Hey everybody.
I just wanted to do my duty and point out that the next release should
be drawing closer.
I'll try to pitch in but I'm not sure how much time I'll actually find.
We are clearly not there at the the moment but I think we should start
to prioritize. I'd really
like the neural net
On 11/05/2013 03:32 PM, Lars Buitinck wrote:
> 2013/11/6 Olivier Grisel :
>> Thanks for bringing this PR back. I completely forgot about it. Last
>> time I tried it worked on my box (I cannot reproduce @larsmans' issue
>> either). It would be great if more people could test it and review it.
> I do
>
> > 2) Will there be a interface for online learning in OnevsRestClassifier?
>
> If you care to implement it, then we're happy to accept a patch. Are
> you doing multi-label classification?
No I am primarily working on multiclass classification with constantly
increasing number of classes
--
2013/11/6 Olivier Grisel :
> Thanks for bringing this PR back. I completely forgot about it. Last
> time I tried it worked on my box (I cannot reproduce @larsmans' issue
> either). It would be great if more people could test it and review it.
I don't have time to check, please merge if you're conv
Thanks for bringing this PR back. I completely forgot about it. Last
time I tried it worked on my box (I cannot reproduce @larsmans' issue
either). It would be great if more people could test it and review it.
--
Olivier
---
2013/11/5 Jim <[email protected]>:
> 1) This might sound like a basic question, but when performing a partial_fit
> in SGD Classifiers, does the new data to train on have to be in the
> categories that are already in the set? Or conversely, if I come across/ want
> to add a few documents(in a docum
1) This might sound like a basic question, but when performing a partial_fit
in SGD Classifiers, does the new data to train on have to be in the
categories that are already in the set? Or conversely, if I come across/ want
to add a few documents(in a document classifier example) in a new categor
Hi Andy,
The issue is targeted in PR #2185
(https://github.com/scikit-learn/scikit-learn/pull/2355). There are a few
proposed ways of solving that problem, but it still needs to run a few more
benchmarks to determine the best one. Is it the problem you mean?
Thanks,
Karol
On Nov 5, 2013, at 1
Hey everybody.
I'm currently running 0.14.1 and I still have some trouble with MiniBatchKMeans
with the default reassignment settings.
Is that a known issue? About 200 of my 300 centers are unused and every couple
of iterations the same number are reassigned.
Setting reassignment ratio to 0 solve
