Sam Havens

Is the party line position that the way to make this happen is to write a JDBC driver in Clojure? EDIT: Looks like there is a PR, #5439, which is...

@gilbsgilbs to the second point, how can we help? I can't imagine the testing usage costing more than a few pennies a month. I feel like a one time payment...

Thanks for the link, I will check it out! For tokenizing and predicting, I was planning on doing something like the German ULMFiT (there's a link somewhere on the fast.ai...

I thought tokenizing that way would let me get away with not using spaCy...

Yeah so the hard part is that `get_rnn_classifier` transitively pulls in a lot of stuff. Like, way too much to fit into a Lambda function... I couldn't find anything like...

I have never used Travis, but there is probably a way to set it up to run this automatically when a PR is accepted into `master`

@nimf Is there a way to have Travis run the build script when a PR is accepted into master? Should that change (whatever it is) be included here?

As far as qualitative eval, we have had users comment that the chat version of the model is particularly good at translation for a 7B model, at least Romance languages

@vince62s I recommend testing it. We've been very surprised with its multilingual abilities

@batselm in the notebook, about halfway down you see `from process_data import vectorize_kv`. So I looked in `process_data.py` and found it on [line 190](https://github.com/jojonki/key-value-memory-networks/blob/master/process_data.py#L190).