-
We use the MMSA framework for our experiments. The code can be run by placing it in the appropriate folder.
-
The main content of the dataset contains: (1) the content of the SELF_MM published dataset, (2) the glove text word vector, and (3) the unimodal labels generated based on SELF_MM.
NOTE: [1] MMSA: https://github.com/thuiar/MMSA [2] SELF_MM: https://github.com/thuiar/Self-MM [3] glove: http://immortal.multicomp.cs.cmu.edu/ [4] [1]+[2]+[3]: https://www.aliyundrive.com/s/3f18JGqUh2d