...and I asked Jeremy Howard on LinkedIn if we could use the fast.ai NLP course (including a github issue for a license), and this is what he said:
Feel free to use stuff from the NLP course - just provide appropriate credit (would be nice if you told people to come check out fast.ai's other courses too!) Not exactly a license. But at least permission. To be clear, he asks for contact if you want to use them commercially: https://github.com/fastai/course-nlp/issues/17 Anyway, I packaged both Rochelle's and Fast.ai's courses in a Gigantum project that is good for RStudio and Jupyter. If you're curious, you can quickly import it by logging into this cloud server: https://try.gigantum.com Then import: gigantum.com/tinydav/nlp-workshop-expo If you want to do serious work, it makes more sense to set up Docker lcoally. Also let me know if you want write access on my project - otherwise, you'll end up with your own copy. I've also set up a GitHub mirror that I can push to if folks want. It's not currently up to date, but please ping me if you want that updated (it's updated as of right now): https://github.com/davclark/nlp-workshop-expo A final point is that (I think?) Rochelle's course doesn't use any GPU-accelerated libraries. The Fast.ai course does - but you'll need to switch the base in gigantum to a cuda-enabled one if you want to use a GPU - maybe in a branch? We don't currently have any cuda bases for R (hasn't come up), but if anyone wants that, lemme know. Best, Dav On Tue, Jul 23, 2019 at 11:53 AM Dav Clark <davcl...@gmail.com> wrote: > I just dropped this issue asking about license preference: > https://github.com/rochelleterman/FSUtext/issues/1 > > What's the policy on Carpentries lessons borrowing from outside projects > with various licenses or no license specified? Or, in a more limited scope: > what's the plan for license for one or more NLP lessons? > > Best, > D > > On Fri, Jul 19, 2019 at 2:27 PM Rochelle Terman <rter...@gmail.com> wrote: > >> Hi all, >> >> If anyone is interested, I put together a Software Carpentry-style 'text >> as data' workshop for Social Scientists. Latest version here. >> <https://github.com/rochelleterman/FSUtext> It uses R. Feel free to >> adapt. >> >> Cheers, Rochelle >> >> On Fri, Jul 19, 2019 at 3:54 PM Dav Clark <davcl...@gmail.com> wrote: >> >>> I am starting to think about my own text-oriented workshop for social >>> scientists that would use Gigantum to help support beginners thinking more >>> about big-picture aspects of data projects. From there, we'd transition to >>> explain how you could do similar things "by hand" (e.g. using git + LFS >>> directly). >>> >>> Ziyaad did you ever create an NLP working group? In any case, I'm happy >>> to contribute to core carpentry-style materials (and I can easily adapt >>> them to my own purposes later). And of course, if anyone wants to do stuff >>> with Gigantum, I'm excited to help with that! But my sense is that this >>> community is pretty committed to the command line ;) >>> >>> In terms of concepts, I'm interested in things that will help learners >>> grok general ideas. So, understanding the transformation of texts to a >>> matrix of frequency counts, perhaps some basics of what a matrix is, and >>> what you can do with linear systems, and perhaps also the idea of an >>> abstract "space" like word2vec. >>> >>> And as far as tools, I'm happy to do R or Python (and use RStudio or >>> Jupyter). I see one vote for Spacy - so if that's a way to get someone >>> on-board that sounds good to me! I've not done NLP for a few years, and >>> spacy looks like magical python easy (cf. XKCD <https://xkcd.com/353/>). >>> >>> Best, >>> Dav >>> >>> On Fri, Jul 19, 2019 at 9:11 AM <ba...@staffmail.ed.ac.uk> wrote: >>> >>>> Hi all, >>>> >>>> just to let you know ... I'm planning to prepare a carpentries style >>>> course on text mining... which is a type of NLP (i.e. Natural Language >>>> Processing). :-) >>>> >>>> Bea Alex (Edinburgh) >>>> >> >> -- >> Rochelle Terman, Ph.D. >> Provost Postdoctoral Fellow >> Department of Political Science >> University of Chicago >> >> http://www.rochelleterman.com >> *The Carpentries <https://carpentries.topicbox.com/latest>* / discuss / >> see discussions <https://carpentries.topicbox.com/groups/discuss> + >> participants <https://carpentries.topicbox.com/groups/discuss/members> + >> delivery >> options <https://carpentries.topicbox.com/groups/discuss/subscription> >> Permalink >> <https://carpentries.topicbox.com/groups/discuss/Tee8f460df4206334-M57efa7254ba1e0cb07d357da> >> ------------------------------------------ The Carpentries: discuss Permalink: https://carpentries.topicbox.com/groups/discuss/T8155b8ebe88ca076-M1fc70b6f300b9d220f740152 Delivery options: https://carpentries.topicbox.com/groups/discuss/subscription
