-1
We should stick with always using squash. It maintains PR reference in commit
history. It is also common practice.
If you want to included commits from multiple contributors in a single PR, open
a branch and make PRs to that branch. Then only use rebase when merging that
branch to master.
Neither TF nor Pytorch uses mailing lists though. In fact I can't think of any
deep learning community that uses mailing lists. Mailing list is an obsolete
legacy for old projects. No point in bringing it into new projects.
Thanks,
Eric
On 2018/06/18 18:42:12, Sebastian wrote:
> I am puzzled
g the situation.
>
> The CCache/EFS failure lasted for 12 hours and was an error - these things
> happen when you run a service. This is not a blame-game.
>
> -Marco
>
> [1]:
> https://github.com/apache/incubator-mxnet/issues?q=is%3Aopen+is%3Aissue+label%3AFlaky
>
> On Fri, Jun 1
Hi Marco de Abreu,
CI has been totally broken recently. It randomly fails for no good reason more
often than it passes. For example the ccache/efs failure has been really
annoying.
Looks like there has been many changes to Jenkins and Docker lately. Do you
think we should revert all of the
also one of them is
> > > working on
> > > > the Memory allocation issue in Scala not as a part of their day
> > job;
> > > This
> > > > is where I find value in managing the project features on JIRA,
> >
Hi,
Since all of MXNet's development happens on Github, I think it's sufficient to
use Github Issues and Github Projects for tracking. There are also many other
plugins you can add to Github if issues and projects are not enough.
It's very easy to cross reference PRs and issues for tracking.
I see several issues with the design. I've commented in the document but for
record here:
1. cpp-package is almost only used for inference. since you are planning a
rewrite that's almost certainly non-backward-compatible, we might as well
create a new interface that's inference only.
2. The
-1
If you make MXNet a submodule of keras, then you should PR that to keras.
If you want something like mxnet.keras, then you should do a full rewrite that
only keeps the keras interface.
On 2018/03/23 05:49:07, sandeep krishnamurthy
wrote:
> Hello MXNet
We want as few dependencies as possible.
CMake alone is enough trouble for our users. We don't want to burden them with
other stuff.
On 2018/03/06 17:21:15, kellen sunderland wrote:
> Short term solution sounds good to me Chris. Converting the CI should be
>
One potential problem is with libstdc++. Specifically with vectors. Our
operator interface uses vectors, which breaks when core lib and extension are
compiled with different compiler version
On 2018/03/06 22:45:16, Chris Olivier wrote:
> The static init of your module
-1
JIRA is ancient and arcane. This adds unnecessary overhead.
On 2018/03/03 06:11:12, CodingCat wrote:
> This vote passes with 6 +1 votes (6 bindings) and no 0 or -1 votes.
>
> Binding +1:
> Chris Olivier
> Indhu Bharathi
> Suneel Marthi
> Yuan Tang
> Marco de Abreu
>
Since committers voted for +1. We consider this vote passed.
Thanks,
Eric
On 2017-11-19 12:51, "Eric Xie"<j...@apache.org> wrote:
> Hi all,
> I'm starting this thread to vote on turning off protected master. The reasons
> are:
>
> 1. Since we turned on protect
I'm fine with a 3rdparty folder. Not sure about apache legal.
On 2017-11-17 10:25, Chris Olivier wrote:
> All,
>
> I often find it desirable to have a method for 3rdparty packages to be
> included (possibly optionally) in a 3rdparty directory. We do this with
> 'cub'
.com>
> > wrote:
> >
> > > +1
> > >
> > >
> > > On Sun, Nov 19, 2017 at 12:52 PM Zha, Sheng <zhash...@amazon.com> wrote:
> > >
> > >> +1
> > >>
> > >> Best regards,
> > >> -sz
> > >
.
>
>
>
>
> On Tue, Oct 31, 2017 at 1:17 PM, Eric Xie <j...@apache.org> wrote:
>
> > The number of pending PRs is growing very fast. At the current rate it
> > will reach 200 before we can fix jenkins.
> >
> > Stability is not the only goal. Master
The number of pending PRs is growing very fast. At the current rate it will
reach 200 before we can fix jenkins.
Stability is not the only goal. Master will be most stable if we don't push
anything, but that's not what we want.
Committers should be responsible for their commits. Good judgement
16 matches
Mail list logo