Code for the CoNLL 2019 paper "Compositional Generalization in Image Captioning" by Mitja Nikolaus, Mostafa Abdou, Matthew Lamm, Rahul Aralikatte and Desmond Elliott
Hi there,
I have just read your paper regarding compositional generalization in image captioning. Very nice work, thanks! I have a small regarding the beam size for re-ranking, I didn't find this detail in the paper. May I know the beam size?
Also, may I know the full split information used as the upper bound in Table 2 (Full) of the paper.