Skip to content

Corpus level and sentence level BLEU calculation for machine translation

Notifications You must be signed in to change notification settings

cshanbo/Smooth_BLEU

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

23 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

BLEU

Out-of-the-box Python script for sentence level and corpus level BLEU calculation We recommend users to use nltk-based BLEU calculation script by installing nltk first.

Run python bleu.py -h or python nltk_bleu.py -h to see the help information

Usage

python bleu.py -h to see the help information

  1. input FILES

    • python bleu.py -t translation.file -t reference.file
    • Set the parameter --sentence-level and other related parameters, such as --smooth-epsilon, --smooth to print sentence-level BLEU score.
  2. input Strings

    python bleu.py -t "test bleu calculation" -r "test blue calculation"
    
    BLEU = 90.36,  66.7/0.0/0.0/0.0 (BP=1.0,  ratio=1.0,  hyp_len=3,  ref_len=3)

The bleu.py is updated from this gist.

nltk-based

Instead of using bleu.py, one can use nltk_bleu.py based on nltk.

for example: python nltk_bleu.py -r reference.file -t translation.file

Similarly, one can use --sentence-level or -sl to print sentence-level BLEU score.

TODO

  1. bleu.py correctness check
  2. better wrapper for nltk_bleu.py
  3. NLTK installation included?

About

Corpus level and sentence level BLEU calculation for machine translation

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published