# Twitter-scratch-roBERTa-base

This is a RoBERTa-base model trained from scratch on ~58M tweets, as described and evaluated in the [_TweetEval_ benchmark (Findings of EMNLP 2020)](https://arxiv.org/pdf/2010.12421.pdf). 
To evaluate this and other LMs on Twitter-specific data, please refer to the [Tweeteval official repository](https://github.com/cardiffnlp/tweeteval).

## Preprocess Text 
Replace usernames and links for placeholders: "@user" and "http".
```python
def preprocess(text):
    new_text = []
    for t in text.split(" "):
        t = '@user' if t.startswith('@') and len(t) > 1 else t
        t = 'http' if t.startswith('http') else t
        new_text.append(t)
    return " ".join(new_text)
```

## Example Masked Language Model 

```python
from transformers import pipeline, AutoTokenizer
import numpy as np

MODEL = "cardiffnlp/twitter-scratch-roberta-base"
fill_mask = pipeline("fill-mask", model=MODEL, tokenizer=MODEL)
tokenizer = AutoTokenizer.from_pretrained(MODEL)

def print_candidates():
    for i in range(5):
        token = tokenizer.decode(candidates[i]['token'])
        score = np.round(candidates[i]['score'], 4)
        print(f"{i+1}) {token} {score}")

texts = [
 "I am so <mask> 😊",
 "I am so <mask> 😢" 
]
for text in texts:
    t = preprocess(text)
    print(f"{'-'*30}\n{t}")
    candidates = fill_mask(t)
    print_candidates()
```

Output: 

```
------------------------------
I am so <mask> 😊
1)  happy 0.530
2)  grateful 0.083
3)  excited 0.078
4)  thankful 0.053
5)  blessed 0.041
------------------------------
I am so <mask> 😢
1)  sad 0.439
2)  sorry 0.088
3)  tired 0.045
4)  hurt 0.026
5)  upset 0.026
```


### BibTeX entry and citation info

Please cite the [reference paper](https://aclanthology.org/2020.findings-emnlp.148/) if you use this model.

```bibtex
@inproceedings{barbieri-etal-2020-tweeteval,
    title = "{T}weet{E}val: Unified Benchmark and Comparative Evaluation for Tweet Classification",
    author = "Barbieri, Francesco  and
      Camacho-Collados, Jose  and
      Espinosa Anke, Luis  and
      Neves, Leonardo",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2020",
    month = nov,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2020.findings-emnlp.148",
    doi = "10.18653/v1/2020.findings-emnlp.148",
    pages = "1644--1650"
}
```