code-docstring-corpus icon indicating copy to clipboard operation
code-docstring-corpus copied to clipboard

Published 20 hours ago • verified publisher icon EdinburghNLP

blue score

Open jigyasa06 opened this issue 7 years ago • 2 comments

hi i followed all your steps, took your dataset and done all the pre processing and tokenization but i did not got the same results as you got. I got BLUE score as 0. can you please tell me why is it so? @Avmb @rsennrich

jigyasa06 avatar Oct 07 '18 16:10 jigyasa06

output_code_documentation.txt

PLease see this file one's this is code documentation file after using translate.py output_code_generation.txt

this is code generation and i want to know in the nmt_output you have written that you have done post-processing what is that all about? @rsennrich @Avmb Please let me know as soon as possible

jigyasa06 avatar Oct 07 '18 16:10 jigyasa06

Hi, it's difficult for us to determine what went wrong.

My suggestion is that you first train a translation model between natural languages, for instance English to German. There are tutorials to do this for Nematus http://data.statmt.org/wmt17_systems/training/ or Marian https://github.com/marian-nmt/marian-examples .

Once managed to set up a system, do translation and get BLEU scores, you can swap in the code-docstring-corpus files and re-run your pipeline.

Avmb avatar Nov 14 '18 14:11 Avmb