Fork me on GitHub
#data-science
<
2020-01-11
>
blueberry00:01:24

@neo2551 Although I don't use Windows, Neanderthal worked for me on Windows when I tried it there without any special GPU setup other than installing GPU drivers and CUDA, which is provided by Nvidia as an automatic installer. For AMD/OpenCL it was even easier: as long as the drivers were installed (a click-through process), everything was available to Clojure/Neanderthal. Of course, YMMV, but there was no need for fiddling with the environment or any special setup in my case.

gigasquid02:01:18

I wrote a quick blog on the hugging face gpt2 interop stuff https://gigasquidsoftware.com/

👀 12
metal 8
👏 12
Daniel Slutsky15:01:13

Clojurists Together is collecting proposals for projects to fund in the next 3 months. https://www.clojuriststogether.org/news/q1-2020-survey-results/

The call for proposals for new projects will close on Wednesday, January 15th, 2019 at 11:59pm PST.

fabrao17:01:08

Hello, anyone could run libpython-clj in Windows? This asking about python3.6m with python3.6.m.dll?

fabrao17:01:30

what is this m version?}

orestis18:01:10

Oh my, I started the spaCy tutorial yesterday - I used Python for 10 years but now that I got used to the REPL, it was super frustrating to go back to typing into a file, then running the code from the command line. Not to mention that for every section that involved loading the medium English dataset, I had to wait 15 seconds, every single time!

ben09:01:48

ipython has an enhanced repl, which I would recommend

ben09:01:22

It’s not quite as nice as clj, but good enough for most things

orestis18:01:30

I can see the value of doing the interop with libpython-clj now.

David Pham19:01:13

@orestis you also have a repl in python (less powerful) or you can try Jupiter

orestis19:01:45

Yeah I’m not defining functions in Pythons repl :) Jupiter is on my list of things to try.

David Pham07:01:32

You can define in your script and emacs evaluate them in you repl

chrisn22:01:29

@fabrao - Good question. That is the standard python shared library used by conda and cpython distributions. I imagine it has the same name on windows.