verissimomanoel/RobertaTwitterBR

This is a RoBERTa Twitter in Portuguese model trained on ~7M tweets . The results will be posted in the future . The model has been trained on 7M tweets and is trained on tweets from around 7,000 tweets . It is based on the data from Twitter RoberTa in Portuguese and the data was collected from 7M Tweets . The result will be published in the next few days and the results are expected to be posted later this week . It's a good idea to train a model on Twitter using a tokenizer or a model based on tweets sent by the model . The tokenizer was trained on over 7 million tweets and the model is now trained on these tweets .,