Thanks for your great work!
In Oscar paper, you mentioned beam size = 5 during inference.
However, in your CIDEr optimization finetune command, you use beam size = 1.
Is that mean you did not use beam search in training and only use beam search in testing?
Thanks for your great work! In Oscar paper, you mentioned beam size = 5 during inference. However, in your CIDEr optimization finetune command, you use beam size = 1. Is that mean you did not use beam search in training and only use beam search in testing?