed 5 years ago
@ggarimagoyal Are the images converted properly? I have mentioned the steps here https://github.com/samkit-jain/Handwriting-Recognition/issues/2#issuecomment-366901594 Because of some prior commitments I do not have the time to help you debug the issue. Request you to debug yourself or experiment with various models or a smaller set of images (like trying with only 4 letters). I'll get back to you on the issue as soon as I get some time and test it on my machine.
Codebase has been updated and you shouldn't be facing this anymore. Feel free to reopen the issue if not problem not solved.
Hello Sir, in letter_recognition.py it is predicting letter "l" for every letter drawn. Help me with this. For making the fs2 directory, I have used the images in NIST dataset (using images from train folder given for each character in NIST dataset) from the link you have provided earlier in another issue.