CANCELLED - WebML CG Teleconference – 17 October 2019 - 14:00-15:00 UTC+0

Hi All,

2019-10-17 teleconference is CANCELLED.

The topics planned for the meeting need a bit more time before we can have a productive group discussion around them.

Please use this time to make progress with the open GH issues, notably:

- conv2d issue #28 https://github.com/webmachinelearning/webnn/issues/28 would benefit from the group's input before spec PR landing
- explainer https://github.com/webmachinelearning/webnn/blob/master/explainer.md accepts contributions; I made a minor contribution https://github.com/webmachinelearning/webnn/pull/30 to start this effort -- the structure of the document is advisory and we can adjust is as we see fit

The agenda for the next scheduled meeting is being drafted at https://github.com/webmachinelearning/meetings/blob/master/telcons/2019-10-31-agenda.md and will be announced closer to the meeting.

Thanks,

-Anssi (WebML CG chair)

On 10 Oct 2019, at 11.48, Kostiainen, Anssi <anssi.kostiainen@intel.com<mailto:anssi.kostiainen@intel.com>> wrote:

https://github.com/webmachinelearning/meetings/blob/master/telcons/2019-10-17-agenda.md


WebML CG Teleconference – 17 October 2019 - 14:00-15:00 UTC+0

Logistics

* Chair: Anssi
* Scribe: ?
* IRC: irc://irc.w3.org:6667/#webmachinelearning

* IRC web client: http://irc.w3.org/?channels=#webmachinelearning

* Call-in details: https://lists.w3.org/Archives/Member/internal-webmachinelearning/2019Apr/0000.html

* Minutes: https://www.w3.org/2019/10/17-webmachinelearning-minutes.html


Agenda

* WebNN interop investigation

 * Review POC requirements for buffer sharing between dedicated ML hardware and GPU (Ningxin, Paul)

* On-device training (exploratory topic)

 * Note: currently out of scope, a future opportunity subject to group agreement
 * Discuss key use cases, feasibility, platform support (Core ML 3, ONNX 1.7 ...)

Thanks,

-Anssi (WebML CG chair)