You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now, we report several MRR measures but it seems the important ones are not well presented:
(i) The Mb model file contains MRR measure that is irrelevant to the Mb model but is MRR of the Mb+clues logregression, is that right? Instead, please take another look at our answer scoring model files. They report cross-validation results, performance on the whole train set, and parameters used when training the model (like the mythical learning rate parameter in our case). Can we get that, please?
(ii) I added print('Mb MRR on train:', mrr(M,b,trainlist)). Can you confirm if this make sense, is this value really Mb MRR on whole train?
The text was updated successfully, but these errors were encountered:
Right now, we report several MRR measures but it seems the important ones are not well presented:
(i) The Mb model file contains MRR measure that is irrelevant to the Mb model but is MRR of the Mb+clues logregression, is that right? Instead, please take another look at our answer scoring model files. They report cross-validation results, performance on the whole train set, and parameters used when training the model (like the mythical learning rate parameter in our case). Can we get that, please?
(ii) I added print('Mb MRR on train:', mrr(M,b,trainlist)). Can you confirm if this make sense, is this value really Mb MRR on whole train?
The text was updated successfully, but these errors were encountered: