The failure was caused by the merge of the memory optimized
implementation of the dot product that probably causes slight changes
in the way the data is generated.

I think the real issue is that this test makes ARPACK very sensitive
to small changes: would it be possible to change the test to use a
dataset that is better conditioned to make ARPACK more robust to small
input changes?

Here is the full log:

https://jenkins.shiningpanda-ci.com/scikit-learn/job/python-2.7-numpy-1.6.2-scipy-0.10.1/2705/consoleFull

The traceback is:

======================================================================
ERROR: sklearn.cluster.bicluster.tests.test_spectral.test_perfect_checkerboard
----------------------------------------------------------------------
Traceback (most recent call last):
  File 
"/home/slave/virtualenvs/cpython-2.7/lib/python2.7/site-packages/nose/case.py",
line 197, in runTest
    self.test(*self.arg)
  File 
"/home/slave/jenkins/workspace/python-2.7-numpy-1.6.2-scipy-0.10.1/sklearn/cluster/bicluster/tests/test_spectral.py",
line 178, in test_perfect_checkerboard
    model.fit(S)
  File 
"/home/slave/jenkins/workspace/python-2.7-numpy-1.6.2-scipy-0.10.1/sklearn/cluster/bicluster/spectral.py",
line 126, in fit
    self._fit(X)
  File 
"/home/slave/jenkins/workspace/python-2.7-numpy-1.6.2-scipy-0.10.1/sklearn/cluster/bicluster/spectral.py",
line 444, in _fit
    u, v = self._svd(normalized_data, n_sv, n_discard)
  File 
"/home/slave/jenkins/workspace/python-2.7-numpy-1.6.2-scipy-0.10.1/sklearn/cluster/bicluster/spectral.py",
line 142, in _svd
    u, _, vt = svds(array, k=n_components, ncv=self.n_svd_vecs)
  File 
"/home/slave/virtualenvs/cpython-2.7/lib/python2.7/site-packages/scipy/sparse/linalg/eigen/arpack/arpack.py",
line 1597, in svds
    eigvals, eigvec = eigensolver(XH_X, k=k, tol=tol ** 2)
  File 
"/home/slave/virtualenvs/cpython-2.7/lib/python2.7/site-packages/scipy/sparse/linalg/eigen/arpack/arpack.py",
line 1545, in eigsh
    params.iterate()
  File 
"/home/slave/virtualenvs/cpython-2.7/lib/python2.7/site-packages/scipy/sparse/linalg/eigen/arpack/arpack.py",
line 578, in iterate
    raise ArpackError(self.info, infodict=self.iterate_infodict)
ArpackError: ARPACK error 3: No shifts could be applied during a cycle
of the Implicitly restarted Arnoldi iteration. One possibility is to
increase the size of NCV relative to NEV.



-- 
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel

------------------------------------------------------------------------------
Introducing Performance Central, a new site from SourceForge and 
AppDynamics. Performance Central is your source for news, insights, 
analysis and resources for efficient Application Performance Management. 
Visit us today!
http://pubads.g.doubleclick.net/gampad/clk?id=48897511&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to