search

ORNL / icat

Interactive machine learning dashboard for textual data exploration
https://ornl.github.io/icat/
BSD 3-Clause "New" or "Revised" License
4 stars 0 forks source link

JOSS review: include intermediate example notebook #46

Closed WarmCyan closed 3 weeks ago

WarmCyan commented 2 months ago

Added from @jhagerer https://github.com/openjournals/joss-reviews/issues/6873#issuecomment-2299852874: We go from a trivial "simple_example" notebook that's just expecting default anchors all the way to "lm_similarity_example" which adds a BERT transformer-based anchor (which has quite a few dependencies to get working.) It may be helpful to include an intermediate example that shows adding a custom anchor but doesn't require transformers to work. (word2vec was suggested as a possibility.)

jhagerer commented 2 months ago

@WarmCyan I would be happy with any example which does not need cuda/nvidia and gives me as a reviewer more to see and explore than the simple example.

A quick win solution could be to just change lm_similarity_example.ipynb such that it does not use cuda anymore. How about that?

WarmCyan commented 1 month ago

@jhagerer good suggestion, I modified that notebook to first check if cuda is available, and use cpu otherwise.

jhagerer commented 3 weeks ago

OK, I think I have something running now to work with.

Please close the ticket. I'll take some time next weekend to try out and understand the tool more in depth.