BERTweet: A pre-trained language model for English Tweets

BERTweet is the first public large-scale language model pre-trained for English Tweets. BERTweet is trained based on the RoBERTa pre-training procedure. The corpus used to pre-train BERTweet consists of 850M English Tweets (16B word tokens ~ 80GB), containing 845M Tweets streamed from 01/2012 to 08/2019 and 5M Tweets related to the COVID-19 pandemic. The general architecture and experimental results of BERTweet can be found in our paper:

@inproceedings{bertweet,
title     = {{BERTweet: A pre-trained language model for English Tweets}},
author    = {Dat Quoc Nguyen and Thanh Vu and Anh Tuan Nguyen},
booktitle = {Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations},
pages     = {9--14},
year      = {2020}
}

Please CITE our paper when BERTweet is used to help produce published results or is incorporated into other software.

For further information or requests, please go to BERTweet's homepage!

Pre-trained models

Model #params Arch. Pre-training data
vinai/bertweet-base 135M base 850M English Tweets (cased)
vinai/bertweet-covid19-base-cased 135M base 23M COVID-19 English Tweets (cased)
vinai/bertweet-covid19-base-uncased 135M base 23M COVID-19 English Tweets (uncased)
vinai/bertweet-large 355M large 873M English Tweets (cased)

Example usage

import torch
from transformers import AutoModel, AutoTokenizer 

bertweet = AutoModel.from_pretrained("vinai/bertweet-base")

# For transformers v4.x+: 
tokenizer = AutoTokenizer.from_pretrained("vinai/bertweet-base", use_fast=False)

# For transformers v3.x: 
# tokenizer = AutoTokenizer.from_pretrained("vinai/bertweet-base")

# INPUT TWEET IS ALREADY NORMALIZED!
line = "SC has first two presumptive cases of coronavirus , DHEC confirms HTTPURL via @USER :crying_face:"

input_ids = torch.tensor([tokenizer.encode(line)])

with torch.no_grad():
    features = bertweet(input_ids)  # Models outputs are now tuples
    
## With TensorFlow 2.0+:
# from transformers import TFAutoModel
# bertweet = TFAutoModel.from_pretrained("vinai/bertweet-base")

Normalize raw input Tweets

Before applying fastBPE to the pre-training corpus of 850M English Tweets, we tokenized these Tweets using TweetTokenizer from the NLTK toolkit and used the emoji package to translate emotion icons into text strings (here, each icon is referred to as a word token). We also normalized the Tweets by converting user mentions and web/url links into special tokens @USER and HTTPURL, respectively. Thus it is recommended to also apply the same pre-processing step for BERTweet-based downstream applications w.r.t. the raw input Tweets. BERTweet provides this pre-processing step by enabling the normalization argument. This argument currently only supports models "vinai/bertweet-base", "vinai/bertweet-covid19-base-cased" and "vinai/bertweet-covid19-base-uncased".

  • Install emoji: pip3 install emoji
import torch
from transformers import AutoTokenizer

# Load the AutoTokenizer with a normalization mode if the input Tweet is raw
tokenizer = AutoTokenizer.from_pretrained("vinai/bertweet-base", normalization=True)

# from transformers import BertweetTokenizer
# tokenizer = BertweetTokenizer.from_pretrained("vinai/bertweet-base", normalization=True)

line = "SC has first two presumptive cases of coronavirus, DHEC confirms https://postandcourier.com/health/covid19/sc-has-first-two-presumptive-cases-of-coronavirus-dhec-confirms/article_bddfe4ae-5fd3-11ea-9ce4-5f495366cee6.html?utm_medium=social&utm_source=twitter&utm_campaign=user-share… via @postandcourier"

input_ids = torch.tensor([tokenizer.encode(line)])
Downloads last month
187,642
Hosted inference API
Fill-Mask
Mask token: <mask>
This model can be loaded on the Inference API on-demand.