You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by "Robin Anil (JIRA)" <ji...@apache.org> on 2013/06/04 15:51:20 UTC
[jira] [Updated] (MAHOUT-1240) Randomized testing and Serialization
of NonZeros
[ https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Robin Anil updated MAHOUT-1240:
-------------------------------
Status: Patch Available (was: Open)
> Randomized testing and Serialization of NonZeros
> ------------------------------------------------
>
> Key: MAHOUT-1240
> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
> Project: Mahout
> Issue Type: Bug
> Reporter: Robin Anil
> Assignee: Robin Anil
> Fix For: 0.8
>
> Attachments: MAHOUT-1240.patch
>
>
> Currently the nonZero iterator does not guarantee nonZero iteration for certain vectors (RASV, SASV) for performance reason. However vector view iterator adds a zero check.. To be correct we have to either remove the check or do correct non zero serialization everywhere. However this means going over the vectors in two passes. Given that is pretty fast already, I am fixing the logic bug. We can tackle the speed up for the next release.
> This also adds a randomized test for serialization that catches all such bugs.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
Re: Mahout - neural network algorithm?
Posted by Mimi Tam <mi...@ieee.org>.
Thank you so much, Yexi.
-----Original Message-----
From: Yexi Jiang
Sent: Thursday, June 06, 2013 12:29 PM
To: dev@mahout.apache.org
Subject: Re: Mahout - neural network algorithm?
You can find it at
https://github.com/yxjiang/hama/tree/trunk/ml/src/main/java/org/apache/hama/ml/perception
.
2013/6/6 Yexi Jiang <ye...@gmail.com>
> I just implemented one based on the BSP model, it is easy to adapt to the
> MapReduce style. If you think it is worth trying, I can do it.
>
>
> 2013/6/6 Ankur Gupta <gm...@gmail.com>
>
>> I have previously worked with java, and Mahout is one project where I
>> would
>> love to contribute, be it in Java.
>>
>> On Thu, Jun 6, 2013 at 6:49 PM, Robin Anil <ro...@gmail.com> wrote:
>>
>> > Ankur you will be pleasantly surprised how easy it is to develop an
>> > algorithm using Mahout's vector classes.
>> >
>> > Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
>> >
>> >
>> > On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com>
>> wrote:
>> >
>> > > We currently do not have any C++ code.
>> > >
>> > > The build complexity and maintenance and portability issues probably
>> > > outweigh any advantages that I see at this time.
>> > >
>> > > We would love to have you contribute, but for now I think it is
>> > > safest/easiest to assume Java.
>> > >
>> > >
>> > > On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
>> > > >wrote:
>> > >
>> > > > Hi
>> > > > I want to contribute to mahout but one thing that holds me back is
>> > that I
>> > > > want to develop in C++ (It is just the comfort level). Is there any
>> > part
>> > > of
>> > > > mahout which is developed in C++ or can be developed in C++. If so
>> > kindly
>> > > > let me know so that I can start working on understanding that and
>> soon
>> > > > start contributing.
>> > > >
>> > > > Thanks & Regards,
>> > > > Ankur
>> > > >
>> > > >
>> > > >
>> > > > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
>> > > wrote:
>> > > >
>> > > > > No.
>> > > > >
>> > > > > Mahout doesn't have any multi-level neural network implementation
>> at
>> > > this
>> > > > > time.
>> > > > >
>> > > > >
>> > > > >
>> > > > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
>> > > > >
>> > > > > > I'd like to phrase the question a bit differently.
>> > > > > >
>> > > > > > Does Mahout have a neural network algorithm (did it adopt
>> Neuroph's
>> > > or
>> > > > > native)?
>> > > > > >
>> > > > > > Thanks.
>> > > > > >
>> > > > > >
>> > > > > > -----Original Message----- From: Mimi Tam
>> > > > > > Sent: Tuesday, June 04, 2013 10:08 AM
>> > > > > > To: dev@mahout.apache.org
>> > > > > > Subject: Question on Mahout (adopted FANN libraries?) if I
>> may...
>> > > > > >
>> > > > > > Sorry, I am a late comer.
>> > > > > >
>> > > > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural
>> > > > > > Network
>> > with
>> > > > > > backpropagation learning on Hadoop ever implemented and is part
>> of
>> > > > Mahout
>> > > > > > now?
>> > > > > >
>> > > > > > Is this using the FANN libraries?
>> > > > > >
>> > > > > > Many Thanks...Mimi
>> > > > > >
>> > > > > >
>> > > > > > -----Original Message----- From: Robin Anil (JIRA)
>> > > > > > Sent: Tuesday, June 04, 2013 9:51 AM
>> > > > > > To: dev@mahout.apache.org
>> > > > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
>> > > > > Serialization
>> > > > > > of NonZeros
>> > > > > >
>> > > > > >
>> > > > > > [
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
>> > > > > > ]
>> > > > > >
>> > > > > > Robin Anil updated MAHOUT-1240:
>> > > > > > -------------------------------
>> > > > > >
>> > > > > > Status: Patch Available (was: Open)
>> > > > > >
>> > > > > >> Randomized testing and Serialization of NonZeros
>> > > > > >> ------------------------------------------------
>> > > > > >>
>> > > > > >> Key: MAHOUT-1240
>> > > > > >> URL:
>> > > https://issues.apache.org/jira/browse/MAHOUT-1240
>> > > > > >> Project: Mahout
>> > > > > >> Issue Type: Bug
>> > > > > >> Reporter: Robin Anil
>> > > > > >> Assignee: Robin Anil
>> > > > > >> Fix For: 0.8
>> > > > > >>
>> > > > > >> Attachments: MAHOUT-1240.patch
>> > > > > >>
>> > > > > >>
>> > > > > >> Currently the nonZero iterator does not guarantee nonZero
>> > iteration
>> > > > for
>> > > > > certain vectors (RASV, SASV) for performance reason. However
>> vector
>> > > view
>> > > > > iterator adds a zero check.. To be correct we have to either
>> remove
>> > the
>> > > > > check or do correct non zero serialization everywhere. However
>> this
>> > > means
>> > > > > going over the vectors in two passes. Given that is pretty fast
>> > > already,
>> > > > I
>> > > > > am fixing the logic bug. We can tackle the speed up for the next
>> > > release.
>> > > > > >> This also adds a randomized test for serialization that
>> > > > > >> catches
>> > all
>> > > > > such bugs.
>> > > > > >
>> > > > > > --
>> > > > > > This message is automatically generated by JIRA.
>> > > > > > If you think it was sent incorrectly, please contact your JIRA
>> > > > > > administrators
>> > > > > > For more information on JIRA, see:
>> > > > > http://www.atlassian.com/software/jira
>> > > > > >
>> > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
>>
>>
>> --
>> Ankur Gupta,
>> Software Engineer
>> Samsung India Software Operations Pvt. Ltd.
>> Bangalore, Karnataka
>> India
>>
>
>
>
> --
> ------
> Yexi Jiang,
> ECS 251, yjian004@cs.fiu.edu
> School of Computer and Information Science,
> Florida International University
> Homepage: http://users.cis.fiu.edu/~yjian004/
>
>
--
------
Yexi Jiang,
ECS 251, yjian004@cs.fiu.edu
School of Computer and Information Science,
Florida International University
Homepage: http://users.cis.fiu.edu/~yjian004/
Re: Mahout - neural network algorithm?
Posted by Yexi Jiang <ye...@gmail.com>.
You can find it at
https://github.com/yxjiang/hama/tree/trunk/ml/src/main/java/org/apache/hama/ml/perception
.
2013/6/6 Yexi Jiang <ye...@gmail.com>
> I just implemented one based on the BSP model, it is easy to adapt to the
> MapReduce style. If you think it is worth trying, I can do it.
>
>
> 2013/6/6 Ankur Gupta <gm...@gmail.com>
>
>> I have previously worked with java, and Mahout is one project where I
>> would
>> love to contribute, be it in Java.
>>
>> On Thu, Jun 6, 2013 at 6:49 PM, Robin Anil <ro...@gmail.com> wrote:
>>
>> > Ankur you will be pleasantly surprised how easy it is to develop an
>> > algorithm using Mahout's vector classes.
>> >
>> > Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
>> >
>> >
>> > On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com>
>> wrote:
>> >
>> > > We currently do not have any C++ code.
>> > >
>> > > The build complexity and maintenance and portability issues probably
>> > > outweigh any advantages that I see at this time.
>> > >
>> > > We would love to have you contribute, but for now I think it is
>> > > safest/easiest to assume Java.
>> > >
>> > >
>> > > On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
>> > > >wrote:
>> > >
>> > > > Hi
>> > > > I want to contribute to mahout but one thing that holds me back is
>> > that I
>> > > > want to develop in C++ (It is just the comfort level). Is there any
>> > part
>> > > of
>> > > > mahout which is developed in C++ or can be developed in C++. If so
>> > kindly
>> > > > let me know so that I can start working on understanding that and
>> soon
>> > > > start contributing.
>> > > >
>> > > > Thanks & Regards,
>> > > > Ankur
>> > > >
>> > > >
>> > > >
>> > > > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
>> > > wrote:
>> > > >
>> > > > > No.
>> > > > >
>> > > > > Mahout doesn't have any multi-level neural network implementation
>> at
>> > > this
>> > > > > time.
>> > > > >
>> > > > >
>> > > > >
>> > > > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
>> > > > >
>> > > > > > I'd like to phrase the question a bit differently.
>> > > > > >
>> > > > > > Does Mahout have a neural network algorithm (did it adopt
>> Neuroph's
>> > > or
>> > > > > native)?
>> > > > > >
>> > > > > > Thanks.
>> > > > > >
>> > > > > >
>> > > > > > -----Original Message----- From: Mimi Tam
>> > > > > > Sent: Tuesday, June 04, 2013 10:08 AM
>> > > > > > To: dev@mahout.apache.org
>> > > > > > Subject: Question on Mahout (adopted FANN libraries?) if I
>> may...
>> > > > > >
>> > > > > > Sorry, I am a late comer.
>> > > > > >
>> > > > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network
>> > with
>> > > > > > backpropagation learning on Hadoop ever implemented and is part
>> of
>> > > > Mahout
>> > > > > > now?
>> > > > > >
>> > > > > > Is this using the FANN libraries?
>> > > > > >
>> > > > > > Many Thanks...Mimi
>> > > > > >
>> > > > > >
>> > > > > > -----Original Message----- From: Robin Anil (JIRA)
>> > > > > > Sent: Tuesday, June 04, 2013 9:51 AM
>> > > > > > To: dev@mahout.apache.org
>> > > > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
>> > > > > Serialization
>> > > > > > of NonZeros
>> > > > > >
>> > > > > >
>> > > > > > [
>> > > > > >
>> > > > >
>> > > >
>> > >
>> >
>> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
>> > > > > > ]
>> > > > > >
>> > > > > > Robin Anil updated MAHOUT-1240:
>> > > > > > -------------------------------
>> > > > > >
>> > > > > > Status: Patch Available (was: Open)
>> > > > > >
>> > > > > >> Randomized testing and Serialization of NonZeros
>> > > > > >> ------------------------------------------------
>> > > > > >>
>> > > > > >> Key: MAHOUT-1240
>> > > > > >> URL:
>> > > https://issues.apache.org/jira/browse/MAHOUT-1240
>> > > > > >> Project: Mahout
>> > > > > >> Issue Type: Bug
>> > > > > >> Reporter: Robin Anil
>> > > > > >> Assignee: Robin Anil
>> > > > > >> Fix For: 0.8
>> > > > > >>
>> > > > > >> Attachments: MAHOUT-1240.patch
>> > > > > >>
>> > > > > >>
>> > > > > >> Currently the nonZero iterator does not guarantee nonZero
>> > iteration
>> > > > for
>> > > > > certain vectors (RASV, SASV) for performance reason. However
>> vector
>> > > view
>> > > > > iterator adds a zero check.. To be correct we have to either
>> remove
>> > the
>> > > > > check or do correct non zero serialization everywhere. However
>> this
>> > > means
>> > > > > going over the vectors in two passes. Given that is pretty fast
>> > > already,
>> > > > I
>> > > > > am fixing the logic bug. We can tackle the speed up for the next
>> > > release.
>> > > > > >> This also adds a randomized test for serialization that catches
>> > all
>> > > > > such bugs.
>> > > > > >
>> > > > > > --
>> > > > > > This message is automatically generated by JIRA.
>> > > > > > If you think it was sent incorrectly, please contact your JIRA
>> > > > > > administrators
>> > > > > > For more information on JIRA, see:
>> > > > > http://www.atlassian.com/software/jira
>> > > > > >
>> > > > >
>> > > > >
>> > > >
>> > >
>> >
>>
>>
>>
>> --
>> Ankur Gupta,
>> Software Engineer
>> Samsung India Software Operations Pvt. Ltd.
>> Bangalore, Karnataka
>> India
>>
>
>
>
> --
> ------
> Yexi Jiang,
> ECS 251, yjian004@cs.fiu.edu
> School of Computer and Information Science,
> Florida International University
> Homepage: http://users.cis.fiu.edu/~yjian004/
>
>
--
------
Yexi Jiang,
ECS 251, yjian004@cs.fiu.edu
School of Computer and Information Science,
Florida International University
Homepage: http://users.cis.fiu.edu/~yjian004/
That's fantastic! --- Re: Mahout - neural network algorithm?
Posted by Mimi Tam <mi...@ieee.org>.
That's fantastic! Absolutely, it is totally worth trying.
Please provide me with any info that you can share about your
implementation/architecture.
Looking forward to hearing from you again soon.
Many Thanks...Mimi
-----Original Message-----
From: Yexi Jiang
Sent: Thursday, June 06, 2013 10:57 AM
To: dev@mahout.apache.org
Subject: Re: Mahout - neural network algorithm?
I just implemented one based on the BSP model, it is easy to adapt to the
MapReduce style. If you think it is worth trying, I can do it.
2013/6/6 Ankur Gupta <gm...@gmail.com>
> I have previously worked with java, and Mahout is one project where I
> would
> love to contribute, be it in Java.
>
> On Thu, Jun 6, 2013 at 6:49 PM, Robin Anil <ro...@gmail.com> wrote:
>
> > Ankur you will be pleasantly surprised how easy it is to develop an
> > algorithm using Mahout's vector classes.
> >
> > Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
> >
> >
> > On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com>
> wrote:
> >
> > > We currently do not have any C++ code.
> > >
> > > The build complexity and maintenance and portability issues probably
> > > outweigh any advantages that I see at this time.
> > >
> > > We would love to have you contribute, but for now I think it is
> > > safest/easiest to assume Java.
> > >
> > >
> > > On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
> > > >wrote:
> > >
> > > > Hi
> > > > I want to contribute to mahout but one thing that holds me back is
> > that I
> > > > want to develop in C++ (It is just the comfort level). Is there any
> > part
> > > of
> > > > mahout which is developed in C++ or can be developed in C++. If so
> > kindly
> > > > let me know so that I can start working on understanding that and
> soon
> > > > start contributing.
> > > >
> > > > Thanks & Regards,
> > > > Ankur
> > > >
> > > >
> > > >
> > > > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
> > > wrote:
> > > >
> > > > > No.
> > > > >
> > > > > Mahout doesn't have any multi-level neural network implementation
> at
> > > this
> > > > > time.
> > > > >
> > > > >
> > > > >
> > > > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> > > > >
> > > > > > I'd like to phrase the question a bit differently.
> > > > > >
> > > > > > Does Mahout have a neural network algorithm (did it adopt
> Neuroph's
> > > or
> > > > > native)?
> > > > > >
> > > > > > Thanks.
> > > > > >
> > > > > >
> > > > > > -----Original Message----- From: Mimi Tam
> > > > > > Sent: Tuesday, June 04, 2013 10:08 AM
> > > > > > To: dev@mahout.apache.org
> > > > > > Subject: Question on Mahout (adopted FANN libraries?) if I
> > > > > > may...
> > > > > >
> > > > > > Sorry, I am a late comer.
> > > > > >
> > > > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network
> > with
> > > > > > backpropagation learning on Hadoop ever implemented and is part
> of
> > > > Mahout
> > > > > > now?
> > > > > >
> > > > > > Is this using the FANN libraries?
> > > > > >
> > > > > > Many Thanks...Mimi
> > > > > >
> > > > > >
> > > > > > -----Original Message----- From: Robin Anil (JIRA)
> > > > > > Sent: Tuesday, June 04, 2013 9:51 AM
> > > > > > To: dev@mahout.apache.org
> > > > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> > > > > Serialization
> > > > > > of NonZeros
> > > > > >
> > > > > >
> > > > > > [
> > > > > >
> > > > >
> > > >
> > >
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > > > > > ]
> > > > > >
> > > > > > Robin Anil updated MAHOUT-1240:
> > > > > > -------------------------------
> > > > > >
> > > > > > Status: Patch Available (was: Open)
> > > > > >
> > > > > >> Randomized testing and Serialization of NonZeros
> > > > > >> ------------------------------------------------
> > > > > >>
> > > > > >> Key: MAHOUT-1240
> > > > > >> URL:
> > > https://issues.apache.org/jira/browse/MAHOUT-1240
> > > > > >> Project: Mahout
> > > > > >> Issue Type: Bug
> > > > > >> Reporter: Robin Anil
> > > > > >> Assignee: Robin Anil
> > > > > >> Fix For: 0.8
> > > > > >>
> > > > > >> Attachments: MAHOUT-1240.patch
> > > > > >>
> > > > > >>
> > > > > >> Currently the nonZero iterator does not guarantee nonZero
> > iteration
> > > > for
> > > > > certain vectors (RASV, SASV) for performance reason. However
> > > > > vector
> > > view
> > > > > iterator adds a zero check.. To be correct we have to either
> > > > > remove
> > the
> > > > > check or do correct non zero serialization everywhere. However
> > > > > this
> > > means
> > > > > going over the vectors in two passes. Given that is pretty fast
> > > already,
> > > > I
> > > > > am fixing the logic bug. We can tackle the speed up for the next
> > > release.
> > > > > >> This also adds a randomized test for serialization that catches
> > all
> > > > > such bugs.
> > > > > >
> > > > > > --
> > > > > > This message is automatically generated by JIRA.
> > > > > > If you think it was sent incorrectly, please contact your JIRA
> > > > > > administrators
> > > > > > For more information on JIRA, see:
> > > > > http://www.atlassian.com/software/jira
> > > > > >
> > > > >
> > > > >
> > > >
> > >
> >
>
>
>
> --
> Ankur Gupta,
> Software Engineer
> Samsung India Software Operations Pvt. Ltd.
> Bangalore, Karnataka
> India
>
--
------
Yexi Jiang,
ECS 251, yjian004@cs.fiu.edu
School of Computer and Information Science,
Florida International University
Homepage: http://users.cis.fiu.edu/~yjian004/
Re: Mahout - neural network algorithm?
Posted by Yexi Jiang <ye...@gmail.com>.
I just implemented one based on the BSP model, it is easy to adapt to the
MapReduce style. If you think it is worth trying, I can do it.
2013/6/6 Ankur Gupta <gm...@gmail.com>
> I have previously worked with java, and Mahout is one project where I would
> love to contribute, be it in Java.
>
> On Thu, Jun 6, 2013 at 6:49 PM, Robin Anil <ro...@gmail.com> wrote:
>
> > Ankur you will be pleasantly surprised how easy it is to develop an
> > algorithm using Mahout's vector classes.
> >
> > Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
> >
> >
> > On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com>
> wrote:
> >
> > > We currently do not have any C++ code.
> > >
> > > The build complexity and maintenance and portability issues probably
> > > outweigh any advantages that I see at this time.
> > >
> > > We would love to have you contribute, but for now I think it is
> > > safest/easiest to assume Java.
> > >
> > >
> > > On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
> > > >wrote:
> > >
> > > > Hi
> > > > I want to contribute to mahout but one thing that holds me back is
> > that I
> > > > want to develop in C++ (It is just the comfort level). Is there any
> > part
> > > of
> > > > mahout which is developed in C++ or can be developed in C++. If so
> > kindly
> > > > let me know so that I can start working on understanding that and
> soon
> > > > start contributing.
> > > >
> > > > Thanks & Regards,
> > > > Ankur
> > > >
> > > >
> > > >
> > > > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
> > > wrote:
> > > >
> > > > > No.
> > > > >
> > > > > Mahout doesn't have any multi-level neural network implementation
> at
> > > this
> > > > > time.
> > > > >
> > > > >
> > > > >
> > > > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> > > > >
> > > > > > I'd like to phrase the question a bit differently.
> > > > > >
> > > > > > Does Mahout have a neural network algorithm (did it adopt
> Neuroph's
> > > or
> > > > > native)?
> > > > > >
> > > > > > Thanks.
> > > > > >
> > > > > >
> > > > > > -----Original Message----- From: Mimi Tam
> > > > > > Sent: Tuesday, June 04, 2013 10:08 AM
> > > > > > To: dev@mahout.apache.org
> > > > > > Subject: Question on Mahout (adopted FANN libraries?) if I may...
> > > > > >
> > > > > > Sorry, I am a late comer.
> > > > > >
> > > > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network
> > with
> > > > > > backpropagation learning on Hadoop ever implemented and is part
> of
> > > > Mahout
> > > > > > now?
> > > > > >
> > > > > > Is this using the FANN libraries?
> > > > > >
> > > > > > Many Thanks...Mimi
> > > > > >
> > > > > >
> > > > > > -----Original Message----- From: Robin Anil (JIRA)
> > > > > > Sent: Tuesday, June 04, 2013 9:51 AM
> > > > > > To: dev@mahout.apache.org
> > > > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> > > > > Serialization
> > > > > > of NonZeros
> > > > > >
> > > > > >
> > > > > > [
> > > > > >
> > > > >
> > > >
> > >
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > > > > > ]
> > > > > >
> > > > > > Robin Anil updated MAHOUT-1240:
> > > > > > -------------------------------
> > > > > >
> > > > > > Status: Patch Available (was: Open)
> > > > > >
> > > > > >> Randomized testing and Serialization of NonZeros
> > > > > >> ------------------------------------------------
> > > > > >>
> > > > > >> Key: MAHOUT-1240
> > > > > >> URL:
> > > https://issues.apache.org/jira/browse/MAHOUT-1240
> > > > > >> Project: Mahout
> > > > > >> Issue Type: Bug
> > > > > >> Reporter: Robin Anil
> > > > > >> Assignee: Robin Anil
> > > > > >> Fix For: 0.8
> > > > > >>
> > > > > >> Attachments: MAHOUT-1240.patch
> > > > > >>
> > > > > >>
> > > > > >> Currently the nonZero iterator does not guarantee nonZero
> > iteration
> > > > for
> > > > > certain vectors (RASV, SASV) for performance reason. However vector
> > > view
> > > > > iterator adds a zero check.. To be correct we have to either remove
> > the
> > > > > check or do correct non zero serialization everywhere. However this
> > > means
> > > > > going over the vectors in two passes. Given that is pretty fast
> > > already,
> > > > I
> > > > > am fixing the logic bug. We can tackle the speed up for the next
> > > release.
> > > > > >> This also adds a randomized test for serialization that catches
> > all
> > > > > such bugs.
> > > > > >
> > > > > > --
> > > > > > This message is automatically generated by JIRA.
> > > > > > If you think it was sent incorrectly, please contact your JIRA
> > > > > > administrators
> > > > > > For more information on JIRA, see:
> > > > > http://www.atlassian.com/software/jira
> > > > > >
> > > > >
> > > > >
> > > >
> > >
> >
>
>
>
> --
> Ankur Gupta,
> Software Engineer
> Samsung India Software Operations Pvt. Ltd.
> Bangalore, Karnataka
> India
>
--
------
Yexi Jiang,
ECS 251, yjian004@cs.fiu.edu
School of Computer and Information Science,
Florida International University
Homepage: http://users.cis.fiu.edu/~yjian004/
Re: Mahout - neural network algorithm?
Posted by Ankur Gupta <gm...@gmail.com>.
I have previously worked with java, and Mahout is one project where I would
love to contribute, be it in Java.
On Thu, Jun 6, 2013 at 6:49 PM, Robin Anil <ro...@gmail.com> wrote:
> Ankur you will be pleasantly surprised how easy it is to develop an
> algorithm using Mahout's vector classes.
>
> Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
>
>
> On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com> wrote:
>
> > We currently do not have any C++ code.
> >
> > The build complexity and maintenance and portability issues probably
> > outweigh any advantages that I see at this time.
> >
> > We would love to have you contribute, but for now I think it is
> > safest/easiest to assume Java.
> >
> >
> > On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
> > >wrote:
> >
> > > Hi
> > > I want to contribute to mahout but one thing that holds me back is
> that I
> > > want to develop in C++ (It is just the comfort level). Is there any
> part
> > of
> > > mahout which is developed in C++ or can be developed in C++. If so
> kindly
> > > let me know so that I can start working on understanding that and soon
> > > start contributing.
> > >
> > > Thanks & Regards,
> > > Ankur
> > >
> > >
> > >
> > > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
> > wrote:
> > >
> > > > No.
> > > >
> > > > Mahout doesn't have any multi-level neural network implementation at
> > this
> > > > time.
> > > >
> > > >
> > > >
> > > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> > > >
> > > > > I'd like to phrase the question a bit differently.
> > > > >
> > > > > Does Mahout have a neural network algorithm (did it adopt Neuroph's
> > or
> > > > native)?
> > > > >
> > > > > Thanks.
> > > > >
> > > > >
> > > > > -----Original Message----- From: Mimi Tam
> > > > > Sent: Tuesday, June 04, 2013 10:08 AM
> > > > > To: dev@mahout.apache.org
> > > > > Subject: Question on Mahout (adopted FANN libraries?) if I may...
> > > > >
> > > > > Sorry, I am a late comer.
> > > > >
> > > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network
> with
> > > > > backpropagation learning on Hadoop ever implemented and is part of
> > > Mahout
> > > > > now?
> > > > >
> > > > > Is this using the FANN libraries?
> > > > >
> > > > > Many Thanks...Mimi
> > > > >
> > > > >
> > > > > -----Original Message----- From: Robin Anil (JIRA)
> > > > > Sent: Tuesday, June 04, 2013 9:51 AM
> > > > > To: dev@mahout.apache.org
> > > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> > > > Serialization
> > > > > of NonZeros
> > > > >
> > > > >
> > > > > [
> > > > >
> > > >
> > >
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > > > > ]
> > > > >
> > > > > Robin Anil updated MAHOUT-1240:
> > > > > -------------------------------
> > > > >
> > > > > Status: Patch Available (was: Open)
> > > > >
> > > > >> Randomized testing and Serialization of NonZeros
> > > > >> ------------------------------------------------
> > > > >>
> > > > >> Key: MAHOUT-1240
> > > > >> URL:
> > https://issues.apache.org/jira/browse/MAHOUT-1240
> > > > >> Project: Mahout
> > > > >> Issue Type: Bug
> > > > >> Reporter: Robin Anil
> > > > >> Assignee: Robin Anil
> > > > >> Fix For: 0.8
> > > > >>
> > > > >> Attachments: MAHOUT-1240.patch
> > > > >>
> > > > >>
> > > > >> Currently the nonZero iterator does not guarantee nonZero
> iteration
> > > for
> > > > certain vectors (RASV, SASV) for performance reason. However vector
> > view
> > > > iterator adds a zero check.. To be correct we have to either remove
> the
> > > > check or do correct non zero serialization everywhere. However this
> > means
> > > > going over the vectors in two passes. Given that is pretty fast
> > already,
> > > I
> > > > am fixing the logic bug. We can tackle the speed up for the next
> > release.
> > > > >> This also adds a randomized test for serialization that catches
> all
> > > > such bugs.
> > > > >
> > > > > --
> > > > > This message is automatically generated by JIRA.
> > > > > If you think it was sent incorrectly, please contact your JIRA
> > > > > administrators
> > > > > For more information on JIRA, see:
> > > > http://www.atlassian.com/software/jira
> > > > >
> > > >
> > > >
> > >
> >
>
--
Ankur Gupta,
Software Engineer
Samsung India Software Operations Pvt. Ltd.
Bangalore, Karnataka
India
Re: Mahout - neural network algorithm?
Posted by Robin Anil <ro...@gmail.com>.
Ankur you will be pleasantly surprised how easy it is to develop an
algorithm using Mahout's vector classes.
Robin Anil | Software Engineer | +1 312 869 2602 | Google Inc.
On Thu, Jun 6, 2013 at 2:39 PM, Ted Dunning <te...@gmail.com> wrote:
> We currently do not have any C++ code.
>
> The build complexity and maintenance and portability issues probably
> outweigh any advantages that I see at this time.
>
> We would love to have you contribute, but for now I think it is
> safest/easiest to assume Java.
>
>
> On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm.ankurgupta@gmail.com
> >wrote:
>
> > Hi
> > I want to contribute to mahout but one thing that holds me back is that I
> > want to develop in C++ (It is just the comfort level). Is there any part
> of
> > mahout which is developed in C++ or can be developed in C++. If so kindly
> > let me know so that I can start working on understanding that and soon
> > start contributing.
> >
> > Thanks & Regards,
> > Ankur
> >
> >
> >
> > On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com>
> wrote:
> >
> > > No.
> > >
> > > Mahout doesn't have any multi-level neural network implementation at
> this
> > > time.
> > >
> > >
> > >
> > > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> > >
> > > > I'd like to phrase the question a bit differently.
> > > >
> > > > Does Mahout have a neural network algorithm (did it adopt Neuroph's
> or
> > > native)?
> > > >
> > > > Thanks.
> > > >
> > > >
> > > > -----Original Message----- From: Mimi Tam
> > > > Sent: Tuesday, June 04, 2013 10:08 AM
> > > > To: dev@mahout.apache.org
> > > > Subject: Question on Mahout (adopted FANN libraries?) if I may...
> > > >
> > > > Sorry, I am a late comer.
> > > >
> > > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
> > > > backpropagation learning on Hadoop ever implemented and is part of
> > Mahout
> > > > now?
> > > >
> > > > Is this using the FANN libraries?
> > > >
> > > > Many Thanks...Mimi
> > > >
> > > >
> > > > -----Original Message----- From: Robin Anil (JIRA)
> > > > Sent: Tuesday, June 04, 2013 9:51 AM
> > > > To: dev@mahout.apache.org
> > > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> > > Serialization
> > > > of NonZeros
> > > >
> > > >
> > > > [
> > > >
> > >
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > > > ]
> > > >
> > > > Robin Anil updated MAHOUT-1240:
> > > > -------------------------------
> > > >
> > > > Status: Patch Available (was: Open)
> > > >
> > > >> Randomized testing and Serialization of NonZeros
> > > >> ------------------------------------------------
> > > >>
> > > >> Key: MAHOUT-1240
> > > >> URL:
> https://issues.apache.org/jira/browse/MAHOUT-1240
> > > >> Project: Mahout
> > > >> Issue Type: Bug
> > > >> Reporter: Robin Anil
> > > >> Assignee: Robin Anil
> > > >> Fix For: 0.8
> > > >>
> > > >> Attachments: MAHOUT-1240.patch
> > > >>
> > > >>
> > > >> Currently the nonZero iterator does not guarantee nonZero iteration
> > for
> > > certain vectors (RASV, SASV) for performance reason. However vector
> view
> > > iterator adds a zero check.. To be correct we have to either remove the
> > > check or do correct non zero serialization everywhere. However this
> means
> > > going over the vectors in two passes. Given that is pretty fast
> already,
> > I
> > > am fixing the logic bug. We can tackle the speed up for the next
> release.
> > > >> This also adds a randomized test for serialization that catches all
> > > such bugs.
> > > >
> > > > --
> > > > This message is automatically generated by JIRA.
> > > > If you think it was sent incorrectly, please contact your JIRA
> > > > administrators
> > > > For more information on JIRA, see:
> > > http://www.atlassian.com/software/jira
> > > >
> > >
> > >
> >
>
Re: Mahout - neural network algorithm?
Posted by Ted Dunning <te...@gmail.com>.
We currently do not have any C++ code.
The build complexity and maintenance and portability issues probably
outweigh any advantages that I see at this time.
We would love to have you contribute, but for now I think it is
safest/easiest to assume Java.
On Thu, Jun 6, 2013 at 11:14 AM, Ankur Gupta <gm...@gmail.com>wrote:
> Hi
> I want to contribute to mahout but one thing that holds me back is that I
> want to develop in C++ (It is just the comfort level). Is there any part of
> mahout which is developed in C++ or can be developed in C++. If so kindly
> let me know so that I can start working on understanding that and soon
> start contributing.
>
> Thanks & Regards,
> Ankur
>
>
>
> On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com> wrote:
>
> > No.
> >
> > Mahout doesn't have any multi-level neural network implementation at this
> > time.
> >
> >
> >
> > On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> >
> > > I'd like to phrase the question a bit differently.
> > >
> > > Does Mahout have a neural network algorithm (did it adopt Neuroph's or
> > native)?
> > >
> > > Thanks.
> > >
> > >
> > > -----Original Message----- From: Mimi Tam
> > > Sent: Tuesday, June 04, 2013 10:08 AM
> > > To: dev@mahout.apache.org
> > > Subject: Question on Mahout (adopted FANN libraries?) if I may...
> > >
> > > Sorry, I am a late comer.
> > >
> > > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
> > > backpropagation learning on Hadoop ever implemented and is part of
> Mahout
> > > now?
> > >
> > > Is this using the FANN libraries?
> > >
> > > Many Thanks...Mimi
> > >
> > >
> > > -----Original Message----- From: Robin Anil (JIRA)
> > > Sent: Tuesday, June 04, 2013 9:51 AM
> > > To: dev@mahout.apache.org
> > > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> > Serialization
> > > of NonZeros
> > >
> > >
> > > [
> > >
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > > ]
> > >
> > > Robin Anil updated MAHOUT-1240:
> > > -------------------------------
> > >
> > > Status: Patch Available (was: Open)
> > >
> > >> Randomized testing and Serialization of NonZeros
> > >> ------------------------------------------------
> > >>
> > >> Key: MAHOUT-1240
> > >> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
> > >> Project: Mahout
> > >> Issue Type: Bug
> > >> Reporter: Robin Anil
> > >> Assignee: Robin Anil
> > >> Fix For: 0.8
> > >>
> > >> Attachments: MAHOUT-1240.patch
> > >>
> > >>
> > >> Currently the nonZero iterator does not guarantee nonZero iteration
> for
> > certain vectors (RASV, SASV) for performance reason. However vector view
> > iterator adds a zero check.. To be correct we have to either remove the
> > check or do correct non zero serialization everywhere. However this means
> > going over the vectors in two passes. Given that is pretty fast already,
> I
> > am fixing the logic bug. We can tackle the speed up for the next release.
> > >> This also adds a randomized test for serialization that catches all
> > such bugs.
> > >
> > > --
> > > This message is automatically generated by JIRA.
> > > If you think it was sent incorrectly, please contact your JIRA
> > > administrators
> > > For more information on JIRA, see:
> > http://www.atlassian.com/software/jira
> > >
> >
> >
>
Re: Mahout - neural network algorithm?
Posted by Ankur Gupta <gm...@gmail.com>.
Hi
I want to contribute to mahout but one thing that holds me back is that I
want to develop in C++ (It is just the comfort level). Is there any part of
mahout which is developed in C++ or can be developed in C++. If so kindly
let me know so that I can start working on understanding that and soon
start contributing.
Thanks & Regards,
Ankur
On Thu, Jun 6, 2013 at 1:20 PM, Ted Dunning <te...@gmail.com> wrote:
> No.
>
> Mahout doesn't have any multi-level neural network implementation at this
> time.
>
>
>
> On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
>
> > I'd like to phrase the question a bit differently.
> >
> > Does Mahout have a neural network algorithm (did it adopt Neuroph's or
> native)?
> >
> > Thanks.
> >
> >
> > -----Original Message----- From: Mimi Tam
> > Sent: Tuesday, June 04, 2013 10:08 AM
> > To: dev@mahout.apache.org
> > Subject: Question on Mahout (adopted FANN libraries?) if I may...
> >
> > Sorry, I am a late comer.
> >
> > Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
> > backpropagation learning on Hadoop ever implemented and is part of Mahout
> > now?
> >
> > Is this using the FANN libraries?
> >
> > Many Thanks...Mimi
> >
> >
> > -----Original Message----- From: Robin Anil (JIRA)
> > Sent: Tuesday, June 04, 2013 9:51 AM
> > To: dev@mahout.apache.org
> > Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and
> Serialization
> > of NonZeros
> >
> >
> > [
> >
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> > ]
> >
> > Robin Anil updated MAHOUT-1240:
> > -------------------------------
> >
> > Status: Patch Available (was: Open)
> >
> >> Randomized testing and Serialization of NonZeros
> >> ------------------------------------------------
> >>
> >> Key: MAHOUT-1240
> >> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
> >> Project: Mahout
> >> Issue Type: Bug
> >> Reporter: Robin Anil
> >> Assignee: Robin Anil
> >> Fix For: 0.8
> >>
> >> Attachments: MAHOUT-1240.patch
> >>
> >>
> >> Currently the nonZero iterator does not guarantee nonZero iteration for
> certain vectors (RASV, SASV) for performance reason. However vector view
> iterator adds a zero check.. To be correct we have to either remove the
> check or do correct non zero serialization everywhere. However this means
> going over the vectors in two passes. Given that is pretty fast already, I
> am fixing the logic bug. We can tackle the speed up for the next release.
> >> This also adds a randomized test for serialization that catches all
> such bugs.
> >
> > --
> > This message is automatically generated by JIRA.
> > If you think it was sent incorrectly, please contact your JIRA
> > administrators
> > For more information on JIRA, see:
> http://www.atlassian.com/software/jira
> >
>
>
Re: Mahout - neural network algorithm?
Posted by Ted Dunning <te...@gmail.com>.
No.
Mahout doesn't have any multi-level neural network implementation at this time.
On Jun 4, 2013, at 4:49 PM, Mimi Tam wrote:
> I'd like to phrase the question a bit differently.
>
> Does Mahout have a neural network algorithm (did it adopt Neuroph's or native)?
>
> Thanks.
>
>
> -----Original Message----- From: Mimi Tam
> Sent: Tuesday, June 04, 2013 10:08 AM
> To: dev@mahout.apache.org
> Subject: Question on Mahout (adopted FANN libraries?) if I may...
>
> Sorry, I am a late comer.
>
> Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
> backpropagation learning on Hadoop ever implemented and is part of Mahout
> now?
>
> Is this using the FANN libraries?
>
> Many Thanks...Mimi
>
>
> -----Original Message----- From: Robin Anil (JIRA)
> Sent: Tuesday, June 04, 2013 9:51 AM
> To: dev@mahout.apache.org
> Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and Serialization
> of NonZeros
>
>
> [
> https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
> ]
>
> Robin Anil updated MAHOUT-1240:
> -------------------------------
>
> Status: Patch Available (was: Open)
>
>> Randomized testing and Serialization of NonZeros
>> ------------------------------------------------
>>
>> Key: MAHOUT-1240
>> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
>> Project: Mahout
>> Issue Type: Bug
>> Reporter: Robin Anil
>> Assignee: Robin Anil
>> Fix For: 0.8
>>
>> Attachments: MAHOUT-1240.patch
>>
>>
>> Currently the nonZero iterator does not guarantee nonZero iteration for certain vectors (RASV, SASV) for performance reason. However vector view iterator adds a zero check.. To be correct we have to either remove the check or do correct non zero serialization everywhere. However this means going over the vectors in two passes. Given that is pretty fast already, I am fixing the logic bug. We can tackle the speed up for the next release.
>> This also adds a randomized test for serialization that catches all such bugs.
>
> --
> This message is automatically generated by JIRA.
> If you think it was sent incorrectly, please contact your JIRA
> administrators
> For more information on JIRA, see: http://www.atlassian.com/software/jira
>
Mahout - neural network algorithm?
Posted by Mimi Tam <mi...@ieee.org>.
I'd like to phrase the question a bit differently.
Does Mahout have a neural network algorithm (did it adopt Neuroph's or
native)?
Thanks.
-----Original Message-----
From: Mimi Tam
Sent: Tuesday, June 04, 2013 10:08 AM
To: dev@mahout.apache.org
Subject: Question on Mahout (adopted FANN libraries?) if I may...
Sorry, I am a late comer.
Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
backpropagation learning on Hadoop ever implemented and is part of Mahout
now?
Is this using the FANN libraries?
Many Thanks...Mimi
-----Original Message-----
From: Robin Anil (JIRA)
Sent: Tuesday, June 04, 2013 9:51 AM
To: dev@mahout.apache.org
Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and Serialization
of NonZeros
[
https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robin Anil updated MAHOUT-1240:
-------------------------------
Status: Patch Available (was: Open)
> Randomized testing and Serialization of NonZeros
> ------------------------------------------------
>
> Key: MAHOUT-1240
> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
> Project: Mahout
> Issue Type: Bug
> Reporter: Robin Anil
> Assignee: Robin Anil
> Fix For: 0.8
>
> Attachments: MAHOUT-1240.patch
>
>
> Currently the nonZero iterator does not guarantee nonZero iteration for
> certain vectors (RASV, SASV) for performance reason. However vector view
> iterator adds a zero check.. To be correct we have to either remove the
> check or do correct non zero serialization everywhere. However this means
> going over the vectors in two passes. Given that is pretty fast already, I
> am fixing the logic bug. We can tackle the speed up for the next release.
> This also adds a randomized test for serialization that catches all such
> bugs.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA
administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
Question on Mahout (adopted FANN libraries?) if I may...
Posted by Mimi Tam <mi...@ieee.org>.
Sorry, I am a late comer.
Did Mahout/Mahout-36 [GSOC] Proposal to implement Neural Network with
backpropagation learning on Hadoop ever implemented and is part of Mahout
now?
Is this using the FANN libraries?
Many Thanks...Mimi
-----Original Message-----
From: Robin Anil (JIRA)
Sent: Tuesday, June 04, 2013 9:51 AM
To: dev@mahout.apache.org
Subject: [jira] [Updated] (MAHOUT-1240) Randomized testing and Serialization
of NonZeros
[
https://issues.apache.org/jira/browse/MAHOUT-1240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Robin Anil updated MAHOUT-1240:
-------------------------------
Status: Patch Available (was: Open)
> Randomized testing and Serialization of NonZeros
> ------------------------------------------------
>
> Key: MAHOUT-1240
> URL: https://issues.apache.org/jira/browse/MAHOUT-1240
> Project: Mahout
> Issue Type: Bug
> Reporter: Robin Anil
> Assignee: Robin Anil
> Fix For: 0.8
>
> Attachments: MAHOUT-1240.patch
>
>
> Currently the nonZero iterator does not guarantee nonZero iteration for
> certain vectors (RASV, SASV) for performance reason. However vector view
> iterator adds a zero check.. To be correct we have to either remove the
> check or do correct non zero serialization everywhere. However this means
> going over the vectors in two passes. Given that is pretty fast already, I
> am fixing the logic bug. We can tackle the speed up for the next release.
> This also adds a randomized test for serialization that catches all such
> bugs.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA
administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira