Mailman 3 May 2020 - scikit-learn

major league hacking summer internship program
by Thomas J Fan June 1, 2020

June 1, 2020

I can commit to reviewing. Diving into their program, it looks like they are hiring supervisers through: https://raise.dev/Apply/?ref=mlh which is titled "Software Developer Coach". By looking at their https://fellowship.mlh.io/students they have about 9 weeks of actual contributing. Given they have an engineer to help, maybe they can work on the documenting the production aspects: 1. Roadmap item 19: Documentation and tooling for model lifecycle management 2. Roadmap item 21: Document good practices to detect temporal distrubiton drift Regards, Thomas > On Thursday, May 28, 2020 at 5:36 PM, Andreas C. Mueller <andreasmuellerml(a)gmail.com (mailto:andreasmuellerml@gmail.com)> wrote: > > Hi Folks. > > > So this program sounds pretty cool. They preselected some people for an ML work group, who will be doing daily standups together and pair programming, > and who might move around between some related projects over the 12 weeks of the program. > > > They made sure to get a diverse set of students and they have an engineer that will supervised them. > > > They would probably have 2-3 students working on sklearn. > > > They don’t expect having one big feature but they do expect some guidance on what issues to work on. > > > Also, the program starts on Monday, and they start contributing to OSS projects about a week after that. > > > Ideally we’d tell them if we’re in or not before Monday, and have a tentative list of issues / projects. > > > > > > What do you all think? Also, if we want to do it, who would have cycles for some reviewing? > > > This seems to be well organized and they seem to have put quite some thought into it, but we do need to do a little bit of work on our end. > > > I can try picking some issues but I probably can’t commit a lot of reviewing time. > > > > > > Cheers, > > > Andy > > > > > > > > > From: scikit-learn <scikit-learn-bounces+t3kcit=gmail.com(a)python.org> On Behalf Of Adrin > Sent: Tuesday, May 19, 2020 3:42 PM > To: Scikit-learn mailing list <scikit-learn(a)python.org> > Subject: Re: [scikit-learn] major league hacking summer internship program > > > > > > > > Sounds pretty cool to me. > > > > > > > > On Tue, May 19, 2020 at 6:45 PM <t3kcit(a)gmail.com (mailto:t3kcit@gmail.com)> wrote: > > > > > > Hey Folks. > > > > > > This program reached out to me: > > > > > > https://news.mlh.io/mlh-fellowship-the-future-of-tech-internships-05-04-2020 > > > > > > > > > > > > What do you think? > > > > > > Sounds like GSOC but with extra mentorship, so it might be a good fit for us? > > > > > > I would say it depends on what level of involvement they require from us. > > > > > > > > > > > > Best, > > > > > > Andy > > > > > > > > > > > > > > > > _______________________________________________ > > scikit-learn mailing list > > scikit-learn(a)python.org (mailto:scikit-learn@python.org) > > https://mail.python.org/mailman/listinfo/scikit-learn > > > > > > > _______________________________________________ > scikit-learn mailing list > scikit-learn(a)python.org > https://mail.python.org/mailman/listinfo/scikit-learn

3 2

[GridSearchCV] Reduction of elapsed time at the second interation
by Pedro Cardoso May 30, 2020

May 30, 2020

Hello fellows, i am knew at slkearn and I have a question about GridSearchCV: I am running the following code at a jupyter notebook : ----------------------*code*------------------------------- opt_models = dict() for feature in [features1, features2, features3, features4]: cmb = CMB(x_train, y_train, x_test, y_test, feature) cmb.fit() cmb.predict() opt_models[str(feature)]=cmb.get_best_model() ------------------------------------------------------- The CMB class is just a class that contains different classification models (SVC, decision tree, etc...). When cmb.fit() is running, a gridSearchCV is performed at the SVC model (which is within the cmb instance) in order to tune the hyperparameters C, gamma, and kernel. The SCV model is implemented using the sklearn.svm.SVC class. Here is the output of the first and second iteration of the for loop: ---------------------*output*------------------------------------- -> 1st iteration Fitting 5 folds for each of 12 candidates, totalling 60 fits [Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers. [Parallel(n_jobs=-1)]: Done 1 tasks | elapsed: 6.1s [Parallel(n_jobs=-1)]: Done 2 tasks | elapsed: 6.1s [Parallel(n_jobs=-1)]: Done 3 tasks | elapsed: 6.1s [Parallel(n_jobs=-1)]: Done 4 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 5 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 6 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 7 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 8 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 9 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 10 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 11 tasks | elapsed: 6.2s [Parallel(n_jobs=-1)]: Done 12 tasks | elapsed: 6.3s [Parallel(n_jobs=-1)]: Done 13 tasks | elapsed: 6.3s [Parallel(n_jobs=-1)]: Done 14 tasks | elapsed: 6.3s [Parallel(n_jobs=-1)]: Done 15 tasks | elapsed: 6.4s [Parallel(n_jobs=-1)]: Done 16 tasks | elapsed: 6.4s [Parallel(n_jobs=-1)]: Done 17 tasks | elapsed: 6.4s [Parallel(n_jobs=-1)]: Done 18 tasks | elapsed: 6.4s [Parallel(n_jobs=-1)]: Done 19 tasks | elapsed: 6.5s [Parallel(n_jobs=-1)]: Done 20 tasks | elapsed: 6.5s [Parallel(n_jobs=-1)]: Done 21 tasks | elapsed: 6.5s [Parallel(n_jobs=-1)]: Done 22 tasks | elapsed: 6.6s [Parallel(n_jobs=-1)]: Done 23 tasks | elapsed: 6.7s [Parallel(n_jobs=-1)]: Done 24 tasks | elapsed: 6.7s [Parallel(n_jobs=-1)]: Done 25 tasks | elapsed: 6.7s [Parallel(n_jobs=-1)]: Done 26 tasks | elapsed: 6.8s [Parallel(n_jobs=-1)]: Done 27 tasks | elapsed: 6.8s [Parallel(n_jobs=-1)]: Done 28 tasks | elapsed: 6.9s [Parallel(n_jobs=-1)]: Done 29 tasks | elapsed: 6.9s [Parallel(n_jobs=-1)]: Done 30 tasks | elapsed: 6.9s [Parallel(n_jobs=-1)]: Done 31 tasks | elapsed: 7.0s [Parallel(n_jobs=-1)]: Done 32 tasks | elapsed: 7.0s [Parallel(n_jobs=-1)]: Done 33 tasks | elapsed: 7.0s [Parallel(n_jobs=-1)]: Done 34 tasks | elapsed: 7.0s [Parallel(n_jobs=-1)]: Done 35 tasks | elapsed: 7.1s [Parallel(n_jobs=-1)]: Done 36 tasks | elapsed: 7.1s [Parallel(n_jobs=-1)]: Done 37 tasks | elapsed: 7.2s [Parallel(n_jobs=-1)]: Done 38 tasks | elapsed: 7.2s [Parallel(n_jobs=-1)]: Done 39 tasks | elapsed: 7.2s [Parallel(n_jobs=-1)]: Done 40 tasks | elapsed: 7.2s [Parallel(n_jobs=-1)]: Done 41 tasks | elapsed: 7.3s [Parallel(n_jobs=-1)]: Done 42 tasks | elapsed: 7.3s [Parallel(n_jobs=-1)]: Done 43 tasks | elapsed: 7.3s [Parallel(n_jobs=-1)]: Done 44 tasks | elapsed: 7.4s [Parallel(n_jobs=-1)]: Done 45 tasks | elapsed: 7.4s [Parallel(n_jobs=-1)]: Done 46 tasks | elapsed: 7.5s -> 2nd iteration Fitting 5 folds for each of 12 candidates, totalling 60 fits [Parallel(n_jobs=-1)]: Using backend LokyBackend with 4 concurrent workers. [Parallel(n_jobs=-1)]: Done 1 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Batch computation too fast (0.0260s.) Setting batch_size=14. [Parallel(n_jobs=-1)]: Done 2 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 3 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 4 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 5 tasks | elapsed: 0.0s [Parallel(n_jobs=-1)]: Done 60 out of 60 | elapsed: 0.7s finished --------------------------------------------------------------------------------------------------------------------- As you can see, the first iteration gets a elapsed time much larger than the 2nd iteration. Does it make sense? I am afraid that the model is doing some kind of cache or shortcut from the 1st iteration, and consequently could decrease the model training/performance? I already read the sklearn documentation and I didn't saw any warning/note about this kind of behaviour. Thank you very much for your time :)

2 2

sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable
by Chamila Wijayarathna May 30, 2020

May 30, 2020

Hello all, I hope I am writing to the correct mailing list about this issue that I am having. Please apologize me if I am not. I am attempting to use a pipeline to feed an ensemble voting classifier as I want the ensemble learner to use models that train on different feature sets. For this purpose, I followed the tutorial available at [1]. Following is the code that I could develop so far. *y = df1.indexx = preprocessing.scale(df1)phy_features = ['A', 'B', 'C']phy_transformer = Pipeline(steps=[('imputer', SimpleImputer(strategy='median')), ('scaler', StandardScaler())])phy_processer = ColumnTransformer(transformers=[('phy', phy_transformer, phy_features)])fa_features = ['D', 'E', 'F']fa_transformer = Pipeline(steps=[('imputer', SimpleImputer(strategy='median')), ('scaler', StandardScaler())])fa_processer = ColumnTransformer(transformers=[('fa', fa_transformer, fa_features)])pipe_phy = Pipeline(steps=[('preprocessor', phy_processer ),('classifier', SVM)])pipe_fa = Pipeline(steps=[('preprocessor', fa_processer ),('classifier', SVM)])ens = VotingClassifier(estimators=[pipe_phy, pipe_fa])cv = KFold(n_splits=10, random_state=None, shuffle=True)for train_index, test_index in cv.split(x): x_train, x_test = x[train_index], x[test_index] y_train, y_test = y[train_index], y[test_index] ens.fit(x_train,y_train) print(ens.score(x_test, y_test))* However, when running the code, I am getting an error saying *TypeError: argument of type 'ColumnTransformer' is not iterable*, at the line *ens.fit(x_train,y_train).* What is the reason for this and how can I fix it? Thank you, Chamila

3 6

major league hacking summer internship program
by t3kcit＠gmail.com May 29, 2020

May 29, 2020

Hey Folks. This program reached out to me: https://news.mlh.io/mlh-fellowship-the-future-of-tech-internships-05-04-2020 What do you think? Sounds like GSOC but with extra mentorship, so it might be a good fit for us? I would say it depends on what level of involvement they require from us. Best, Andy

4 4

Class weight SVC
by Francesco basciani May 27, 2020

May 27, 2020

Hi, i have a question regarding the class weights in SVC. I have an imbalanced binary classification problem. In my case the ratio between the positive class and the negative class is 4:1. I just want to know if setting class weight to: class_weight = {1: 0.25, 0: 1} is the same to setting it to: class_weight = {1: 1, 0: 4}. Because in my case i obtain differents results using the two definition of the class weight Inviato da Posta<https://go.microsoft.com/fwlink/?LinkId=550986> per Windows 10

2 1

Random Binning Features
by sai_ng May 27, 2020

May 27, 2020

Hey folks ! Hope you're all doing well. I'm developing Random Fourier Feature implementation in c++ for a repository. Scikits implementation on RBFSampler has been really helpful, and I must say that I'm charmed but how compact, yet powerful each line of code is. I'm writing this mail because I couldn't find your implementation of Random Binning Features, is it under development?. I tried searching in the issues but, to no avail. I noticed you've put few of your algorithms on a different repository for ex: https://scikit-learn-extra.readthedocs.io/en/latest/generated/sklearn_extra…. Overall, I'd like to know if it's under development or has there been any draft/proposal or is it already implemented. I'd greatly appreciate if you could point me to other sources (if not here) which have successfully implemented it in code (preferably python/c++) "Hit me back, Just to chat, Your biggest fan, This is stan" ~ Eminem: Stan

2 1

Welcome to the TC Adrin!
by Andreas C. Mueller May 26, 2020

May 26, 2020

Dear all! It's my pleasure to announce that Adrin has been accepted into the Technical Committee unanimously! Thank you Adrin for all the hard work so far and for your contributions to the community! Best, Andy Ps: yes I have a new email address

2 1

scikit-learn monthly meeting May 25th
by Chiara Marmo May 22, 2020

May 22, 2020

Hi all, The next scikit-learn monthly meeting will take place on Monday May 25th at 12PM UTC: https://www.timeanddate.com/worldclock/meetingdetails.html?year=2020&month=… While these meetings are mainly for core-devs to discuss the current topics, we're also happy to welcome non-core devs and other projects maintainers! Feel free to join, using the following link: https://meet.google.com/xhq-yoga-rtf <https://anaconda.zoom.us/j/94399382811?pwd=cXBtQ2lTVEtVbFpVTkE3TVFxdEhqZz09> If you plan to attend and you would like to discuss something specific about your contribution please add your name (or github pseudo) in the " Contributors <https://hackmd.io/4VeWX5H9Tlmz132WAD-Q0w#Contributors>" section, of the public pad: https://hackmd.io/4VeWX5H9Tlmz132WAD-Q0w <https://hackmd.io/5c6LxpnWSzeaBwJfuX5gPA> *@core devs, please make sure to update your notes before the week-end.* Best, Chiara

2 1

Notes core-dev meeting May 25th
by Chiara Marmo May 19, 2020

May 19, 2020

Dear core-devs, I've taken the liberty to build a skeleton for the meeting notes. Please, have a look and let me know if this is useful. For each section, please, consider to add no more than two issues or PRs per person to be discussed, this will allow everybody speak and close the meeting in a reasonable time. Also, it could be really useful to have all the contributions before the week-end as everybody could take a look to the referenced discussions. A more detailed reminder will be sent soon with a call for contribution from the community and the details for the connection. Thanks for your attention. Best, Chiara

4 4

ANN: scikit-learn 0.23.1 release
by Adrin May 19, 2020

May 19, 2020

We're happy to announce the 0.23.1 release which fixes a few issues affecting many users, namely: K-Means should be faster for small sample sizes, and the representation of third-party estimators was fixed. You can check this version out using: pip install -U scikit-learn You can see the changelog here: https://scikit-learn.org/stable/whats_new/v0.23.html#version-0-23-1 The conda-forge builds will be available shortly, which you can then install using: conda install -c conda-forge scikit-learn On behalf of the scikit-learn development community, Adrin

1 0