README.md

---
language:
- en
license: mit
datasets:
- cardiffnlp/super_tweeteval
pipeline_tag: sentence-similarity
---
# cardiffnlp/twitter-roberta-large-similarity-latest

This is a RoBERTa-large model trained on 154M tweets until the end of December 2022 and finetuned for tweet similarity (regression on two texts) on the _TweetSIM_ dataset of [SuperTweetEval](https://huggingface.co/datasets/cardiffnlp/super_tweeteval).
The original Twitter-based RoBERTa model can be found [here](https://huggingface.co/cardiffnlp/twitter-roberta-large-2022-154m).

## Example
```python
from transformers import AutoModelForSequenceClassification, AutoTokenizer

model_name = "cardiffnlp/twitter-roberta-large-similarity-latest"
model = AutoModelForSequenceClassification.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)


text_1 = 'Looooooool what is this story #TalksWithAsh'
text_2 = 'For someone who keeps saying long story short, the story is quite long iyah #TalksWithAsh'

text_input = f"{text_1} </s> {text_2}"

model(**tokenizer(text_input, return_tensors="pt")).logits
>>tensor([[2.9565]])
```


## Citation Information
Please cite the [reference paper](https://arxiv.org/abs/2310.14757) if you use this model.

```bibtex
@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}
}
```