Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Suneel Marthi
Downloads fine for me @ssc, I am on a horribly slow wifi connection on a Trans-Atlantic flight - it worked for me. On Sat, Oct 21, 2017 at 1:37 AM, Meghna Baijal wrote: > I verified again, on an ubuntu instance this time. Worked without the ‘g’ > > sha512sum --check ./apache-mxnet-src-0.12.0.rc

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Meghna Baijal
I verified again, on an ubuntu instance this time. Worked without the ‘g’ sha512sum --check ./apache-mxnet-src-0.12.0.rc0-incubating.tar.gz.sha512 On Fri, Oct 20, 2017 at 10:16 PM, Sebastian wrote: > Hmm, still doesn't work. Looks like the download of > https://dist.apache.org/repos/dist/dev/i

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Sebastian
Hmm, still doesn't work. Looks like the download of https://dist.apache.org/repos/dist/dev/incubator/mxnet/0.12.0.rc0 /apache-mxnet-src-0.12.0.rc0-incubating.tar.gz.sha512 has some problems on my machine, I can't view the file contents, maybe a problem with the content-type? On 21.10.2017 07:

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Meghna Baijal
Hi Sebastian, Can you try the following: gsha512sum --check ./apache-mxnet-src-0.12.0.rc0-incubating.tar.gz.sha512 This worked for me. Thanks, Meghna Baijal On Oct 20, 2017 9:53 PM, "Sebastian" wrote: Verifying the sha512 sig didn't work for me, what am I doing wrong here? sha512sum -c apache

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Sebastian
Verifying the sha512 sig didn't work for me, what am I doing wrong here? sha512sum -c apache-mxnet-src-0.12.0.rc0-incubating.tar.gz.sha512 sha512sum: apache-mxnet-src-0.12.0.rc0-incubating.tar.gz.sha512: no properly formatted SHA512 checksum lines found On 21.10.2017 02:36, Suneel Marthi wrote

Re: Build cancellations

2017-10-20 Thread Meghna Baijal
Hello All, Jenkins CI seems to be stable. If your build was terminated by someone over the last couple of days, you can trigger it now by pushing an empty commit to the same branch or manually in Jenkins if you have access. Thanks, Meghna Baijal On Thu, Oct 19, 2017 at 12:04 PM, Chris Olivier wr

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
Ok, just looking for anything that can cut a task out if possible. I do support not using Apache Jenkins server anyMore — it’s really not been working out for various reasons. But having a person full time is something that Steffen would have to address, I imagine. On Fri, Oct 20, 2017 at 6:03 PM

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Mu Li
I didn't see the clear advantage of CodePipline over pure jenkins, because we don't need to deploy here. On Fri, Oct 20, 2017 at 5:34 PM, Chris Olivier wrote: > CodePipeline, then. You can point it to Jenkins instances. > > > On Fri, Oct 20, 2017 at 4:49 PM Mu Li wrote: > > > AWS CodeBuild is

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Indhu
+1 On Fri, Oct 20, 2017, 5:01 PM Meghna Baijal wrote: > This is the vote to release Apache MXNet (incubating) version 0.12.0. > > Voting will start now (Friday, October 20, 2017 11:55PM UTC) and > > close Tuesday, October 24, 2017 11:55PM UTC. > > > Link to release notes: > > > https://cwiki.ap

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Suneel Marthi
+1 binding 1. Checked Sigs and hashes 2. has -incubating in artifact name On Fri, Oct 20, 2017 at 8:34 PM, Chris Olivier wrote: > +1 > > > On Fri, Oct 20, 2017 at 5:01 PM Meghna Baijal > wrote: > > > This is the vote to release Apache MXNet (incubating) version 0.12.0. > > > > Voting will sta

Re: [VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Chris Olivier
+1 On Fri, Oct 20, 2017 at 5:01 PM Meghna Baijal wrote: > This is the vote to release Apache MXNet (incubating) version 0.12.0. > > Voting will start now (Friday, October 20, 2017 11:55PM UTC) and > > close Tuesday, October 24, 2017 11:55PM UTC. > > > Link to release notes: > > > https://cwiki.

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
CodePipeline, then. You can point it to Jenkins instances. On Fri, Oct 20, 2017 at 4:49 PM Mu Li wrote: > AWS CodeBuild is not an option. It doesn't support GPU instances, mac os x, > and windows. Not even mention the edge devices. > > On Fri, Oct 20, 2017 at 4:07 PM, Chris Olivier > wrote: >

[VOTE] Release MXNet version 0.12.0.rc0

2017-10-20 Thread Meghna Baijal
This is the vote to release Apache MXNet (incubating) version 0.12.0. Voting will start now (Friday, October 20, 2017 11:55PM UTC) and close Tuesday, October 24, 2017 11:55PM UTC. Link to release notes: https://cwiki.apache.org/confluence/display/MXNET/MXNet+0.12.0+Release+Notes Link to rele

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Mu Li
AWS CodeBuild is not an option. It doesn't support GPU instances, mac os x, and windows. Not even mention the edge devices. On Fri, Oct 20, 2017 at 4:07 PM, Chris Olivier wrote: > Why don;t we look into fully managed AWS CodeBuild? It maintains > everything. It's also compatible with Jenkins. >

Re: New Apache MXNet logo idea

2017-10-20 Thread Seb Kiureghian
Sorry, posting to the correct thread: https://imgur.com/a/aADkA On Fri, Oct 20, 2017 at 4:06 PM, Chris Olivier wrote: > Can you resend the link? One below doesn't work for me. > > On Fri, Oct 20, 2017 at 1:02 PM, sandeep krishnamurthy < > sandeep.krishn...@gmail.com> wrote: > > > Looks nice. Rab

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
I believe that Mu already started that discussion about using old mxnet.io Jenkins server. I expect deciding whether to replace would hinge in large part upon what it would be replaced with. On Fri, Oct 20, 2017 at 4:30 PM, sandeep krishnamurthy < sandeep.krishn...@gmail.com> wrote: > Chris: If

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread sandeep krishnamurthy
Chris: If the community decides to go with separate setup, then there will be a tech design discussion and CodeCommit / Jenkins / Travis such proposals will be covered and discussed. Thanks, Sandeep On Fri, Oct 20, 2017 at 4:22 PM, Seb Kiureghian wrote: > But the feather can definitely be added

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Seb Kiureghian
But the feather can definitely be added once MXNet graduates. On Fri, Oct 20, 2017 at 4:21 PM, Seb Kiureghian wrote: > The feather can only be used by Top Level Projects. > > On Fri, Oct 20, 2017 at 4:19 PM, Chris Olivier > wrote: > >> When the word Apache is in the Hadoop logo (not always), it

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Seb Kiureghian
The feather can only be used by Top Level Projects. On Fri, Oct 20, 2017 at 4:19 PM, Chris Olivier wrote: > When the word Apache is in the Hadoop logo (not always), it includes the > feather and color scheme. > > On Fri, Oct 20, 2017 at 4:18 PM, Chris Olivier > wrote: > >> Thanks. >> >> Is ther

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
When the word Apache is in the Hadoop logo (not always), it includes the feather and color scheme. On Fri, Oct 20, 2017 at 4:18 PM, Chris Olivier wrote: > Thanks. > > Is there any way to work the feather into it? > > i.e. https://goo.gl/images/BU4dnG > > On Fri, Oct 20, 2017 at 4:11 PM, Seb Kiu

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
Thanks. Is there any way to work the feather into it? i.e. https://goo.gl/images/BU4dnG On Fri, Oct 20, 2017 at 4:11 PM, Seb Kiureghian wrote: > https://imgur.com/a/aADkA > > On Fri, Oct 20, 2017 at 4:07 PM, Chris Olivier > wrote: > > > Why don;t we look into fully managed AWS CodeBuild? It

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Seb Kiureghian
https://imgur.com/a/aADkA On Fri, Oct 20, 2017 at 4:07 PM, Chris Olivier wrote: > Why don;t we look into fully managed AWS CodeBuild? It maintains > everything. It's also compatible with Jenkins. > > On Fri, Oct 20, 2017 at 1:51 PM, Tianqi Chen > wrote: > > > +1 > > > > Tianqi > > On Fri, Oct

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Chris Olivier
Why don;t we look into fully managed AWS CodeBuild? It maintains everything. It's also compatible with Jenkins. On Fri, Oct 20, 2017 at 1:51 PM, Tianqi Chen wrote: > +1 > > Tianqi > On Fri, Oct 20, 2017 at 1:39 PM Mu Li wrote: > > > +1 > > > > > > It seems that the Apache CI is quite overloade

Re: New Apache MXNet logo idea

2017-10-20 Thread Chris Olivier
Can you resend the link? One below doesn't work for me. On Fri, Oct 20, 2017 at 1:02 PM, sandeep krishnamurthy < sandeep.krishn...@gmail.com> wrote: > Looks nice. Rabbit :-) > +1 > > > On Fri, Oct 20, 2017 at 11:02 AM, Seb Kiureghian > wrote: > > > Can we start a vote on the logo? > > > > __

Re: New Apache MXNet logo idea

2017-10-20 Thread Seb Kiureghian
I can create the same logo in white so it will fit with the /master website. Longer term, I think the site could use another redesign. The blue background makes it difficult to read text on the homepage. On Fri, Oct 20, 2017 at 2:03 PM, Mu Li wrote: > Clearly better than the previous logos we

Re: New Apache MXNet logo idea

2017-10-20 Thread Mu Li
Clearly better than the previous logos we had. One concern is that the logo color is inconsistent with the current doc theme: https://mxnet.incubator.apache.org/versions/master/index.html. We need to change either the logo or the doc theme. On Fri, Oct 20, 2017 at 1:02 PM, sandeep krishnamurthy <

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, and more tasks such

Re: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread Mu Li
+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, and more tasks such as pip build. It means a lot of requests to the Infra team. We can

Re: New Apache MXNet logo idea

2017-10-20 Thread sandeep krishnamurthy
Looks nice. Rabbit :-) +1 On Fri, Oct 20, 2017 at 11:02 AM, Seb Kiureghian wrote: > Can we start a vote on the logo? > > > From: Lupesko, Hagay > Sent: Monday, October 2, 2017 12:08:49 AM > To: dev@mxnet.incubator.apache.org; Madan Jampani; sebou...@gmail.com >

Fwd: [Proposal] Stabilizing Apache MXNet CI build system

2017-10-20 Thread sandeep krishnamurthy
Hello all, I am hereby opening up a discussion thread on how we can stabilize Apache MXNet CI build system. Problems: Recently, we have seen following issues with Apache MXNet CI build systems: 1. Apache Jenkins master is overloaded and we see issues like - unable to trigger bui

Re: New Apache MXNet logo idea

2017-10-20 Thread Seb Kiureghian
Can we start a vote on the logo? From: Lupesko, Hagay Sent: Monday, October 2, 2017 12:08:49 AM To: dev@mxnet.incubator.apache.org; Madan Jampani; sebou...@gmail.com Subject: Re: New Apache MXNet logo idea Like it! It’s cute, fun and playful. For me it also assoc

MXNet Slack channel

2017-10-20 Thread danny.stephan...@ieee.org
Please subscribe me to Mxnet slack channel. Thx

Re: Does mxnet NDArray have autodiff ?

2017-10-20 Thread YiZhi Liu
Currently, it hasn't. We'll add soon. 2017-10-20 17:43 GMT+08:00 TongKe Xue : > Hi! > > [I'm referring to the Scala API] > > 1. mxnet symbol has auto diff: > https://mxnet.incubator.apache.org/tutorials/r/symbol.html > > 2. from Googling around, I can't find NDArray auto diff anywhere > >

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
As with every change, there need to be steps. The python API of gluon is a first step toward moving API to the set of the new operator. I think exchange and compilation are next step it is relatively easy to do so without start implementing these ops (we canonicalize to core ops and convert them ba

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
The core ops do not mean implementation. They are merely a spec that we want to agree on what attributes we need. - The current list of the proposal spec is here: http://nnvm.tvmlang.org/top.html (I think eventually MXNet's document should have one such page) - They are is modeled after numpy and

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Chris Olivier
Thansk for that explanation! What is meant by the "core ops"? Are these the gluon python ops? I know what the "legacy ops" 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

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
The major concern is engineering overhead. The exchange formats are not serialization format like json. So we need to do the conversion back and force, and there are quite a lot of caveat there. The existing ones like onnx is also evolving very fast and there is no backward compatibility so far. I

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
First of all, I agree that supporting the format as an important step of adding influence. We are going to do it in either options. The overhead of engineering is not that much difference. I also do not argue for specific types of APIs(gluon vs symbolic) we should use. MXNet Sym API contains two

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Chris Olivier
I am not implementing it, but I prefer C++ for this reason. But then again, 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 wrote: > I do not have a preference in terms of specifically having to do thi

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
I do not have a preference in terms of specifically having to do this in particular language, and merely mean python would be an easier path. Most other languages users can still get onnx model to mxnet's format(with a script) and load it from there. Tianqi On Fri, Oct 20, 2017 at 8:06 AM, Chris

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Chris Olivier
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 wrote: > Given the instability nature of the current status of onnx, and the ease of > development, python would be favorable as the initial choice > > On Fri, Oct 20, 2017 at

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Tianqi Chen
Given the instability nature of the current status of onnx, and the ease of development, python would be favorable as the initial choice On Fri, Oct 20, 2017 at 8:00 AM, Chris Olivier wrote: > Also, what language is the expectation here? Is the ONNX serialization to > be developed in the C++ la

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Chris Olivier
Also, what language is the expectation here? Is the ONNX serialization to be developed in the C++ layer or python? Seems like what I saw before was python. On Fri, Oct 20, 2017 at 1:05 AM, Lupesko, Hagay wrote: > Tianqi, > > “I would want the code to be in the repo as long as we reach the > co

Does mxnet NDArray have autodiff ?

2017-10-20 Thread TongKe Xue
Hi! [I'm referring to the Scala API] 1. mxnet symbol has auto diff: https://mxnet.incubator.apache.org/tutorials/r/symbol.html 2. from Googling around, I can't find NDArray auto diff anywhere 3. however, NDArray tracks dependencies -- which, in theory, should be enough for doing autodif

Re: The Exchange Layer Support of MXNet

2017-10-20 Thread Lupesko, Hagay
Tianqi, “I would want the code to be in the repo as long as we reach the consensus.” +1 “The reason why I am seeing this decision so seriously is that it will affect how we can influence the design of the exchange format we act on” IMO, the most important first thing to do in order to influence