Skip to content

Code for the official implementation of Dynamic Routing Transformer Network for Multimodal Sarcasm Detection (ACL'23), used for reproducibility investigation purposes

Notifications You must be signed in to change notification settings

elequaranta/DynRT_reproducibility

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Reproducibility Study of: Dynamic Routing Tranformer Network for Multimodal Sarcasm Detection

Research study for course CS421: Natural Language Processing @ UIC.

The paper analyzed can be found here, and its original repo here.

Study performed by:

Setup

In order to reproduce the study that we have performed, follow these steps:

  1. Create a subfolder of source called dataset_image.
  2. Download the image data from here.
  3. Put all the installed images in the folder just created (i.e. in source/dataset_image/).
  4. Into run.ipynb change the cell just under the voice Preprocessing with the path where you have stored the repo and, if you are not using Gogle Colab delete or comment the first cell.
  5. Run all cells of file run.ipynb.

Acknowledgements

Thanks for the original repo from https://github.com/TIAN-viola/DynRT

Thanks for the dataset from https://github.com/headacheboy/data-of-multimodal-sarcasm-detection

Thanks for the RoBERTa model from https://huggingface.co/roberta-base/

Thanks for the TRAR from https://github.com/rentainhe/TRAR-VQA

About

Code for the official implementation of Dynamic Routing Transformer Network for Multimodal Sarcasm Detection (ACL'23), used for reproducibility investigation purposes

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 84.1%
  • Python 11.4%
  • TeX 4.3%
  • Makefile 0.2%