Fork me on GitHub
#data-science
<
2017-08-30
>
gdeer8119:08:24

wow, you're going to write a library and a conj talk in two months? you're going to be a busy boy

bpiel19:08:11

well. I started a few months ago, but yes, I will continue to be a busy boy

gdeer8119:08:52

There are a couple of people at work that are dabbling in tensor flow for product recognition and things like that, I'll reach out to them and see if they'd be willing to try to port their projects to clojure to get some concrete feedback. their main concern will probably be that they don't know clojure, but we'll see how it goes

bpiel19:08:45

are they working in python?

bpiel19:08:10

I wouldn't ask anyone to port their TF work to clojure at this point.

gdeer8119:08:46

I think they are in python

gdeer8119:08:16

but probably in java

gdeer8119:08:36

I'll ask them before I say anything else

bpiel19:08:51

ok thanks

bpiel19:08:01

I'd have similar questions for anyone who'd like to use TF with some other language (whatever they prefer) than python: what would convince them to try switching to the non-python library?

bpiel19:08:27

I've gotten a few good response so far, but would love to get many more

gdeer8119:08:37

is your project going to tie into Sayid so we can debug tensor flow in emacs?

bpiel19:08:36

ha, no. The population who'd be interested in clojure+tf is narrow enough. I don't need to eliminate non-emacsers as well.

bpiel19:08:58

BUT, I am thinking a lot about debugging tools (of course)

gdeer8119:08:12

you're a good man Mr. Piel, I'm gonna miss you at the conj this year

bpiel19:08:49

ha thanks. I try. Too bad you can't make it!