[..]
Interesting. What is the order of magnitude of the decrease in speed
at fit time?
IMHO it's negligible
here are some timings for::
rs = np.random.RandomState(13)
X = rs.rand(5, 100)
y = rs.randint(2, size=5)
from sklearn.tree import tree
clf =
2011/11/4 Peter Prettenhofer peter.prettenho...@gmail.com:
[..]
Interesting. What is the order of magnitude of the decrease in speed
at fit time?
IMHO it's negligible
here are some timings for::
rs = np.random.RandomState(13)
X = rs.rand(5, 100)
y = rs.randint(2,
I have just submitted a PR to brian's branch :)
On 4 November 2011 11:13, Peter Prettenhofer
peter.prettenho...@gmail.com wrote:
Gilles,
I was not aware of your work in _tree.pyx. Looks great! Still, I
didn't touch any line in `find_best_split` so the merging/rebase
should be quite
On Thu, Nov 03, 2011 at 11:40:30PM +0100, Peter Prettenhofer wrote:
I created an experimental branch which uses numpy arrays (as Gael
suggested) instead of the composite structure to represent the tree.
Great work (as usual)! Thanks heaps.
What I really like about this (in addition to the
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
Hi everybody.
I was thinking about putting some work into making a multi layer
perceptron implementation
for sklearn. I think it would be a good addition to the other, mostly
linear, classifiers
in sklearn. Together with the decision trees /
On 11/04/2011 02:59 PM, Lars Buitinck wrote:
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
My question is: has anyone started with a mlp implementation yet?
I was just working on one :)
I have the predict function for an arbitrary number of hidden layers
(classifier case) and some
On Fri, Nov 4, 2011 at 4:54 PM, Andreas Müller amuel...@ais.uni-bonn.de wrote:
On 11/04/2011 03:49 PM, Andreas Müller wrote:
On 11/04/2011 03:42 PM, Alexandre Passos wrote:
On Fri, Nov 4, 2011 at 10:34, Lars Buitinck l.j.buiti...@uva.nl wrote:
2011/11/4 Alexandre Passos alexandre...@gmail.com:
ranks = np.argsort(np.sum(estimator.coef_ ** 2, axis=0))
My question is: Why the summation of the squared weight matrix is used?
What is the logic behind it?
This is used for handling estimators that assign several weights to
the same feature. Indeed, if several weights are assigned to a each
On Fri, Nov 4, 2011 at 10:49 AM, Andreas Müller
amuel...@ais.uni-bonn.de wrote:
On 11/04/2011 03:42 PM, Alexandre Passos wrote:
On Fri, Nov 4, 2011 at 10:34, Lars Buitinck l.j.buiti...@uva.nl wrote:
2011/11/4 Alexandre Passos alexandre...@gmail.com:
I have a question: why not just use Theano
Hey Frederic.
I don't have a good understanding of scikit.learn, but I think that
all the hyper-parameter selection is a hot research topic for now. How
do you plan to include this in the current scikit.learn interface of
the fit method?
Depends on what you think of when you say hyper
For a simple mlp, I think theano will not beat a hand implemented version.
I think you'd be in for a rather rude surprise, at least on your first
attempt. :)
It'll not be my first attempt but I must confess, I never benchmarked
my labs GPU mlp against yours ;)
Afaik, torch7 is faster than
I think it make sense to have a pure cython implementation in
scikit-learn without having runtime dependency on a compiler nor CUDA
/ OpenCL and have advanced, theano based neural networks (with more
parameter auto-tuning and pluggable exotic objective functions) in
pylearn.
I think there is room
On Sat, Nov 5, 2011 at 12:44 AM, Olivier Grisel
olivier.gri...@ensta.org wrote:
I think it make sense to have a pure cython implementation in
scikit-learn without having runtime dependency on a compiler nor CUDA
/ OpenCL and have advanced, theano based neural networks (with more
parameter
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
For a simple mlp, I think theano will not beat a hand implemented version.
I think you'd be in for a rather rude surprise, at least on your first
attempt. :)
It'll not be my first attempt but I must confess, I never benchmarked
my labs GPU
In my case I don't use RPROP (I don't know what it is and I just use a
simple backprop) and I use Leon Bottou's trick to perform a burn-in on
the first 10k samples with a grid search of learning rate parameters
and then select the most effective learning rate and multiply it by 2
(it brings
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
On 11/04/2011 02:49 PM, Olivier Grisel wrote:
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
Hi everybody.
I was thinking about putting some work into making a multi layer
perceptron implementation
for sklearn. I think it would be a good
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
In my case I don't use RPROP (I don't know what it is and I just use a
simple backprop) and I use Leon Bottou's trick to perform a burn-in on
the first 10k samples with a grid search of learning rate parameters
and then select the most
On Fri, Nov 04, 2011 at 04:33:39PM +0100, Andreas Müller wrote:
Hey Frederic.
I don't have a good understanding of scikit.learn, but I think that
all the hyper-parameter selection is a hot research topic for now. How
do you plan to include this in the current scikit.learn interface of
the
On 11/04/2011 06:33 PM, David Warde-Farley wrote:
On Fri, Nov 04, 2011 at 06:12:48PM +0100, Andreas Müller wrote:
In my case I don't use RPROP (I don't know what it is and I just use a
simple backprop) and I use Leon Bottou's trick to perform a burn-in on
the first 10k samples with a grid
2011/11/4 Andreas Müller amuel...@ais.uni-bonn.de:
On 11/04/2011 06:33 PM, David Warde-Farley wrote:
On Fri, Nov 04, 2011 at 06:12:48PM +0100, Andreas Müller wrote:
In my case I don't use RPROP (I don't know what it is and I just use a
simple backprop) and I use Leon Bottou's trick to perform
On 2011-11-04, at 13:44, Andreas Müller amuel...@ais.uni-bonn.de wrote:
On 11/04/2011 06:33 PM, David Warde-Farley wrote:
On Fri, Nov 04, 2011 at 06:12:48PM +0100, Andreas Müller wrote:
In my case I don't use RPROP (I don't know what it is and I just use a
simple backprop) and I use Leon
Hi,
Just to be sure, I agree that there is space for both an MLP in
scikit.learn and Theano/Pylearn or similar. I was curious on how you
planned to solve the hyper-parameter selection problem. That is not
an easy problem. The answer was interesting. I didn't know about the
Leon Bottou trick.
I
On Fri, Nov 04, 2011 at 07:22:15PM +0100, Andreas Müller wrote:
One of the reasons I want an MLP in sklearn is so it is easier
to compare with other learning algorithms on a wide range of
tasks.
I guess that this is one of the most compeling reasons to have them in. I
tend to believe the MLPs
On Fri, Nov 4, 2011 at 12:25 PM, Mathieu Blondel math...@mblondel.org wrote:
Another possibility is to host a Theanos-based implementation as a
side project on github and make the API scikit-learn compatible.
# In general, I don't really buy the why implement X if it already
exists in Y
On Fri, Nov 04, 2011 at 07:22:15PM +0100, Andreas Müller wrote:
Of course there are many other possibilities like pretraining,
deeper networks, different learning rate schedules etc..
You are right, this is somewhat of an active research field
Though I have not seen conclusive evidence
25 matches
Mail list logo
