metadata

language:
  - en
license: mit
datasets:
  - cardiffnlp/super_tweeteval
pipeline_tag: text-classification

cardiffnlp/twitter-roberta-base-tempo-wic-latest

This is a RoBERTa-base model trained on 154M tweets until the end of December 2022 and finetuned for meaning shift detection (binary classification) on the TempoWIC dataset of SuperTweetEval. The original Twitter-based RoBERTa model can be found here.

Labels

"id2label": { "0": "no", "1": "yes" }

Example

from transformers import pipeline
text_1 = "We don't like the search and frisk so this bitch in neutral"
text_2 = "who the fuck is listening to mike bloomberg railing in 'bernie bros'? mr stop and frisk, literally turned the police on occupy wall street, rnc protestors, and new york muslims. get the fuck out"
text_input = f"{text_1}</s>{text_2}"

pipe = pipeline('text-classification', model="cardiffnlp/twitter-roberta-base-tempo-wic-latest")
pipe(text_input)
>> [{'label': 'yes', 'score': 0.9994196891784668}]

Citation Information

Please cite the reference paper if you use this model.

@inproceedings{antypas2023supertweeteval,
  title={SuperTweetEval: A Challenging, Unified and Heterogeneous Benchmark for Social Media NLP Research},
  author={Dimosthenis Antypas and Asahi Ushio and Francesco Barbieri and Leonardo Neves and Kiamehr Rezaee and Luis Espinosa-Anke and Jiaxin Pei and Jose Camacho-Collados},
  booktitle={Findings of the Association for Computational Linguistics: EMNLP 2023},
  year={2023}
}