Welcome!
This is the community forum for my apps Pythonista and Editorial.
For individual support questions, you can also send an email. If you have a very short question or just want to say hello — I'm @olemoritz on Twitter.
Feature request - machine learning
-
Hi
First off... Pythonista is amazing!
Was wondering if there are any plans to add any machine learning modules? I’m trying to learn machine learning using skleqen on my laptop. Would be great to do this on my iOS devices too, when I’m away from home. As more and more people are becoming interested in AI , I think this would put pythonista at the top if the added machine learning modules. -
you can use tensorslow in pythonista
-
@JonB i guess that is a pun? Or is there really a tensorflow implementation in pure python (hence the tensor-slow). Just to be sure.
-
@JonB what ?? Tensorflow can work on Pythonista ? How did you manage to make it working ? I tried with stash but C-based modules were needed....
-
@biob I already done multiple Deep Learning experiences with Pythonista (numpy), I can show you some of them if you want to see how it works, just mp me !
-
-
nice :-)
-
@JonB said:
An alternative might be to create a webview and run TF.js there. Theoretically this will allow the ML to still run on GPU. I have not tested this though...
-
@biob probably not what you're looking for but omz posted a demo of using CoreML for image classification: https://gist.github.com/omz/a7c5f310e1c8b829a5a613cd556863d4
You could certainly play around with more CoreML libraries that way.. probably not the easiest route for learning though.
I think your best bet for now is to stick with numpy.. you'll have to do more of your own implementation that way than using something like sklearn but you'll probably end up understanding more in the end. Do you have any end goals in mind?
-
Tensorslow seems to be a pure numpy implementation of tensorflow, for whatever that is worth. There is an associated blog where they basically develop it from scratch, teaching you the basics of neural networks on the way.