WebML CG Teleconference – 16 April 2020 - 15:00-16:00 UTC+0

https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md


WebML CG Teleconference – 16 April 2020 - 15:00-16:00 UTC+0
San Francisco (U.S.A. - California)     Thu, 16 April 2020      07:00 PST       UTC-7 hours
Boston (U.S.A. - Massachusetts) Thu, 16 April 2020      10:00 EST       UTC-4 hours
London (United Kingdom - England)       Thu, 16 April 2020      15:00 GMT       UTC+0 hours
Berlin (Germany)        Thu, 16 April 2020      16:00 CET       UTC+1 hours
Helsinki (Finland)      Thu, 16 April 2020      17:00 EET       UTC+2 hours
Shanghai (China)        Thu, 16 April 2020      23:00 CST       UTC+8 hours
Tokyo (Japan)   Fri, 17 April 2020      00:00 JST       UTC+9 hours
Corresponding UTC (GMT) Thu, 16 April 2020      15:00 UTC
<https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md#logistics>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 (login with your W3C account<https://www.w3.org/Help/Account/>)
  *   Minutes: https://www.w3.org/2020/04/16-webmachinelearning-minutes.html


<https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md#agenda>Agenda
<https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md#new-ops-for-webnn-api>New ops for WebNN API

Discuss & decide new ops to add to the WebNN API. Currently specified are conv2d and matMul.

Priority on ops that accelerate models for tasks that address identified use cases: https://webmachinelearning.github.io/webnn/#usecases


<https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md#inference-api-to-load-and-run-a-model>Inference API to load and run a model

Discuss next steps:

  *   Logistics: repo migration & naming 🚲https://github.com/webmachinelearning/webnn/issues/51

  *   Spec draft: start with spec text, use cases discussed in explainer
  *   Prototyping: Chromium and polyfilling plan
  *   Open: model formats

<https://github.com/webmachinelearning/meetings/blob/master/telcons/2020-04-16-agenda.md#virtual-workshop-agenda-building>Virtual workshop agenda building

Gather feedback on priority topics and talks for the virtual workshop pre-recorded offline mode:

https://bit.ly/webml-workshop-talks


The virtual workshop scope coverage target 30-70% of the cancelled in-person event.

Infrastucture discussion. W3C's MIT host provides us Zoom. Proposal to dogfood outselves with the MIT's Zoom instance to evaluate, use for our upcoming CG calls (this one still on WebEx).