linkedin / detext

DeText: A Deep Neural Text Understanding Framework for Ranking and Classification Tasks
BSD 2-Clause "Simplified" License
1.26k stars 134 forks source link

Documentation? #33

Closed multinucliated closed 4 years ago

multinucliated commented 4 years ago

I can't find much documentation about it.

sbourke commented 4 years ago

It could be useful to have some examples on open datasets to build a model run from start to finish. I'm going through the code to work out what does what.

But generally looks really interesting.

multinucliated commented 4 years ago

@sbourke that's correct, even we can submit or provide sample codes if we have something to start with.

xwli-chelsea commented 4 years ago

Thanks for the comments @multinucliated @sbourke ! Pease refer to the TRAINING.md for details on the training data format/preparation and training params.

We are working on to add some detailed hands-on tutorial notebooks with some public datasets to show users end-to-end training examples including data preparation. Please stay tuned! :)

vlasvlasvlas commented 4 years ago

this, is important. thanks!

StarWang commented 4 years ago

Hi guys, we've hosted a tutorial on how to use DeText in a KDD workshop. Helpful materials (notebooks, slides, etc.) can be found at the workshop page

xwli-chelsea commented 4 years ago

Closing this as examples and references are added. Thanks!