Thanks a lot, that worked.
I think after doing that, I can again reshape it to the default output of
safe_sparse_dot(X_test, coef)
On Sun, Dec 15, 2013 at 2:46 AM, Joel Nothman <joel.noth...@gmail.com>wrote:
> Hi Manoj,
>
> I think one solution is to:
> * roll the features axis to be first: coef = np.rollaxis(coef, 1)
> * reshape so that the last two axes become one: coef =
> coef.reshape(len(coef.shape[0]), -1)
> * perform out = safe_sparse_dot(X_test, coef)
> * for sample i, target index j, alpha index k, the result is in out[i, j *
> n_alphas + k], which is not especially user-friendly, but is fine for
> internal use...
> * where the result is sparse, it may be efficiently stored with BSR sparse
> format, but I don't think you can coerce it to that format before or during
> the sparse dot if the coef is dense
>
> Cheers,
>
> Joel
>
>
> On Sun, Dec 15, 2013 at 3:51 AM, Manoj Kumar <
> manojkumarsivaraj...@gmail.com> wrote:
>
>> Hello,
>>
>> I have been trying to implement MultiTaskElasticNetCV in this PR,
>> https://github.com/scikit-learn/scikit-learn/pull/2598
>>
>> I have got a number of test failures, one of them which include
>> ValueError: Could Not interpret dimensions.
>>
>> I traced this to L924 in the PR. safe_sparse_dot(X_test, coef). This is
>> because I have stored coef as a 3-D numpy array,
>> a] First dimension - order/tasks
>> b] Second dimension - features
>> c] Third dimension - alpha.
>> Hence if X_test is a sparse matrix doing X_test*coef would raise a
>> ValueError.
>>
>> The only alternative, I can think of is using a 'for' loop across the
>> order or tasks, which seems a 'non-numpy way' of doing it. Is there a
>> better way of doing this?
>>
>> Awaiting a reply.
>> Thanks.
>> --
>> Regards,
>> Manoj Kumar,
>> Mech Undergrad
>> http://manojbits.wordpress.com
>>
>>
>> ------------------------------------------------------------------------------
>> Rapidly troubleshoot problems before they affect your business. Most IT
>> organizations don't have a clear picture of how application performance
>> affects their revenue. With AppDynamics, you get 100% visibility into your
>> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics
>> Pro!
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
> ------------------------------------------------------------------------------
> Rapidly troubleshoot problems before they affect your business. Most IT
> organizations don't have a clear picture of how application performance
> affects their revenue. With AppDynamics, you get 100% visibility into your
> Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics
> Pro!
> http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
--
Regards,
Manoj Kumar,
Mech Undergrad
http://manojbits.wordpress.com
------------------------------------------------------------------------------
Rapidly troubleshoot problems before they affect your business. Most IT
organizations don't have a clear picture of how application performance
affects their revenue. With AppDynamics, you get 100% visibility into your
Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro!
http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
