Word embeddings

RoBERTa

Robustly Optimized BERT Pretraining Approach, improved version of BERT with longer pre-training on more data and optimized hyperparameters.

← Indietro