Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
We'll use the CoLA dataset from the GLUE benchmark,
since it's a simple binary text classification task, and just take the training split for now.