y_true (label) in CTC

[tuanphan09]

Hi, I've just learnt CTC loss, and as I know it allows various length of labels as long as it's not longer than label_len. For that reason, I don't understand why you needed to pad '-' for the labels (your comment doesn't make sense btw):

# due to the explanation of ctc_loss, try to not add "-" for blank
while len(lexicon) < label_len:
     lexicon += "-"

and why you added '-' symbol in your vocabulary (characters):

characters = '0123456789'+string.ascii_lowercase+'-'
label_classes = len(characters)+1

EDIT: I fought that you need to pad to the label to make the code run well. Last question, eg. label='12345---' and label_len=5, CTC just uses label[:label_len] for caculating the loss, right?

[sbillburg]

I have tried not to add "-" for blank, but it didn't work well.

I think your method may make sense

[tuanphan09]

After a few day of reading, I fought that we can add any symbols to the label, that doesn't matter. We do it because we want the input to have same size. When training we give CTC: padded-label and true label_len, the algorithm actualy doens't use the padding part (like the EDIT part in the first question above).

For that reason, you can make your model more efficiently by fixing this:

label_classes = len(characters)