mxnet-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dominic Divakaruni <dominic.divakar...@gmail.com>
Subject Re: Board report due
Date Wed, 05 Jul 2017 18:15:58 GMT
I've made some updated and also posted to this
https://docs.google.com/document/d/1PGhs96klZB6DXhpK9_biPh4-aCm8-bWwFzexnOW_GMA/edit?usp=sharing

1.     Your project name: Apache MXNet

2.     A brief description of your project, which assumes no knowledge of
the project or necessarily of its field

MXNet is an open-source deep learning framework that allows you to define,
train, and deploy deep neural networks on a wide array of devices, from
cloud infrastructure to mobile devices. It is highly scalable, allowing for
fast model training, and supports a flexible programming model and multiple
languages. MXNet allows you to mix symbolic and imperative programming
flavors to maximize both efficiency and productivity. MXNet is built on a
dynamic dependency scheduler that automatically parallelizes both symbolic
and imperative operations on the fly. A graph optimization layer on top of
that makes symbolic execution fast and memory efficient. The MXNet library
is portable and lightweight, and it scales to multiple GPUs and multiple
machines.

3.     A list of the three most important issues to address in the move
towards graduation.

3.1.  Migrate code (GitHub) and website to Apache.

3.2.  Grow the community:

3.2.1.     Expand reference material including – new machine learning
research published based on MXNet, tutorials, documented use cases and
architecture documentation.

3.2.2.     Improving user-experience –for example improved error messages

3.2.3.     Improved support for various programming languages

3.2.4.     Establish a dependable, Apache-way consistent release process.

3.3.  Features:

3.3.1.     Capability (such as low precision support and quantization) that
allows models to run efficiently on mobile and edge devices. Integrations
with mobile and edge device acceleration drivers.

3.3.2.     Accelerate performance on CPUs and GPUs.

4.     Any issues that the Incubator PMC or ASF Board might wish/need to be
aware of: None.

5.     How has the community developed since the last report.

5.1.  On 5/27 MXNet published a comprehensive edit and makeover of the
documentation including tutorials, how-to’s, APIs and architecture guides.
This was a broad effort that involved over 40 contributors.

5.2.  The PMC voted in a new contributor who has been helping with the code
migration and setup of the test infrastructure. We are making slow but
steady progress towards getting the GitHub code migrated. The target date
for migration is 7/17. Website migration will happen next.

5.3.  Slack and dev@ are being used more actively.

5.4.  Two presentations/workshops on Apache MXNet at the O’Reilly AI Conf
on 6/27 and 6/28

5.5.  A new blog post published on 6/23 showing users how to Build a
Real-time Object Classification System with ApacheMXNet on Raspberry Pi.
https://aws.amazon.com/blogs/ai/build-a-real-time-object-classification-system-with-apache-mxnet-on-raspberry-pi/



6.     How has the project developed since the last report.

6.1.  Since the last report 42 authors have pushed 326 commits to master.

6.2.  Documentation- Architecture guides, How To’s, Tutorials, and APIs
have been improved.

6.3.  More features (e.g. operators) requested by the user community has
been added.

6.4.   A new Perl language binding for MXNet was added.

7.     How does the podling rate their own maturity. Maturity = Low.


On Wed, Jul 5, 2017 at 7:24 AM, Suneel Marthi <smarthi@apache.org> wrote:

> Dom,
>
> Its much easier to comment/modify if u created a google doc and send a
> editable link out. Please do that.
>
> On Tue, Jul 4, 2017 at 6:06 PM, Dominic Divakaruni <
> dominic.divakaruni@gmail.com> wrote:
>
> > Hello all,
> > Hope those of us in the US are having a great 4th of July! I've taken a
> > stab at a draft of the report. Section 6 needs to be updated. Please
> pitch
> > in with your updates
> >
> > 1.     Your project name: Apache MXNet
> >
> >
> >
> > 2.     A brief description of your project, which assumes no knowledge of
> > the project or necessarily of its field: MXNet is an open-source deep
> > learning framework that allows you to define, train, and deploy deep
> neural
> > networks on a wide array of devices, from cloud infrastructure to mobile
> > devices. It is highly scalable, allowing for fast model training, and
> > supports a flexible programming model and multiple languages. MXNet
> allows
> > you to mix symbolic and imperative programming flavors to maximize both
> > efficiency and productivity. MXNet is built on a dynamic dependency
> > scheduler that automatically parallelizes both symbolic and imperative
> > operations on the fly. A graph optimization layer on top of that makes
> > symbolic execution fast and memory efficient. The MXNet library is
> portable
> > and lightweight, and it scales to multiple GPUs and multiple machines.
> >
> > 3.     A list of the three most important issues to address in the move
> > towards graduation.
> >
> > 3.1.  Migrate code (GitHub) and website to Apache.
> >
> > 3.2.  Grow the community:
> >
> > 3.2.1.     Expand reference material including – new machine learning
> > research published based on MXNet, tutorials, documented use cases and
> > architecture documentation.
> >
> > 3.2.2.     Improving user-experience –for example improved error messages
> >
> > 3.2.3.     Improved support for various programming languages
> >
> > 3.2.4.     Establish a dependable, Apache-way consistent release process.
> >
> > 3.3.  Features:
> >
> > 3.3.1.     Capability (such as low precision support and quantization)
> that
> > allows models to run efficiently on mobile and edge devices. Integrations
> > with mobile and edge device acceleration drivers.
> >
> > 3.3.2.     Accelerate performance on CPUs and GPUs.
> >
> >
> >
> > 4.     Any issues that the Incubator PMC or ASF Board might wish/need to
> be
> > aware of:
> >
> > None.
> >
> >
> >
> > 5.     How has the community developed since the last report.
> >
> > 5.1.  On 5/27 MXNet published a comprehensive edit and makeover of the
> > documentation including tutorials, how-to’s, APIs and architecture
> guides.
> > This was a broad effort that involved over 40 contributors.
> >
> > 5.2.  The PMC voted in a new contributor who has been helping with the
> code
> > migration and setup of the test infrastructure. We are making slow but
> > steady progress towards getting the GitHub code migrated. The target date
> > for migration is 7/17. Website migration will happen next.
> >
> > 5.3.  Slack and dev@ are being used more actively.
> >
> > 6.     How has the project developed since the last report.
> >
> > 6.1.  Since the last report 42 authors have pushed 326 commits to master.
> >
> > 6.2.  (Previous Update)  On master, 502 files have changed and there have
> > been 26,246 additions and 12,188 deletions. Count of Closed Issues = 62,
> > Count of new Issues = 146, Count of Merged Pull Requests = 161, Count of
> > Proposed Pull Requests = 27.
> >
> > 6.3.  (Previous Update) The API Documentation has improved.
> >
> > 6.4.  (Previous Update) More features (e.g. operators) requested by the
> > user community has been added.
> >
> > 6.5.  (Previous update) Hardware acceleration like cuDDN6 integration and
> > MKL ML package integration was completed.
> >
> > 6.6.  (Previous Update) A new Perl language binding for MXNet was added.
> >
> >
> >
> > 7.     How does the podling rate their own maturity: Maturity = Low.
> >
> >
> > On Mon, Jul 3, 2017 at 11:39 PM, Henri Yandell <bayard@apache.org>
> wrote:
> >
> > > In case the relentless automated pinging hasn't given it away, we've a
> > > board report due.
> > >
> > > Hen
> > >
> >
> >
> >
> > --
> >
> >
> > Dominic Divakaruni
> > 206.475.9200 Cell
> >
>



-- 


Dominic Divakaruni
206.475.9200 Cell

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message