Re: Java API for MXNet

2017-08-16 Thread Tianqi Chen
My two cents here: Start with something simple and share foundation with scala, focus on the API and usability. Once the java API get stabilized, we can then discuss if such issue arise. I feel maven is quite automatic in terms of handling scala deps and that won't create an issue except for

Re: Licensing update per code move

2017-07-06 Thread Tianqi Chen
w or agreed to in writing, software > * distributed under the License is distributed on an "AS IS" BASIS, > * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. > * See the License for the specific language governing permissions and > * limitations under th

Re: IMPORTANT: Link accounts to Apache today 7/12

2017-07-12 Thread Tianqi Chen
My apache id is tqchen: I also find MFA part is not checked.. Tianqi On Wed, Jul 12, 2017 at 8:16 PM, Indhu wrote: > Done. > > Apache user ID: indhub > > Thanks, > Indu > > On Wed, Jul 12, 2017 at 3:02 PM Ly Nguyen wrote: > > > As mentioned in

Re: Licensing update per code move

2017-07-06 Thread Tianqi Chen
I did notice that there is a shorter version of the header. I think we could use that ? Tianqi On Thu, Jul 6, 2017 at 8:58 AM Suneel Marthi wrote: > Yes its absolutely necessary that every file have a Apache license header - > and every project that comes into Apache does

Re: Formalize Committer Proposal and Application Procedure

2017-08-08 Thread Tianqi Chen
We totally understand the apache way of doing things and the projects has always been welcoming users as contributors from days of DMLC. This happens in MXNet, as well as other project, for example XGBoost ( https://github.com/dmlc/xgboost) , another project that originated from DMLC have most of

Re: Formalize Committer Proposal and Application Procedure

2017-08-04 Thread Tianqi Chen
> > Could you provide an example that provides a likely (imaginary if you'd > like) candidate? Mu's a pretty bad example for a new committer :) From the > attached doc I walk away thinking that I need to contribute for 2 years > before I can become a committer. For example, I think

Re: Formalize Committer Proposal and Application Procedure

2017-08-04 Thread Tianqi Chen
FYI here is the comitter checklist from Apache Mesos http://mesos.apache.org/documentation/latest/committer-candidate-checklist/ which I mainly adopted from Tianqi On Fri, Aug 4, 2017 at 8:14 AM, Madan Jampani wrote: > There is a middle ground here. Instead of saying

Re: Formalize Committer Proposal and Application Procedure

2017-08-04 Thread Tianqi Chen
My experience from the existing open-source project we have is that the developers are willing to contribute back as long as the software they use are hold up to a standard. I do not meant to say that the contributions of the language, documentations and others do not count as contributions to

Re: Announcement of DMLC/TVM Our Deep Learning Compilation Stack

2017-08-17 Thread Tianqi Chen
t hardware. Thanks Tianqi and the > > efforts of all the contributors. > > > > On Thu, Aug 17, 2017 at 3:41 PM, Tianqi Chen <tqc...@cs.washington.edu> > > wrote: > > > > > Hi Guys: > > >I am super excited to announce DMLC/TVM our deep lear

Re: Announcement of DMLC/TVM Our Deep Learning Compilation Stack

2017-08-17 Thread Tianqi Chen
all the contributors. > > > > On Thu, Aug 17, 2017 at 3:41 PM, Tianqi Chen <tqc...@cs.washington.edu> > > wrote: > > > > > Hi Guys: > > >I am super excited to announce DMLC/TVM our deep learning > compilation > > > stack. There

Re: Blog

2017-05-21 Thread Tianqi Chen
We might want to still keep dmlc blog, as it also contains announcement for other dmlc projects, but move the mxnet related announcements to new blog location. Tianqi On Sat, May 20, 2017 at 10:38 PM, Mu Li wrote: > Currently, we have one at http://dmlc.ml/. We can transfer

Re: Blog

2017-05-21 Thread Tianqi Chen
method by which drafts are > written, mailing list informed, and folk on the list can offer editorial > feedback. > > If you ever need someone to test blogs for newbies, I can definitely help > there :) > > Hen > > > On Sun, May 21, 2017 at 9:54 AM, Tianqi Chen <tqc...

Re: MXNet: Run PR builds on Apache Jenkins only after the commit is reviewed

2017-09-11 Thread Tianqi Chen
I agree that have the CI is useful, at least make sure that lint stage is done. Tianqi On Mon, Sep 11, 2017 at 7:49 PM, Hagay Lupesko wrote: > Build success or failure is a great feedback mechanism, of equal importance > to code review. Do we really want to delay it until

Re: Improving and rationalizing unit tests

2017-10-16 Thread Tianqi Chen
I would be great if there is a chance of a few testcase to reflect these principles, so we have a concrete discussion basis. Having seeded random number is good, but usually it is not the cause of non deterministic error( most of which already resolved by having a relaxed tolerance level).

Re: What's everyone working on?

2017-09-25 Thread Tianqi Chen
I am primarily working on deep learning compilation https://github.com/dmlc/tvm and hopefully you will hear the related updates in MXNet soon :) Tianqi On Mon, Sep 25, 2017 at 12:45 PM, Joern Kottmann wrote: > Hello all, > > I am working on the Java API and frequently

Announcement of DMLC/TVM Our Deep Learning Compilation Stack

2017-08-17 Thread Tianqi Chen
Hi Guys: I am super excited to announce DMLC/TVM our deep learning compilation stack. There will be followups on mxnet to add the official support soon. To check what it is, see the announcement http://tvmlang.org/2017/08/17/tvm-release-announcement.html Tianqi

NNVM Compiler

2017-10-09 Thread Tianqi Chen
Hi MXNet Community: As you may have heard, we released nnvm compiler which brings MXNet and other deep learning frameworks to bare metal http://www.tvmlang.org/2017/10/06/nnvm-compiler-announcement.html This opens a new page of deep learning on a various amount of hardwares. While being an

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
operators do not suffer from legacy issues and strictly follows convention of numpy and Gluon. - In that sense, actually nnvm/top's symbolic API is cleaner and more stable, and is the final form we want to migrate into. Tianqi > On 10/18/17, 14:13, "Tianqi Chen" <workc...@gmail.com

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
il.com> wrote: > My $0.02: > > NNVM is not currently an Apache module. It’s under dmlc umbrella, whose > direction and governance is unclear. For this reason, I am inclined to > support new effort being places in Apache MXNet > > > -Chris > > On Wed, O

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
onstruct the graph, go back up to MXNet to expose it as a Module. This > seems to complex and does not add any benefit. In whatever way we construct > the MXNet Module object, NNVM will always be the underlying in memory IR > that is being executed, so why not take the simpler route? > >

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
around a group of dmlc > modules? > > > On Wed, Oct 18, 2017 at 6:58 PM Tianqi Chen <tqc...@cs.washington.edu> > wrote: > > > To better answer Hagay's question, I would like to dive down a bit deeper > > on the relation between MXNet, NNVM and model exchange format like

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
olic), but to do it as an encapsulated implementation detail, so > the implementation can be migrated to NNVM or another implementation in the > future, if at that point it seems like the right thing to do. > > Interested in hearing other opinions though… > > Haga

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
> consume or export an ONNX format. How would that work with NNVM? Not sure > users care about the implementation, as long as it doesn’t add another > layer of complexity to the workflow. > > Regards, > Dom > > > > On Oct 18, 2017, at 6:29 PM, Tianqi Chen <tqc...@c

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
down to a bare metal module. And this is the solution nnvm/top compiler pipeline offer, which I believe will become a standard practice of deployment and where all systems go to Tianqi On Wed, Oct 18, 2017 at 10:03 PM, Tianqi Chen <tqc...@cs.washington.edu> wrote: > OK, ther

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-18 Thread Tianqi Chen
I am strongly recommending going through the nnvm/top. One major reason in here is that the support of nnvm/top layer NOT ONLY mean compatibility of model format with onnx. These are the major benefits: - More hardware backends to mxnet, including opencl, metal, Raspberry Pi, web browser. These

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
onnx 0.1 as Roshani mentioned. Which is contributed by Zhi Zhang, another Apache MXNet committer. Zhi already provided code to alleviate this process. Built code on the existing effort would actually make the problem easier. On Thu, Oct 19, 2017 at 11:55 AM, Tianqi Chen <tqc...@cs.washingto

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
just move all of these dmlc modules into the Apache repository > right now and have the correct discussions on dev? What's the argument > against this? IMHO, I thought that's what was going to be done originally. > > On Thu, Oct 19, 2017 at 12:14 PM, Tianqi Chen <tqc...@cs.washingto

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
is is a good first step that hopefully allows MXNet users to > easily leverage ONNX, but still leave a clear path to update the > implementation later on if it makes sense. > > How do we move forward with a decision? > > On 10/19/17, 12:14, "Tianqi Chen" <workc...@

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
etter way of doing this' > > > > It may just be misinterpretation or misunderstanding amongst folks here > due > > to language barrier. > > > > > > On Thu, Oct 19, 2017 at 3:48 PM, Tianqi Chen <tqc...@cs.washington.edu> > > wrote: > > &

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
codebase and moves on without being tied down by the decisions of a > non-compatible community. > > Hen > > > > On Thu, Oct 19, 2017 at 11:59 AM, Tianqi Chen <tqc...@cs.washington.edu> > wrote: > > > Here are the detailed points(sorry for resenting it over again

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
erimental, with whatever way Roshani thinks makes > sense. > > Once there is code in place, and MXNet users and contributors are > able to > > check it out, we can consider future steps. > > > > Does this proposal make sense to folks? > &

Re: CODEOWNERS file being removed

2017-11-29 Thread Tianqi Chen
Codeowner feature is used to automatically trigger the review request for the people. All the committers are already listed as global owners, which is consistent with Apache's policy Tianqi On Wed, Nov 29, 2017 at 1:37 PM, Hen wrote: > Was there more discussion than Justin's

Re: Futex

2017-11-23 Thread Tianqi Chen
If they are not bottleneck. Then resorting to standard library solution is always preferred Tianqi On Thu, Nov 23, 2017 at 11:01 AM, Chris Olivier wrote: > Was doing some timing with futexes (we used them a lot in a previous life > in database engines) and they're

Re: 1.0 release problems and TVM

2017-11-28 Thread Tianqi Chen
As far as i know Mxnet is not using tvm for now. So you can simply not ship that jar Tianqi On Tue, Nov 28, 2017 at 4:38 PM Chris Olivier wrote: > tvm has a jar in it that (along with other things) is blocking 1,0 release. > Can someone on tvm team please address this? >

Re: 1.0 release problems and TVM

2017-11-28 Thread Tianqi Chen
I mean Mxnet is not using tvm as a direct dependency. The support is being done in nnvm compiler, which is not part of the release Tianqi On Tue, Nov 28, 2017 at 4:39 PM Tianqi Chen <tqc...@cs.washington.edu> wrote: > As far as i know Mxnet is not using tvm for now. So you can simply n

Re: 1.0 release problems and TVM

2017-11-28 Thread Tianqi Chen
ating/ps-lite/tests/test_simple_app.cc > 19. apache-mxnet-src-1.0.0.rc0-incubating/dmlc-core/tracker/yarn > /src/main/java/org/apache/hadoop/yarn/dmlc/ApplicationMaster.java > 20. ./apache-mxnet-src-1.0.0.rc0-incubating/nnvm/tvm/apps/androi > d_rpc/gradle/wrapper/gradle-wrapper.jar

Re: 1.0 release problems and TVM

2017-11-28 Thread Tianqi Chen
t; which has TVM related commits. > > So one of the solution could be to go back in nnvm which doesn't include > TVM. > > On Tue, Nov 28, 2017 at 4:41 PM, Tianqi Chen <tqc...@cs.washington.edu> > wrote: > > > I mean Mxnet is not using tvm as a direct dependency

Re: AWS contributing ONNX-MXNet

2017-11-17 Thread Tianqi Chen
I have watched the issue for around two days. Here are my two cents. First of all, there is no legal constraint to enforce you do anything, but as you said(which I fully agree on), we need to assume others have best intentions and give goodwill - It is great to reuse code, that is what

Re: [RFQ] Deprecate amalgamation

2017-11-20 Thread Tianqi Chen
We could resort to a middle ground. Instead of having an amalgamation script that generates a single file, simply have a file that includes everything and compiles that one. Which should also work. The javascript port can likely be superseded with some form of support in nnvm compiler, which

Re: Protected master needs to be turned off

2017-11-20 Thread Tianqi Chen
+1 until new CI is implemented. Tianqi On Mon, Nov 20, 2017 at 11:11 AM, Eric Xie wrote: > A lot of people seems to be confused, so let's clarify the separation of > roles/responsibilities: > > 1. The committers that merge code are responsible for code quality and > tests

Re: [RFQ] Deprecate amalgamation

2017-11-21 Thread Tianqi Chen
s an embedded library that can't use the shared > library + headers or specific language bindings? > > Can't we call emscripten from CMake? I'm not familiar with our JS > bindings, but I don't see why we can't compile for emscripten as for > any other platform. > > Pedr

Re: MXNet External Sources (was Re: [VOTE] Apache MXNet (incubating) 1.0.0 release RC1)

2017-12-02 Thread Tianqi Chen
Many of them are stand-alone project on their own that MXNet happens to depend on. Major open source projects other than MXNet depend on these projects(e.g. XGBoost depend on dmlc-core). So, for now, they can be viewed as external dependencies like boost. DMLC have talked about migrating more

Re: Protected master needs to be turned off

2017-12-01 Thread Tianqi Chen
I think we are still using CI to catch bugs. And we should take caution when merging something that CI did not catch up due to the response time. This do not contradict what comitter can do with their best judgement. For example, I would normally switch CI off and merge simple typo fixes. If the

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Tianqi Chen
+1 Tianqi On Fri, Oct 20, 2017 at 1:39 PM Mu Li wrote: > +1 > > > It seems that the Apache CI is quite overloaded these days, and MXNet's CI > pipeline is too complex to run there. In addition, we may need to add more > devices, e.g. macpro and rasbperry pi, into the server,

Re: Accidental commits to 1.2.0 branch using Maven apache-release profile

2018-06-09 Thread Tianqi Chen
You should never force push to master branch because it messes up people's commit history. Ideally, you shall never push to the release branch and PR is preferred. Given no people will continue work on the release branch, the impact of squash commits is minimalized and is fine in this case, given

Re: Accidental commits to 1.2.0 branch using Maven apache-release profile

2018-06-09 Thread Tianqi Chen
This would only happen if somebody pushed some changes on these new commits pushed 1.2.0 branch. That is if I checked out 1.2.0 before the change is pushed, and rebase against the code after squash, it will be fine. The problem will only happen if someone checked out 1.2.0 after these commits get

Re: Accidental commits to 1.2.0 branch using Maven apache-release profile

2018-06-09 Thread Tianqi Chen
One potential problem is that, if a release is tagged to a certain hash code, and that specific commit get squashed, the previous tagged release is no longer valid, so at least make sure that did not happen Tianqi On Sat, Jun 9, 2018 at 1:23 PM, Tianqi Chen wrote: > You should never force p

Re: Vote to stop using JIRA

2018-06-08 Thread Tianqi Chen
I would suggest we have a separate discussion issue about transparency. First of all, I agree that transparency is important. This can be achieved by public roadmaps, RFCs, that do not have particularly tie to JIRA or github issues. Having a clear guideline on that would be great, and it is great

Re: ICLA?

2018-06-07 Thread Tianqi Chen
ers and signed ICLA, just hadn't back in > September 2017. > > Hen > > > On Wed, Jun 6, 2018 at 10:14 PM, Tianqi Chen > wrote: > > > Maybe I was wrong, The existing code in the DMLC repo is merged by > original > > committers(who are now Apache MXNet comi

Re: users@mxnet

2018-06-18 Thread Tianqi Chen
t; wrote: > >>> > >>> users@ mailing lists have great societal advantages that one > >>> shouldn't ignore... > >>> > >>> And it's not like this is the only project with "multiple" > >&g

Re: users@mxnet

2018-06-15 Thread Tianqi Chen
tle as our community? I really hope we are > thinking a bit beyond that... > > On Fri, Jun 15, 2018, 21:22 Tianqi Chen wrote: > > > I remember last time during the mxnet meetup in Seattle, we did a survey, > > and most users preferred the current discuss forum

Re: users@mxnet

2018-06-15 Thread Tianqi Chen
ferent feelings about this. Should I cast a VOTE? > > > On Fri, Jun 15, 2018, 21:27 Tianqi Chen wrote: > > > I do think we are targeting all the community, but we must also agree > that > > the voice of users from the meetup is a representative sample of users'

Re: users@mxnet

2018-06-15 Thread Tianqi Chen
I don't want to argue here, as "Apache way" also says VOTE should not be a way to enforce our opinion, and consensus need to be reached through discussion Thanks! Tianqi On Fri, Jun 15, 2018 at 9:41 PM, Tianqi Chen wrote: > Then who should represent the users who are usi

Re: users@mxnet

2018-06-15 Thread Tianqi Chen
I remember last time during the mxnet meetup in Seattle, we did a survey, and most users preferred the current discuss forum. So I would say we stick with that given the user community prefers that Tianqi On Fri, Jun 15, 2018 at 9:13 PM, Sergio Fernández wrote: > Then, if everybody agree,

Re: users@mxnet

2018-06-15 Thread Tianqi Chen
beyond that... > > On Fri, Jun 15, 2018, 21:22 Tianqi Chen wrote: > > > I remember last time during the mxnet meetup in Seattle, we did a survey, > > and most users preferred the current discuss forum. So I would say we > stick > > with that given the user community prefers tha

Re: users@mxnet

2018-06-18 Thread Tianqi Chen
quot;strong evidence" that this is NOT the case, I would > recommend we create the list. > > > On Jun 16, 2018, at 12:28 AM, Tianqi Chen > wrote: > > > > So unless there is a strong evidence that our community users prefers the > > mail-list, I would recommend we keep the current

Re: About Becoming a Committer

2018-06-15 Thread Tianqi Chen
First of all, Apache allows each community to define its standard of comittership -- which I think is super valuable as different projects have different backgrounds and challenges. Having such flexibility instead of enforcing a global standard is one of the reasons why many Apache projects

Re: Reverting pull request

2018-06-15 Thread Tianqi Chen
+1 We would be stuck at local minimums if we just keep reverting the PR that brings improvements in the long term Tianqi On Fri, Jun 15, 2018 at 2:15 PM, Mu Li wrote: > Why reverting instead of fixing the bugs? Static memory aims to reduce > memory allocation, it's a key feature to bridge

Re: Reverting pull request

2018-06-15 Thread Tianqi Chen
been tested by many > > people. > > Other people's work now depends on this PR. Reverting it at this point > can > > cause a lot of problems for many other people. > > > > Best, > > Da > > > > On 6/15/18, 2:18 PM, "workc...@gmail.c

Re: Make cmake default

2018-06-01 Thread Tianqi Chen
In light of this, maybe it is time to deprecate amalgamation? GIven there is already support for Pi, and we could use TVM to compile to javascript with WebGL Tianqi On Fri, Jun 1, 2018 at 3:24 PM, Rahul Huilgol wrote: > +1 > > Let's move to CMake. It has much better support, and it's not worth

Re: ICLA?

2018-06-06 Thread Tianqi Chen
Maybe I was wrong, The existing code in the DMLC repo is merged by original committers(who are now Apache MXNet comitters and signed CLA). So I assume there is no such problem. Similar to the case that only comitter to the Apache MXNet repo has to sign CLA Tianqi On Wed, Jun 6, 2018 at 6:04 PM,

Re: [LAZY VOTE] Test coverage of PRs

2018-06-20 Thread Tianqi Chen
While I think test coverage is a nice information to have. I would object to using this as a metric to decide whether a PR should be merged. Code-cov act as a mere coverage of APIs, which is a useful aspect, it can be misleading in many cases, especially when such change involves cross-language

Re: DMLC [Was: Request for suggestions- Supporting onnx in mxnet]

2017-10-19 Thread Tianqi Chen
gt; Or probably it makes people feel better we suggest to do ONNX <-> > > MXNet/Gluon <-> MXNet/Operator, instead of ONNX <-> MXNet/Operator. > > > > The comment "Something is rotten." doesn't make any sense to me. > > > > Best > > Mu >

The Exchange Layer Support of MXNet

2017-10-19 Thread Tianqi Chen
I will start forking the previous discussion and it has gone awry and I hope to start a pure technical discussion thread. I said in another email that we could do a vote to settle this issue. I now think that I was wrong and would like to apology for my rush proposal on this. I hope to reopen

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
se. > Once there is code in place, and MXNet users and contributors are able to > check it out, we can consider future steps. > > Does this proposal make sense to folks? > > On 10/18/17, 23:01, "Tianqi Chen" <workc...@gmail.com on behalf of > tqc...@cs.washington.e

Re: Request for suggestions- Supporting onnx in mxnet

2017-10-19 Thread Tianqi Chen
w can we export Gluon model to ONNX? > > Hagay > > On 10/19/17, 13:26, "Tianqi Chen" <workc...@gmail.com on behalf of > tqc...@cs.washington.edu> wrote: > > Again my recommendation is to go through mxnet/gluon (which in that > case > core operator se

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
mean by “offers compilation”? And since MXNet Sym is built on > > top of NNVM, why will we not have all of that if we go down the route of > > implementing the conversion on top of MXNet Sym? > > > > Hagay > > > > On 10/19/17, 20:43, "Tianqi Chen" <w

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
t; are -- things like BatchNormProp/BatchNormOp. WHat > are the "new ones"? > > I apologize for not knowing this, but what are some of the "pain points" > with the legacy ops? Not in great detail, but would just like to get a > little perspective. > >

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
back to the ops we have). The final step is to move the implementation toward the new ops, which can happen gradually. Tianqi On Fri, Oct 20, 2017 at 8:58 AM, Tianqi Chen <tqc...@cs.washington.edu> wrote: > The core ops do not mean implementation. They are merely a spec that we > w

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
I am biased towards C++ in general :) > Are there reasons to not do it in C++? Maybe there are... > > On Fri, Oct 20, 2017 at 8:08 AM, Tianqi Chen <tqc...@cs.washington.edu> > wrote: > > > I do not have a preference in terms of specifically having to do this in > >

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
Olivier <cjolivie...@gmail.com> wrote: > How would this impact use by other languages such as Scala, Perl, C++, etc? > > On Fri, Oct 20, 2017 at 8:01 AM, Tianqi Chen <tqc...@cs.washington.edu> > wrote: > > > Given the instability nature of the cur

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
great advantage of a clear core set of operators. To facilitate discussion technically. I would suggest we break our points down. I tried to do that in this email on things we agree on and have opions. So we can find common ground and move forward. Tianqi > On 10/19/17, 20:43, "Tianqi C

Re: Suggestions on how to increase community involvement on Apache MXNet incubating?

2018-01-04 Thread Tianqi Chen
I think the main point being raised is an easy way to know what is happening and proposals. Github has quite a lot of nice features with respect to this: - Issues make perfect sense in terms of discussing a single-point of feature/bug and get feedback from anyone who can comment, additionally

Re: Increase indentation limit from 100 to 120 characters

2018-01-05 Thread Tianqi Chen
An argument against such change would be the coding style standard is people already get used to it, and there is less benefit of making the change. PEP and Google C style suggest 80 chars as limit, I usually write with that in mind and try to break multiple arguments into multiple lines when

Re: Increase indentation limit from 100 to 120 characters

2018-01-08 Thread Tianqi Chen
gt; > >On Sat, Jan 6, 2018 at 12:00 PM, kellen sunderland < > > >kellen.sunderl...@gmail.com> wrote: > > > > > >> Just a note that I don't think Pedro was suggesting the change for > > Python > > >> or Scala. How would folks fe

Re: Increase indentation limit from 100 to 120 characters

2018-01-08 Thread Tianqi Chen
8cb1cdc/tests/cpp/operator/batchnorm_test.cc#L963 > > > > On Sat, Jan 6, 2018 at 12:00 PM, kellen sunderland < > > kellen.sunderl...@gmail.com> wrote: > > > > > Just a note that I don't think Pedro was suggesting the change for > Python > > > or S

Re: Reduce 99% of your memory leaks with this simple trick!

2018-01-15 Thread Tianqi Chen
rBlock for a > >> > single operator, > >> > > >> > We store the exception_ptr in the operator. We then propagate the > >> > exception_ptr down to all the vars that the > >> > > >> > Operator writes to. Similarly, if an operator

Re: dmlc packages into 3rdpary

2018-01-19 Thread Tianqi Chen
I think it is fine either way as it won’t affect the build status of the projects. Tianqi On Fri, Jan 19, 2018 at 11:43 AM Chris Olivier wrote: > Is the general consensus to move the dmlc packages into 3rdparty? > > If so, I can submit a PR that does this. > > I have no

Re: [VOTE] Release MXNet version 1.1.0.RC1

2018-02-10 Thread Tianqi Chen
+1 tianqi On Fri, Feb 9, 2018 at 11:11 PM YiZhi Liu wrote: > Hi everyone, > > As we have updated the LICENSE file which caused the failure vote for > 1.1.0.RC0, I'd like to propose a vote to release Apache MXNet > (incubating) version 1.1.0.RC1. Voting will start now

Re: A question about fusion of TVM and MXNet/mshadow

2018-08-08 Thread Tianqi Chen
It is true that most of the current GPU code depends on mshadow. Porting the operator code entirely over to TVM will take quite a huge effort. So a more gradual path forward is to could be drop-in TVM to support cases that it optimizes well(ARM, AMDGPU, accelerators) while keeping the old

Re: Help with understanding docs

2018-08-08 Thread Tianqi Chen
As far as I understand, these are packages that are build on top of gluon and runs on MXNet, it is up to the maintainers of these packages as well the the comitters to decide whether they want to put it into MXNet master. I personally feel it is great to see ecosystem build and content on top of

Re: [VOTE] Subscribe dev@ to Github Activities

2018-07-17 Thread Tianqi Chen
+1, most of issue and PR activities are about development, and they belong to dev. It also helps us to recognizes contributors who are actively contributing but less vocal via emails -- there are many of them. Tianqi On Tue, Jul 17, 2018 at 8:47 AM, Anirudh wrote: > -1 > > The low signal to

Re: [VOTE] Subscribe dev@ to Github Activities

2018-07-18 Thread Tianqi Chen
that valuable discussion does not happen on Github. The > point is that mails about it will be dwarfed by the other activity on > Github. > > On Wed, Jul 18, 2018 at 10:30 AM, Tianqi Chen > wrote: > > > Being Apache is about being inclusive to the new contributors. Ap

Re: [VOTE] Subscribe dev@ to Github Activities

2018-07-18 Thread Tianqi Chen
Being Apache is about being inclusive to the new contributors. Apache encourages the use of Github, and currently, the community is doing so. I don't think it is a good idea to use political terms to force proliferate our contributors --- we are all Apache contributors. Instead, we should make

Re: [VOTE] Subscribe dev@ to Github Activities

2018-07-18 Thread Tianqi Chen
Discussions happened on Github are highly valuable, as a matter of fact, we have quite a lot of proliferating contributors who discuss things on GitHub when they contribute. We need to be inclusive to these contributors, to welcome and recognize these discussions. The filtering solutions seem to

Re: [DISCUSS] Build OSX builds in CI (possibly with TravisCI).

2018-09-05 Thread Tianqi Chen
ave any particular > > work item that I can help. > > > > Best, > > > > Lin > > > > On Wed, Sep 5, 2018 at 9:51 AM Tianqi Chen > > wrote: > > > > > is it possible for us to get a MacBook and hook it to the current > Jenkins > &g

Re: [DISCUSS] Build OSX builds in CI (possibly with TravisCI).

2018-09-05 Thread Tianqi Chen
is it possible for us to get a MacBook and hook it to the current Jenkins CI? Travis OSX usually build from scratch and that was pretty slow Tianqi On Wed, Sep 5, 2018 at 8:49 AM kellen sunderland < kellen.sunderl...@gmail.com> wrote: > Great you feel that way Lin, please feel free to

Re: [DISCUSS] Build OSX builds in CI (possibly with TravisCI).

2018-09-05 Thread Tianqi Chen
and it seems to work quite well. See for example this build: > https://travis-ci.org/KellenSunderland/incubator-mxnet/builds/424768656 > > On Wed, Sep 5, 2018 at 7:10 PM Tianqi Chen > wrote: > > > Travis it self is stateless, which means ccache is not likely going to > > w

Re: FW: Success at Apache: The Apache Way for Executives

2018-07-10 Thread Tianqi Chen
Totally agree with what being said here, as community strives to move forward it is important to be inclusive and communicative. The same principle also applies beyond this mail-list, as we also need be inclusive and welcoming to contributors who contribute via github, write issues and use

Re: [DISCUSS] Subscribe dev@ to Github Activities?

2018-07-12 Thread Tianqi Chen
+1 On Thu, Jul 12, 2018 at 11:10 AM, Sheng Zha wrote: > Hi all, > > Should we subscribe dev list to github updates on mxnet repo? Both github > issues/PRs and the dev list are intended for technical discussions and in > that aspect largely share the same goal. Since MXNet has most activity >

Re: Deprecate python 2

2018-07-12 Thread Tianqi Chen
I think we just need to be on side with the community. Most community plans to deprecate python2 at beginning of 2019 and drop support at 2020 Tianqi On Thu, Jul 12, 2018 at 2:51 PM, Pedro Larroy wrote: > Hi > > I would like to know your opinion in regards to deprecating and removing > Python

Re: Deprecate python 2

2018-07-12 Thread Tianqi Chen
FYI https://python3statement.org/ On Thu, Jul 12, 2018 at 3:00 PM, Marco de Abreu < marco.g.ab...@googlemail.com.invalid> wrote: > CentOS 7, for example, does not offer stable Python 3 support. We're using > an unstable version in our CI to verify it none the less. I think that's a > hard stop

Re: Cython Document iteration

2018-03-09 Thread Tianqi Chen
The cython part was not actively maintained and currently disabled, but the code structure can likely be reused Tianqi On Fri, Mar 9, 2018 at 10:31 AM, Xingjian SHI wrote: > Great! Also, what’s the current status of Cython support in MXNet? I think > we have the option

Re: [VOTE] Disconnect all non-C API's from mxnet versioning

2018-03-12 Thread Tianqi Chen
+1 On Mon, Mar 12, 2018 at 9:32 AM, Chris Olivier wrote: > It has been proposed that all Non-C API's follow separate versioning from > the main mxnet C API/releases. > > A +1 vote is in *favor of* using a different versioning for all > non-C-API's, with each API (Scala,

Re: [RESULT][VOTE] tracking code changes with JIRA by associating pull requests

2018-03-08 Thread Tianqi Chen
To summarize: - JIRA might be more maintainer friendly(have more features for tracking progress etc.) - Github issues are more community friendly (it has been the way most people contribute, and have lower bar of entry). Tianqi On Thu, Mar 8, 2018 at 11:05 AM, Xingjian SHI

Re: Cython Document iteration

2018-03-14 Thread Tianqi Chen
, but still can use ctypes to implement most things, and even support cases when cython is not available. Tianqi On Wed, Mar 14, 2018 at 8:38 AM, Tianqi Chen <tqc...@cs.washington.edu> wrote: > Is the complexity of adding the Cython code generation and compilation > step considered? Will it

Re: Cython Document iteration

2018-03-14 Thread Tianqi Chen
Is the complexity of adding the Cython code generation and compilation step considered? Will it work in all platforms? The complexity in terms of compiler support for pybind11 is same as cython. If the users can build pybind, then likely cython works(because cython compiles to c that have same

Re: Cython Document iteration

2018-03-14 Thread Tianqi Chen
Let me share one anecdote: Minerva, one of the predecessor of MxNet uses cython to wrap C++ API, then we decided to switch to C API in MxNet, partly due to some pains expressed by the MXNet members in Minerva team. Tianqi On Wed, Mar 14, 2018 at 8:40 AM, Tianqi Chen <tqc...@cs.washington.

Re: Meetup in Seattle

2018-04-02 Thread Tianqi Chen
I can do a quick talk on thoughts about new accelerator support and other new directions in deep learning system. Thanks Tianqi On Sat, Mar 31, 2018 at 3:52 PM, Steffen Rochel wrote: > Reminder to sign up for our meetup on 4/24. Please let me know if you are >

Re: MXNet C++ package improvements

2018-03-23 Thread Tianqi Chen
api adds more value than a bigger one On Fri, Mar 23, 2018 at 12:11 PM Tianqi Chen <tqc...@cs.washington.edu> wrote: > +1 for hour-glass CAPI design. It is there for very good reason. For > example, if you are windows user an want to build your app using mingw, > current way works.

  1   2   >