rsennrich subword-nmt issues

rsennrich / subword-nmt

Unsupervised Word Segmentation for Neural Machine Translation and Text Generation

MIT License

2.19k stars 465 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Using stats.get without lambda to get most_frequent

#20 alvations closed 7 years ago
2
Motivation for the change in final `<\w>` addition in BPE

#19 alvations closed 7 years ago
2
Use cache when applying the learnt bpe

#18 dmesq closed 7 years ago
2
`cache` parameter in the apply_bpe.py script not being used

#17 dmesq closed 7 years ago
2
Final vocabulary size is not equal to character vocabulary plus num_operations ?

#16 huache closed 7 years ago
7
The given sed pattern may not be enough to remove @@

#15 ozancaglayan closed 7 years ago
2
Allow passing in a word - count file

#14 mboyanov closed 7 years ago
1
Different BPE models produced based on python version and/or stdin/argparse

#13 ozancaglayan closed 7 years ago
6
Desired Final Vocabulary size

#12 philip30 closed 7 years ago
2
Encoding on file fixed

#11 rmeertens closed 7 years ago
5
using python3 print function

#10 rmeertens closed 7 years ago
1
Default char n-gram order for chrf should be set to 6 and beta set to 2

#9 alvations closed 8 years ago
2
Frequency threshold?

#8 emjotde closed 8 years ago
1
what is num_operation?!

#7 amirj closed 8 years ago
4
Vocabulary size

#6 casperkaae closed 8 years ago
2
replace_pair in learn_bpe.py throws exception for strings that end with escape char (backslash)

#5 dblandan closed 8 years ago
8
Fixes #3

#4 He-Ro closed 8 years ago
1
chrF fails on short references.

#3 He-Ro closed 8 years ago
0
strip

#2 kyunghyuncho closed 8 years ago
1
use strip

#1 kyunghyuncho closed 8 years ago
0