Mailman 3 January 2019 - scikit-learn

Re: [scikit-learn] Google Cloud ML Error
by Bill Ross Jan. 25, 2019

Jan. 25, 2019

I'm a kibitzer who never ran it myself, just a compulsive debugger looking at a basic possibility. Bill <div>-------- Original message --------</div><div>From: Liam Geron <liam(a)chatdesk.com> </div><div>Date:01/25/2019 12:16 PM (GMT-08:00) </div><div>To: Scikit-learn mailing list <scikit-learn(a)python.org> </div><div>Subject: Re: [scikit-learn] Google Cloud ML Error </div><div> </div>As in updated the sklearn module or the joblib module? I'm currently running sklearn on 0.19.1 and joblib on 0.13.1. Do I need to be running them on a specific version? On Fri, Jan 25, 2019 at 2:35 PM Bill Ross <ross(a)cgl.ucsf.edu> wrote: Have you updated the project since this: Since joblib is involved here as well, I'd look at that checkin. Joblib expects there to be a model, maybe it is just configure to look in the wrong place. On 1/25/19 10:54 AM, Liam Geron wrote: No such luck, the file doesn't seem to exist. Here's the output on my local: "ls: /tmp/model/0001/model.joblib: No such file or directory" and "/tmp/model/0001/model.joblib: cannot open `/tmp/model/0001/model.joblib' (No such file or directory)" and on the cloud shell: "ls: cannot access '/tmp/model/0001/model.joblib': No such file or directory" and "/bin/sh: 1: file: not found". On Fri, Jan 25, 2019 at 1:29 PM Bill Ross <ross(a)cgl.ucsf.edu> wrote: Dumb generic cross-check from supporting compchem code in the day: What do these give? Might yield a clue, e.g. all model files seeing this got corrupted somehow. $ file /tmp/model/0001/model.joblib $ ls -l /tmp/model/0001/model.joblib On 1/25/19 9:26 AM, Liam Geron wrote: Hi scikit learn contributors, I am currently attempting to transfer our preexisting models into cloud ML for scalability, however I am encountering bugs while running through some tutorial code found here (https://github.com/GoogleCloudPlatform/cloudml-samples/blob/master/sklearn/…). On both my local machine in a virtual environment and on the cloud shell I'm encountering errors when it comes to version creation and online prediction. For version creation on my local machine and on the cloud shell I'm encountering this error: "ERROR: (gcloud.ml-engine.versions.create) Bad model detected with error: "Failed to load model: Could not load the model: /tmp/model/0001/model.joblib. 32. (Error code: 0)"" with Python 3.6.4 (local) and Python 3.5.6 (cloud shell) when running the command: "gcloud ml-engine versions create $VERSION_NAME \ --model $MODEL_NAME \ --config config.yaml" Any help would be greatly appreciated. Thank you, Liam Geron _______________________________________________ scikit-learn mailing list scikit-learn(a)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 _______________________________________________ scikit-learn mailing list scikit-learn(a)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

1 0

Google Cloud ML Error
by Liam Geron Jan. 25, 2019

Jan. 25, 2019

Hi scikit learn contributors, I am currently attempting to transfer our preexisting models into cloud ML for scalability, however I am encountering bugs while running through some tutorial code found here ( https://github.com/GoogleCloudPlatform/cloudml-samples/blob/master/sklearn/… ). On both my local machine in a virtual environment and on the cloud shell I'm encountering errors when it comes to version creation and online prediction. For version creation on my local machine and on the cloud shell I'm encountering this error: *"ERROR: (gcloud.ml-engine.versions.create) Bad model detected with error: "Failed to load model: Could not load the model: /tmp/model/0001/model.joblib. 32. (Error code: 0)""* with Python 3.6.4 (local) and Python 3.5.6 (cloud shell) when running the command: *"gcloud ml-engine versions create $VERSION_NAME \* * --model $MODEL_NAME \* * --config config.yaml"* Any help would be greatly appreciated. Thank you, Liam Geron

2 4

cluster.affinity_propagation doesn't accept sparse?
by Neal Becker Jan. 23, 2019

Jan. 23, 2019

It appears affinity propagation would appear to accept sparse similarity input: X = check_array(X, accept_sparse='csr') But if I try it, I get: ~/.local/lib/python3.7/site- packages/sklearn/cluster/affinity_propagation_.py in affinity_propagation(S, preference, convergence_iter, max_iter, damping, copy, verbose, return_n_iter) 137 138 # Place preference on the diagonal of S --> 139 S.flat[::(n_samples + 1)] = preference 140 141 A = np.zeros((n_samples, n_samples)) ~/.local/lib/python3.7/site-packages/scipy/sparse/base.py in __getattr__(self, attr) 687 return self.getnnz() 688 else: --> 689 raise AttributeError(attr + " not found") 690 691 def transpose(self, axes=None, copy=False): AttributeError: flat not found

1 0

affinity propagation not giving desired answer
by Neal Becker Jan. 23, 2019

Jan. 23, 2019

I am not too familiar with affinity propagation, but just trying it out. The problem is to cluster using a distance metric that is euclidean distance but with a limit. When the distance is greater than some threshold than the metric is -Inf. In other words, a point can be accepted into a cluster only if the distance from the point to the cluster center is less than some threshold. It seems my test with affinity propagation will sometimes produce a correct result, but other times the result seems to violate the condition. In the example code, a couple of outlier points seem to be in clusters that are not close at all. I've tried playing with parameters (such as preference) without eliminating the problem. Any suggestions? --------- import numpy as np from sklearn.cluster import AffinityPropagation # from randomgen import RandomGenerator, Xoroshiro128 # rs = RandomGenerator (Xoroshiro128 (0)) from numpy.random import RandomState rs = RandomState(3) pts = rs.uniform (-5, 5, (50,2)) import seaborn as sns import matplotlib.pyplot as plt def distance (ax, ay, bx, by): d = (ax - bx)**2 + (ay - by)**2 if d > 1: return -1e6 else: return -d d = np.empty ((pts.shape[0], pts.shape[0])) for i in range(pts.shape[0]): for j in range(pts.shape[0]): d[i,j] = distance(pts[i,0], pts[i,1], pts[j,0], pts[j,1]) preference = -20 #np.mean (d[d > -1e6]) print ('preference:', preference) clustering = AffinityPropagation(affinity='precomputed', verbose=True, preference=preference) res = clustering.fit(d) c = clustering colors = np.array(sns.color_palette("hls", np.max(c.labels_)+1)) print('n_clusters:', np.max(c.labels_)+1) centers = pts[c.cluster_centers_indices_] plt.scatter (pts[:,0], pts[:,1], c=colors[c.labels_]) plt.scatter (centers[:,0], centers[:,1], marker='X', s=100, c=colors) plt.show()

1 0

Affinity Propagation is the best algo for without choosing the number of cluster?
by lampahome Jan. 23, 2019

Jan. 23, 2019

I search for clustering algo to cluster into groups without considering about number of groups. I found AP algo which needn't choose the number of clusters. In my experiments, AP cluster well without choosing any parameters. But I'm not sure any corner case which will caused clustering worse. Does anyone try AP and found some side-effect? or the way to tune parameters? thx

1 0

PR #13003: [MRG] Add Tensor Sketch algorithm to Kernel Approximation module
by Daniel López-Sánchez Jan. 22, 2019

Jan. 22, 2019

Dear all, I recently posted a PR <https://github.com/scikit-learn/scikit-learn/pull/13003> which adds the Tensor Sketch algorithm [1] to the Kernel Approximation module of Scikit-learn. I believe this new feature makes the Kernel Approximation module more complete by providing a data-independent method for polynomial kernel approximation, as the currently included methods either require access to training data (Nystroem) or do not work with polynomial kernels. The implementation has been tested to provide the same results as the original Matlab implementation provided by the author of [1]. I would appreciate any feedback you can provide, Regards, [1] Pham, N., & Pagh, R. (2013, August). Fast and scalable polynomial kernels via explicit feature maps. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining (pp. 239-247). ACM. *Daniel López Sánchez* <https://github.com/lopelh> lope(a)usal.es / (+34) 687174328 BISITE Research Group (http://bisite.usal.es <http://bisite.usal.es/en>) Edificio I+D+i Universidad de Salamanca, C/ Espejo S/N, 37007 Salamanca, Spain

2 1

Any clustering algo to cluster multiple timing series data?
by lampahome Jan. 21, 2019

Jan. 21, 2019

Cluster algo cluster samples by calculating the euclidean distance. I wonder if any clustering algo can cluster the timing series data? EX: Every items has there sold numbers of everyday. Item,Day1,Day2,Day3,Day4,Day5 A,1,5,1,5,1 B,10,50,10,50,10, C,4,70,30,10,50 The difference ratio of A and B are 500%,20%,500%,20%, I want to make A&B be the same cluster, C is another one. If I don't want to calculate the difference ratio of each samples Is there anyway to cluster by the difference ratio of each samples? thx

3 5

ANN: scikit-image 0.14.2
by Stefan van der Walt Jan. 18, 2019

Jan. 18, 2019

Announcement: scikit-image 0.14.2 ================================= This release handles an incompatibility between scikit-image and NumPy 1.16.0, released on January 13th 2019. It contains the following changes from 0.14.1: API changes ----------- - ``skimage.measure.regionprops`` no longer removes singleton dimensions from label images (#3284). To recover the old behavior, replace ``regionprops(label_image)`` calls with ``regionprops(np.squeeze(label_image))`` Bug fixes --------- - Address deprecation of NumPy ``_validate_lengths`` (backport of #3556) - Correctly handle the maximum number of lines in Hough transforms (backport of #3514) - Correctly implement early stopping criterion for rank kernel noise filter (backport of #3503) - Fix ``skimage.measure.regionprops`` for 1x1 inputs (backport of #3284) Enhancements ------------ - Rewrite of ``local_maxima`` with flood-fill (backport of #3022, #3447) Build Process & Testing ----------------------- - Dedicate a ``--pre`` build in appveyor (backport of #3222) - Avoid Travis-CI failure regarding ``skimage.lookfor`` (backport of #3477) - Stop using the ``pytest.fixtures`` decorator (#3558) - Filter out DeprecationPendingWarning for matrix subclass (#3637) - Fix matplotlib test warnings and circular import (#3632) Contributors & Reviewers ------------------------ - François Boulogne - Emmanuelle Gouillart - Lars Grüter - Mark Harfouche - Juan Nunez-Iglesias - Egor Panfilov - Stefan van der Walt

1 0

Scipy 2019 Tutorial
by Andreas Mueller Jan. 18, 2019

Jan. 18, 2019

Hey Folks. The scipy tutorial chairs just pinged me about submitting a tutorial. I'm planning to, and wanted to ask if anyone is interested in co-teaching with me. I might transition from the "scipy tutorial" materials (evolved over maybe 5 years) to my own materials, but not sure yet. Nicolas said he'd potentially be interested but I wanted to ask around who else is coming and might be interested. Cheers, Andy

1 0

Non-core developers at the sprint
by Gael Varoquaux Jan. 16, 2019

Jan. 16, 2019

Dear users and developers, We have a sprint coming up in Paris Feb 25th to March 1st: https://github.com/scikit-learn/scikit-learn/wiki/Upcoming-events Looking at the list of people who are coming, I am noticing that we have mostly core developers. While the priority of the sprint is to work on the big picture rather than onboarding, I am worried that there might be some self-selection happening. I am sure that some excellent people, who are contributors yet not core contributors could come. I would like to invite people who already have contributed and want to get more involved in the project to contact us to join the sprint. Specifically, we are willing to fund accommodation and travel for one or two participants. Please send a short message to Guillaume Lemaître <guillaume.lemaitre(a)inria.fr> and myself presenting what you have contributed and what you would like to contribute, as well as your funding needs. We will curate this list and core contributors will settle on who we can accommodate. Cheers, Gaël -- Gael Varoquaux Senior Researcher, INRIA Parietal NeuroSpin/CEA Saclay , Bat 145, 91191 Gif-sur-Yvette France Phone: ++ 33-1-69-08-79-68 http://gael-varoquaux.info http://twitter.com/GaelVaroquaux

1 0