Clojurians
#clojure-nlp
<
2018-06-08
>

This page is not created by, affiliated with, or supported by Slack Technologies, Inc.

rustam.gilaztdinov07:06:29

@d4hines No, flare is dead. Take a look at here -- https://github.com/gigasquid/clojure-mxnet

novel09:06:17

@rustam.gilaztdinov why do consider flare to be dead? the commits are as recent as end of march.

rustam.gilaztdinov09:06:04

And what about last commits? And this library doesn't support cuda. What's more important -- entire library is really OOP driven, not about data. Look at examples in clojure-mxnet, trust me, this is really what you looking for

novel09:06:44

looking now at clojure-mxnet and you’re right it looks indeed nice. But this doesn’t mean flare is dead. also I thought cuda is supported by flare via neanderthal.

rustam.gilaztdinov09:06:59

Neanderthal support cuda, you right. But flare — doesn’t. Why do I decide that flare is dead? Because of I opened issues, and nobody responds to them. No updates, no answers -- is that a sign of dead library? I think, yes.

novel09:06:55

I don’t see any open issues for flare.

rustam.gilaztdinov09:06:30

Cause I close them.

novel09:06:59

and thanks for pointing towards clojure-mxnet, will give it a try

d4hines12:06:03

@rustam.gilaztdinov, thanks for the suggestion. I'll check that out.

d4hines13:06:27

However, I wonder if I'm asking the right question. The examples I've seen of deep-learning chatbots are fascinating, but the responses are always nonsensical, and far removed from anything I would dare put before a customer. For example, http://neuralconvo.huggingface.co/. I'd very much appreciate the opinions/guidance of smart people like ya'll, but is this an appropriate channel for very chatbot-specific questions?