Andriy Mulyar

The implementation works only on a Linux machine. On Fri, Mar 13, 2020, 1:24 AM David Rodriguez wrote: > I have this exacto problem could you get this running? >...

No sorry. On Thu, Dec 3, 2020, 3:29 AM jaehong-park-78 wrote: > Can you make it work on Windows platform? > > — > You are receiving this because you...

The [0,5] is a standard annotation range for STS tasks with 0 representing complete dissimilarity and 5 total similarity. On Wed, Jan 8, 2020, 4:46 PM Chandrak1907 wrote: > How...

Please contact the official organizers in regards to obtaining data. On Thu, Mar 12, 2020, 7:40 AM WinnaYuan wrote: > Hi, > > I attend the 2019 n2c2 task3, and...

Fine-tune on your task specific data. Best of luck! On Sun, Jan 12, 2020, 9:05 PM Chandrak1907 wrote: > I used this model on quora qa data set ( >...

If you would like to implement and merge I will happily accept!

The parameters are shared in the clinical model download. Look in your ~/.cache directory. Andriy On Mon, Dec 30, 2019, 3:26 AM WinnaYuan wrote: > Hi, can you share all...

I never got around to publishing the training parameters due to not having time to clean it up. The training parameters were identical to the pytorch transformers sentence pair CLS...

This happens because switching the sentence order will influence which pre-trained attention weights each token enters. In theory, (semantic) similarity should be symmetric - the way BERT works makes this...

Currently pull request #68 begins preliminary work mentioned above. ### Ideas for further improvements: 1) Method in `Dataset` that will allow to compare gold and predicted over a whole corpus...