Skip to content
This repository has been archived by the owner on Dec 29, 2022. It is now read-only.

Deprecate non-standard BLEU scripts #365

Open
kpu opened this issue Jul 6, 2020 · 0 comments
Open

Deprecate non-standard BLEU scripts #365

kpu opened this issue Jul 6, 2020 · 0 comments

Comments

@kpu
Copy link

kpu commented Jul 6, 2020

multi-bleu.perl has been deprecated for years now because it encourages people to use non-standard tokenization. This repository contains another non-standard BLEU implementation that a user might not notice they are using: https://github.com/google/seq2seq/blob/7f485894d412e8d81ce0e07977831865e44309ce/seq2seq/metrics/bleu.py

A paper by @mjpost https://www.aclweb.org/anthology/W18-6319/ shows how much this can vary.

Please put fat warnings on non-standard BLEU scripts that they are not appropriate for publication, as Moses has. And remove multi-bleu.perl from examples.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant