Tired of training multimodal models with one caption <--> one image? Excited that our paper on multi-sentence, multi-image documents was accepted to
#EMNLP2019
ðŸ‡ðŸ‡°ðŸ‡ðŸ‡°!! Joint w/ Lillian Lee +
@dmimno
:D
Preprint, data, and code:
(Camera ready coming soon!)