SigmaWe/DocAsRef

Dump results

forrestbao opened this issue · 1 comments

forrestbao commented
  1. Additional metrics to test
    • [high priority] BERTScore-sentence, MNLI, {RoBERTa, DeBERTA}, Entail - Contradict, top-k and top-p
    • [medium] BERTScore-original, DeBERTa -- to see how much language models impact it
    • [low] BERTScore-sentence, cosine, DeBERTa -- again, to see the impact of language models
  2. how to print result into Google Sheets
    Done SigmaWe/EvalBase#4
forrestbao commented

Please run BERTScore sentence, non-MNLI just simple cosine, but with sentence weighting when you have time. @TURX