For image captioning, Bleu score can by calculated by using pycocoevalcap repo or NLTK Python library like:
`from utils.coco_caption.pycocoevalcap.bleu.bleu import Bleu # // by pycocoevalcap
.
.
.
from nltk.translate.bleu_score import corpus_bleu # // by nltk
.
.
.`
What is the difference between them?
