Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

I would like to revisit a bit about the release topic to shed light about
the release process.
The TVM PPMC has always been focusing on producing high quality releases.

- The project has produced five major (non-apache) releases prior joining
Apache
- While most incubator projects start with a DISCLAIMER-WIP, the TVM
community strives to keep the release quality high and uses the DISCLAIMER
   in the beginning of the first release[1]. The first release contains a
few rough edges and gets resolved very quickly thanks to the feedback from
Justin and other IPMC members.
   The second release is super smooth.
- Multiple PMC members work together to generate these releases.

TQ

- [1]
https://lists.apache.org/thread.html/ad19634352f97d0a5b2bec866bcdecd556ca414cc6d4487b2282ce10%40%3Cgeneral.incubator.apache.org%3E

On Fri, Aug 28, 2020 at 8:54 PM Tianqi Chen  wrote:

> We have also talked to Apache INFRA about the use of the
> https://ci.tvm.ai/ due to the need of special setups in terms of GPU
> based resources etc.
>
> So far we have received warm help from the INFRA and they are super
> helpful in helping setting up CI hooks so that the community can enjoy quick
> development with the volunteer maintained CI. Running and serving the
> community well, we would certainly happy to use a different name if asked by
> the trademark.
>
> TQ
>
> On Fri, Aug 28, 2020 at 8:49 PM Tianqi Chen  wrote:
>
>> tvmlang.org was an old url by the project that eventually got deprecated
>> at some time point, it now redirects to the https://tvm.apache.org/
>>
>> The PMC makes sure to redirect to the static Apache project when possible
>> for official project related information.
>>
>> Due to the mechanism of the discourse website(need an active server), and
>> CI (tracking is not used and is removed). It is necessary to use a
>> standalone domain name
>> - These domains are maintained by volunteers from TVM PMC members
>> (tqchen, yizhi, haichen, jroesch, ziheng, lianminzheng, zhiics, masahi,
>> thierry from different organizations)
>>   as a thirdparty service to help the community.
>> - Given that these domains are not used to represent the official apache
>> project website(no confusion), and the usage voted by the community [2],
>>   the PPMC believes it is fair to use these domains to serve the
>> community, just like the use of thirdparty conference website for some of
>> the ASF projects' developer conferences.
>> - Of course we would love to get approval/feedback from the trademark. We
>> have started a trademark thread on the community discourse forum domain
>> [1],
>>and would be more than happy to follow up the guidance of the
>> trademark discussion.
>>
>> Again the hopefully outcome is not to discourage the use of these
>> services. The PMC members strive to make sure that these thirdparty
>> services are sustained,
>> and remain "non-blocker" even when they go down.
>>
>> - Introduce multiple volunteers from different organizations in the PMC
>> to do the work.
>> - Make sure the messages are backed up to mail-list.
>> - In the case of CI, have a clear workflow(JenkinsFile) and docker source
>> so that any committer could have spin up CI when necessary.
>> - Every development activities (also) happens on dev@
>>
>>
>> TQ
>>
>> ---
>> - [1]
>> https://lists.apache.org/thread.html/rd23004bf526975fceebd1b0bae7120dc1935bb61b7cdd8243fbc36dd%40%3Ctrademarks.apache.org%3E
>> - [2]
>> https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E
>>
>>
>> On Fri, Aug 28, 2020 at 8:20 PM Justin Mclean 
>> wrote:
>>
>>> Hi,
>>>
>>> Sorry my mistake, I meant to say tvm.ai not tvm.io, but as I said
>>> subdomains on tvm.ai are not redirecting. For instance discuss.tvm.ai
>>> stays as it is. Other sub-domains include ci.tvm.ai , tracking.tvm.ai
>>> and docs.tvm.ai that I’ve found. There might be others, of those four,
>>> only docs redirects.
>>>
>>> I notice that forum [1] also states "Open source community on end to end
>>> stack for deploying deep learning workloads to hardware backends
>>> tvmlang.org.” Who controls tvmlang.org? It not clear on that page that
>>> this is a space for discussion about an Apache project.
>>>
>>> Thanks,
>>> Justin
>>>
>>> 1. https://discuss.tvm.ai
>>> -
>>> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
>>> For additional commands, e-mail: general-h...@incubator.apache.org
>>>
>>>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

We have also talked to Apache INFRA about the use of the https://ci.tvm.ai/ due
to the need of special setups in terms of GPU based resources etc.

So far we have received warm help from the INFRA and they are super helpful
in helping setting up CI hooks so that the community can enjoy quick
development with the volunteer maintained CI. Running and serving the
community well, we would certainly happy to use a different name if asked by
the trademark.

TQ

On Fri, Aug 28, 2020 at 8:49 PM Tianqi Chen  wrote:

> tvmlang.org was an old url by the project that eventually got deprecated
> at some time point, it now redirects to the https://tvm.apache.org/
>
> The PMC makes sure to redirect to the static Apache project when possible
> for official project related information.
>
> Due to the mechanism of the discourse website(need an active server), and
> CI (tracking is not used and is removed). It is necessary to use a
> standalone domain name
> - These domains are maintained by volunteers from TVM PMC members (tqchen,
> yizhi, haichen, jroesch, ziheng, lianminzheng, zhiics, masahi, thierry from
> different organizations)
>   as a thirdparty service to help the community.
> - Given that these domains are not used to represent the official apache
> project website(no confusion), and the usage voted by the community [2],
>   the PPMC believes it is fair to use these domains to serve the
> community, just like the use of thirdparty conference website for some of
> the ASF projects' developer conferences.
> - Of course we would love to get approval/feedback from the trademark. We
> have started a trademark thread on the community discourse forum domain
> [1],
>and would be more than happy to follow up the guidance of the trademark
> discussion.
>
> Again the hopefully outcome is not to discourage the use of these
> services. The PMC members strive to make sure that these thirdparty
> services are sustained,
> and remain "non-blocker" even when they go down.
>
> - Introduce multiple volunteers from different organizations in the PMC to
> do the work.
> - Make sure the messages are backed up to mail-list.
> - In the case of CI, have a clear workflow(JenkinsFile) and docker source
> so that any committer could have spin up CI when necessary.
> - Every development activities (also) happens on dev@
>
>
> TQ
>
> ---
> - [1]
> https://lists.apache.org/thread.html/rd23004bf526975fceebd1b0bae7120dc1935bb61b7cdd8243fbc36dd%40%3Ctrademarks.apache.org%3E
> - [2]
> https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E
>
>
> On Fri, Aug 28, 2020 at 8:20 PM Justin Mclean 
> wrote:
>
>> Hi,
>>
>> Sorry my mistake, I meant to say tvm.ai not tvm.io, but as I said
>> subdomains on tvm.ai are not redirecting. For instance discuss.tvm.ai
>> stays as it is. Other sub-domains include ci.tvm.ai , tracking.tvm.ai
>> and docs.tvm.ai that I’ve found. There might be others, of those four,
>> only docs redirects.
>>
>> I notice that forum [1] also states "Open source community on end to end
>> stack for deploying deep learning workloads to hardware backends
>> tvmlang.org.” Who controls tvmlang.org? It not clear on that page that
>> this is a space for discussion about an Apache project.
>>
>> Thanks,
>> Justin
>>
>> 1. https://discuss.tvm.ai
>> -
>> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
>> For additional commands, e-mail: general-h...@incubator.apache.org
>>
>>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

tvmlang.org was an old url by the project that eventually got deprecated at
some time point, it now redirects to the https://tvm.apache.org/

The PMC makes sure to redirect to the static Apache project when possible
for official project related information.

Due to the mechanism of the discourse website(need an active server), and
CI (tracking is not used and is removed). It is necessary to use a
standalone domain name
- These domains are maintained by volunteers from TVM PMC members (tqchen,
yizhi, haichen, jroesch, ziheng, lianminzheng, zhiics, masahi, thierry from
different organizations)
  as a thirdparty service to help the community.
- Given that these domains are not used to represent the official apache
project website(no confusion), and the usage voted by the community [2],
  the PPMC believes it is fair to use these domains to serve the community,
just like the use of thirdparty conference website for some of the ASF
projects' developer conferences.
- Of course we would love to get approval/feedback from the trademark. We
have started a trademark thread on the community discourse forum domain
[1],
   and would be more than happy to follow up the guidance of the trademark
discussion.

Again the hopefully outcome is not to discourage the use of these services.
The PMC members strive to make sure that these thirdparty services are
sustained,
and remain "non-blocker" even when they go down.

- Introduce multiple volunteers from different organizations in the PMC to
do the work.
- Make sure the messages are backed up to mail-list.
- In the case of CI, have a clear workflow(JenkinsFile) and docker source
so that any committer could have spin up CI when necessary.
- Every development activities (also) happens on dev@


TQ

---
- [1]
https://lists.apache.org/thread.html/rd23004bf526975fceebd1b0bae7120dc1935bb61b7cdd8243fbc36dd%40%3Ctrademarks.apache.org%3E
- [2]
https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E


On Fri, Aug 28, 2020 at 8:20 PM Justin Mclean 
wrote:

> Hi,
>
> Sorry my mistake, I meant to say tvm.ai not tvm.io, but as I said
> subdomains on tvm.ai are not redirecting. For instance discuss.tvm.ai
> stays as it is. Other sub-domains include ci.tvm.ai , tracking.tvm.ai and
> docs.tvm.ai that I’ve found. There might be others, of those four, only
> docs redirects.
>
> I notice that forum [1] also states "Open source community on end to end
> stack for deploying deep learning workloads to hardware backends
> tvmlang.org.” Who controls tvmlang.org? It not clear on that page that
> this is a space for discussion about an Apache project.
>
> Thanks,
> Justin
>
> 1. https://discuss.tvm.ai
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

Sorry my mistake, I meant to say tvm.ai not tvm.io, but as I said subdomains on 
tvm.ai are not redirecting. For instance discuss.tvm.ai stays as it is. Other 
sub-domains include ci.tvm.ai , tracking.tvm.ai and docs.tvm.ai that I’ve 
found. There might be others, of those four, only docs redirects.

I notice that forum [1] also states "Open source community on end to end stack 
for deploying deep learning workloads to hardware backends tvmlang.org.” Who 
controls tvmlang.org? It not clear on that page that this is a space for 
discussion about an Apache project.

Thanks,
Justin

1. https://discuss.tvm.ai
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [VOTE] Graduate Apache TVM as TLP

[Ming Wen]

my -1 non-binding.
see discuss thread about my concerns.


Justin Mclean  于 2020年8月29日周六 上午5:59写道：

> HI,
>
> I’m -1 (binding), please see the discussion thread for my concerns.
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Justin. To further address your listed points:

P0: tvm.io

The original project only uses the domain tvm.ai, which is redirected to
the https://tvm.apache.org/.
Other tvm related domain names have nothing to do with ApacheTVM. Note that
because TVM as a three letter abbreviation is quite common and could have
other
meanings. The project uses the name ApacheTVM during the podling name
search.

We agree with the need to discuss with trademark in terms of the domain
names and will do it now as what we did for the other branding related
requests [1].

We agree that customary the community usually makes more releases to
graduate. The TVM community is working on some major features right now and
the PMC would like to focus
effort on producing a high quality major release, but of course because of
the focus on the quality and uncertainty in terms of the cutting edge
technology we are developing,
it might take longer to produce the release. In the meantime, the community
could try to produce minor maintenance releases if there is a strong
feeling for such demonstration.

After discussion with our mentors we agree that it is not a hard blocker.
>From the community, and quality point of view the project is already
running like an Apache project.
That is why we bring up the case.

Thank you!
TQ

---
- [1]
https://lists.apache.org/thread.html/r8559979c58fb92f08f3e253522dc227a51313dafff2ec19076e1093c%40%3Ctrademarks.apache.org%3E

On Fri, Aug 28, 2020 at 7:41 PM Justin Mclean 
wrote:

> Hi,
>
> Thanks for the response. I think the docker issue is resolved.
>
> I still have concerns about use of tvm.io. I have discovered several
> other subdomains (six in total) that do not redirect to the Apache site,
> there might be others. It's still unclear who has control of this domain
> name. The issue here, is not the use of the forum but of the domain name
> and who it is controlled by, branding policy is clear on the use of such
> names. I suggest you discuss this with trademarks / branding.
>
> I still have concerns about the low number of release (i..e two) and there
> have been only a singe release manager. It’s generally expected that a
> project make 3 or 4 releases with a couple of different release managers,
> this certainly isn’t a hard rule but two in this case doesn’t seem enough.
> How much effort would it be to make another release with a different
> release manager?
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

Thanks for the response. I think the docker issue is resolved.

I still have concerns about use of tvm.io. I have discovered several other 
subdomains (six in total) that do not redirect to the Apache site, there might 
be others. It's still unclear who has control of this domain name. The issue 
here, is not the use of the forum but of the domain name and who it is 
controlled by, branding policy is clear on the use of such names. I suggest you 
discuss this with trademarks / branding.

I still have concerns about the low number of release (i..e two) and there have 
been only a singe release manager. It’s generally expected that a project make 
3 or 4 releases with a couple of different release managers, this certainly 
isn’t a hard rule but two in this case doesn’t seem enough. How much effort 
would it be to make another release with a different release manager?

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Ming.

Certainly, different PMC can take a different approach towards release.
We are not suggesting that every PMC should take our approach.
In the particular case of TVM, the community has been working on several
major features including Ansor, uTVM and it does not make sense to cut
the release say right now.

Again, the PPMC wants to produce high quality releases according to feature
plans coordinated with the community, rather than creating more releases.
Notably the last release was pretty recent as well.

While there is one release manager in the past two releases, many PPMC
members(including myself) participated in the release to make sure that
release is in high quality.
I imagine I could have added my name to the release manager as well :) Also
the PMC contains several Apache members who have previous experiences in
cutting releases in TLP projects. So we are very confident that the PMC can
continue to do so.

TQ

On Fri, Aug 28, 2020 at 7:25 PM Ming Wen  wrote:

> I think only two releases and only one release manager are blocking issues.
>  Considering that TVM merges 150 PRs a month on average, but only two
> releases within a year and a half, that is, users use codes that are much
> behind the master, which is not friendly to the community.
>
> Second, there is no conflict between high-quality releases and frequent
> releases. Many Apache projects have maintained frequent and high-quality
> releases.
>
> Third, since there is a good release document, why is there no other ppmc
> to serve as release manager?
>
> Tianqi Chen  于 2020年8月29日周六 上午9:51写道：
>
> > Thanks Matt!
> >
> > That was indeed the approach we used before. The main problem of this
> > approach is
> > - There could be certain upstream changes (say tensorflow get updated)
> that
> > may not retrigger the rebuild
> > - The CI instance itself can quickly get crowded by the historical cached
> > images and causes disk overflow problems.
> > - Overall we find it is very hard to reproduce the CI error when there
> is a
> > problem because of the potential mismatch(due to stale build)
> >
> > So the community switched to using a stable binary tag as for more stable
> > env without blocking the CI, while periodically updating
> > these images when this is a dependency change.
> >
> > TQ
> >
> > On Fri, Aug 28, 2020 at 6:47 PM Matt Sicker  wrote:
> >
> > > Is there even a need to upload the Docker images for CI? Docker
> > recognizes
> > > layers by checksums, so CI build agents will cache the image layers
> > > regardless of whether or not you upload them to Docker Hub. Layers get
> > > rebuilt when the Docker file changes or you force an update.
> > >
> > > On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:
> > >
> > > > Thanks Justin and Henry for the discussion thread.
> > > >
> > > >
> > > >
> > > > Please allow me to dissect and summarize again some of the discussion
> > > >
> > > > points:)
> > > >
> > > >
> > > >
> > > > C0: Docker Image
> > > >
> > > > - There was a mis-understanding that the docker image like ci-gpu
> > > >
> > > >   contains tvm. They do not, and instead contain an environment to
> > > >
> > > >   build and run test cases.
> > > >
> > > > - The thirdparty binary cache is used to speed up the test build.
> > Anyone
> > > >
> > > >   could build these binaries from a TVM source.
> > > >
> > > > - To avoid confusion about the branding, we have renamed the docker
> hub
> > > >
> > > > name to
> > > >
> > > >   a different name that does not contain tvm so that it is clear as a
> > > >
> > > > thirdparty.
> > > >
> > > >
> > > >
> > > > C1: Pointing to Apache Release
> > > >
> > > > - The PPMC 100% agree that we should produce high-quality Apache
> > > >
> > > >   compatible releases, and refer to them in the documents.
> > > >
> > > > - The installation documentation points to the official release
> > download
> > > >
> > > > page [1].
> > > >
> > > > - There is a for developer section, which gives instructions to
> > > developers
> > > >
> > > > about
> > > >
> > > >   how to clone the code from the git repo.
> > > >
> > > >
> > > >
> > > > C2: Number of releases
> > > >
> > > > - The PPMC wants to produce high quality releases according to
> feature
> > > >
> > > >   plan coordinated with the community, rather than creating more
> > > releases.
> > > >
> > > > - The release process is well documented[2],
> > > >
> > > > - Per incubator policy[2], number of releases is not a hard bar for
> > > >
> > > > graduation.
> > > >
> > > >   "Podlings do not need to actually publish a release...". Instead
> the
> > > most
> > > >
> > > > important
> > > >
> > > >   thing is the demonstration of being able to cut apache release.
> > > >
> > > > - Besides the release process being well documented and there are
> > already
> > > >
> > > >   two high-quality releases(without WIP). The PMC contains several
> > Apache
> > > >
> > > > members
> > > >
> > > >   who have previous experiences in cutting releases in TLP 

Re: [DISCUSS] Apache TVM Graduation

[Ming Wen]

I think only two releases and only one release manager are blocking issues.
 Considering that TVM merges 150 PRs a month on average, but only two
releases within a year and a half, that is, users use codes that are much
behind the master, which is not friendly to the community.

Second, there is no conflict between high-quality releases and frequent
releases. Many Apache projects have maintained frequent and high-quality
releases.

Third, since there is a good release document, why is there no other ppmc
to serve as release manager?

Tianqi Chen  于 2020年8月29日周六 上午9:51写道：

> Thanks Matt!
>
> That was indeed the approach we used before. The main problem of this
> approach is
> - There could be certain upstream changes (say tensorflow get updated) that
> may not retrigger the rebuild
> - The CI instance itself can quickly get crowded by the historical cached
> images and causes disk overflow problems.
> - Overall we find it is very hard to reproduce the CI error when there is a
> problem because of the potential mismatch(due to stale build)
>
> So the community switched to using a stable binary tag as for more stable
> env without blocking the CI, while periodically updating
> these images when this is a dependency change.
>
> TQ
>
> On Fri, Aug 28, 2020 at 6:47 PM Matt Sicker  wrote:
>
> > Is there even a need to upload the Docker images for CI? Docker
> recognizes
> > layers by checksums, so CI build agents will cache the image layers
> > regardless of whether or not you upload them to Docker Hub. Layers get
> > rebuilt when the Docker file changes or you force an update.
> >
> > On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:
> >
> > > Thanks Justin and Henry for the discussion thread.
> > >
> > >
> > >
> > > Please allow me to dissect and summarize again some of the discussion
> > >
> > > points:)
> > >
> > >
> > >
> > > C0: Docker Image
> > >
> > > - There was a mis-understanding that the docker image like ci-gpu
> > >
> > >   contains tvm. They do not, and instead contain an environment to
> > >
> > >   build and run test cases.
> > >
> > > - The thirdparty binary cache is used to speed up the test build.
> Anyone
> > >
> > >   could build these binaries from a TVM source.
> > >
> > > - To avoid confusion about the branding, we have renamed the docker hub
> > >
> > > name to
> > >
> > >   a different name that does not contain tvm so that it is clear as a
> > >
> > > thirdparty.
> > >
> > >
> > >
> > > C1: Pointing to Apache Release
> > >
> > > - The PPMC 100% agree that we should produce high-quality Apache
> > >
> > >   compatible releases, and refer to them in the documents.
> > >
> > > - The installation documentation points to the official release
> download
> > >
> > > page [1].
> > >
> > > - There is a for developer section, which gives instructions to
> > developers
> > >
> > > about
> > >
> > >   how to clone the code from the git repo.
> > >
> > >
> > >
> > > C2: Number of releases
> > >
> > > - The PPMC wants to produce high quality releases according to feature
> > >
> > >   plan coordinated with the community, rather than creating more
> > releases.
> > >
> > > - The release process is well documented[2],
> > >
> > > - Per incubator policy[2], number of releases is not a hard bar for
> > >
> > > graduation.
> > >
> > >   "Podlings do not need to actually publish a release...". Instead the
> > most
> > >
> > > important
> > >
> > >   thing is the demonstration of being able to cut apache release.
> > >
> > > - Besides the release process being well documented and there are
> already
> > >
> > >   two high-quality releases(without WIP). The PMC contains several
> Apache
> > >
> > > members
> > >
> > >   who have previous experiences in cutting releases in TLP projects.
> > >
> > >   We are very confident that the PMC can continue to do so.
> > >
> > >
> > >
> > > C3: The discourse forum
> > >
> > > - The discourse forum, like the use of github, is mainly a way for user
> > >
> > > community to
> > >
> > >   engage with the project.
> > >
> > > - The usage of discourse forum is voted by the community[3]
> > >
> > > - The community follows the public archive principle, which means
> > >
> > > everything is archived
> > >
> > >   to mail-lists. Again, everything happens (also) happens on a
> mail-list.
> > >
> > > - The discourse forum is currently maintained by a few PMC members as
> > >
> > > volunteers,
> > >
> > >   as a thirdparty service to the community.
> > >
> > > - Again, the development happens in github, everything mirrored in dev@
> .
> > >
> > > The existence of the forum
> > >
> > >   would not block the development.
> > >
> > >
> > >
> > > Finally, to summarize. I believe one of the main reasons behind these
> > >
> > > issues are trust rather than procedural.
> > >
> > > As a ASF member, I am certainly strive to make sure the related
> volunteer
> > >
> > > maintained resources
> > >
> > > will continue beyond a certain individual by many ways including:
> > >
> > >
> > >
> > 

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Matt!

The main reason we used thirdparty(as stated above) is that the TVM PPMC
believes it is important to
protect the Apache brand and release policy(artifact being compatible
apache).
Given some of the optional test packages (e.g. pytorch/tensorflow) are not
what we could control,
and it is hard to validate their compatibility. There are also similar
issues around specific
runtime environments like cuda. Note that CI images are not artifacts that
we release, but merely
as a way to setup the ci build env.

So to keep things simple and clean from branding/licensing concerns, we use
a thirdparty and
clearly distinguish them from ASF artifacts. To the development, there is
no strict location requirement as
long as we can point to a cache. And the cache location can be changed at
any time point by any committer.

TQ

On Fri, Aug 28, 2020 at 7:13 PM Matt Sicker  wrote:

> Makes sense. Do you think the recent changes to Docker Hub will affect
> this? It might be a good idea to move the images into the Apache org, even
> if they’re not for end users.
>
> On Fri, Aug 28, 2020 at 20:52 Tianqi Chen  wrote:
>
> > Thanks Matt!
> >
> >
> >
> > That was indeed the approach we used before. The main problem of this
> >
> > approach is
> >
> > - There could be certain upstream changes (say tensorflow get updated)
> that
> >
> > may not retrigger the rebuild
> >
> > - The CI instance itself can quickly get crowded by the historical cached
> >
> > images and causes disk overflow problems.
> >
> > - Overall we find it is very hard to reproduce the CI error when there
> is a
> >
> > problem because of the potential mismatch(due to stale build)
> >
> >
> >
> > So the community switched to using a stable binary tag as for more stable
> >
> > env without blocking the CI, while periodically updating
> >
> > these images when this is a dependency change.
> >
> >
> >
> > TQ
> >
> >
> >
> > On Fri, Aug 28, 2020 at 6:47 PM Matt Sicker  wrote:
> >
> >
> >
> > > Is there even a need to upload the Docker images for CI? Docker
> > recognizes
> >
> > > layers by checksums, so CI build agents will cache the image layers
> >
> > > regardless of whether or not you upload them to Docker Hub. Layers get
> >
> > > rebuilt when the Docker file changes or you force an update.
> >
> > >
> >
> > > On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:
> >
> > >
> >
> > > > Thanks Justin and Henry for the discussion thread.
> >
> > > >
> >
> > > >
> >
> > > >
> >
> > > > Please allow me to dissect and summarize again some of the discussion
> >
> > > >
> >
> > > > points:)
> >
> > > >
> >
> > > >
> >
> > > >
> >
> > > > C0: Docker Image
> >
> > > >
> >
> > > > - There was a mis-understanding that the docker image like ci-gpu
> >
> > > >
> >
> > > >   contains tvm. They do not, and instead contain an environment to
> >
> > > >
> >
> > > >   build and run test cases.
> >
> > > >
> >
> > > > - The thirdparty binary cache is used to speed up the test build.
> > Anyone
> >
> > > >
> >
> > > >   could build these binaries from a TVM source.
> >
> > > >
> >
> > > > - To avoid confusion about the branding, we have renamed the docker
> hub
> >
> > > >
> >
> > > > name to
> >
> > > >
> >
> > > >   a different name that does not contain tvm so that it is clear as a
> >
> > > >
> >
> > > > thirdparty.
> >
> > > >
> >
> > > >
> >
> > > >
> >
> > > > C1: Pointing to Apache Release
> >
> > > >
> >
> > > > - The PPMC 100% agree that we should produce high-quality Apache
> >
> > > >
> >
> > > >   compatible releases, and refer to them in the documents.
> >
> > > >
> >
> > > > - The installation documentation points to the official release
> > download
> >
> > > >
> >
> > > > page [1].
> >
> > > >
> >
> > > > - There is a for developer section, which gives instructions to
> >
> > > developers
> >
> > > >
> >
> > > > about
> >
> > > >
> >
> > > >   how to clone the code from the git repo.
> >
> > > >
> >
> > > >
> >
> > > >
> >
> > > > C2: Number of releases
> >
> > > >
> >
> > > > - The PPMC wants to produce high quality releases according to
> feature
> >
> > > >
> >
> > > >   plan coordinated with the community, rather than creating more
> >
> > > releases.
> >
> > > >
> >
> > > > - The release process is well documented[2],
> >
> > > >
> >
> > > > - Per incubator policy[2], number of releases is not a hard bar for
> >
> > > >
> >
> > > > graduation.
> >
> > > >
> >
> > > >   "Podlings do not need to actually publish a release...". Instead
> the
> >
> > > most
> >
> > > >
> >
> > > > important
> >
> > > >
> >
> > > >   thing is the demonstration of being able to cut apache release.
> >
> > > >
> >
> > > > - Besides the release process being well documented and there are
> > already
> >
> > > >
> >
> > > >   two high-quality releases(without WIP). The PMC contains several
> > Apache
> >
> > > >
> >
> > > > members
> >
> > > >
> >
> > > >   who have previous experiences in cutting releases in TLP projects.
> >
> > > >
> >
> > > >   We are very confident that the 

Re: [DISCUSS] Apache TVM Graduation

[Matt Sicker]

Makes sense. Do you think the recent changes to Docker Hub will affect
this? It might be a good idea to move the images into the Apache org, even
if they’re not for end users.

On Fri, Aug 28, 2020 at 20:52 Tianqi Chen  wrote:

> Thanks Matt!
>
>
>
> That was indeed the approach we used before. The main problem of this
>
> approach is
>
> - There could be certain upstream changes (say tensorflow get updated) that
>
> may not retrigger the rebuild
>
> - The CI instance itself can quickly get crowded by the historical cached
>
> images and causes disk overflow problems.
>
> - Overall we find it is very hard to reproduce the CI error when there is a
>
> problem because of the potential mismatch(due to stale build)
>
>
>
> So the community switched to using a stable binary tag as for more stable
>
> env without blocking the CI, while periodically updating
>
> these images when this is a dependency change.
>
>
>
> TQ
>
>
>
> On Fri, Aug 28, 2020 at 6:47 PM Matt Sicker  wrote:
>
>
>
> > Is there even a need to upload the Docker images for CI? Docker
> recognizes
>
> > layers by checksums, so CI build agents will cache the image layers
>
> > regardless of whether or not you upload them to Docker Hub. Layers get
>
> > rebuilt when the Docker file changes or you force an update.
>
> >
>
> > On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:
>
> >
>
> > > Thanks Justin and Henry for the discussion thread.
>
> > >
>
> > >
>
> > >
>
> > > Please allow me to dissect and summarize again some of the discussion
>
> > >
>
> > > points:)
>
> > >
>
> > >
>
> > >
>
> > > C0: Docker Image
>
> > >
>
> > > - There was a mis-understanding that the docker image like ci-gpu
>
> > >
>
> > >   contains tvm. They do not, and instead contain an environment to
>
> > >
>
> > >   build and run test cases.
>
> > >
>
> > > - The thirdparty binary cache is used to speed up the test build.
> Anyone
>
> > >
>
> > >   could build these binaries from a TVM source.
>
> > >
>
> > > - To avoid confusion about the branding, we have renamed the docker hub
>
> > >
>
> > > name to
>
> > >
>
> > >   a different name that does not contain tvm so that it is clear as a
>
> > >
>
> > > thirdparty.
>
> > >
>
> > >
>
> > >
>
> > > C1: Pointing to Apache Release
>
> > >
>
> > > - The PPMC 100% agree that we should produce high-quality Apache
>
> > >
>
> > >   compatible releases, and refer to them in the documents.
>
> > >
>
> > > - The installation documentation points to the official release
> download
>
> > >
>
> > > page [1].
>
> > >
>
> > > - There is a for developer section, which gives instructions to
>
> > developers
>
> > >
>
> > > about
>
> > >
>
> > >   how to clone the code from the git repo.
>
> > >
>
> > >
>
> > >
>
> > > C2: Number of releases
>
> > >
>
> > > - The PPMC wants to produce high quality releases according to feature
>
> > >
>
> > >   plan coordinated with the community, rather than creating more
>
> > releases.
>
> > >
>
> > > - The release process is well documented[2],
>
> > >
>
> > > - Per incubator policy[2], number of releases is not a hard bar for
>
> > >
>
> > > graduation.
>
> > >
>
> > >   "Podlings do not need to actually publish a release...". Instead the
>
> > most
>
> > >
>
> > > important
>
> > >
>
> > >   thing is the demonstration of being able to cut apache release.
>
> > >
>
> > > - Besides the release process being well documented and there are
> already
>
> > >
>
> > >   two high-quality releases(without WIP). The PMC contains several
> Apache
>
> > >
>
> > > members
>
> > >
>
> > >   who have previous experiences in cutting releases in TLP projects.
>
> > >
>
> > >   We are very confident that the PMC can continue to do so.
>
> > >
>
> > >
>
> > >
>
> > > C3: The discourse forum
>
> > >
>
> > > - The discourse forum, like the use of github, is mainly a way for user
>
> > >
>
> > > community to
>
> > >
>
> > >   engage with the project.
>
> > >
>
> > > - The usage of discourse forum is voted by the community[3]
>
> > >
>
> > > - The community follows the public archive principle, which means
>
> > >
>
> > > everything is archived
>
> > >
>
> > >   to mail-lists. Again, everything happens (also) happens on a
> mail-list.
>
> > >
>
> > > - The discourse forum is currently maintained by a few PMC members as
>
> > >
>
> > > volunteers,
>
> > >
>
> > >   as a thirdparty service to the community.
>
> > >
>
> > > - Again, the development happens in github, everything mirrored in dev@
> .
>
> > >
>
> > > The existence of the forum
>
> > >
>
> > >   would not block the development.
>
> > >
>
> > >
>
> > >
>
> > > Finally, to summarize. I believe one of the main reasons behind these
>
> > >
>
> > > issues are trust rather than procedural.
>
> > >
>
> > > As a ASF member, I am certainly strive to make sure the related
> volunteer
>
> > >
>
> > > maintained resources
>
> > >
>
> > > will continue beyond a certain individual by many ways including:
>
> > >
>
> > >
>
> > >
>
> > > - Introduce multiple volunteers from different 

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Matt!

That was indeed the approach we used before. The main problem of this
approach is
- There could be certain upstream changes (say tensorflow get updated) that
may not retrigger the rebuild
- The CI instance itself can quickly get crowded by the historical cached
images and causes disk overflow problems.
- Overall we find it is very hard to reproduce the CI error when there is a
problem because of the potential mismatch(due to stale build)

So the community switched to using a stable binary tag as for more stable
env without blocking the CI, while periodically updating
these images when this is a dependency change.

TQ

On Fri, Aug 28, 2020 at 6:47 PM Matt Sicker  wrote:

> Is there even a need to upload the Docker images for CI? Docker recognizes
> layers by checksums, so CI build agents will cache the image layers
> regardless of whether or not you upload them to Docker Hub. Layers get
> rebuilt when the Docker file changes or you force an update.
>
> On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:
>
> > Thanks Justin and Henry for the discussion thread.
> >
> >
> >
> > Please allow me to dissect and summarize again some of the discussion
> >
> > points:)
> >
> >
> >
> > C0: Docker Image
> >
> > - There was a mis-understanding that the docker image like ci-gpu
> >
> >   contains tvm. They do not, and instead contain an environment to
> >
> >   build and run test cases.
> >
> > - The thirdparty binary cache is used to speed up the test build. Anyone
> >
> >   could build these binaries from a TVM source.
> >
> > - To avoid confusion about the branding, we have renamed the docker hub
> >
> > name to
> >
> >   a different name that does not contain tvm so that it is clear as a
> >
> > thirdparty.
> >
> >
> >
> > C1: Pointing to Apache Release
> >
> > - The PPMC 100% agree that we should produce high-quality Apache
> >
> >   compatible releases, and refer to them in the documents.
> >
> > - The installation documentation points to the official release download
> >
> > page [1].
> >
> > - There is a for developer section, which gives instructions to
> developers
> >
> > about
> >
> >   how to clone the code from the git repo.
> >
> >
> >
> > C2: Number of releases
> >
> > - The PPMC wants to produce high quality releases according to feature
> >
> >   plan coordinated with the community, rather than creating more
> releases.
> >
> > - The release process is well documented[2],
> >
> > - Per incubator policy[2], number of releases is not a hard bar for
> >
> > graduation.
> >
> >   "Podlings do not need to actually publish a release...". Instead the
> most
> >
> > important
> >
> >   thing is the demonstration of being able to cut apache release.
> >
> > - Besides the release process being well documented and there are already
> >
> >   two high-quality releases(without WIP). The PMC contains several Apache
> >
> > members
> >
> >   who have previous experiences in cutting releases in TLP projects.
> >
> >   We are very confident that the PMC can continue to do so.
> >
> >
> >
> > C3: The discourse forum
> >
> > - The discourse forum, like the use of github, is mainly a way for user
> >
> > community to
> >
> >   engage with the project.
> >
> > - The usage of discourse forum is voted by the community[3]
> >
> > - The community follows the public archive principle, which means
> >
> > everything is archived
> >
> >   to mail-lists. Again, everything happens (also) happens on a mail-list.
> >
> > - The discourse forum is currently maintained by a few PMC members as
> >
> > volunteers,
> >
> >   as a thirdparty service to the community.
> >
> > - Again, the development happens in github, everything mirrored in dev@.
> >
> > The existence of the forum
> >
> >   would not block the development.
> >
> >
> >
> > Finally, to summarize. I believe one of the main reasons behind these
> >
> > issues are trust rather than procedural.
> >
> > As a ASF member, I am certainly strive to make sure the related volunteer
> >
> > maintained resources
> >
> > will continue beyond a certain individual by many ways including:
> >
> >
> >
> > - Introduce multiple volunteers from different organizations in the PMC
> to
> >
> > do the work.
> >
> > - Make sure the messages are backed up to mail-list.
> >
> > - Help INFRA to manage certain services instead if they are willing.
> >
> > However, note in cases like CI
> >
> >   INFRA is also happy to let the PMC make specific choices.
> >
> >
> >
> > While many of the arguments in the above points boils down to "what if
> the
> >
> > few people did something evil via loophole X".
> >
> > And similar arguments can be applied to committer nominations in general.
> >
> > As I may quote from our committer
> >
> > nomination message "We like to work on trust rather than unnecessary
> >
> > constraints".
> >
> > This is the message that keeps bringing me back and makes me proud as a
> ASF
> >
> > member.
> >
> >
> >
> > There is no "one way" to the Apache way[3]. IMHO, the best 

Re: [DISCUSS] Apache TVM Graduation

[Matt Sicker]

Is there even a need to upload the Docker images for CI? Docker recognizes
layers by checksums, so CI build agents will cache the image layers
regardless of whether or not you upload them to Docker Hub. Layers get
rebuilt when the Docker file changes or you force an update.

On Fri, Aug 28, 2020 at 20:12 Tianqi Chen  wrote:

> Thanks Justin and Henry for the discussion thread.
>
>
>
> Please allow me to dissect and summarize again some of the discussion
>
> points:)
>
>
>
> C0: Docker Image
>
> - There was a mis-understanding that the docker image like ci-gpu
>
>   contains tvm. They do not, and instead contain an environment to
>
>   build and run test cases.
>
> - The thirdparty binary cache is used to speed up the test build. Anyone
>
>   could build these binaries from a TVM source.
>
> - To avoid confusion about the branding, we have renamed the docker hub
>
> name to
>
>   a different name that does not contain tvm so that it is clear as a
>
> thirdparty.
>
>
>
> C1: Pointing to Apache Release
>
> - The PPMC 100% agree that we should produce high-quality Apache
>
>   compatible releases, and refer to them in the documents.
>
> - The installation documentation points to the official release download
>
> page [1].
>
> - There is a for developer section, which gives instructions to developers
>
> about
>
>   how to clone the code from the git repo.
>
>
>
> C2: Number of releases
>
> - The PPMC wants to produce high quality releases according to feature
>
>   plan coordinated with the community, rather than creating more releases.
>
> - The release process is well documented[2],
>
> - Per incubator policy[2], number of releases is not a hard bar for
>
> graduation.
>
>   "Podlings do not need to actually publish a release...". Instead the most
>
> important
>
>   thing is the demonstration of being able to cut apache release.
>
> - Besides the release process being well documented and there are already
>
>   two high-quality releases(without WIP). The PMC contains several Apache
>
> members
>
>   who have previous experiences in cutting releases in TLP projects.
>
>   We are very confident that the PMC can continue to do so.
>
>
>
> C3: The discourse forum
>
> - The discourse forum, like the use of github, is mainly a way for user
>
> community to
>
>   engage with the project.
>
> - The usage of discourse forum is voted by the community[3]
>
> - The community follows the public archive principle, which means
>
> everything is archived
>
>   to mail-lists. Again, everything happens (also) happens on a mail-list.
>
> - The discourse forum is currently maintained by a few PMC members as
>
> volunteers,
>
>   as a thirdparty service to the community.
>
> - Again, the development happens in github, everything mirrored in dev@.
>
> The existence of the forum
>
>   would not block the development.
>
>
>
> Finally, to summarize. I believe one of the main reasons behind these
>
> issues are trust rather than procedural.
>
> As a ASF member, I am certainly strive to make sure the related volunteer
>
> maintained resources
>
> will continue beyond a certain individual by many ways including:
>
>
>
> - Introduce multiple volunteers from different organizations in the PMC to
>
> do the work.
>
> - Make sure the messages are backed up to mail-list.
>
> - Help INFRA to manage certain services instead if they are willing.
>
> However, note in cases like CI
>
>   INFRA is also happy to let the PMC make specific choices.
>
>
>
> While many of the arguments in the above points boils down to "what if the
>
> few people did something evil via loophole X".
>
> And similar arguments can be applied to committer nominations in general.
>
> As I may quote from our committer
>
> nomination message "We like to work on trust rather than unnecessary
>
> constraints".
>
> This is the message that keeps bringing me back and makes me proud as a ASF
>
> member.
>
>
>
> There is no "one way" to the Apache way[3]. IMHO, the best thing I love
>
> about the ASF is indeed the people, and the
>
> trust we give to each PMC as long as things are operated under the Apache
>
> Way. The TVM PPMC has been working hard to
>
> build a healthy, diverse and independent community. And strives to protect
>
> the Apache brand and uphold apache release policy.
>
> Again, we welcome constructive feedback for continued improvement.
>
>
>
> Thank you!
>
>
>
> TQ
>
>
>
> -
>
> - [1]
>
> https://tvm.apache.org/docs/install/from_source.html#install-from-source
>
> - [2] https://incubator.apache.org/guides/graduation.html
>
> - [3]
>
>
> https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E
>
> - [4] https://www.apache.org/theapacheway/
>
>
>
> On Fri, Aug 28, 2020 at 5:15 PM Henry Saputra 
>
> wrote:
>
>
>
> > Sure, that is easy to fix and good suggestion. But, hope it is not a
>
> > blocker issue.
>
> >
>
> > On Fri, Aug 28, 2020, 5:09 PM Justin Mclean 
>
> > wrote:
>
> >
>
> > > Hi,
>
> > 

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Justin and Henry for the discussion thread.

Please allow me to dissect and summarize again some of the discussion
points:)

C0: Docker Image
- There was a mis-understanding that the docker image like ci-gpu
  contains tvm. They do not, and instead contain an environment to
  build and run test cases.
- The thirdparty binary cache is used to speed up the test build. Anyone
  could build these binaries from a TVM source.
- To avoid confusion about the branding, we have renamed the docker hub
name to
  a different name that does not contain tvm so that it is clear as a
thirdparty.

C1: Pointing to Apache Release
- The PPMC 100% agree that we should produce high-quality Apache
  compatible releases, and refer to them in the documents.
- The installation documentation points to the official release download
page [1].
- There is a for developer section, which gives instructions to developers
about
  how to clone the code from the git repo.

C2: Number of releases
- The PPMC wants to produce high quality releases according to feature
  plan coordinated with the community, rather than creating more releases.
- The release process is well documented[2],
- Per incubator policy[2], number of releases is not a hard bar for
graduation.
  "Podlings do not need to actually publish a release...". Instead the most
important
  thing is the demonstration of being able to cut apache release.
- Besides the release process being well documented and there are already
  two high-quality releases(without WIP). The PMC contains several Apache
members
  who have previous experiences in cutting releases in TLP projects.
  We are very confident that the PMC can continue to do so.

C3: The discourse forum
- The discourse forum, like the use of github, is mainly a way for user
community to
  engage with the project.
- The usage of discourse forum is voted by the community[3]
- The community follows the public archive principle, which means
everything is archived
  to mail-lists. Again, everything happens (also) happens on a mail-list.
- The discourse forum is currently maintained by a few PMC members as
volunteers,
  as a thirdparty service to the community.
- Again, the development happens in github, everything mirrored in dev@.
The existence of the forum
  would not block the development.

Finally, to summarize. I believe one of the main reasons behind these
issues are trust rather than procedural.
As a ASF member, I am certainly strive to make sure the related volunteer
maintained resources
will continue beyond a certain individual by many ways including:

- Introduce multiple volunteers from different organizations in the PMC to
do the work.
- Make sure the messages are backed up to mail-list.
- Help INFRA to manage certain services instead if they are willing.
However, note in cases like CI
  INFRA is also happy to let the PMC make specific choices.

While many of the arguments in the above points boils down to "what if the
few people did something evil via loophole X".
And similar arguments can be applied to committer nominations in general.
As I may quote from our committer
nomination message "We like to work on trust rather than unnecessary
constraints".
This is the message that keeps bringing me back and makes me proud as a ASF
member.

There is no "one way" to the Apache way[3]. IMHO, the best thing I love
about the ASF is indeed the people, and the
trust we give to each PMC as long as things are operated under the Apache
Way. The TVM PPMC has been working hard to
build a healthy, diverse and independent community. And strives to protect
the Apache brand and uphold apache release policy.
Again, we welcome constructive feedback for continued improvement.

Thank you!

TQ

-
- [1]
https://tvm.apache.org/docs/install/from_source.html#install-from-source
- [2] https://incubator.apache.org/guides/graduation.html
- [3]
https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E
- [4] https://www.apache.org/theapacheway/

On Fri, Aug 28, 2020 at 5:15 PM Henry Saputra 
wrote:

> Sure, that is easy to fix and good suggestion. But, hope it is not a
> blocker issue.
>
> On Fri, Aug 28, 2020, 5:09 PM Justin Mclean 
> wrote:
>
> > Hi,
> >
> > The "install from source” page should probably point people to the source
> > releases rather than the latest code in GitHub.
> >
> > Thanks,
> > Justin
> > -
> > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > For additional commands, e-mail: general-h...@incubator.apache.org
> >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Hi Justin:

Again to further clarify the Docker image.

First of all, the PPMC is 100% with you that it is NOT fine to
release docker images with the Apache brand, or not following the
apache official release tag.

The test cache binary used by TVM CI is a different thing.
As the docker image itself is being used for test CI environment purposes.
For example ci-cpu contains the gcc compiler and deps like python,
tensorflow
in order to run the tests. The docker image itself does not contain TVM in
any form.

In order to use the ci-cpu, a developer would need to download a source,
enter the ci-cpu environment, and then type make to build tvm inside.


Q0: Why tag mismatch from Apache release tag

Whenever a ci dependency environment changes(e.g. we want to test against a
new version of tensorflow)
during development. We need to generate a new CI image for the test.
To generate a binary cache for the particular purpose of testing, we use a
new tag.
That is why the tag does not follow the Apache release tag because they are
different thing.

Q1: Why use thirdparty docker

The CI test environment can contain additional dependencies, e.g.
tensorflow/pytorch that may or maynot
be 100% Apache compatible. Because they are only used by the test cases and
are not dependency of the tvm itself.
We choose to use thirdparty binary caches, to simplify trademark concerns.
As per previous comment, we believe this is allowed

Q2: Branding

As mentioned in the above post, we agree that the name tvmai as a
thirdparty binary cache can be confusing,
and will move it to a name that is clearly thirdparty.

Q3: Pointing people to source releases

The PPMC is very mindful to only refer to the official Apache release as
official releases.

The install from source page indeed points people to the official Apache
release!
Note that the mention of github source explicitly contains "Developers
keyword".

TQ

---

- [1]
https://tvm.apache.org/docs/install/from_source.html#install-from-source


On Fri, Aug 28, 2020 at 5:14 PM Henry Saputra 
wrote:

> Look like TQ has given you answer and proposed solution about the Docker
> images.
>
> Hope that helps.
>
> On Fri, Aug 28, 2020, 5:01 PM Justin Mclean 
> wrote:
>
> > Hi,
> >
> > > I am sure there are many other binary packaging out there that use
> Apache
> > > project names, for example this one for Apache Spark:
> > > https://hub.docker.com/r/bitnami/spark/
> >
> > What other Apache projects do (Spark in particular for various reasons)
> > are not always good examples to follow. But you'll note in this case it:
> > a) Calls it Apache Spark
> > b) Includes a link to space.apache.org
> > b) In this case it seem clear it is a 3rd party, unlike TVM.
> >
> > IMO It should also have a trademark disclaimer and I’m not sure that the
> > way they have set up latest would be compliant with ASF policy. Ignoring
> > the latest tag you note that the tags also match Apache releases.
> >
> > > Are you concern about the name? The Docker image hub clearly stated
> this
> > is
> > > not official Docker image from Apache TVM podling. I think that is
> enough
> > > to indicate this is not done by Apache TVM as community.
> >
> > The docker images, if you found them on their own, give no indication
> that
> > is an Apache project or that this are not official releases. If I search
> > for “docker tvm” I get this link [1]. There is very little information on
> > that page, no recognition of Apache trademarks and no disclaimers.There
> are
> > also versions tagged that are not Apache releases.
> >
> > Why didn’t the project use Infra’a docker offering? [3]
> >
> > > And I am kind of surprise about your comment and question about the
> > number
> > > of downloads.
> >
> > A large number of downloads may indicate that users are using the docker
> > images rather than the offical sources. This may be fine if they are
> > released versions, but is probably not be fine if they are not. Hence the
> > latest tag being problematic.
> >
> > I find it odd in the TVM case there are more releases in Docker using the
> > same numbering system as the Apache releases. Why this discrepancy?
> >
> > Thanks,
> > Justin
> >
> > 1. https://hub.docker.com/u/tvmai
> > 2. https://www.apache.org/foundation/marks/#products
> > 3. https://hub.docker.com/u/apache
> > -
> > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > For additional commands, e-mail: general-h...@incubator.apache.org
> >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Henry Saputra]

Sure, that is easy to fix and good suggestion. But, hope it is not a
blocker issue.

On Fri, Aug 28, 2020, 5:09 PM Justin Mclean 
wrote:

> Hi,
>
> The "install from source” page should probably point people to the source
> releases rather than the latest code in GitHub.
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Henry Saputra]

Look like TQ has given you answer and proposed solution about the Docker
images.

Hope that helps.

On Fri, Aug 28, 2020, 5:01 PM Justin Mclean 
wrote:

> Hi,
>
> > I am sure there are many other binary packaging out there that use Apache
> > project names, for example this one for Apache Spark:
> > https://hub.docker.com/r/bitnami/spark/
>
> What other Apache projects do (Spark in particular for various reasons)
> are not always good examples to follow. But you'll note in this case it:
> a) Calls it Apache Spark
> b) Includes a link to space.apache.org
> b) In this case it seem clear it is a 3rd party, unlike TVM.
>
> IMO It should also have a trademark disclaimer and I’m not sure that the
> way they have set up latest would be compliant with ASF policy. Ignoring
> the latest tag you note that the tags also match Apache releases.
>
> > Are you concern about the name? The Docker image hub clearly stated this
> is
> > not official Docker image from Apache TVM podling. I think that is enough
> > to indicate this is not done by Apache TVM as community.
>
> The docker images, if you found them on their own, give no indication that
> is an Apache project or that this are not official releases. If I search
> for “docker tvm” I get this link [1]. There is very little information on
> that page, no recognition of Apache trademarks and no disclaimers.There are
> also versions tagged that are not Apache releases.
>
> Why didn’t the project use Infra’a docker offering? [3]
>
> > And I am kind of surprise about your comment and question about the
> number
> > of downloads.
>
> A large number of downloads may indicate that users are using the docker
> images rather than the offical sources. This may be fine if they are
> released versions, but is probably not be fine if they are not. Hence the
> latest tag being problematic.
>
> I find it odd in the TVM case there are more releases in Docker using the
> same numbering system as the Apache releases. Why this discrepancy?
>
> Thanks,
> Justin
>
> 1. https://hub.docker.com/u/tvmai
> 2. https://www.apache.org/foundation/marks/#products
> 3. https://hub.docker.com/u/apache
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

The "install from source” page should probably point people to the source 
releases rather than the latest code in GitHub.

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

> I am sure there are many other binary packaging out there that use Apache
> project names, for example this one for Apache Spark:
> https://hub.docker.com/r/bitnami/spark/

What other Apache projects do (Spark in particular for various reasons) are not 
always good examples to follow. But you'll note in this case it:
a) Calls it Apache Spark
b) Includes a link to space.apache.org
b) In this case it seem clear it is a 3rd party, unlike TVM.

IMO It should also have a trademark disclaimer and I’m not sure that the way 
they have set up latest would be compliant with ASF policy. Ignoring the latest 
tag you note that the tags also match Apache releases.

> Are you concern about the name? The Docker image hub clearly stated this is
> not official Docker image from Apache TVM podling. I think that is enough
> to indicate this is not done by Apache TVM as community.

The docker images, if you found them on their own, give no indication that is 
an Apache project or that this are not official releases. If I search for 
“docker tvm” I get this link [1]. There is very little information on that 
page, no recognition of Apache trademarks and no disclaimers.There are also 
versions tagged that are not Apache releases. 

Why didn’t the project use Infra’a docker offering? [3]

> And I am kind of surprise about your comment and question about the number
> of downloads.

A large number of downloads may indicate that users are using the docker images 
rather than the offical sources. This may be fine if they are released 
versions, but is probably not be fine if they are not. Hence the latest tag 
being problematic.

I find it odd in the TVM case there are more releases in Docker using the same 
numbering system as the Apache releases. Why this discrepancy?

Thanks,
Justin

1. https://hub.docker.com/u/tvmai
2. https://www.apache.org/foundation/marks/#products
3. https://hub.docker.com/u/apache
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Henry Saputra]

Thanks TQ, I understand the situation for the current Docker images.

The reason I mentioned about having official Docker images for TVM in the
ASF Docker hub to help differentiate between community/ external one, and
the one officially come from Apache TVM (P)PMC.

- Henry

On Fri, Aug 28, 2020, 4:50 PM Tianqi Chen  wrote:

> Thanks Henry.
>
> In this case, we choose the thirdparty for binary testing cache in
> particular to simplify the trademark concerns
> (since we do not want to vet the license of the test env). Which I believe
> is allowed.
>
> In terms of the docker image source, the community strives to release
> source releases that are 100% Apache compatible.
> The users can optionally build these docker images from the source.
>
> TQ
>
>
> On Fri, Aug 28, 2020 at 4:44 PM Henry Saputra 
> wrote:
>
> > The Apache does have official Docker hub [1].
> >
> > So, the TVM PPMC could, in the near future, create Docker images for
> Apache
> > TVM for official distribution to help people quickly consume and run the
> > source code releases.
> >
> >
> >
> > [1] https://hub.docker.com/u/apache
> >
> >
> > On Fri, Aug 28, 2020, 4:29 PM Tianqi Chen  wrote:
> >
> > > Thanks Justin!
> > >
> > > Docker Images and Downloads Stats
> > >
> > > - These images are indeed only used for testing purposes.
> > >   Note that there are more than 150 PRs every month, results in more
> than
> > >   tens of thousands of tests being triggered.
> > >   Every now and then the CI machine need to pull the fresh new image
> > after
> > >   the cache is cleared. So it is understandable that they have more
> than
> > 4K
> > > Downloads
> > > - As a additional note. The image DOES NOT contain TVM in any form. It
> > only
> > > contains
> > >   the build env that makes it easy to build tvm. So even if someone
> want
> > to
> > > download
> > >   the
> > >
> > > Advertisement of Docker Page in the Docs
> > > - As you rightfully points out, the installation page advertises the
> > docker
> > > page[1],
> > >   however note that the majority of the docker page contains
> instructions
> > > to build
> > >   the image from source, which is totally fine because source are 100%
> > > Apache compatible.
> > > - When docker image is referred in other place of the documents, they
> > > refers to the docker image
> > >   build from the Apache source release[2].
> > > - Base on the same rationale of priorizing Apache compatible branding
> > over
> > > thirdparties,
> > >   we always first mention how build from source first.
> > > - When reference to thirdparty image is mentioned, a clear disclaimer
> is
> > > being used
> > >   indicating that they are not apache release.
> > >
> > > - [1] https://tvm.apache.org/docs/install/index.html
> > > - [2] https://tvm.apache.org/docs/install/docker.html
> > >
> > > On Fri, Aug 28, 2020 at 4:24 PM Justin Mclean <
> jus...@classsoftware.com>
> > > wrote:
> > >
> > > > Hi,
> > > >
> > > > On that installation pages it does says "We provide docker utility
> > > scripts
> > > > to help developers to setup development environment. They are also
> > > helpful
> > > > run through TVM demo and tutorials.” and "You can find some
> un-official
> > > > prebuilt images in https://hub.docker.com/r/tvmai/.” IMO this
> doesn’t
> > > > match with them being just CI images used for testing.
> > > >
> > > > With the documentation I note in one place it says "to use the
> official
> > > > Docker Image.”
> > > >
> > > > Thanks,
> > > > Justin
> > > > -
> > > > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > > > For additional commands, e-mail: general-h...@incubator.apache.org
> > > >
> > > >
> > >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Henry.

In this case, we choose the thirdparty for binary testing cache in
particular to simplify the trademark concerns
(since we do not want to vet the license of the test env). Which I believe
is allowed.

In terms of the docker image source, the community strives to release
source releases that are 100% Apache compatible.
The users can optionally build these docker images from the source.

TQ


On Fri, Aug 28, 2020 at 4:44 PM Henry Saputra 
wrote:

> The Apache does have official Docker hub [1].
>
> So, the TVM PPMC could, in the near future, create Docker images for Apache
> TVM for official distribution to help people quickly consume and run the
> source code releases.
>
>
>
> [1] https://hub.docker.com/u/apache
>
>
> On Fri, Aug 28, 2020, 4:29 PM Tianqi Chen  wrote:
>
> > Thanks Justin!
> >
> > Docker Images and Downloads Stats
> >
> > - These images are indeed only used for testing purposes.
> >   Note that there are more than 150 PRs every month, results in more than
> >   tens of thousands of tests being triggered.
> >   Every now and then the CI machine need to pull the fresh new image
> after
> >   the cache is cleared. So it is understandable that they have more than
> 4K
> > Downloads
> > - As a additional note. The image DOES NOT contain TVM in any form. It
> only
> > contains
> >   the build env that makes it easy to build tvm. So even if someone want
> to
> > download
> >   the
> >
> > Advertisement of Docker Page in the Docs
> > - As you rightfully points out, the installation page advertises the
> docker
> > page[1],
> >   however note that the majority of the docker page contains instructions
> > to build
> >   the image from source, which is totally fine because source are 100%
> > Apache compatible.
> > - When docker image is referred in other place of the documents, they
> > refers to the docker image
> >   build from the Apache source release[2].
> > - Base on the same rationale of priorizing Apache compatible branding
> over
> > thirdparties,
> >   we always first mention how build from source first.
> > - When reference to thirdparty image is mentioned, a clear disclaimer is
> > being used
> >   indicating that they are not apache release.
> >
> > - [1] https://tvm.apache.org/docs/install/index.html
> > - [2] https://tvm.apache.org/docs/install/docker.html
> >
> > On Fri, Aug 28, 2020 at 4:24 PM Justin Mclean 
> > wrote:
> >
> > > Hi,
> > >
> > > On that installation pages it does says "We provide docker utility
> > scripts
> > > to help developers to setup development environment. They are also
> > helpful
> > > run through TVM demo and tutorials.” and "You can find some un-official
> > > prebuilt images in https://hub.docker.com/r/tvmai/.” IMO this doesn’t
> > > match with them being just CI images used for testing.
> > >
> > > With the documentation I note in one place it says "to use the official
> > > Docker Image.”
> > >
> > > Thanks,
> > > Justin
> > > -
> > > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > > For additional commands, e-mail: general-h...@incubator.apache.org
> > >
> > >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Henry Saputra]

The Apache does have official Docker hub [1].

So, the TVM PPMC could, in the near future, create Docker images for Apache
TVM for official distribution to help people quickly consume and run the
source code releases.



[1] https://hub.docker.com/u/apache


On Fri, Aug 28, 2020, 4:29 PM Tianqi Chen  wrote:

> Thanks Justin!
>
> Docker Images and Downloads Stats
>
> - These images are indeed only used for testing purposes.
>   Note that there are more than 150 PRs every month, results in more than
>   tens of thousands of tests being triggered.
>   Every now and then the CI machine need to pull the fresh new image after
>   the cache is cleared. So it is understandable that they have more than 4K
> Downloads
> - As a additional note. The image DOES NOT contain TVM in any form. It only
> contains
>   the build env that makes it easy to build tvm. So even if someone want to
> download
>   the
>
> Advertisement of Docker Page in the Docs
> - As you rightfully points out, the installation page advertises the docker
> page[1],
>   however note that the majority of the docker page contains instructions
> to build
>   the image from source, which is totally fine because source are 100%
> Apache compatible.
> - When docker image is referred in other place of the documents, they
> refers to the docker image
>   build from the Apache source release[2].
> - Base on the same rationale of priorizing Apache compatible branding over
> thirdparties,
>   we always first mention how build from source first.
> - When reference to thirdparty image is mentioned, a clear disclaimer is
> being used
>   indicating that they are not apache release.
>
> - [1] https://tvm.apache.org/docs/install/index.html
> - [2] https://tvm.apache.org/docs/install/docker.html
>
> On Fri, Aug 28, 2020 at 4:24 PM Justin Mclean 
> wrote:
>
> > Hi,
> >
> > On that installation pages it does says "We provide docker utility
> scripts
> > to help developers to setup development environment. They are also
> helpful
> > run through TVM demo and tutorials.” and "You can find some un-official
> > prebuilt images in https://hub.docker.com/r/tvmai/.” IMO this doesn’t
> > match with them being just CI images used for testing.
> >
> > With the documentation I note in one place it says "to use the official
> > Docker Image.”
> >
> > Thanks,
> > Justin
> > -
> > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > For additional commands, e-mail: general-h...@incubator.apache.org
> >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

To follow up the comment on docker image, as I mentioned in my previous
email.

We are changing the use of docker hub name to a new name that does not
contain tvm https://github.com/apache/incubator-tvm/pull/6360
The tvmai docker hub will be removed after the PR is merged
As I stated in my previous email. The PPMC strives to protect the Apache
brand, and we appreciate your comment on the docker binary cache name.

We would certainly love further constructive feedback on this topic.

Thanks
TQ

On Fri, Aug 28, 2020 at 4:34 PM Henry Saputra 
wrote:

> Hi Justin,
>
> I am sure there are many other binary packaging out there that use Apache
> project names, for example this one for Apache Spark:
> https://hub.docker.com/r/bitnami/spark/
>
> Are you concern about the name? The Docker image hub clearly stated this is
> not official Docker image from Apache TVM podling. I think that is enough
> to indicate this is not done by Apache TVM as community.
>
> And I am kind of surprise about your comment and question about the number
> of downloads. We could not regulate how many people try to dowload the
> images.
>
> - Henry
>
>
> On Fri, Aug 28, 2020, 4:16 PM Justin Mclean 
> wrote:
>
> > Hi,
> >
> > > Then why do they have 40K+ downloads? It also doesn’t explain the
> > version numbers.
> >
> > Sorry 4K I misread that, but even that seem high for something that is
> > used for testing.
> >
> > Thanks,
> > Justin
> > -
> > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> > For additional commands, e-mail: general-h...@incubator.apache.org
> >
> >
>

Re: [DISCUSS] Apache TVM Graduation

[Henry Saputra]

Hi Justin,

I am sure there are many other binary packaging out there that use Apache
project names, for example this one for Apache Spark:
https://hub.docker.com/r/bitnami/spark/

Are you concern about the name? The Docker image hub clearly stated this is
not official Docker image from Apache TVM podling. I think that is enough
to indicate this is not done by Apache TVM as community.

And I am kind of surprise about your comment and question about the number
of downloads. We could not regulate how many people try to dowload the
images.

- Henry


On Fri, Aug 28, 2020, 4:16 PM Justin Mclean 
wrote:

> Hi,
>
> > Then why do they have 40K+ downloads? It also doesn’t explain the
> version numbers.
>
> Sorry 4K I misread that, but even that seem high for something that is
> used for testing.
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Justin!

Docker Images and Downloads Stats

- These images are indeed only used for testing purposes.
  Note that there are more than 150 PRs every month, results in more than
  tens of thousands of tests being triggered.
  Every now and then the CI machine need to pull the fresh new image after
  the cache is cleared. So it is understandable that they have more than 4K
Downloads
- As a additional note. The image DOES NOT contain TVM in any form. It only
contains
  the build env that makes it easy to build tvm. So even if someone want to
download
  the

Advertisement of Docker Page in the Docs
- As you rightfully points out, the installation page advertises the docker
page[1],
  however note that the majority of the docker page contains instructions
to build
  the image from source, which is totally fine because source are 100%
Apache compatible.
- When docker image is referred in other place of the documents, they
refers to the docker image
  build from the Apache source release[2].
- Base on the same rationale of priorizing Apache compatible branding over
thirdparties,
  we always first mention how build from source first.
- When reference to thirdparty image is mentioned, a clear disclaimer is
being used
  indicating that they are not apache release.

- [1] https://tvm.apache.org/docs/install/index.html
- [2] https://tvm.apache.org/docs/install/docker.html

On Fri, Aug 28, 2020 at 4:24 PM Justin Mclean 
wrote:

> Hi,
>
> On that installation pages it does says "We provide docker utility scripts
> to help developers to setup development environment. They are also helpful
> run through TVM demo and tutorials.” and "You can find some un-official
> prebuilt images in https://hub.docker.com/r/tvmai/.” IMO this doesn’t
> match with them being just CI images used for testing.
>
> With the documentation I note in one place it says "to use the official
> Docker Image.”
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

On that installation pages it does says "We provide docker utility scripts to 
help developers to setup development environment. They are also helpful run 
through TVM demo and tutorials.” and "You can find some un-official prebuilt 
images in https://hub.docker.com/r/tvmai/.” IMO this doesn’t match with them 
being just CI images used for testing.

With the documentation I note in one place it says "to use the official Docker 
Image.” 

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

> Then why do they have 40K+ downloads? It also doesn’t explain the version 
> numbers.

Sorry 4K I misread that, but even that seem high for something that is used for 
testing.

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

> - As you can see that the domain redirects to https://tvm.apache.org/.
>  So hopefully there is no confusion here.

In some cases it does not. e.g. discuss.tvm.ai IMO this is a graduation blocker.

> P2: releases
> - The PPMC has a well documented release process[1]
>  that anyone can go and produce a release.
> - While we understand that it is good to produce more releases,
>  hopefully it is not a hard blocker, as the most important thing
>  for the PMC to be able to make high-quality Apache releases.

IMO this is a graduation blocker. A project meeds to make several releases with 
different release managers. Even if it is well documented only having one 
release manager is a large risk. If the process is well documented and easy to 
do then why haven’t you produced more releases?


> P2: Docker image
> 
> - These docker images do not contain tvm, instead they contain CI
> environment
>  to set up dependencies(e.g. compilers libraries) for running integration
> tests.

Then why do they have 40K+ downloads? It also doesn’t explain the version 
numbers.

> - These images are not Apache releases, there is a clear disclaimer
>  that they are thirdparty and not part of the ASF release[1].

If they are not Apache releases then they cannot use the TVM name.

> - The PPMC also clearly understand the need to protect Apache brand, and
>  clearly differentiate Apache release from thirdparty ones by adding
> disclaimers[3] to places when they are referred to.

You need to do more than just add a disclaimer.

>  name that does not contain tvm. Please note again these binaries are only
>  used for integration test purposes, the community do not intend to
> advertise them as Apache releases.

Except they are [1]

Thanks,
Justin

1. https://tvm.apache.org/docs/install/index.html
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Justin for bringing up these points!

P0: tvm.ai domain

- As you can see that the domain redirects to https://tvm.apache.org/.
  So hopefully there is no confusion here.

P2: releases
- The PPMC has a well documented release process[1]
  that anyone can go and produce a release.
- While we understand that it is good to produce more releases,
  hopefully it is not a hard blocker, as the most important thing
  for the PMC to be able to make high-quality Apache releases.

P1: Recognizing reviewers and form of contribution
- As said in the above email, the PPMC considers all forms of
contributions, including
  but not limited to code reviews, code, community building, documentation.
- Just like the (un-official) maturity model, which serves the goal to help
community growth,
  it is not a gatekeeper in anyform. There is nothing that prevents a PMC
member to discuss
  the case of non-code contributors, bringing them in as reviewer (to keep
them in the loop of these perspectives),
  or as committer.
- As a fact, community interaction and contributions documentation and
building are actively being discussed
  and leads to some of the latest committer nominations

P2: Docker image

- These docker images do not contain tvm, instead they contain CI
environment
  to set up dependencies(e.g. compilers libraries) for running integration
tests.
- These images are not Apache releases, there is a clear disclaimer
  that they are thirdparty and not part of the ASF release[1].
  We would be more than happy to improve disclaimer.
- These images can be built by anyone from the source in the TVM repo (not
depending on a single person).
  The docker hub image mainly serves as a cache to make sure the CI does
not have to
  rebuilt them each time.
- The docker caches are maintained by community volunteers, not wearing ASF
hat.
  In an event, say the cache image stops to work.
  The community can always self-operate, by picking a different cache, and
using the same script.
- We view it as similar cases where INFRA allows each project to use
thirdparty services
  for test and integration purposes. The bottom line though is not to
release non-official Apache
  artifacts as Apache.

P3: Trademark and Licensing

- The PPMC fully understands the importance of producing release
  that are 100% Apache compatible and has been doing so in the past two
releases[2].
- The PPMC also clearly understand the need to protect Apache brand, and
  clearly differentiate Apache release from thirdparty ones by adding
disclaimers[3] to places when they are referred to.
- To make things simple, the TVM community only produce source release[2]
  that are 100% Apache compatible.
- We agree that the use of tvmai as a docker hub org name for test cache is
a gray area,
  and might cause confusion for branding despite the disclaimer.
  To ensure better clarity, we will move these binary caches to a different
org
  name that does not contain tvm. Please note again these binaries are only
  used for integration test purposes, the community do not intend to
advertise them as Apache releases.

I think most of the concerns are around whether the community can protect
the Apache brand
and produce high quality apache releases. These are very important
perspectives, and
the PPMC is very committed to these goals,  and making sure that the
project can continue to thrive beyond
the support of a single organization, or participation of single individual:

- The release process is clearly documented.
- The test infra is completely checked into as source of the repo (100%
Apache), anyone can take over and reproduce.
- The documents, tutorials and other things points to
https://tvm.apache.org/.
- When necessary thirdparty items are mentioned, a clear disclaimer is
placed.
- A PMC member only nominates candidates from other organizations to
encourage diverse sustained community growth.
- Engage trademark for permission in terms of community driven event[4]

While we think that the community is matured enough to handle these
perspectives.
We would be more than happy to receive constructive feedback on how we can
continue to do better on these areas.

--
- [1] https://tvm.apache.org/docs/contribute/release_process
- [2] https://tvm.apache.org/download
- [3] https://tvm.apache.org/docs/install/docker.html
- [4]
https://lists.apache.org/thread.html/r8559979c58fb92f08f3e253522dc227a51313dafff2ec19076e1093c%40%3Ctrademarks.apache.org%3E

On Fri, Aug 28, 2020 at 3:08 PM Justin Mclean 
wrote:

> Hi,
>
> I’m also concerned that the person who seems responsible for these docker
> images seems to be the proposed chair of the project.
>
> Thanks,
> Justin
> -
> To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
> For additional commands, e-mail: general-h...@incubator.apache.org
>
>

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

I’m also concerned that the person who seems responsible for these docker 
images seems to be the proposed chair of the project.

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [DISCUSS] Apache TVM Graduation

[Justin Mclean]

Hi,

I also have several concerns.

1.  Use of the tvm.ai domain - has permission been obtained from trademark to 
use this domain? Who has control of it?

2.  With only 2 releases with 1 release manager I don’t think the project is 
mature enough to graduate,

3. Branding/trademarks on these docker images [1] I’m concerned that users are 
using these rather than the official releases. I note they include version 
numbers that are not Apache releases. Do these contain unreleased code?

4. I also have concerns about the review position. It would seem that this  
gives a path for people to contribute code to become committers but not other 
forms of contribution.


1. https://hub.docker.com/u/tvmai
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [VOTE] Graduate Apache TVM as TLP

[Justin Mclean]

HI,

I’m -1 (binding), please see the discussion thread for my concerns.

Thanks,
Justin
-
To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org
For additional commands, e-mail: general-h...@incubator.apache.org

Re: [VOTE] Graduate Apache TVM as TLP

[Kevin Ratnasekera]

+1 (binding)

On Fri, Aug 28, 2020 at 10:21 PM Felix Cheung 
wrote:

> +1 (binding)
>
>
> On Thu, Aug 27, 2020 at 10:02 AM Tianqi Chen  wrote:
>
> > Dear Incubator:
> >
> >
> >
> > Following a heated discussion with great support from our mentors,
> >
> > committers and community members.
> >
> > The Apache(incubating) TVM community has formally voted for graduation
> and
> >
> > bring the case to the IPMC.
> >
> >
> >
> > Please also see the community discussion thread[1] and voting thread[2]:
> >
> >
> >
> > This is a formal voting thread about Apache TVM's graduation, please
> Vote:
> >
> > [ ] +1 - Recommend graduation of Apache TVM as a TLP
> >
> > [ ]  0 - I don't feel strongly about it, but don't object
> >
> > [ ] -1 - Do not recommend graduation of Apache TVM because...
> >
> >
> >
> > The VOTE will open for at least 72 hours.
> >
> >
> >
> > --
> >
> > The TVM project has been an Apache incubator project for nearly 1.5 year
> >
> > now. In the past one and half year, the community grew healthily under
> the
> >
> > Apache way. Some highlights include:
> >
> >
> >
> > - A successful developer conference that we are continuing to host this
> > year
> >
> > - Great community growth, as of now, the community contains 16 PPMC
> >
> > members, 31 committers, from a diverse list of organizations. We are
> >
> > actively growing the list monthly.
> >
> > - Active contributions: ~ 150 PRs merged each month.
> >
> >
> >
> > The community has produced two formal apache releases. While we could
> also
> >
> > wait until more releases. We feel that the community is mature enough
> that
> >
> > we can push for graduation as it is, and continue to push for high
> quality
> >
> > releases concurrently.
> >
> >
> >
> > For reference, we also put together a maturity evaluation doc[3] under
> the
> >
> > Apache maturity model.
> >
> >
> >
> > Some additional note about the resolution below: the current PPMC will be
> >
> > transitioned to the PMC. We have invited all the mentors in the current
> >
> > PPMC who like to stay involved.
> >
> >
> >
> > -
> >
> >
> >
> > Establish the Apache TVM Project
> >
> >
> >
> > WHEREAS, the Board of Directors deems it to be in the best interests of
> >
> > the Foundation and consistent with the Foundation's purpose to establish
> >
> > a Project Management Committee charged with the creation and maintenance
> >
> > of open-source software, for distribution at no charge to the public,
> >
> > related to compilation of machine learning models to run on a wide range
> of
> >
> > hardware platforms...
> >
> >
> >
> > NOW, THEREFORE, BE IT RESOLVED, that a Project Management Committee
> >
> > (PMC), to be known as the "Apache TVM Project", be and hereby is
> >
> > established pursuant to Bylaws of the Foundation; and be it further
> >
> >
> >
> > RESOLVED, that the Apache TVM Project be and hereby is responsible for
> the
> >
> > creation and maintenance of software related to compilation of machine
> >
> > learning models to run on a wide range of hardware platforms; and be it
> >
> > further
> >
> >
> >
> > RESOLVED, that the office of "Vice President, Apache TVM" be and
> >
> > hereby is created, the person holding such office to serve at the
> >
> > direction of the Board of Directors as the chair of the Apache TVM
> >
> > Project, and to have primary responsibility for management of the
> >
> > projects within the scope of responsibility of the Apache TVM
> >
> > Project; and be it further
> >
> >
> >
> > RESOLVED, that the persons listed immediately below be and hereby are
> >
> > appointed to serve as the initial members of the Apache TVM Project:
> >
> >
> >
> >  * Tianqi Chen 
> >
> >  * Timothy Chen 
> >
> >  * Zhi Chen 
> >
> >  * Byung-Gon Chun 
> >
> >  * Ziheng Jiang 
> >
> >  * Furkan Kamaci 
> >
> >  * YiZhi Liu 
> >
> >  * Masahiro Masuda 
> >
> >  * Thierry Moreau 
> >
> >  * Jared Roesch 
> >
> >  * Henry Saputra 
> >
> >  * Haichen Shen 
> >
> >  * Markus Weimer 
> >
> >  * Eddie Yan 
> >
> >  * Lianmin Zheng 
> >
> >
> >
> > NOW, THEREFORE, BE IT FURTHER RESOLVED, that Tianqi Chen be appointed to
> >
> > the office of Vice President, Apache TVM, to serve in accordance
> >
> > with and subject to the direction of the Board of Directors and the
> >
> > Bylaws of the Foundation until death, resignation, retirement, removal
> >
> > or disqualification, or until a successor is appointed; and be it
> >
> > further
> >
> >
> >
> > RESOLVED, that the Apache TVM Project be and hereby is tasked with
> >
> > the migration and rationalization of the Apache Incubator TVM
> >
> > podling; and be it further
> >
> >
> >
> > RESOLVED, that all responsibilities pertaining to the Apache Incubator
> >
> > TVM  podling encumbered upon the Apache Incubator PMC are hereafter
> >
> > Discharged.
> >
> >
> >
> > - [1]
> >
> >
> >
> https://lists.apache.org/thread.html/r91b8f469c6a54769869bb2435b7334a28bcff885ae078ab5612dae00%40%3Cdev.tvm.apache.org%3E
> >

Re: [VOTE] Graduate Apache TVM as TLP

[Felix Cheung]

+1 (binding)


On Thu, Aug 27, 2020 at 10:02 AM Tianqi Chen  wrote:

> Dear Incubator:
>
>
>
> Following a heated discussion with great support from our mentors,
>
> committers and community members.
>
> The Apache(incubating) TVM community has formally voted for graduation and
>
> bring the case to the IPMC.
>
>
>
> Please also see the community discussion thread[1] and voting thread[2]:
>
>
>
> This is a formal voting thread about Apache TVM's graduation, please Vote:
>
> [ ] +1 - Recommend graduation of Apache TVM as a TLP
>
> [ ]  0 - I don't feel strongly about it, but don't object
>
> [ ] -1 - Do not recommend graduation of Apache TVM because...
>
>
>
> The VOTE will open for at least 72 hours.
>
>
>
> --
>
> The TVM project has been an Apache incubator project for nearly 1.5 year
>
> now. In the past one and half year, the community grew healthily under the
>
> Apache way. Some highlights include:
>
>
>
> - A successful developer conference that we are continuing to host this
> year
>
> - Great community growth, as of now, the community contains 16 PPMC
>
> members, 31 committers, from a diverse list of organizations. We are
>
> actively growing the list monthly.
>
> - Active contributions: ~ 150 PRs merged each month.
>
>
>
> The community has produced two formal apache releases. While we could also
>
> wait until more releases. We feel that the community is mature enough that
>
> we can push for graduation as it is, and continue to push for high quality
>
> releases concurrently.
>
>
>
> For reference, we also put together a maturity evaluation doc[3] under the
>
> Apache maturity model.
>
>
>
> Some additional note about the resolution below: the current PPMC will be
>
> transitioned to the PMC. We have invited all the mentors in the current
>
> PPMC who like to stay involved.
>
>
>
> -
>
>
>
> Establish the Apache TVM Project
>
>
>
> WHEREAS, the Board of Directors deems it to be in the best interests of
>
> the Foundation and consistent with the Foundation's purpose to establish
>
> a Project Management Committee charged with the creation and maintenance
>
> of open-source software, for distribution at no charge to the public,
>
> related to compilation of machine learning models to run on a wide range of
>
> hardware platforms...
>
>
>
> NOW, THEREFORE, BE IT RESOLVED, that a Project Management Committee
>
> (PMC), to be known as the "Apache TVM Project", be and hereby is
>
> established pursuant to Bylaws of the Foundation; and be it further
>
>
>
> RESOLVED, that the Apache TVM Project be and hereby is responsible for the
>
> creation and maintenance of software related to compilation of machine
>
> learning models to run on a wide range of hardware platforms; and be it
>
> further
>
>
>
> RESOLVED, that the office of "Vice President, Apache TVM" be and
>
> hereby is created, the person holding such office to serve at the
>
> direction of the Board of Directors as the chair of the Apache TVM
>
> Project, and to have primary responsibility for management of the
>
> projects within the scope of responsibility of the Apache TVM
>
> Project; and be it further
>
>
>
> RESOLVED, that the persons listed immediately below be and hereby are
>
> appointed to serve as the initial members of the Apache TVM Project:
>
>
>
>  * Tianqi Chen 
>
>  * Timothy Chen 
>
>  * Zhi Chen 
>
>  * Byung-Gon Chun 
>
>  * Ziheng Jiang 
>
>  * Furkan Kamaci 
>
>  * YiZhi Liu 
>
>  * Masahiro Masuda 
>
>  * Thierry Moreau 
>
>  * Jared Roesch 
>
>  * Henry Saputra 
>
>  * Haichen Shen 
>
>  * Markus Weimer 
>
>  * Eddie Yan 
>
>  * Lianmin Zheng 
>
>
>
> NOW, THEREFORE, BE IT FURTHER RESOLVED, that Tianqi Chen be appointed to
>
> the office of Vice President, Apache TVM, to serve in accordance
>
> with and subject to the direction of the Board of Directors and the
>
> Bylaws of the Foundation until death, resignation, retirement, removal
>
> or disqualification, or until a successor is appointed; and be it
>
> further
>
>
>
> RESOLVED, that the Apache TVM Project be and hereby is tasked with
>
> the migration and rationalization of the Apache Incubator TVM
>
> podling; and be it further
>
>
>
> RESOLVED, that all responsibilities pertaining to the Apache Incubator
>
> TVM  podling encumbered upon the Apache Incubator PMC are hereafter
>
> Discharged.
>
>
>
> - [1]
>
>
> https://lists.apache.org/thread.html/r91b8f469c6a54769869bb2435b7334a28bcff885ae078ab5612dae00%40%3Cdev.tvm.apache.org%3E
>
> - [2]
>
>
> https://lists.apache.org/thread.html/rd5b8eefe49af09a2d0913758a5e5737b3fdb9072bc0becf4a2b2c7ee%40%3Cdev.tvm.apache.org%3E
>
> - [3]
>
>
> https://docs.google.com/document/d/18nyAH-fcptVezAxPQe6H3FeTKPRkujOp1tc1YRSPLok/edit?usp=sharing
>
> <
> https://docs.google.com/document/d/18nyAH-fcptVezAxPQe6H3FeTKPRkujOp1tc1YRSPLok/edit?usp=sharing
> >
>
>

Re: [VOTE] Graduate Apache TVM as TLP

[Dave Fisher]

I’m changing my vote to +1 (binding)

Best of luck TVM.

Regards,
Dave

> On Aug 27, 2020, at 10:29 AM, Dave Fisher  wrote:
> 
> I’m -1 (binding), because this is being rushed. I’ll be willing to change 
> after there is actual discussion on general@.
> 
>> On Aug 27, 2020, at 10:02 AM, Tianqi Chen  wrote:
>> 
>> Dear Incubator:
>> 
>> Following a heated discussion with great support from our mentors,
>> committers and community members.
>> The Apache(incubating) TVM community has formally voted for graduation and
>> bring the case to the IPMC.
>> 
>> Please also see the community discussion thread[1] and voting thread[2]:
>> 
>> This is a formal voting thread about Apache TVM's graduation, please Vote:
>> [ ] +1 - Recommend graduation of Apache TVM as a TLP
>> [ ]  0 - I don't feel strongly about it, but don't object
>> [ ] -1 - Do not recommend graduation of Apache TVM because...
>> 
>> The VOTE will open for at least 72 hours.
>> 
>> --
>> The TVM project has been an Apache incubator project for nearly 1.5 year
>> now. In the past one and half year, the community grew healthily under the
>> Apache way. Some highlights include:
>> 
>> - A successful developer conference that we are continuing to host this year
>> - Great community growth, as of now, the community contains 16 PPMC
>> members, 31 committers, from a diverse list of organizations. We are
>> actively growing the list monthly.
>> - Active contributions: ~ 150 PRs merged each month.
>> 
>> The community has produced two formal apache releases. While we could also
>> wait until more releases. We feel that the community is mature enough that
>> we can push for graduation as it is, and continue to push for high quality
>> releases concurrently.
>> 
>> For reference, we also put together a maturity evaluation doc[3] under the
>> Apache maturity model.
>> 
>> Some additional note about the resolution below: the current PPMC will be
>> transitioned to the PMC. We have invited all the mentors in the current
>> PPMC who like to stay involved.
>> 
>> -
>> 
>> Establish the Apache TVM Project
>> 
>> WHEREAS, the Board of Directors deems it to be in the best interests of
>> the Foundation and consistent with the Foundation's purpose to establish
>> a Project Management Committee charged with the creation and maintenance
>> of open-source software, for distribution at no charge to the public,
>> related to compilation of machine learning models to run on a wide range of
>> hardware platforms...
>> 
>> NOW, THEREFORE, BE IT RESOLVED, that a Project Management Committee
>> (PMC), to be known as the "Apache TVM Project", be and hereby is
>> established pursuant to Bylaws of the Foundation; and be it further
>> 
>> RESOLVED, that the Apache TVM Project be and hereby is responsible for the
>> creation and maintenance of software related to compilation of machine
>> learning models to run on a wide range of hardware platforms; and be it
>> further
>> 
>> RESOLVED, that the office of "Vice President, Apache TVM" be and
>> hereby is created, the person holding such office to serve at the
>> direction of the Board of Directors as the chair of the Apache TVM
>> Project, and to have primary responsibility for management of the
>> projects within the scope of responsibility of the Apache TVM
>> Project; and be it further
>> 
>> RESOLVED, that the persons listed immediately below be and hereby are
>> appointed to serve as the initial members of the Apache TVM Project:
>> 
>> * Tianqi Chen 
>> * Timothy Chen 
>> * Zhi Chen 
>> * Byung-Gon Chun 
>> * Ziheng Jiang 
>> * Furkan Kamaci 
>> * YiZhi Liu 
>> * Masahiro Masuda 
>> * Thierry Moreau 
>> * Jared Roesch 
>> * Henry Saputra 
>> * Haichen Shen 
>> * Markus Weimer 
>> * Eddie Yan 
>> * Lianmin Zheng 
>> 
>> NOW, THEREFORE, BE IT FURTHER RESOLVED, that Tianqi Chen be appointed to
>> the office of Vice President, Apache TVM, to serve in accordance
>> with and subject to the direction of the Board of Directors and the
>> Bylaws of the Foundation until death, resignation, retirement, removal
>> or disqualification, or until a successor is appointed; and be it
>> further
>> 
>> RESOLVED, that the Apache TVM Project be and hereby is tasked with
>> the migration and rationalization of the Apache Incubator TVM
>> podling; and be it further
>> 
>> RESOLVED, that all responsibilities pertaining to the Apache Incubator
>> TVM  podling encumbered upon the Apache Incubator PMC are hereafter
>> Discharged.
>> 
>> - [1]
>> https://lists.apache.org/thread.html/r91b8f469c6a54769869bb2435b7334a28bcff885ae078ab5612dae00%40%3Cdev.tvm.apache.org%3E
>> - [2]
>> https://lists.apache.org/thread.html/rd5b8eefe49af09a2d0913758a5e5737b3fdb9072bc0becf4a2b2c7ee%40%3Cdev.tvm.apache.org%3E
>> - [3]
>> https://docs.google.com/document/d/18nyAH-fcptVezAxPQe6H3FeTKPRkujOp1tc1YRSPLok/edit?usp=sharing
>> 
> 

Re: [DISCUSS] Apache TVM Graduation

[Tianqi Chen]

Thanks Dave, Henry, Ming for the discussions so far.

Please allow me to try to dissect and summarize the
discussion points and answers so far:

D0: Release process
- The PPMC has a well document release process[1] and has proj

D1: Use of github and dev@
- We follow the principle “everything happens (also) happens on dev@”
- Important dev related github issues(with tag RFC, COMMUNITY, VOTE) are
synced to dev@
- Decisions and discussions still and will continue happening via the dev@
list,
  but we also understand that different people in the community may choose
to use other
  interface to interact, just like existing practice of other ASF projects’
usage of JIRA,
  knowing that things will be mirrored to dev@
- The use of github for community conversations, while still following

D1: Recognition of reviewer
- A reviewer is not an official role, and PPMC made it very clear via
explicit disclaimer, no vote is carried out.
- It is more like a social mechanism, rather than a formal status. There is
no power granted to a reviewer,
  except for putting the person’s name on a list.
- A committer can also choose to solicit reviews from anyone, everyone is
welcomed to review.
- Just like the (un-official) maturity model, the PPMC believes it provides
useful (optional)
  information for committers and PPMC, with a goal to help growth and
recognition of candidates early, rather than a gatekeeper.
- The idea of recognizing reviewers was discussed in IPMC during the
initial proposal[2], and voted on by the TVM community[3].

D2: Community
- A PPMC member strives to only nominate a person from another org. As a
result,
  we might see people who interact more with the broader community being
nominated early,
  and the order of nomination does not strictly follow the order of raw
stats(which we
  think are related but do not represent merit).
- All contributions are considered.
- We understand that different communities can set different calibrations
for merit[4].
  The PPMC has been doing so in an organic way via healthy Apache
discussions, recognizing
  potential candidates early, and welcoming contributors from diverse
backgrounds.
- Regardless of the procedural choices in the above. The PPMC actively
discusses,
  and brings in new committers from diverse backgrounds. The community
grows healthily
  and is quite vibrant. New committers are being brought up every month and
the list continues to grow.


There is no “one way” to the Apache Way[5]. IMHO, one of the best parts
about ASF is the trust
bestowed to PMC under the common ground of the Apache way. The PPMC does
not intend to suggest
any of the above applies to other communities. What is common though, is
that the PPMC will continue
to encourage open communications(everything also happens in dev@), welcome
new contributors from diverse backgrounds
collectively, and build a healthy, independent and diverse community under
the Apache Way.

Thank you!
-
- [1] https://tvm.apache.org/docs/contribute/release_process
- [2]
https://lists.apache.org/thread.html/e2b1fe9ca76422ec80b146a6b120091f2419e2f1c27d57080f39cf6f%40%3Cgeneral.incubator.apache.org%3E
- [3]
https://lists.apache.org/thread.html/c34b728f01d1030146594e47e0706cd1990ed731d06e3c179b7d501a%40%3Cdev.tvm.apache.org%3E
- [4] https://apache.org/foundation/governance/pmcs.html#merit
- [5] https://www.apache.org/theapacheway/

On Thu, Aug 27, 2020 at 4:42 PM Henry Saputra 
wrote:

> Thanks for the comments, Ming Wen.
>
> About the content of the list, it is not a problem and also not unusual on
> changes sent to dev@ list.
> Some Apache projects like MetaModel and during the early of Apache Flink
> also have Github change sent to dev@ list.
> Also some Apache projects have all JIRA updates send to dev@ list that
> make
> it also looks noisy. For example Apache Kylin and DLab.
> You can check them out, and probably more Apache projects have similar
> patterns of usage.
>
> If it is getting too noisy we could always move them to commits@ or other
> separate list. It is up to the community.
> But definitely not an issue or blocker.
>
> - Henry
>
> On Thu, Aug 27, 2020 at 4:34 PM Ming Wen  wrote:
>
> > > As per apache way, everything happens (also happens) on dev@tvm. We
> try
> > to
> > > set up the mechanism to enable maximum participation from the
> community,
> > > while complying with ASF rules :)
> > > The emails from the dev@tvm are also forwarded to the github thread.
> >
> > The mailing list seems to be a backup of github issues, even for
> something
> > as important as announcing a new committer [1] [2].
> > I am not sure if this is a problem, but I have not seen such an operation
> > in other apache projects.
> >
> > [1]
> >
> >
> https://lists.apache.org/thread.html/r17bdb633ec3ed1f8f80502705f9cc8cc642b189ab99cc2958847c1ac%40%3Cdev.tvm.apache.org%3E
> > [2] https://github.com/apache/incubator-tvm/pull/4636
> >
> > Thanks,
> > Ming Wen, Apache APISIX & Apache SkyWalking
> > Twitter: _WenMing
> >
> >
> >