1 paper
Abhinav Suresh, Henning Schlömer, Baran Hashemi +1
Due to their inherent capabilities of capturing non-local dependencies, Transformer neural networks have quickly been established as the paradigmatic architecture for large languag…