WebML WG Teleconference – 4 April 2024 - 14:00-15:00 UTC

Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md


WebML WG Teleconference – 4 April 2024 - 14:00-15:00 UTC
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#webml-wg-teleconference--4-april-2024---1400-1500-utc>

⏰ This is a semi-annual reminder of the Daylight Saving Time (DST) change<https://www.timeanddate.com/time/dst/2024a.html>. Please note the adjusted time for the locations annotated with "⏰ adjusted for DST".

See the timezone table ...
San Francisco   Thu, 4 April 2024       07:00   UTC-7 hours (⏰ adjusted for DST)
Boston  Thu, 4 April 2024       10:00   UTC-4 hours (⏰ adjusted for DST)
London  Thu, 4 April 2024       15:00   UTC+1 hours (⏰ adjusted for DST)
Berlin  Thu, 4 April 2024       16:00   UTC+2 hours (⏰ adjusted for DST)
Helsinki        Thu, 4 April 2024       17:00   UTC+3 hours (⏰ adjusted for DST)
Shanghai        Thu, 4 April 2024       22:00   UTC+8 hours
Tokyo   Thu, 4 April 2024       23:00   UTC+9 hours
UTC     Thu, 4 April 2024       14:00 UTC

Other locations: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20240404T14


Logistics
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#logistics>

  *   Chair: Anssi
  *   Scribe: ? (see howto<https://github.com/webmachinelearning/meetings/blob/main/scribe-howto.md>, volunteers welcome!)
  *   IRC: irc://irc.w3.org:6667/#webmachinelearning

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

  *   Joining instructions: https://lists.w3.org/Archives/Member/internal-webmachinelearning/2023Jun/0000.html

  *   Minutes: https://www.w3.org/2024/04/04-webmachinelearning-minutes.html


Agenda
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#agenda>
ℹ️ NPU support discussion
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%E2%84%B9%EF%B8%8F-npu-support-discussion>

The goal of this discussion is to revisit the problem space for NPU support for WebNN, review the high-level proposal for the key elements of the design, gather feedback and signals from the group regarding interest, scope, priority.

NPU device type and support for quantized models have been explored in the group prior and have been awaiting implementation experience.

  *   ⨀ WebNN should support NPU and QDQ operations webmachinelearning/webnn#623<https://github.com/webmachinelearning/webnn/issues/623>

Related:

  *   ⨀ WebNN should support int8 quantized models webmachinelearning/webnn#128<https://github.com/webmachinelearning/webnn/issues/128>
  *   ⨀ API simplification: context types, context options, createContext() webmachinelearning/webnn#302<https://github.com/webmachinelearning/webnn/issues/302>

ℹ️ Hybrid AI exploration update
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%E2%84%B9%EF%B8%8F-hybrid-ai-exploration-update>

A brief update from the project team on the proof-of-concept informed by the group's feedback.

  *   ⨀ Hybrid AI exploration webmachinelearning/proposals#5<https://github.com/webmachinelearning/proposals/issues/5>

ℹ️ Open issues and PRs
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%E2%84%B9%EF%B8%8F-open-issues-and-prs>

Discuss open issues and review PRs based on your feedback:

  *   ⨀ All open issues<https://github.com/webmachinelearning/webnn/issues>
  *   ⛙ All open pull requests<https://github.com/webmachinelearning/webnn/pulls>

A curated shortlist of recently discussed issues below. If any are addressed prior to the meeting, we'll debrief the group on the changes to keep everyone up to date with significant developments. Please refer to the triage guidance<https://github.com/webmachinelearning/webnn/blob/main/docs/IssueTriage.md> for more on our triage practices.

🏷️ Debrief on PRs merged recently
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%EF%B8%8F-debrief-on-prs-merged-recently>

The editors and PR authors will debrief the group on substantive PRs that got merged in the last few weeks, answer question from the group.

  *   ⛙ Recently merged PRs<https://github.com/webmachinelearning/webnn/pulls?q=is%3Apr+is%3Amerged>

🏷️ bug<https://github.com/webmachinelearning/webnn/labels/bug>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%EF%B8%8F-bug>

  *   Need clarify scale factor for resample2d

     *   ⨀ webmachinelearning/webnn#610<https://github.com/webmachinelearning/webnn/issues/610>
     *   ☰ Overflow handling convention preferences?
  *   Synchronously validate input operands/activations

     *   ⨀ webmachinelearning/webnn#572<https://github.com/webmachinelearning/webnn/issues/572>
     *   ☰ Introduce structure in op creation?

🏷️ question<https://github.com/webmachinelearning/webnn/labels/question>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%EF%B8%8F-question>

  *   Allow no-op graphs?

     *   ⨀ webmachinelearning/webnn#614<https://github.com/webmachinelearning/webnn/issues/614>
     *   ☰ Review issue discussion
  *   Graph with no input

     *   ⨀ webmachinelearning/webnn#615<https://github.com/webmachinelearning/webnn/issues/615>
     *   ☰ Emerging consensus use cases exists, proposal to keep support
  *   Can an MLGraphBuilder be reused?

     *   ⨀ webmachinelearning/webnn#567<https://github.com/webmachinelearning/webnn/issues/567>
     *   ☰ Review issue discussion
  *   Consider adopting new broadcasting rules

     *   ⨀ webmachinelearning/webnn#590<https://github.com/webmachinelearning/webnn/issues/590>
     *   ☰ Gentle nudge: feedback on proposed options for broadcasting rules: 1) NumPy's 2) XLA's 3) keep current 4) adjust current
  *   Is "validate graph resources" backwards?

     *   ⨀ webmachinelearning/webnn#602<https://github.com/webmachinelearning/webnn/issues/602>
     *   ⛙ PR expected using a reflexive test
     *   ☰ Review PR
  *   Need clarify the usage of axes=[0,1] for resample2d

     *   ⨀ webmachinelearning/webnn#624<https://github.com/webmachinelearning/webnn/issues/624>
     *   ☰ Implementation experience informed question

🏷️ feature request<https://github.com/webmachinelearning/webnn/labels/feature%20request>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%EF%B8%8F-feature-request>

  *   Gaussian error linear unit (GELU) activation
     *   ⨀ webmachinelearning/webnn#626<https://github.com/webmachinelearning/webnn/issues/626>
     *   ☰ Required by well-known transform models

🏷️ conventions<https://github.com/webmachinelearning/webnn/labels/conventions>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-04-04-wg-agenda.md#%EF%B8%8F-conventions>

No major issues at this time.