Pre-training Text Representations
method
ID: pretrainingtextrepresentations
Related Papers
CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data
AbstractPre-training text representations have led to significant improvements in many areas of natural language processing. The quality of these models benefits greatly from the size of the pretraining corpora as …...