Santosh Gupta

If you would like to fork your own version, there's three lines where you have to change '1-' to '~' I made these changes to my own fork https://github.com/Santosh-Gupta/BertSum/blob/master/src/models/encoder.py https://github.com/Santosh-Gupta/BertSum/blob/master/src/models/data_loader.py...

yes

Thanks, looking forward to it! The reason why I am interested in doing it in Numpy is that Keras/Tensorflow isn't very great for sparse training. My use case is training...

It looks like this is exactly what I was looking for. I'm not familiar with a lot of computer science terms, but it uses sqlite as the datastore, so I'm...

Looks like you can't use it for training, oh well https://github.com/plasticityai/magnitude/issues/32

But looking at the code gives me a strong idea of how I could implement it in this library, by using SQLite database as the vector store, and just copying...

Sounds good!

Interesting. I am assuming that the full computation graph is already built for training and inference. Couldn't you save the computation graph then? Why do you need _set_inputs

> Try the api : https://pocket-derm.firebaseapp.com/ Pretty neat but it looks like it costs money. How did you make this?

I am unable test these yet but in the meanwhile try instantiating the model object and then trying to save it `thisModel = doc.qa_embed.model` `thisModel.save("model.h5")` let me know what the...