Re: assimilation of mshadow into the MXNet codebase
Hi, > Here's the second update. At the moment we are only missing ICLAs from 20 > (out of 70) contributors, accounting for 31 (out of 913) commits left. IMO That's still a significant number of commits and people. > Regarding whether the contributors are employed by a company that requires > CCLA for contribution, I have no way of verifying the contributors' > employment status at the time of contribution, and not enough bandwidth to > verify the individual company policies on such contribution. As such, I will > solely rely on the ICLAs. The risk is that the IPMC may require more than that, I can’t predict how other IPMC members will vote in this case. Intel have supplied CCLAs in the past and there are people on that list not covered by a CCLA could be a concern along with the missing ICLAs. Thanks, Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, Here's the second update. At the moment we are only missing ICLAs from 20 (out of 70) contributors, accounting for 31 (out of 913) commits left. 3 @zhenlinluo 3 @jpauwels 3 @hjk41 3 @DrustZ 2 @zhangchen-qinyinghua 2 @yinghu5 2 @reyoung 1 @xinyu-intel 1 @xingmingjie 1 @qiaohaijun 1 @loveisp 1 @lebeg 1 @kaleidoscopical 1 @jason-xuan 1 @happynear 1 @glingyan 1 @asitstands 1 @antoine-wdg-rmz 1 @alextnewman 1 @Harmonicahappy Regarding whether the contributors are employed by a company that requires CCLA for contribution, I have no way of verifying the contributors' employment status at the time of contribution, and not enough bandwidth to verify the individual company policies on such contribution. As such, I will solely rely on the ICLAs. Given the current status, I think the rest of the 31 commits is manageable for me even if I end up having to revert and rework all of them. Let me know if you have any concern on starting the IP clearance process. Otherwise I think we can start it on general@incubator soon. Cheers, Sheng On Sun, Jul 26, 2020 at 8:59 PM Justin Mclean wrote: > Hi, > > In the case of Intel and other companies, it may be that their employee > contracts do not allow employees to contribute to OS projects. It more > likely that the contributor doesn’t own copyright of the code but their > employer does. A CCLA give a clear indication that the contributors are > intact allowed to contrive code and own the copyright of their > contributions. We have CCLAs from Intel on file from other contributions so > it would seem that Intel requires this. > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, In the case of Intel and other companies, it may be that their employee contracts do not allow employees to contribute to OS projects. It more likely that the contributor doesn’t own copyright of the code but their employer does. A CCLA give a clear indication that the contributors are intact allowed to contrive code and own the copyright of their contributions. We have CCLAs from Intel on file from other contributions so it would seem that Intel requires this. Thanks, Justin
Re: assimilation of mshadow into the MXNet codebase
Justin, Are you OK with proceeding? Regards, Sheng On Sun, Jul 26, 2020 at 8:30 PM Tianqi Chen wrote: > As long as we have CLA covering for the majority of the code(which I > believe so), I think we should be good. > Just like the case of Apache only requires iCLA from committers. > > The rationale is that normal contributions are already in the form of ALv2, > in the case of a(unlikely) dispute, the community can quickly rewrite the > code(since that is non-majority). > > TQ > > On Sun, Jul 26, 2020 at 5:49 PM Sheng Zha wrote: > > > Hi Justin, > > > > Thanks, that's a good point. I think we have already received CCLA from > > Intel. I will take that into account when providing the next update. > > > > Regards, > > Sheng > > > > On Sun, Jul 26, 2020 at 5:39 PM Justin Mclean > > wrote: > > > > > Hi, > > > > > > > Several peoples in below list are from Intel and I have added them > into > > > CC. > > > > > > Has Intel signed a CCLA? And if so does it list people who are allowed > to > > > contribute to this project? Are there any others on that list who > > > employer’s may need to also sign CCLAs if we don’t have them? > > > > > > Thanks, > > > Justin > > >
Re: assimilation of mshadow into the MXNet codebase
As long as we have CLA covering for the majority of the code(which I believe so), I think we should be good. Just like the case of Apache only requires iCLA from committers. The rationale is that normal contributions are already in the form of ALv2, in the case of a(unlikely) dispute, the community can quickly rewrite the code(since that is non-majority). TQ On Sun, Jul 26, 2020 at 5:49 PM Sheng Zha wrote: > Hi Justin, > > Thanks, that's a good point. I think we have already received CCLA from > Intel. I will take that into account when providing the next update. > > Regards, > Sheng > > On Sun, Jul 26, 2020 at 5:39 PM Justin Mclean > wrote: > > > Hi, > > > > > Several peoples in below list are from Intel and I have added them into > > CC. > > > > Has Intel signed a CCLA? And if so does it list people who are allowed to > > contribute to this project? Are there any others on that list who > > employer’s may need to also sign CCLAs if we don’t have them? > > > > Thanks, > > Justin >
Re: assimilation of mshadow into the MXNet codebase
Hi Justin, Thanks, that's a good point. I think we have already received CCLA from Intel. I will take that into account when providing the next update. Regards, Sheng On Sun, Jul 26, 2020 at 5:39 PM Justin Mclean wrote: > Hi, > > > Several peoples in below list are from Intel and I have added them into > CC. > > Has Intel signed a CCLA? And if so does it list people who are allowed to > contribute to this project? Are there any others on that list who > employer’s may need to also sign CCLAs if we don’t have them? > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, > Several peoples in below list are from Intel and I have added them into CC. Has Intel signed a CCLA? And if so does it list people who are allowed to contribute to this project? Are there any others on that list who employer’s may need to also sign CCLAs if we don’t have them? Thanks, Justin
RE: assimilation of mshadow into the MXNet codebase
Several peoples in below list are from Intel and I have added them into CC.
Sheng, you can contact with them for ICLA.
Thanks,
--Patric
> -Original Message-
> From: Sheng Zha
> Sent: Monday, July 27, 2020 5:33 AM
> To: Justin Mclean
> Cc: d...@mxnet.apache.org; Wall Michael ; Bob Paulin
> ; wei...@apache.org; jason...@apache.org; Chen, Ciyong
>
> Subject: Re: assimilation of mshadow into the MXNet codebase
>
> Hi,
>
> Here's an update on this issue. We are still missing the ICLAs from 32 (out
> of 70)
> mshadow contributors, accounting for a total of 62 (out of 913) commits. (@ap-
> hynninen passed away a few years ago and is not included). I reached out to
> them through email and other channels to collect ICLA for mshadow. I will wait
> for a day or two before updating on the progress again, and we can decide then
> whether we are good to start the IP clearance.
>
> The complete list of mshadow contributors' GitHub logins that are missing ICLA
> is here ("#commits @github-login"):
>
> 8 @Lorrainexun
> 6 @tornadomeet
> 5 @asmushetzel
> 3 @zhenlinluo
> 3 @stefanhenneking
> 3 @jpauwels
> 3 @hjk41
> 3 @DrustZ
> 2 @zhangchen-qinyinghua
> 2 @yinghu5
> 2 @reyoung
> 2 @forwchen
> 1 @yupbank
> 1 @yllan
> 1 @xinyu-intel
> 1 @xingmingjie
> 1 @xianyi
> 1 @tdomhan
> 1 @siemanko
> 1 @qiaohaijun
> 1 @maxint
> 1 @loveisp
> 1 @lebeg
> 1 @kdavis-mozilla
> 1 @kaleidoscopical
> 1 @jason-xuan
> 1 @happynear
> 1 @glingyan
> 1 @asitstands
> 1 @antoine-wdg-rmz
> 1 @alextnewman
> 1 @Harmonicahappy
>
> Best,
> Sheng
>
> On Thu, Jul 23, 2020 at 12:28 AM Sheng Zha wrote:
>
> > Hi,
> >
> > No, I don’t think we used ICLAs for mshadow before.
> >
> > Out of the 42 people who made more than 1 commit or more than 10 lines
> > of code change to mshadow, 26 signed ICLA with Apache (and
> > additionally one member is unfortunately deceased...). Would this be a
> > better criteria as “the major ones”? I wasn’t part of the initial code
> > donation or the initial PPMC group, so apologies if the questions were
> > silly.
> >
> > I think the rest of the commits are manageable so that I could do a
> > revert and rework for those commits if/when necessary.
> >
> > Regards,
> > Sheng
> >
> > > On Jul 22, 2020, at 11:50 PM, Justin Mclean
> > >
> > wrote:
> > >
> > > Hi,
> > >
> > >> Thanks for clarifying. All contributors who made more than 10
> > >> commits
> > to msahdow before are committers of MXNet, so their ICLAs should
> > already be on file: tqchen, bingxu, eric.xie, sxjscience, mli,
> > yajiedesign [1]. If you think this is OK, one of the mentors or I can start
> > the
> notification.
> > >
> > >
> > > What about the other 60 contributors? More than 10 commits is not a
> > > line
> > I would feel comfortable with. You need to be able to account for the
> > IP provenance of every line of code, just like in your initial code
> > donation.
> > It would probably be best to make a list all contributors and if they
> > have an ICLA or not. Did the mshadow project use ICLAs? If so that may also
> help.
> > >
> > > Thanks,
> > > Justin
> >
Re: assimilation of mshadow into the MXNet codebase
Hi,
Here's an update on this issue. We are still missing the ICLAs from 32 (out
of 70) mshadow contributors, accounting for a total of 62 (out of 913)
commits. (@ap-hynninen passed away a few years ago and is not included). I
reached out to them through email and other channels to collect ICLA for
mshadow. I will wait for a day or two before updating on the progress
again, and we can decide then whether we are good to start the IP clearance.
The complete list of mshadow contributors' GitHub logins that are missing
ICLA is here ("#commits @github-login"):
8 @Lorrainexun
6 @tornadomeet
5 @asmushetzel
3 @zhenlinluo
3 @stefanhenneking
3 @jpauwels
3 @hjk41
3 @DrustZ
2 @zhangchen-qinyinghua
2 @yinghu5
2 @reyoung
2 @forwchen
1 @yupbank
1 @yllan
1 @xinyu-intel
1 @xingmingjie
1 @xianyi
1 @tdomhan
1 @siemanko
1 @qiaohaijun
1 @maxint
1 @loveisp
1 @lebeg
1 @kdavis-mozilla
1 @kaleidoscopical
1 @jason-xuan
1 @happynear
1 @glingyan
1 @asitstands
1 @antoine-wdg-rmz
1 @alextnewman
1 @Harmonicahappy
Best,
Sheng
On Thu, Jul 23, 2020 at 12:28 AM Sheng Zha wrote:
> Hi,
>
> No, I don’t think we used ICLAs for mshadow before.
>
> Out of the 42 people who made more than 1 commit or more than 10 lines of
> code change to mshadow, 26 signed ICLA with Apache (and additionally one
> member is unfortunately deceased...). Would this be a better criteria as
> “the major ones”? I wasn’t part of the initial code donation or the initial
> PPMC group, so apologies if the questions were silly.
>
> I think the rest of the commits are manageable so that I could do a revert
> and rework for those commits if/when necessary.
>
> Regards,
> Sheng
>
> > On Jul 22, 2020, at 11:50 PM, Justin Mclean
> wrote:
> >
> > Hi,
> >
> >> Thanks for clarifying. All contributors who made more than 10 commits
> to msahdow before are committers of MXNet, so their ICLAs should already be
> on file: tqchen, bingxu, eric.xie, sxjscience, mli, yajiedesign [1]. If you
> think this is OK, one of the mentors or I can start the notification.
> >
> >
> > What about the other 60 contributors? More than 10 commits is not a line
> I would feel comfortable with. You need to be able to account for the IP
> provenance of every line of code, just like in your initial code donation.
> It would probably be best to make a list all contributors and if they have
> an ICLA or not. Did the mshadow project use ICLAs? If so that may also help.
> >
> > Thanks,
> > Justin
>
RE: assimilation of mshadow into the MXNet codebase
Hi Justin, We're doing 1.7.0 source release recently, and the vote on dev@ was passed, vote thread [1], result thread[2]. Seems the discussion on mshadow donation is still not finalized, may I know if you have any concern to proceed the current release under DISCLAIMER-WIP? Thanks, -Ciyong [1] https://lists.apache.org/thread.html/r525a961a10f69bdfb255c64f0be0589bb70efdd880c1be87c81c0c06%40%3Cdev.mxnet.apache.org%3E [2] https://lists.apache.org/thread.html/rbd53614ca01f714d00097a02d906895211336a14ce0e083865cf5144%40%3Cdev.mxnet.apache.org%3E -Original Message- From: Sheng Zha Sent: Thursday, July 23, 2020 3:29 PM To: Justin Mclean Cc: d...@mxnet.apache.org; Wall Michael ; Bob Paulin ; wei...@apache.org; jason...@apache.org Subject: Re: assimilation of mshadow into the MXNet codebase Hi, No, I don’t think we used ICLAs for mshadow before. Out of the 42 people who made more than 1 commit or more than 10 lines of code change to mshadow, 26 signed ICLA with Apache (and additionally one member is unfortunately deceased...). Would this be a better criteria as “the major ones”? I wasn’t part of the initial code donation or the initial PPMC group, so apologies if the questions were silly. I think the rest of the commits are manageable so that I could do a revert and rework for those commits if/when necessary. Regards, Sheng > On Jul 22, 2020, at 11:50 PM, Justin Mclean wrote: > > Hi, > >> Thanks for clarifying. All contributors who made more than 10 commits to >> msahdow before are committers of MXNet, so their ICLAs should already be on >> file: tqchen, bingxu, eric.xie, sxjscience, mli, yajiedesign [1]. If you >> think this is OK, one of the mentors or I can start the notification. > > > What about the other 60 contributors? More than 10 commits is not a line I > would feel comfortable with. You need to be able to account for the IP > provenance of every line of code, just like in your initial code donation. It > would probably be best to make a list all contributors and if they have an > ICLA or not. Did the mshadow project use ICLAs? If so that may also help. > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, No, I don’t think we used ICLAs for mshadow before. Out of the 42 people who made more than 1 commit or more than 10 lines of code change to mshadow, 26 signed ICLA with Apache (and additionally one member is unfortunately deceased...). Would this be a better criteria as “the major ones”? I wasn’t part of the initial code donation or the initial PPMC group, so apologies if the questions were silly. I think the rest of the commits are manageable so that I could do a revert and rework for those commits if/when necessary. Regards, Sheng > On Jul 22, 2020, at 11:50 PM, Justin Mclean wrote: > > Hi, > >> Thanks for clarifying. All contributors who made more than 10 commits to >> msahdow before are committers of MXNet, so their ICLAs should already be on >> file: tqchen, bingxu, eric.xie, sxjscience, mli, yajiedesign [1]. If you >> think this is OK, one of the mentors or I can start the notification. > > > What about the other 60 contributors? More than 10 commits is not a line I > would feel comfortable with. You need to be able to account for the IP > provenance of every line of code, just like in your initial code donation. It > would probably be best to make a list all contributors and if they have an > ICLA or not. Did the mshadow project use ICLAs? If so that may also help. > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, > Thanks for clarifying. All contributors who made more than 10 commits to > msahdow before are committers of MXNet, so their ICLAs should already be on > file: tqchen, bingxu, eric.xie, sxjscience, mli, yajiedesign [1]. If you > think this is OK, one of the mentors or I can start the notification. What about the other 60 contributors? More than 10 commits is not a line I would feel comfortable with. You need to be able to account for the IP provenance of every line of code, just like in your initial code donation. It would probably be best to make a list all contributors and if they have an ICLA or not. Did the mshadow project use ICLAs? If so that may also help. Thanks, Justin
Re: assimilation of mshadow into the MXNet codebase
Thanks for clarifying. All contributors who made more than 10 commits to msahdow before are committers of MXNet, so their ICLAs should already be on file: tqchen, bingxu, eric.xie, sxjscience, mli, yajiedesign [1]. If you think this is OK, one of the mentors or I can start the notification. Regards, Sheng [1] https://github.com/dmlc/mshadow/graphs/contributors On Wed, Jul 22, 2020 at 10:37 PM Justin Mclean wrote: > Hi, > > > Thank you, Justin. Though I’m still uncertain about what the definition > of IP clearance process is. > > The bit you quoted there is for an initial code base, it the second part > of that document you need to look at. > > In short as well as the SGA you need to get signed ICLA from all of the > contributors to the code base. It might be OK to just get the major ones > depending on the type of contributions. You then need to notify the > incubator of the IP clearance and see if they have any questions about it. > > Here’s an example: > > https://lists.apache.org/thread.html/r750880f7295c1a8c31c99e7a40f3466c177bd714254d0c98a506dede%40%3Cgeneral.incubator.apache.org%3E > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, > Thank you, Justin. Though I’m still uncertain about what the definition of IP > clearance process is. The bit you quoted there is for an initial code base, it the second part of that document you need to look at. In short as well as the SGA you need to get signed ICLA from all of the contributors to the code base. It might be OK to just get the major ones depending on the type of contributions. You then need to notify the incubator of the IP clearance and see if they have any questions about it. Here’s an example: https://lists.apache.org/thread.html/r750880f7295c1a8c31c99e7a40f3466c177bd714254d0c98a506dede%40%3Cgeneral.incubator.apache.org%3E Thanks, Justin
Re: assimilation of mshadow into the MXNet codebase
Thank you, Justin. Though I’m still uncertain about what the definition of IP clearance process is, I found the following paragraphs that seem relevant. Sounds like we need three votes from our mentors here for this acceptance. If that’s the case, I can start a vote on it. Regards, Sheng > The Incubator PMC must approve the clearance. This indicates that the project > is happy to receive the code donated. When a new podling is created, this is > done by the identification of existing codebases in the proposal. Otherwise, > the IPMC delegates this decision to the PPMC. > As usual, three binding votes are required. So, Mentors need to be involved > in IP clearance for podlings. If too few binding VOTEs are posted on list, > the VOTE will need to be posted to the general list for ratification. > On Jul 22, 2020, at 6:31 PM, Justin Mclean wrote: > > Hi, > > See also: > https://incubator.apache.org/guides/ip_clearance.html > > Thanks, > Justin
Re: assimilation of mshadow into the MXNet codebase
Hi, See also: https://incubator.apache.org/guides/ip_clearance.html Thanks, Justin
Re: assimilation of mshadow into the MXNet codebase
HI, > Yes and yes. I filed the software grant and received confirmation from > secretary@. As well as the software grant the incline code base needs to go through IP clearance. See [1] option 2. IP clearance involves making sure all all contributors have signed ICLAs and there are no license or other IP issues and getting IP clearance from the incubator. [2] Thanks, Justin 1. https://www.apache.org/foundation/how-it-works/legal.html#incoming-code 2. https://incubator.apache.org/ip-clearance/
Re: assimilation of mshadow into the MXNet codebase
Hi Justin, Yes and yes. I filed the software grant and received confirmation from secretary@. I’m not sure if I should be updating the page, and if so, how. Regards, Sheng > On Jul 22, 2020, at 1:59 AM, Justin Mclean wrote: > > Hi, > > Has the IP clearance process been followed? I don't see it listed on this > page [1] > > Does the current release being voted on contain this code? > > Thanks, > Justin > > 1. https://incubator.apache.org/ip-clearance/
Re: assimilation of mshadow into the MXNet codebase
Hi, Has the IP clearance process been followed? I don't see it listed on this page [1] Does the current release being voted on contain this code? Thanks, Justin 1. https://incubator.apache.org/ip-clearance/
Re: assimilation of mshadow into the MXNet codebase
I found the template in the link Marco provided and filed the software grant to the secretary. Sheng On Sun, Jul 5, 2020 at 10:09 AM Michael Wall wrote: > Yes, to secretary@. Do you need a template? > > Thanks Sheng > > Mike > > On Sun, Jul 5, 2020 at 12:59 PM Sheng Zha wrote: > > > > Hi Michael, > > > > Thanks for offering help. I can represent the code donors and file the > software grant. Should the filing go to secretary@? > > > > Sheng > > > > > On Jul 5, 2020, at 9:50 AM, Michael Wall wrote: > > > > > > Is this being tracked in a ticket anywhere? What help can I offer? > > > > > > Mike > > > > > >> On Fri, Jun 12, 2020 at 6:44 PM Marco de Abreu < > marco.g.ab...@gmail.com> wrote: > > >> > > >> Hi Sheng, > > >> > > >> since this is a "large one off code contribution", the policy [1] > states > > >> that they should be brought in through a software grant. > > >> > > >> Best regards, > > >> Marco > > >> > > >> [1]: https://www.apache.org/foundation/how-it-works/legal.html > > >> > > >>> On Fri, Jun 12, 2020 at 11:41 PM Sheng Zha > wrote: > > >>> > > >>> To mentors, > > >>> > > >>> Do we the PPMC need to fill out IP clearance for this code donation? > > >>> > > >>> -sz > > >>> > > >>> On 2019/04/24 21:19:49, Sheng Zha wrote: > > The community has agreed to donate mshadow to the mxnet code base. I > > >>> will start the migration and build logic changes soon. > > > > -sz > > > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > > I agree it would make development easier to donate mshadow to mxnet > > >>> code base, since mshadow is only used in MXNet. I support donating > the > > >>> mshadow code to mxnet and I started an RFC for this in mshadow [1]. > > > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > > > -sz > > > > > > On 2019/04/06 04:38:19, Tianqi Chen > wrote: > > >> Technically, mshadow is sufficient for MXNet. Adopting other > > >>> libraries ( > > >> eigen or xtensor) will unnecessarily increase the codebase > complexity > > >> without any additional gains. > > >> > > >> Given that mshadow is only used by mxnet. I do support donating it > > >>> into > > >> mxnet codebase. > > >> To respect the original mshadow community. I would recommend > > >>> starting a > > >> community RFC In the mshadow github issue for a week, before we > > >>> start the > > >> migrating process. > > >> Also, I would recommend a rebase merge just like the case of > > >>> MXNet.jl code > > >> base to preserve the contribution history. > > >> > > >> Tianqi > > >> > > >> > > >> On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > >> wrote: > > >> > > >>> Do you have a link to both of these proposals? > > >>> > > >>> On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya < > > >>> anirudhk...@gmail.com> > > >>> wrote: > > >>> > > Hi Pedro, > > > > mshadow is mostly used for tensor arithmetic. There have been > > >>> discussions > > about including it within mxnet. I think it is a good idea. > > > > As a more long term solution using libraries like eigen to > > >>> perform linear > > algebra operations was also suggested by anirudh2290@. I think > > >>> xtensor( > > https://github.com/QuantStack/xtensor ) can also be a candidate > > >>> here. > > > > - > > Anirudh > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > >>> pedro.larroy.li...@gmail.com> > > wrote: > > > > > Hi > > > > > > Some developers have noticed that working in mshadow is > > >>> cumbersome as > > > it's a 3rdparty subrepo. > > > > > > Since mshadow is a bunch of headers which don't have much of > > > independent tests / library functionality, me and other > > >>> developers > > > believe that it would be good to assimilate this code in the > > > repository for ease of contribution and changes without having > > >>> to go > > > trough contortions to test PRs that modify mshadow. > > > > > > Would anybody oppose this change? > > > > > > Thanks and have a nice weekend. > > > > > > Pedro. > > > > > > > >>> > > >> > > > > > > > >>> >
Re: assimilation of mshadow into the MXNet codebase
Yes, to secretary@. Do you need a template? Thanks Sheng Mike On Sun, Jul 5, 2020 at 12:59 PM Sheng Zha wrote: > > Hi Michael, > > Thanks for offering help. I can represent the code donors and file the > software grant. Should the filing go to secretary@? > > Sheng > > > On Jul 5, 2020, at 9:50 AM, Michael Wall wrote: > > > > Is this being tracked in a ticket anywhere? What help can I offer? > > > > Mike > > > >> On Fri, Jun 12, 2020 at 6:44 PM Marco de Abreu > >> wrote: > >> > >> Hi Sheng, > >> > >> since this is a "large one off code contribution", the policy [1] states > >> that they should be brought in through a software grant. > >> > >> Best regards, > >> Marco > >> > >> [1]: https://www.apache.org/foundation/how-it-works/legal.html > >> > >>> On Fri, Jun 12, 2020 at 11:41 PM Sheng Zha wrote: > >>> > >>> To mentors, > >>> > >>> Do we the PPMC need to fill out IP clearance for this code donation? > >>> > >>> -sz > >>> > >>> On 2019/04/24 21:19:49, Sheng Zha wrote: > The community has agreed to donate mshadow to the mxnet code base. I > >>> will start the migration and build logic changes soon. > > -sz > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > I agree it would make development easier to donate mshadow to mxnet > >>> code base, since mshadow is only used in MXNet. I support donating the > >>> mshadow code to mxnet and I started an RFC for this in mshadow [1]. > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > -sz > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > >> Technically, mshadow is sufficient for MXNet. Adopting other > >>> libraries ( > >> eigen or xtensor) will unnecessarily increase the codebase complexity > >> without any additional gains. > >> > >> Given that mshadow is only used by mxnet. I do support donating it > >>> into > >> mxnet codebase. > >> To respect the original mshadow community. I would recommend > >>> starting a > >> community RFC In the mshadow github issue for a week, before we > >>> start the > >> migrating process. > >> Also, I would recommend a rebase merge just like the case of > >>> MXNet.jl code > >> base to preserve the contribution history. > >> > >> Tianqi > >> > >> > >> On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > >> wrote: > >> > >>> Do you have a link to both of these proposals? > >>> > >>> On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya < > >>> anirudhk...@gmail.com> > >>> wrote: > >>> > Hi Pedro, > > mshadow is mostly used for tensor arithmetic. There have been > >>> discussions > about including it within mxnet. I think it is a good idea. > > As a more long term solution using libraries like eigen to > >>> perform linear > algebra operations was also suggested by anirudh2290@. I think > >>> xtensor( > https://github.com/QuantStack/xtensor ) can also be a candidate > >>> here. > > - > Anirudh > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > >>> pedro.larroy.li...@gmail.com> > wrote: > > > Hi > > > > Some developers have noticed that working in mshadow is > >>> cumbersome as > > it's a 3rdparty subrepo. > > > > Since mshadow is a bunch of headers which don't have much of > > independent tests / library functionality, me and other > >>> developers > > believe that it would be good to assimilate this code in the > > repository for ease of contribution and changes without having > >>> to go > > trough contortions to test PRs that modify mshadow. > > > > Would anybody oppose this change? > > > > Thanks and have a nice weekend. > > > > Pedro. > > > > >>> > >> > > > > >>>
Re: assimilation of mshadow into the MXNet codebase
Hi Michael, Thanks for offering help. I can represent the code donors and file the software grant. Should the filing go to secretary@? Sheng > On Jul 5, 2020, at 9:50 AM, Michael Wall wrote: > > Is this being tracked in a ticket anywhere? What help can I offer? > > Mike > >> On Fri, Jun 12, 2020 at 6:44 PM Marco de Abreu >> wrote: >> >> Hi Sheng, >> >> since this is a "large one off code contribution", the policy [1] states >> that they should be brought in through a software grant. >> >> Best regards, >> Marco >> >> [1]: https://www.apache.org/foundation/how-it-works/legal.html >> >>> On Fri, Jun 12, 2020 at 11:41 PM Sheng Zha wrote: >>> >>> To mentors, >>> >>> Do we the PPMC need to fill out IP clearance for this code donation? >>> >>> -sz >>> >>> On 2019/04/24 21:19:49, Sheng Zha wrote: The community has agreed to donate mshadow to the mxnet code base. I >>> will start the migration and build logic changes soon. -sz On 2019/04/07 21:47:39, Sheng Zha wrote: > I agree it would make development easier to donate mshadow to mxnet >>> code base, since mshadow is only used in MXNet. I support donating the >>> mshadow code to mxnet and I started an RFC for this in mshadow [1]. > > [1] https://github.com/dmlc/mshadow/issues/373 > > -sz > > On 2019/04/06 04:38:19, Tianqi Chen wrote: >> Technically, mshadow is sufficient for MXNet. Adopting other >>> libraries ( >> eigen or xtensor) will unnecessarily increase the codebase complexity >> without any additional gains. >> >> Given that mshadow is only used by mxnet. I do support donating it >>> into >> mxnet codebase. >> To respect the original mshadow community. I would recommend >>> starting a >> community RFC In the mshadow github issue for a week, before we >>> start the >> migrating process. >> Also, I would recommend a rebase merge just like the case of >>> MXNet.jl code >> base to preserve the contribution history. >> >> Tianqi >> >> >> On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque >> wrote: >> >>> Do you have a link to both of these proposals? >>> >>> On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya < >>> anirudhk...@gmail.com> >>> wrote: >>> Hi Pedro, mshadow is mostly used for tensor arithmetic. There have been >>> discussions about including it within mxnet. I think it is a good idea. As a more long term solution using libraries like eigen to >>> perform linear algebra operations was also suggested by anirudh2290@. I think >>> xtensor( https://github.com/QuantStack/xtensor ) can also be a candidate >>> here. - Anirudh On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < >>> pedro.larroy.li...@gmail.com> wrote: > Hi > > Some developers have noticed that working in mshadow is >>> cumbersome as > it's a 3rdparty subrepo. > > Since mshadow is a bunch of headers which don't have much of > independent tests / library functionality, me and other >>> developers > believe that it would be good to assimilate this code in the > repository for ease of contribution and changes without having >>> to go > trough contortions to test PRs that modify mshadow. > > Would anybody oppose this change? > > Thanks and have a nice weekend. > > Pedro. > >>> >> > >>>
Re: assimilation of mshadow into the MXNet codebase
Is this being tracked in a ticket anywhere? What help can I offer? Mike On Fri, Jun 12, 2020 at 6:44 PM Marco de Abreu wrote: > > Hi Sheng, > > since this is a "large one off code contribution", the policy [1] states > that they should be brought in through a software grant. > > Best regards, > Marco > > [1]: https://www.apache.org/foundation/how-it-works/legal.html > > On Fri, Jun 12, 2020 at 11:41 PM Sheng Zha wrote: > > > To mentors, > > > > Do we the PPMC need to fill out IP clearance for this code donation? > > > > -sz > > > > On 2019/04/24 21:19:49, Sheng Zha wrote: > > > The community has agreed to donate mshadow to the mxnet code base. I > > will start the migration and build logic changes soon. > > > > > > -sz > > > > > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > > > I agree it would make development easier to donate mshadow to mxnet > > code base, since mshadow is only used in MXNet. I support donating the > > mshadow code to mxnet and I started an RFC for this in mshadow [1]. > > > > > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > > > > > -sz > > > > > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > > > > Technically, mshadow is sufficient for MXNet. Adopting other > > libraries ( > > > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > > > without any additional gains. > > > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it > > into > > > > > mxnet codebase. > > > > > To respect the original mshadow community. I would recommend > > starting a > > > > > community RFC In the mshadow github issue for a week, before we > > start the > > > > > migrating process. > > > > > Also, I would recommend a rebase merge just like the case of > > MXNet.jl code > > > > > base to preserve the contribution history. > > > > > > > > > > Tianqi > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > > wrote: > > > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya < > > anirudhk...@gmail.com> > > > > > > wrote: > > > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > discussions > > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > > > As a more long term solution using libraries like eigen to > > perform linear > > > > > > > algebra operations was also suggested by anirudh2290@. I think > > xtensor( > > > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > > here. > > > > > > > > > > > > > > - > > > > > > > Anirudh > > > > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > > pedro.larroy.li...@gmail.com> > > > > > > > wrote: > > > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > > cumbersome as > > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > > independent tests / library functionality, me and other > > developers > > > > > > > > believe that it would be good to assimilate this code in the > > > > > > > > repository for ease of contribution and changes without having > > to go > > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Hi Sheng, since this is a "large one off code contribution", the policy [1] states that they should be brought in through a software grant. Best regards, Marco [1]: https://www.apache.org/foundation/how-it-works/legal.html On Fri, Jun 12, 2020 at 11:41 PM Sheng Zha wrote: > To mentors, > > Do we the PPMC need to fill out IP clearance for this code donation? > > -sz > > On 2019/04/24 21:19:49, Sheng Zha wrote: > > The community has agreed to donate mshadow to the mxnet code base. I > will start the migration and build logic changes soon. > > > > -sz > > > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > > I agree it would make development easier to donate mshadow to mxnet > code base, since mshadow is only used in MXNet. I support donating the > mshadow code to mxnet and I started an RFC for this in mshadow [1]. > > > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > > > -sz > > > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > > > Technically, mshadow is sufficient for MXNet. Adopting other > libraries ( > > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > > without any additional gains. > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it > into > > > > mxnet codebase. > > > > To respect the original mshadow community. I would recommend > starting a > > > > community RFC In the mshadow github issue for a week, before we > start the > > > > migrating process. > > > > Also, I would recommend a rebase merge just like the case of > MXNet.jl code > > > > base to preserve the contribution history. > > > > > > > > Tianqi > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > wrote: > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya < > anirudhk...@gmail.com> > > > > > wrote: > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > discussions > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > As a more long term solution using libraries like eigen to > perform linear > > > > > > algebra operations was also suggested by anirudh2290@. I think > xtensor( > > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > here. > > > > > > > > > > > > - > > > > > > Anirudh > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > pedro.larroy.li...@gmail.com> > > > > > > wrote: > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > cumbersome as > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > independent tests / library functionality, me and other > developers > > > > > > > believe that it would be good to assimilate this code in the > > > > > > > repository for ease of contribution and changes without having > to go > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
To mentors, Do we the PPMC need to fill out IP clearance for this code donation? -sz On 2019/04/24 21:19:49, Sheng Zha wrote: > The community has agreed to donate mshadow to the mxnet code base. I will > start the migration and build logic changes soon. > > -sz > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > I agree it would make development easier to donate mshadow to mxnet code > > base, since mshadow is only used in MXNet. I support donating the mshadow > > code to mxnet and I started an RFC for this in mshadow [1]. > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > -sz > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > without any additional gains. > > > > > > Given that mshadow is only used by mxnet. I do support donating it into > > > mxnet codebase. > > > To respect the original mshadow community. I would recommend starting a > > > community RFC In the mshadow github issue for a week, before we start the > > > migrating process. > > > Also, I would recommend a rebase merge just like the case of MXNet.jl code > > > base to preserve the contribution history. > > > > > > Tianqi > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > wrote: > > > > > > > Do you have a link to both of these proposals? > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > wrote: > > > > > > > > > Hi Pedro, > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > > discussions > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > As a more long term solution using libraries like eigen to perform > > > > > linear > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > > xtensor( > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > > > > > > > - > > > > > Anirudh > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > pedro.larroy.li...@gmail.com> > > > > > wrote: > > > > > > > > > > > Hi > > > > > > > > > > > > Some developers have noticed that working in mshadow is cumbersome > > > > > > as > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > independent tests / library functionality, me and other developers > > > > > > believe that it would be good to assimilate this code in the > > > > > > repository for ease of contribution and changes without having to go > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Hi Sheng. Do you need some help with this? Do we plan to have this for 1.5? Pedro. On Wed, Apr 24, 2019 at 4:26 PM Pedro Larroy wrote: > > Thanks. Great to read. > > On Wed, Apr 24, 2019 at 2:19 PM Sheng Zha wrote: > > > > The community has agreed to donate mshadow to the mxnet code base. I will > > start the migration and build logic changes soon. > > > > -sz > > > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > > I agree it would make development easier to donate mshadow to mxnet code > > > base, since mshadow is only used in MXNet. I support donating the mshadow > > > code to mxnet and I started an RFC for this in mshadow [1]. > > > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > > > -sz > > > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > > without any additional gains. > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it into > > > > mxnet codebase. > > > > To respect the original mshadow community. I would recommend starting a > > > > community RFC In the mshadow github issue for a week, before we start > > > > the > > > > migrating process. > > > > Also, I would recommend a rebase merge just like the case of MXNet.jl > > > > code > > > > base to preserve the contribution history. > > > > > > > > Tianqi > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > wrote: > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > > wrote: > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > > > discussions > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > As a more long term solution using libraries like eigen to perform > > > > > > linear > > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > > > xtensor( > > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > > > > > > here. > > > > > > > > > > > > - > > > > > > Anirudh > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > pedro.larroy.li...@gmail.com> > > > > > > wrote: > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > > > > > > > cumbersome as > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > independent tests / library functionality, me and other developers > > > > > > > believe that it would be good to assimilate this code in the > > > > > > > repository for ease of contribution and changes without having to > > > > > > > go > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Thanks. Great to read. On Wed, Apr 24, 2019 at 2:19 PM Sheng Zha wrote: > > The community has agreed to donate mshadow to the mxnet code base. I will > start the migration and build logic changes soon. > > -sz > > On 2019/04/07 21:47:39, Sheng Zha wrote: > > I agree it would make development easier to donate mshadow to mxnet code > > base, since mshadow is only used in MXNet. I support donating the mshadow > > code to mxnet and I started an RFC for this in mshadow [1]. > > > > [1] https://github.com/dmlc/mshadow/issues/373 > > > > -sz > > > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > without any additional gains. > > > > > > Given that mshadow is only used by mxnet. I do support donating it into > > > mxnet codebase. > > > To respect the original mshadow community. I would recommend starting a > > > community RFC In the mshadow github issue for a week, before we start the > > > migrating process. > > > Also, I would recommend a rebase merge just like the case of MXNet.jl code > > > base to preserve the contribution history. > > > > > > Tianqi > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > wrote: > > > > > > > Do you have a link to both of these proposals? > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > wrote: > > > > > > > > > Hi Pedro, > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > > discussions > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > As a more long term solution using libraries like eigen to perform > > > > > linear > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > > xtensor( > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > > > > > > > - > > > > > Anirudh > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > pedro.larroy.li...@gmail.com> > > > > > wrote: > > > > > > > > > > > Hi > > > > > > > > > > > > Some developers have noticed that working in mshadow is cumbersome > > > > > > as > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > independent tests / library functionality, me and other developers > > > > > > believe that it would be good to assimilate this code in the > > > > > > repository for ease of contribution and changes without having to go > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
The community has agreed to donate mshadow to the mxnet code base. I will start the migration and build logic changes soon. -sz On 2019/04/07 21:47:39, Sheng Zha wrote: > I agree it would make development easier to donate mshadow to mxnet code > base, since mshadow is only used in MXNet. I support donating the mshadow > code to mxnet and I started an RFC for this in mshadow [1]. > > [1] https://github.com/dmlc/mshadow/issues/373 > > -sz > > On 2019/04/06 04:38:19, Tianqi Chen wrote: > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > eigen or xtensor) will unnecessarily increase the codebase complexity > > without any additional gains. > > > > Given that mshadow is only used by mxnet. I do support donating it into > > mxnet codebase. > > To respect the original mshadow community. I would recommend starting a > > community RFC In the mshadow github issue for a week, before we start the > > migrating process. > > Also, I would recommend a rebase merge just like the case of MXNet.jl code > > base to preserve the contribution history. > > > > Tianqi > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > wrote: > > > > > Do you have a link to both of these proposals? > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > wrote: > > > > > > > Hi Pedro, > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > discussions > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > As a more long term solution using libraries like eigen to perform > > > > linear > > > > algebra operations was also suggested by anirudh2290@. I think xtensor( > > > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > > > > > - > > > > Anirudh > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > pedro.larroy.li...@gmail.com> > > > > wrote: > > > > > > > > > Hi > > > > > > > > > > Some developers have noticed that working in mshadow is cumbersome as > > > > > it's a 3rdparty subrepo. > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > independent tests / library functionality, me and other developers > > > > > believe that it would be good to assimilate this code in the > > > > > repository for ease of contribution and changes without having to go > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > Pedro. > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
There's a flag MSHADOW_STAND_ALONE which supports gemm but not all of them, and looks like an untested codepath. From what I have seen I don't think we use this from MXNet, hence the need for a BLAS implementation. On Sun, Apr 7, 2019 at 6:16 PM Zhao, Patric wrote: > > Agree. > > Recently, we (Tao, Shufan, Pengxin) are trying to integrate the Intel MKL > math functions into mshadow and MXNet. > We have to go through two repos and make lots of tradeoff between them. > If we can move mshadow into MXNet, it will be more flexible to redesign and > refactor parts of legacy code. > > > -Original Message- > > From: Sheng Zha [mailto:zhash...@apache.org] > > Sent: Monday, April 8, 2019 5:48 AM > > To: d...@mxnet.apache.org > > Subject: Re: assimilation of mshadow into the MXNet codebase > > > > mshadow depends on *a* BLAS library, and there's nothing inherent in > > mshadow code base that requires OpenBLAS over MKL. The linked issue > > #11769 seems to be more of a build logic issue. > > > > -sz > > > > On 2019/04/07 18:56:43, Aaron Markham > > wrote: > > > +1 > > > Reduced complexity. Choice of math library... Hopefully you can just > > > install MKL and not be forced into mshadow's dependency on OpenBLAS. > > > This could make Windows setup easier. > > > Maybe this issue will get fixed: #11769. > > > > > > On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > > > > > > > Does merging mshadow into mxnet bring any actual benefit for > > > > customers in sense of performance, portability, or anything else? > > > > > > > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > > > > > > > > wrote: > > > > > > > > > Technically, mshadow is sufficient for MXNet. Adopting other > > > > > libraries ( eigen or xtensor) will unnecessarily increase the > > > > > codebase complexity without any additional gains. > > > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it > > > > > into mxnet codebase. > > > > > To respect the original mshadow community. I would recommend > > > > > starting a community RFC In the mshadow github issue for a week, > > > > > before we start the migrating process. > > > > > Also, I would recommend a rebase merge just like the case of > > > > > MXNet.jl > > > > code > > > > > base to preserve the contribution history. > > > > > > > > > > Tianqi > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > > wrote: > > > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > > > > > > > > > wrote: > > > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > > discussions > > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > > > As a more long term solution using libraries like eigen to > > > > > > > perform > > > > > linear > > > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > > xtensor( > > > > > > > https://github.com/QuantStack/xtensor ) can also be a > > > > > > > candidate > > > > here. > > > > > > > > > > > > > > - > > > > > > > Anirudh > > > > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > > pedro.larroy.li...@gmail.com> > > > > > > > wrote: > > > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > > > > > > > > cumbersome > > > > as > > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > > independent tests / library functionality, me and other > > > > > > > > developers believe that it would be good to assimilate this > > > > > > > > code in the repository for ease of contribution and changes > > > > > > > > without having to > > > > go > > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Thanks Kellen for the explanation, +1 for this! On Sun, Apr 7, 2019 at 6:16 PM Zhao, Patric wrote: > Agree. > > Recently, we (Tao, Shufan, Pengxin) are trying to integrate the Intel MKL > math functions into mshadow and MXNet. > We have to go through two repos and make lots of tradeoff between them. > If we can move mshadow into MXNet, it will be more flexible to redesign > and refactor parts of legacy code. > > > -Original Message- > > From: Sheng Zha [mailto:zhash...@apache.org] > > Sent: Monday, April 8, 2019 5:48 AM > > To: d...@mxnet.apache.org > > Subject: Re: assimilation of mshadow into the MXNet codebase > > > > mshadow depends on *a* BLAS library, and there's nothing inherent in > > mshadow code base that requires OpenBLAS over MKL. The linked issue > > #11769 seems to be more of a build logic issue. > > > > -sz > > > > On 2019/04/07 18:56:43, Aaron Markham > > wrote: > > > +1 > > > Reduced complexity. Choice of math library... Hopefully you can just > > > install MKL and not be forced into mshadow's dependency on OpenBLAS. > > > This could make Windows setup easier. > > > Maybe this issue will get fixed: #11769. > > > > > > On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > > > > > > > Does merging mshadow into mxnet bring any actual benefit for > > > > customers in sense of performance, portability, or anything else? > > > > > > > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > > > > > > > > wrote: > > > > > > > > > Technically, mshadow is sufficient for MXNet. Adopting other > > > > > libraries ( eigen or xtensor) will unnecessarily increase the > > > > > codebase complexity without any additional gains. > > > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it > > > > > into mxnet codebase. > > > > > To respect the original mshadow community. I would recommend > > > > > starting a community RFC In the mshadow github issue for a week, > > > > > before we start the migrating process. > > > > > Also, I would recommend a rebase merge just like the case of > > > > > MXNet.jl > > > > code > > > > > base to preserve the contribution history. > > > > > > > > > > Tianqi > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > > wrote: > > > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > > > > > > > > > wrote: > > > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > > discussions > > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > > > As a more long term solution using libraries like eigen to > > > > > > > perform > > > > > linear > > > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > > xtensor( > > > > > > > https://github.com/QuantStack/xtensor ) can also be a > > > > > > > candidate > > > > here. > > > > > > > > > > > > > > - > > > > > > > Anirudh > > > > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > > pedro.larroy.li...@gmail.com> > > > > > > > wrote: > > > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > > > > > > > > cumbersome > > > > as > > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > > independent tests / library functionality, me and other > > > > > > > > developers believe that it would be good to assimilate this > > > > > > > > code in the repository for ease of contribution and changes > > > > > > > > without having to > > > > go > > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
RE: assimilation of mshadow into the MXNet codebase
Agree. Recently, we (Tao, Shufan, Pengxin) are trying to integrate the Intel MKL math functions into mshadow and MXNet. We have to go through two repos and make lots of tradeoff between them. If we can move mshadow into MXNet, it will be more flexible to redesign and refactor parts of legacy code. > -Original Message- > From: Sheng Zha [mailto:zhash...@apache.org] > Sent: Monday, April 8, 2019 5:48 AM > To: d...@mxnet.apache.org > Subject: Re: assimilation of mshadow into the MXNet codebase > > mshadow depends on *a* BLAS library, and there's nothing inherent in > mshadow code base that requires OpenBLAS over MKL. The linked issue > #11769 seems to be more of a build logic issue. > > -sz > > On 2019/04/07 18:56:43, Aaron Markham > wrote: > > +1 > > Reduced complexity. Choice of math library... Hopefully you can just > > install MKL and not be forced into mshadow's dependency on OpenBLAS. > > This could make Windows setup easier. > > Maybe this issue will get fixed: #11769. > > > > On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > > > > > Does merging mshadow into mxnet bring any actual benefit for > > > customers in sense of performance, portability, or anything else? > > > > > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > > > > > > wrote: > > > > > > > Technically, mshadow is sufficient for MXNet. Adopting other > > > > libraries ( eigen or xtensor) will unnecessarily increase the > > > > codebase complexity without any additional gains. > > > > > > > > Given that mshadow is only used by mxnet. I do support donating it > > > > into mxnet codebase. > > > > To respect the original mshadow community. I would recommend > > > > starting a community RFC In the mshadow github issue for a week, > > > > before we start the migrating process. > > > > Also, I would recommend a rebase merge just like the case of > > > > MXNet.jl > > > code > > > > base to preserve the contribution history. > > > > > > > > Tianqi > > > > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > > wrote: > > > > > > > > > Do you have a link to both of these proposals? > > > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > > > > > > > wrote: > > > > > > > > > > > Hi Pedro, > > > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > > discussions > > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > > > As a more long term solution using libraries like eigen to > > > > > > perform > > > > linear > > > > > > algebra operations was also suggested by anirudh2290@. I think > > > > xtensor( > > > > > > https://github.com/QuantStack/xtensor ) can also be a > > > > > > candidate > > > here. > > > > > > > > > > > > - > > > > > > Anirudh > > > > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > > pedro.larroy.li...@gmail.com> > > > > > > wrote: > > > > > > > > > > > > > Hi > > > > > > > > > > > > > > Some developers have noticed that working in mshadow is > > > > > > > cumbersome > > > as > > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > > independent tests / library functionality, me and other > > > > > > > developers believe that it would be good to assimilate this > > > > > > > code in the repository for ease of contribution and changes > > > > > > > without having to > > > go > > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
mshadow depends on *a* BLAS library, and there's nothing inherent in mshadow code base that requires OpenBLAS over MKL. The linked issue #11769 seems to be more of a build logic issue. -sz On 2019/04/07 18:56:43, Aaron Markham wrote: > +1 > Reduced complexity. Choice of math library... Hopefully you can just > install MKL and not be forced into mshadow's dependency on OpenBLAS. This > could make Windows setup easier. > Maybe this issue will get fixed: #11769. > > On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > > > Does merging mshadow into mxnet bring any actual benefit for customers in > > sense of performance, portability, or anything else? > > > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > > wrote: > > > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > without any additional gains. > > > > > > Given that mshadow is only used by mxnet. I do support donating it into > > > mxnet codebase. > > > To respect the original mshadow community. I would recommend starting a > > > community RFC In the mshadow github issue for a week, before we start the > > > migrating process. > > > Also, I would recommend a rebase merge just like the case of MXNet.jl > > code > > > base to preserve the contribution history. > > > > > > Tianqi > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > wrote: > > > > > > > Do you have a link to both of these proposals? > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > wrote: > > > > > > > > > Hi Pedro, > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > discussions > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > As a more long term solution using libraries like eigen to perform > > > linear > > > > > algebra operations was also suggested by anirudh2290@. I think > > > xtensor( > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > > here. > > > > > > > > > > - > > > > > Anirudh > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > pedro.larroy.li...@gmail.com> > > > > > wrote: > > > > > > > > > > > Hi > > > > > > > > > > > > Some developers have noticed that working in mshadow is cumbersome > > as > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > independent tests / library functionality, me and other developers > > > > > > believe that it would be good to assimilate this code in the > > > > > > repository for ease of contribution and changes without having to > > go > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
I agree it would make development easier to donate mshadow to mxnet code base, since mshadow is only used in MXNet. I support donating the mshadow code to mxnet and I started an RFC for this in mshadow [1]. [1] https://github.com/dmlc/mshadow/issues/373 -sz On 2019/04/06 04:38:19, Tianqi Chen wrote: > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > eigen or xtensor) will unnecessarily increase the codebase complexity > without any additional gains. > > Given that mshadow is only used by mxnet. I do support donating it into > mxnet codebase. > To respect the original mshadow community. I would recommend starting a > community RFC In the mshadow github issue for a week, before we start the > migrating process. > Also, I would recommend a rebase merge just like the case of MXNet.jl code > base to preserve the contribution history. > > Tianqi > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > wrote: > > > Do you have a link to both of these proposals? > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > wrote: > > > > > Hi Pedro, > > > > > > mshadow is mostly used for tensor arithmetic. There have been discussions > > > about including it within mxnet. I think it is a good idea. > > > > > > As a more long term solution using libraries like eigen to perform linear > > > algebra operations was also suggested by anirudh2290@. I think xtensor( > > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > > > - > > > Anirudh > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > pedro.larroy.li...@gmail.com> > > > wrote: > > > > > > > Hi > > > > > > > > Some developers have noticed that working in mshadow is cumbersome as > > > > it's a 3rdparty subrepo. > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > independent tests / library functionality, me and other developers > > > > believe that it would be good to assimilate this code in the > > > > repository for ease of contribution and changes without having to go > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > Would anybody oppose this change? > > > > > > > > Thanks and have a nice weekend. > > > > > > > > Pedro. > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
"Does merging mshadow into mxnet bring any actual benefit for customers in sense of performance, portability, or anything else?" It would improve the contributor experience in that if we find a bug which requires fixes in both repos, we won't have to coordinate 2 PRs. It would also make compilation more straightforward (as others have mentioned). On Sun, Apr 7, 2019 at 11:56 AM Aaron Markham wrote: > +1 > Reduced complexity. Choice of math library... Hopefully you can just > install MKL and not be forced into mshadow's dependency on OpenBLAS. This > could make Windows setup easier. > Maybe this issue will get fixed: #11769. > > On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > > > Does merging mshadow into mxnet bring any actual benefit for customers in > > sense of performance, portability, or anything else? > > > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > > wrote: > > > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries > ( > > > eigen or xtensor) will unnecessarily increase the codebase complexity > > > without any additional gains. > > > > > > Given that mshadow is only used by mxnet. I do support donating it into > > > mxnet codebase. > > > To respect the original mshadow community. I would recommend starting a > > > community RFC In the mshadow github issue for a week, before we start > the > > > migrating process. > > > Also, I would recommend a rebase merge just like the case of MXNet.jl > > code > > > base to preserve the contribution history. > > > > > > Tianqi > > > > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > > wrote: > > > > > > > Do you have a link to both of these proposals? > > > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > > wrote: > > > > > > > > > Hi Pedro, > > > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > > discussions > > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > > > As a more long term solution using libraries like eigen to perform > > > linear > > > > > algebra operations was also suggested by anirudh2290@. I think > > > xtensor( > > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > > here. > > > > > > > > > > - > > > > > Anirudh > > > > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > > pedro.larroy.li...@gmail.com> > > > > > wrote: > > > > > > > > > > > Hi > > > > > > > > > > > > Some developers have noticed that working in mshadow is > cumbersome > > as > > > > > > it's a 3rdparty subrepo. > > > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > > independent tests / library functionality, me and other > developers > > > > > > believe that it would be good to assimilate this code in the > > > > > > repository for ease of contribution and changes without having to > > go > > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > > > Pedro. > > > > > > > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
+1 Reduced complexity. Choice of math library... Hopefully you can just install MKL and not be forced into mshadow's dependency on OpenBLAS. This could make Windows setup easier. Maybe this issue will get fixed: #11769. On Sun, Apr 7, 2019, 00:51 Junru Shao wrote: > Does merging mshadow into mxnet bring any actual benefit for customers in > sense of performance, portability, or anything else? > > On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen > wrote: > > > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > > eigen or xtensor) will unnecessarily increase the codebase complexity > > without any additional gains. > > > > Given that mshadow is only used by mxnet. I do support donating it into > > mxnet codebase. > > To respect the original mshadow community. I would recommend starting a > > community RFC In the mshadow github issue for a week, before we start the > > migrating process. > > Also, I would recommend a rebase merge just like the case of MXNet.jl > code > > base to preserve the contribution history. > > > > Tianqi > > > > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > > wrote: > > > > > Do you have a link to both of these proposals? > > > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > > wrote: > > > > > > > Hi Pedro, > > > > > > > > mshadow is mostly used for tensor arithmetic. There have been > > discussions > > > > about including it within mxnet. I think it is a good idea. > > > > > > > > As a more long term solution using libraries like eigen to perform > > linear > > > > algebra operations was also suggested by anirudh2290@. I think > > xtensor( > > > > https://github.com/QuantStack/xtensor ) can also be a candidate > here. > > > > > > > > - > > > > Anirudh > > > > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > > pedro.larroy.li...@gmail.com> > > > > wrote: > > > > > > > > > Hi > > > > > > > > > > Some developers have noticed that working in mshadow is cumbersome > as > > > > > it's a 3rdparty subrepo. > > > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > > independent tests / library functionality, me and other developers > > > > > believe that it would be good to assimilate this code in the > > > > > repository for ease of contribution and changes without having to > go > > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > > > Would anybody oppose this change? > > > > > > > > > > Thanks and have a nice weekend. > > > > > > > > > > Pedro. > > > > > > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Does merging mshadow into mxnet bring any actual benefit for customers in sense of performance, portability, or anything else? On Fri, Apr 5, 2019 at 9:38 PM Tianqi Chen wrote: > Technically, mshadow is sufficient for MXNet. Adopting other libraries ( > eigen or xtensor) will unnecessarily increase the codebase complexity > without any additional gains. > > Given that mshadow is only used by mxnet. I do support donating it into > mxnet codebase. > To respect the original mshadow community. I would recommend starting a > community RFC In the mshadow github issue for a week, before we start the > migrating process. > Also, I would recommend a rebase merge just like the case of MXNet.jl code > base to preserve the contribution history. > > Tianqi > > > On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque > wrote: > > > Do you have a link to both of these proposals? > > > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > > wrote: > > > > > Hi Pedro, > > > > > > mshadow is mostly used for tensor arithmetic. There have been > discussions > > > about including it within mxnet. I think it is a good idea. > > > > > > As a more long term solution using libraries like eigen to perform > linear > > > algebra operations was also suggested by anirudh2290@. I think > xtensor( > > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > > > - > > > Anirudh > > > > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > > pedro.larroy.li...@gmail.com> > > > wrote: > > > > > > > Hi > > > > > > > > Some developers have noticed that working in mshadow is cumbersome as > > > > it's a 3rdparty subrepo. > > > > > > > > Since mshadow is a bunch of headers which don't have much of > > > > independent tests / library functionality, me and other developers > > > > believe that it would be good to assimilate this code in the > > > > repository for ease of contribution and changes without having to go > > > > trough contortions to test PRs that modify mshadow. > > > > > > > > Would anybody oppose this change? > > > > > > > > Thanks and have a nice weekend. > > > > > > > > Pedro. > > > > > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Technically, mshadow is sufficient for MXNet. Adopting other libraries ( eigen or xtensor) will unnecessarily increase the codebase complexity without any additional gains. Given that mshadow is only used by mxnet. I do support donating it into mxnet codebase. To respect the original mshadow community. I would recommend starting a community RFC In the mshadow github issue for a week, before we start the migrating process. Also, I would recommend a rebase merge just like the case of MXNet.jl code base to preserve the contribution history. Tianqi On Fri, Apr 5, 2019 at 9:25 PM Alfredo Luque wrote: > Do you have a link to both of these proposals? > > On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya > wrote: > > > Hi Pedro, > > > > mshadow is mostly used for tensor arithmetic. There have been discussions > > about including it within mxnet. I think it is a good idea. > > > > As a more long term solution using libraries like eigen to perform linear > > algebra operations was also suggested by anirudh2290@. I think xtensor( > > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > > > - > > Anirudh > > > > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy < > pedro.larroy.li...@gmail.com> > > wrote: > > > > > Hi > > > > > > Some developers have noticed that working in mshadow is cumbersome as > > > it's a 3rdparty subrepo. > > > > > > Since mshadow is a bunch of headers which don't have much of > > > independent tests / library functionality, me and other developers > > > believe that it would be good to assimilate this code in the > > > repository for ease of contribution and changes without having to go > > > trough contortions to test PRs that modify mshadow. > > > > > > Would anybody oppose this change? > > > > > > Thanks and have a nice weekend. > > > > > > Pedro. > > > > > >
Re: assimilation of mshadow into the MXNet codebase
Do you have a link to both of these proposals? On Fri, Apr 5, 2019 at 20:14 Anirudh Acharya wrote: > Hi Pedro, > > mshadow is mostly used for tensor arithmetic. There have been discussions > about including it within mxnet. I think it is a good idea. > > As a more long term solution using libraries like eigen to perform linear > algebra operations was also suggested by anirudh2290@. I think xtensor( > https://github.com/QuantStack/xtensor ) can also be a candidate here. > > - > Anirudh > > > On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy > wrote: > > > Hi > > > > Some developers have noticed that working in mshadow is cumbersome as > > it's a 3rdparty subrepo. > > > > Since mshadow is a bunch of headers which don't have much of > > independent tests / library functionality, me and other developers > > believe that it would be good to assimilate this code in the > > repository for ease of contribution and changes without having to go > > trough contortions to test PRs that modify mshadow. > > > > Would anybody oppose this change? > > > > Thanks and have a nice weekend. > > > > Pedro. > > >
Re: assimilation of mshadow into the MXNet codebase
Hi Pedro, mshadow is mostly used for tensor arithmetic. There have been discussions about including it within mxnet. I think it is a good idea. As a more long term solution using libraries like eigen to perform linear algebra operations was also suggested by anirudh2290@. I think xtensor( https://github.com/QuantStack/xtensor ) can also be a candidate here. - Anirudh On Fri, Apr 5, 2019 at 7:03 PM Pedro Larroy wrote: > Hi > > Some developers have noticed that working in mshadow is cumbersome as > it's a 3rdparty subrepo. > > Since mshadow is a bunch of headers which don't have much of > independent tests / library functionality, me and other developers > believe that it would be good to assimilate this code in the > repository for ease of contribution and changes without having to go > trough contortions to test PRs that modify mshadow. > > Would anybody oppose this change? > > Thanks and have a nice weekend. > > Pedro. >
