I was interested in updating Flux’s model zoo since everything I tried is out of date. But I saw that there are a lot of pull requests which are outstanding.
Is it still maintained? Did I miss an announcement on that?
Thanks.
It’s still maintained (in the sense that we want to keep it going) but yes, we’re a bit behind on the review effort.
One excellent way to help here, if you’re up for it, is to review existing PRs. Even for those who can’t ultimately merge, it makes things quicker for those who can, and if a PR looks like it’s ready to go you can ping me or the other Flux committers to make sure it doesn’t get forgotten.
If folks want to help with this kind of thing, please don’t hesitate to reach out!
6 Likes
OK. Thanks.
I also found this quite discouraging, and nearly gave up on Julia.
For both Flux and Knet, the examples I tried were all out of date and had various errors that prevented them from running. I imagine I am not the only person who gets discouraged, probably the julia ML community would be larger if all the examples worked.
If it is not possible to keep the examples up to date, perhaps they could simply say in the file or README which version they require.
I am tentatively glad that I did not give up, have started to understand how to make things work.
@jaynick I felt the same way when I was exploring Julia & the Flux/Knet examples didn’t work well.
We discuss this explicitly here: Custom XGBoost Loss function w/ Zygote. Julia Computing blog post
I think the solution would be to have a Julia page w/ ML tutorials maintained by a volunteer.
I’m gonna volunteer to start this & can commit to checking all the tutorials @ least 1x per month.
4 Likes
Excellent.
May I suggest that you also periodically announce this (including a call for volunteers to help.)
(Checking examples would be a great activity at my level at some point…, though for this month I’m over-extended trying to come to speed on julia while still keeping my day job in those other languages)
2 Likes
Given the post by Chris https://discourse.julialang.org/t/zygote-error-only-reference-types-can-be-differentiated-with-params/38224/5, it seems that a SciML zoo would be more appropriate.
I’d interested in participating.
This is also of interest for inclusion: https://github.com/GdMacmillan/ml_flux_tutorial on how to use Julia on Google Colab.