Which one is your best model?

[sub1120]

Hey i am working on a run on correction problem, which is similar to punctuation (PERIOD only) restoration task, so i wanted to compare your tool performance with my approach, since you have trained so many models, i am confused which model should i use from this paper. Please help me out!

[ottokart]

Hi, the two best models are T-BRNN and T-BRNN-pre. T-BRNN does not use any external datasets, T-BRNN-pre uses pretrained embeddings and is slightly better.

[sub1120]

Thank you

On Tue 1 Sep, 2020, 7:27 PM Ottokar Tilk, [email protected] wrote:

Hi, the two best models are T-BRNN and T-BRNN-pre. T-BRNN does not use any external datasets, T-BRNN-pre uses pretrained embeddings and is slightly better.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ottokart/punctuator2/issues/65#issuecomment-684873026, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKOB6NFLATYJE2ADKXZFX63SDT4TPANCNFSM4QKSOC3Q .