Multi-task Transformers for Sequence Classification

Installation Steps:

1. Install HuggingFace's transformers library

2. Install multitask_transformers
   

git clone https://github.com/ritvikshrivastava/multitask_transformers.git
cd multitask_transformers
pip install .

Run DistilBert/Bert/RoBERTa Multi-Task for Sarcasm Detection - Argument Relation Detection Task:

cd multitask_transformers/scripts/
./run_sarc_arg.sh

update model type in run_sarc_arg.sh by updating:
--model_name_or_path <pretrained/finetuned model name or path to dir>

Run DistilBert/Bert/RoBERTa Alternative Training Multi-Task for Sarcasm Detection - Argument Relation Detection Task:

cd multitask_transformers/scripts/
./run_alternate_data_multitask.sh

update model type in run_alternate_data_multitask.sh by updating
--model_name_or_path <pretrained/finetuned model name or path to dir>

Citation

Please cite the following paper if you use or ideate with Multi-task Transformers in your work:

https://aclanthology.org/2021.eacl-main.171.pdf

Ghosh, D., Shrivastava, R. and Muresan, S., 2021, April. “Laughing at you or with you”: The Role of Sarcasm in Shaping the Disagreement Space. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume (pp. 1998-2010).

@inproceedings{ghosh2021laughing,
  title={“Laughing at you or with you”: The Role of Sarcasm in Shaping the Disagreement Space},
  author={Ghosh, Debanjan and Shrivastava, Ritvik and Muresan, Smaranda},
  booktitle={Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume},
  pages={1998--2010},
  year={2021}
}