hi, what is the inputs dimensions? thanks, ishay בתאריך יום רביעי, 22 ביולי 2015 בשעה 13:55:32 UTC+3, מאת André L: > > But hey, "python ZCA Whitening" gives an answer", it didnt for me, sorry. > And i spent some hours googling this before posting. > > > 2015-07-22 7:54 GMT-03:00 André Lopes <andrelo...@gmail.com <javascript:>> > : > >> But a Theano implementation of ZCA... i mean, isnt the same thing of a >> python ZCA? Unless you use Shared Variables of course? >> >> Here's the implementation im using, >> >> And trust me, i googled a lot. >> >> import numpy as np >> >> def zca_whitening(inputs): >> #Correlation matrix >> sigma = np.dot(inputs, inputs.T)/inputs.shape[1] >> >> #Singular Value Decomposition >> U,S,V = np.linalg.svd(sigma) >> >> #Whitening constant, it prevents division by zero >> epsilon = 0.1 >> >> #ZCA Whitening matrix >> ZCAMatrix = np.dot(np.dot(U, np.diag(1.0/np.sqrt(np.diag(S) + >> epsilon))), U.T) >> >> #Data whitening >> return np.dot(ZCAMatrix, inputs) >> >> >> >> >> 2015-07-22 7:51 GMT-03:00 Daniel Renshaw <pante...@gmail.com >> <javascript:>>: >> >>> Although there are many examples showing how to use Theano to process >>> MNIST/faces/other images using a CNN, Theano itself is a much more general >>> framework. >>> >>> If your question is along the lines of "how can we *use* machine >>> learning technique X to solve task Y?" (i.e. irrespective of how that >>> technique is implemented) then it would probably be best to find a more >>> general machine learning forum to ask it. >>> >>> However, if your question is "how can machine learning technique X be >>> implemented in Theano?" (i.e. specific to Theano but irrespective of the >>> task, though the task detail may be useful information too) then by all >>> means ask here. >>> >>> Asking "Does anyone have a python implementation of ZCA Whitening?" here >>> may not reach as broad an audience as you might reach elsewhere. Though a >>> simple Google search for "python ZCA Whitening" gives an answer. >>> >>> Asking "Does anyone have a *Theano* implementation of ZCA Whitening?" >>> would be more appropriate here, especially as a quick Google search doesn't >>> seem to give a good result. >>> >>> Daniel >>> >>> >>> On 22 July 2015 at 11:41, André Lopes <andrelo...@gmail.com >>> <javascript:>> wrote: >>> >>>> I suggest you study this : https://caglarift6266.wordpress.com/ >>>> >>>> >>>> 2015-07-22 4:29 GMT-03:00 Sigurd <sigurd.sp...@gmail.com <javascript:>> >>>> : >>>> >>>>> Sorry, but how is this question related to Theano? >>>>> >>>>> -- >>>>> >>>>> --- >>>>> You received this message because you are subscribed to a topic in the >>>>> Google Groups "theano-users" group. >>>>> To unsubscribe from this topic, visit >>>>> https://groups.google.com/d/topic/theano-users/_63ggUdWeEI/unsubscribe >>>>> . >>>>> To unsubscribe from this group and all its topics, send an email to >>>>> theano-users...@googlegroups.com <javascript:>. >>>>> For more options, visit https://groups.google.com/d/optout. >>>>> >>>> >>>> -- >>>> >>>> --- >>>> You received this message because you are subscribed to the Google >>>> Groups "theano-users" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to theano-users...@googlegroups.com <javascript:>. >>>> For more options, visit https://groups.google.com/d/optout. >>>> >>> >>> -- >>> >>> --- >>> You received this message because you are subscribed to a topic in the >>> Google Groups "theano-users" group. >>> To unsubscribe from this topic, visit >>> https://groups.google.com/d/topic/theano-users/_63ggUdWeEI/unsubscribe. >>> To unsubscribe from this group and all its topics, send an email to >>> theano-users...@googlegroups.com <javascript:>. >>> >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> >
-- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to theano-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.