\begin{plug}
And in a similar vein there is also http://sklearn-theano.github.io for
those who want to leverage existing, pre-trained networks as feature
extractors for their scikit-learn pipeline, which can, e.g. be followed by
a simple logistic regression to fine-tune to new types of objects.
\end{plug}
Michael
On Thu, May 7, 2015 at 3:45 PM, Boyuan Deng <bryanhsud...@gmail.com> wrote:
> FYI, in April a project called *"*scikit-neuralnetwork" came into
> existence. It's a library wrapping Pylearn2 and providing a scikit-learn
> compatible interface. https://github.com/aigamedev/scikit-neuralnetwork
>
> As others have stated, there are already some nice Python neural network
> libraries and we don't really need to reinvent a wheel. Pylean2 is mostly
> research-oriented and provides building blocks. I think sknn is able to
> fill that gap and enable sklearn-like experience with neural networks.
>
> Personally I hope in the future our community can officially help with
> sknn and feature it as an "add-on" to sklearn (because Pylearn2 may never
> go into stable release). In that way sklearn can definitely be the
> dominating Python machine learning library.
>
>
> On 05/07/2015 07:10 AM, Joel Nothman wrote:
>
> What Sebastian and Ronnie said. Plus: there are multiple off-the-shelf
> neural net pull requests in the process of review, notably those by Issam
> Laradji for GSoC 2014. Extreme Learning Machines and Multilayer Perceptrons
> should be merged Real Soon Now.
>
>
> On 7 May 2015 at 14:58, Ronnie Ghose <ronnie.gh...@gmail.com> wrote:
>
>> neural nets are already well supported in other python libraries and
>> don't fit the current transformer model that scikit-learn uses
>>
>> On Thu, May 7, 2015 at 12:55 AM, Sebastian Raschka <se.rasc...@gmail.com>
>> wrote:
>>
>>> I am not one of the core developers, just a typical user, but although I
>>> think that neural nets would be a nice addition, I have to admit that I
>>> wouldn't count them as top priority. I think that in applications, neural
>>> networks require far more flexibility for tweaking than the "classic"
>>> off-the-shelve learning algorithms currently implemented in scikit-learn. I
>>> think that it really requires a lot of planning to implement them in a way
>>> that allows a user certain flexibility. To me, neural nets are more of an
>>> "research tool" in contrast to the currently implemented algos in
>>> scikit-learn. I really would like to see some way of implementing
>>> frameworks for neural networks in some useful way in scikit-learn, but I
>>> can understand that it would really require a different API, a lot of
>>> planning, and a lot of work. Also, there are many attempts to implement
>>> neural nets already, like pylearn2, lasagne, and all the other theano
>>> wrappers
>>>
>>>
>>>
>>>
>>> > On May 6, 2015, at 11:15 PM, 赵孽 <snakehunt2...@gmail.com> wrote:
>>> >
>>> > I used to seeking neural network algorithms in sklearn, but I just
>>> found a RBF in it.
>>> > There are plenty of Neural Network algorithms, why dose we only
>>> support RBF which is not even a typical neural network ?
>>> > I thought the neural networks should be the largest family amount
>>> sklearn algoritms, but it is far smaller than embedings, far smaller than
>>> SVMs.
>>> >
>>> ------------------------------------------------------------------------------
>>> > One dashboard for servers and applications across
>>> Physical-Virtual-Cloud
>>> > Widest out-of-the-box monitoring support with 50+ applications
>>> > Performance metrics, stats and reports that give you Actionable
>>> Insights
>>> > Deep dive visibility with transaction tracing using APM Insight.
>>> >
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y_______________________________________________
>>> > Scikit-learn-general mailing list
>>> > Scikit-learn-general@lists.sourceforge.net
>>> > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>> Widest out-of-the-box monitoring support with 50+ applications
>>> Performance metrics, stats and reports that give you Actionable Insights
>>> Deep dive visibility with transaction tracing using APM Insight.
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>> _______________________________________________
>>> Scikit-learn-general mailing list
>>> Scikit-learn-general@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Scikit-learn-general mailing list
>> Scikit-learn-general@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>>
>>
>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Scikit-learn-general mailing list
> Scikit-learn-general@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
>
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
