Language Model Pre-Training

BERT-large-uncased fine-tuned on the following numbers of vaccine-related tweets :

• 500K
• 2M (de-duplicated from 4.7M)
• 3.6M (de-duplicated from 10.9M)

RoBERTa

• 20M (de-duplicated from 64M full dataset)

CT BERT

• 20M

Vaccine Sentiment Dataset

Use of fine-tuned models for classification on the Vaccine Sentiment dataset

UN Vaccine Dataset

Use of fine-tuned models for classification on the UN Vaccine dataset