Commit 411018fc authored by ngimel's avatar ngimel Committed by Myle Ott
Browse files

add count for padding words (#180)

parent 8cba7b14

fairseq/data/dictionary.py

+1 −0
Changes for fairseq/data/dictionary.py: 1 added line, 0 removed lines.
Original line number Diff line number Diff line
@@ -125,6 +125,7 @@ class Dictionary(object): 
            i = 0

            while threshold_nwords % padding_factor != 0:

                new_symbols.append('madeupword{:04d}'.format(i))

                new_count.append(0)

                i += 1

                threshold_nwords += 1