Re: [scikit-learn] Sprint discussion points?

What do you think needs to be raised for discussion? On Tue., 26 Feb. 2019, 12:06 pm Jeremie du Boisberranger, < jeremie.du-boisberran...@inria.fr> wrote: > Not the same, although there are similarities. However asv provides > tools to compare benchmarks across commits, and to publish them in

Re: [scikit-learn] Sprint discussion points?

Not the same, although there are similarities. However asv provides tools to compare benchmarks across commits, and to publish them in html format to follow their evolution through time such as https://pv.github.io/numpy-bench/ Here's the link of the benchmark suite :

Re: [scikit-learn] Sprint discussion points?

Was that the same that Vlad used? https://github.com/scikit-learn/scikit-learn-speed We might want to just replace that, given that it hasn't been touched in 7 years? On 2/26/19 5:22 AM, Jeremie du Boisberranger wrote: I totally forgot to mention it before the sprint started but i'd like

Re: [scikit-learn] Sprint discussion points?

I totally forgot to mention it before the sprint started but i'd like to have a discussion about the integration of a new benchmark suite into the scikit-learn organization. Essentially, I've been working on a benchmark suite for sklearn using the airspeed velocity (asv) framework. The

Re: [scikit-learn] Sprint discussion points?

I'm all for the decorator if you can get numpydoc working with it! ___ scikit-learn mailing list scikit-learn@python.org https://mail.python.org/mailman/listinfo/scikit-learn

Re: [scikit-learn] Sprint discussion points?

One other topic that I kind of forgot about are keyword-only arguments. I like the idea of the decorator that I proposed but Joel didn't like it, I think ;) We might want to think about other Python3 features like type annotations as well. On 2/20/19 4:40 PM, Gael Varoquaux wrote: On Tue,

Re: [scikit-learn] Sprint discussion points?

Something else worth discussing might be the maintenance of scikit-learn-contrib ___ scikit-learn mailing list scikit-learn@python.org https://mail.python.org/mailman/listinfo/scikit-learn

Re: [scikit-learn] Sprint discussion points?

you need to consider my time. I'll comment on relevant issues if I have any thoughts. Hanmin Qin - Original Message - From: Joel Nothman To: Scikit-learn user and developer mailing list , Hanmin Qin Subject: Re: [scikit-learn] Sprint discussion points? Date: 2019-02-21 15:40 @Hanmin

Re: [scikit-learn] Sprint discussion points?

@Hanmin are there particular conversations you are keen to take part in, and particular times that suit you? On Thu., 21 Feb. 2019, 9:13 am Andreas Mueller, wrote: > > > On 2/20/19 4:40 PM, Gael Varoquaux wrote: > > On Tue, Feb 19, 2019 at 06:16:20PM -0500, Andreas Mueller wrote: > >> I put a

Re: [scikit-learn] Sprint discussion points?

On 2/20/19 4:40 PM, Gael Varoquaux wrote: On Tue, Feb 19, 2019 at 06:16:20PM -0500, Andreas Mueller wrote: I put a draft schedule here: https://github.com/scikit-learn/scikit-learn/wiki/Upcoming-events#technical-discussions-schedule I'd like to discuss sample_props. They are important to

Re: [scikit-learn] Sprint discussion points?

On Tue, Feb 19, 2019 at 06:16:20PM -0500, Andreas Mueller wrote: > I put a draft schedule here: > https://github.com/scikit-learn/scikit-learn/wiki/Upcoming-events#technical-discussions-schedule I'd like to discuss sample_props. They are important to me. Should I add them somewhere on the

Re: [scikit-learn] Sprint discussion points?

I messaged them and also tasks Thomas Fan with working with Microsoft to set up azure pipelines. On 2/20/19 12:12 PM, Guillaume Lemaître wrote: @Andy You were the one contacting Travis. On Wed, 20 Feb 2019 at 17:23, Andreas Mueller > wrote: Thanks for bringing

Re: [scikit-learn] Sprint discussion points?

@Andy You were the one contacting Travis. On Wed, 20 Feb 2019 at 17:23, Andreas Mueller wrote: > Thanks for bringing that up. > Did I email travis last time? > > We should also follow up with Microsoft as they promised unlimited > builds... > > > On 2/20/19 3:48 AM, Alexandre Gramfort wrote: >

Re: [scikit-learn] Sprint discussion points?

Sure, we can change it up on Tuesday. I agree having things that we can implement during the week would be good. I was actually kind of optimistic and was hoping we could make some dent into the freezing, and the convergence issues might be less controversial and more a technical challenge. I

Re: [scikit-learn] Sprint discussion points?

Thanks for bringing that up. Did I email travis last time? We should also follow up with Microsoft as they promised unlimited builds... On 2/20/19 3:48 AM, Alexandre Gramfort wrote: we should also see if we can have a lot of CI machines for the 5 days as it's always the blocker to move fast

Re: [scikit-learn] Sprint discussion points?

we should also see if we can have a lot of CI machines for the 5 days as it's always the blocker to move fast during 1 week. my 2c Alex On Wed, Feb 20, 2019 at 7:35 AM Roman Yurchak via scikit-learn wrote: > > Thanks for putting the draft schedule together! > > Personally I will be there 3 days

Re: [scikit-learn] Sprint discussion points?

Thanks for putting the draft schedule together! Personally I will be there 3 days out of 5 and wouldn't want to miss the discussion on euclidean distance issues. Maybe we could adjust the schedule during the sprint (say on Tuesday) based on people's interest and availability? That might be

Re: [scikit-learn] Sprint discussion points?

I don't think I'll be able to stay for the Friday 10am discussion, but have a PR open on "efficient grid search" so should probably be involved. Perhaps the fit_transform discussion can happen without you, Andy? On Wed, 20 Feb 2019 at 10:17, Andreas Mueller wrote: > I put a draft schedule

Re: [scikit-learn] Sprint discussion points?

I put a draft schedule here: https://github.com/scikit-learn/scikit-learn/wiki/Upcoming-events#technical-discussions-schedule it's obviously somewhat opinionated ;) Happy to reprioritize. Basically I wouldn't like to miss any of the big API discussions because coming late to the party. The

Re: [scikit-learn] Sprint discussion points?

> Not sure if Guillaume had ideas about the schedule, given that he seems to be running the show? Mostly running behind the show ... For the moment, we only have a 30 minutes presentation of introduction planned on Monday. For the rest of the week, this is pretty much opened and I think that we

Re: [scikit-learn] Sprint discussion points?

Yeah, sounds good. I didn't want to unilaterally post a schedule, but doing some google form or similar seems a bit heavy-handed? Not sure if Guillaume had ideas about the schedule, given that he seems to be running the show? On 2/19/19 4:17 PM, Joel Nothman wrote: I don't think optics

Re: [scikit-learn] Sprint discussion points?

I don't think optics requires a large meeting, just a few people. I'm happy with your proposal generally, Andy. Do we schedule specific topics at this point? ___ scikit-learn mailing list scikit-learn@python.org

Re: [scikit-learn] Sprint discussion points?

> Does OPTICS require a meeting or is it clear what to do and the work > "just" needs to be done? Definitely needs (some) discussions. On Tue, Feb 19, 2019, 18:25 Andreas Mueller > > On 2/14/19 11:40 AM, Nicolas Hug wrote: > > > >> or we could go as far as to schedule meetings on the different

Re: [scikit-learn] Sprint discussion points?

On 2/14/19 11:40 AM, Nicolas Hug wrote: or we could go as far as to schedule meetings on the different topics. Given the number of issues to discuss this is probably the best approach IMO If we want to schedule meetings we could do one of two things: have a scheduling meeting first

Re: [scikit-learn] Sprint discussion points?

On 2/18/19 3:06 PM, Joel Nothman wrote: And here I was thinking we'd better just push out 0.20.3 this week with what's been listed for it. I wouldn't mind this, just don't expect me to help ;) ___ scikit-learn mailing list scikit-learn@python.org

Re: [scikit-learn] Sprint discussion points?

And here I was thinking we'd better just push out 0.20.3 this week with what's been listed for it. ___ scikit-learn mailing list scikit-learn@python.org https://mail.python.org/mailman/listinfo/scikit-learn

Re: [scikit-learn] Sprint discussion points?

seems not trivial (e.g., #13142 related to BaseMixture and #13124 related to StratifiedKFold). Hanmin Qin - Original Message - From: Olivier Grisel To: Scikit-learn mailing list Subject: Re: [scikit-learn] Sprint discussion points? Date: 2019-02-15 23:06 I would also add generalizing

Re: [scikit-learn] Sprint discussion points?

I would also add generalizing early stopping options to most estimators. This is a bit related to Joel's point on max_iter consistency in LogisticRegression. -- Olivier ___ scikit-learn mailing list scikit-learn@python.org

Re: [scikit-learn] Sprint discussion points?

or we could go as far as to schedule meetings on the different topics. Given the number of issues to discuss this is probably the best approach IMO On 2/14/19 8:31 AM, Andreas Mueller wrote: As I said, I think it's too much and we need to prioritize. We could either rank issues and start

Re: [scikit-learn] Sprint discussion points?

As I said, I think it's too much and we need to prioritize. We could either rank issues and start with some and see how far we get, or we could go as far as to schedule meetings on the different topics. Also, I'll be only arriving Tuesday late morning, I think. On 2/14/19 8:05 AM, Adrin

Re: [scikit-learn] Sprint discussion points?

On 2/13/19 11:28 PM, Joel Nothman wrote: Convergence in logistic regression (https://github.com/scikit-learn/scikit-learn/issues/11536) is indeed one problem (and it presents a general issue of what max_iter means when you have several solvers, or how good defaults are selected). But I was

Re: [scikit-learn] Sprint discussion points?

I've been working on some bias mitigation metrics and methods and that usecase changes the data as well as up/down sampling as a transformer. Almost all those methods also need sample properties for the observations to work. I'm trying to make them "sklearn compatible", but for now it's pretty

Re: [scikit-learn] Sprint discussion points?

I am really interested in the union of the list given by Andy and Joel. I'll like to have some discussions related to the "impute" module. Compare to the other topics, it is not a high priority discussion thought. On Thu, 14 Feb 2019 at 05:31, Joel Nothman wrote: > Convergence in logistic

Re: [scikit-learn] Sprint discussion points?

Convergence in logistic regression ( https://github.com/scikit-learn/scikit-learn/issues/11536) is indeed one problem (and it presents a general issue of what max_iter means when you have several solvers, or how good defaults are selected). But I was sure we had problems with non-determinism on

Re: [scikit-learn] Sprint discussion points?

Do you have a reference for the logistic regression stability? Is it convergence warnings? Happy to discuss the other two issues, though I feel they seem easier than most of what's on my list. I have no idea what's going on with OPTICS tbh, and I'll leave it up to you and the others to

Re: [scikit-learn] Sprint discussion points?

Yes, I was thinking the same. I think there are some other core issues to solve, such as: * euclidean_distances numerical issues * commitment to ARM testing and debugging * logistic regression stability We should also nut out OPTICS issues or remove it from 0.21. I'm still keen on trying to work

[scikit-learn] Sprint discussion points?

Hey all. Should we collect some discussion points for the sprint? There's an unusual amount of core-devs present and I think we should seize the opportunity. Maybe we should create a page in the wiki or add it to the sprint page? Things that are high on my list of priorities are: * slicing