Small multimodal models in a math-and-language domain

Code for the (Dis)embodiment workshop 2022 paper "Can We Use Small Models to Investigate Multimodal Fusion Methods?" (TBD: provide link to article).

Main files

• Information about how to run the experiments described in the paper can be found under data/experiments.md.
• Once the experiments have been run, the code used for generating the plots in the paper can be found under data/results.ipynb.

Acknowledgements

We thank the authors behind VOLTA and LXMERT for providing well-structured code repositories describing their work.