On Sat, Jan 21, 2012 at 11:15 PM, Gael Varoquaux
wrote:
> sklearn.svm.liblinear.decision_function_wrap
> (sklearn/svm/liblinear.c:2775)
> ValueError: ndarray is not Fortran contiguous
I think this error message sometimes also appears if the array is 2d
instead of 1d.
Mathieu
--
2012/1/21 Gael Varoquaux :
> ==
> ERROR: sklearn.svm.tests.test_svm.test_liblinear_set_coef
> --
> Traceback (most recent call last):
> File "/usr/lib/pymodules/p
On Sat, Jan 21, 2012 at 09:27:21AM -0500, Alexandre Passos wrote:
> I deleted it because your answer made it redundant, by saying
> everything that I had said more clearly and with more references.
> I figure someone finding that page on google should see a long, clear
> answer rather than a short
On Sat, Jan 21, 2012 at 09:25, Gael Varoquaux
wrote:
> On Sat, Jan 21, 2012 at 12:52:29AM +0530, Jaidev Deshpande wrote:
>> This is a wonderful resource.
>
> It is.
>
> I have answered your question:
>
> http://metaoptimize.com/qa/questions/8763/compressive-sensing-conditions-for-recovery#8771
>
>
On Sat, Jan 21, 2012 at 12:52:29AM +0530, Jaidev Deshpande wrote:
> This is a wonderful resource.
It is.
I have answered your question:
http://metaoptimize.com/qa/questions/8763/compressive-sensing-conditions-for-recovery#8771
I tried to put as much hand waiving as I could without being incorre
I haven't followed the recent changes to the SVM array ordering, but I
must say that I have the following test failing on my box:
==
ERROR: sklearn.svm.tests.test_svm.test_liblinear_set_coef
---
On 01/21/2012 07:08 AM, Gilles Louppe wrote:
> Yep, I think that your solution would work Olivier. I am buzy this
> week-end but I can push a first draft of this refactoring by the
> beginning of next week.
>
That would be pretty awesome :)
Thanks for tackling this!
Cheers,
Andy
---
