WebML WG Teleconference – 11 January 2024 - 15:00-16:00 UTC

Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md


WebML WG Teleconference – 11 January 2024 - 15:00-16:00 UTC
See the timezone table ...
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#logistics>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/2024/01/11-webmachinelearning-minutes.html


<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#agenda>Agenda
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#%E2%84%B9%EF%B8%8F-web-llm-project-contributions>ℹ️ Web LLM project contributions

A discussion with Tianqi Chen on proposed new WebNN use cases informed by the Web LLM project e.g. hybrid execution of models (custom ops w/ WebGPU + WebNN ops), a JSON schema of the WebNN declaration, others.

References:

  *   Web LLM project<https://webllm.mlc.ai/>
  *    issue: Proposed use cases from Web LLM project webmachinelearning/webnn#480<https://github.com/webmachinelearning/webnn/issues/480>
  *    issue: Custom operations webmachinelearning/webnn#6<https://github.com/webmachinelearning/webnn/issues/6>

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#%E2%84%B9%EF%B8%8F-efficient-webgpu-integration-discussion>ℹ️ Efficient WebGPU integration discussion

During 2023 we've sought for implementation experience and user feedback for the MLCommandEncoder interface that proposes to enable more efficient WebGPU integration. To start the year, we'll revisit the feedback and suggestions to inform this work.

Related:

  *    issue: webmachinelearning/webnn#477<https://github.com/webmachinelearning/webnn/issues/477>
  *    issue: webmachinelearning/webnn#343<https://github.com/webmachinelearning/webnn/issues/343>
  *    issue: webmachinelearning/webnn#302<https://github.com/webmachinelearning/webnn/issues/302>
  *    issue: webmachinelearning/webnn#264<https://github.com/webmachinelearning/webnn/issues/264>
  *    issue: gpuweb/gpuweb#2500<https://github.com/gpuweb/gpuweb/issues/2500>

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#%E2%84%B9%EF%B8%8F-npu-support-discussion>ℹ️ NPU support discussion

Discuss proposals for new features, enhancements, possible API surface changes beneficial for NPU support. Also triage issues with relevant insights.

Related:

  *    issue: webmachinelearning/webnn#482<https://github.com/webmachinelearning/webnn/issues/482>
  *    issue: webmachinelearning/webnn#350<https://github.com/webmachinelearning/webnn/issues/350>
  *    issue: webmachinelearning/webnn#128<https://github.com/webmachinelearning/webnn/issues/128>
  *    issue: webmachinelearning/webnn#33<https://github.com/webmachinelearning/webnn/issues/33>

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#%E2%84%B9%EF%B8%8F-delta-wide-review-and-a-new-cr-snapshot-plan>ℹ️ Delta wide review and a new CR Snapshot plan

Target to publish a new CR Snapshot during Q1'24. Discuss what short-term WebNN issues to address and what new features & enhancements to highlight in the delta wide review to satisfy publication prerequisites.

References:

  *   WebNN publication history<https://www.w3.org/standards/history/webnn/> and diff between CR Snapshot 30 Mar 2023 and latest<https://services.w3.org/htmldiff?doc1=https%3A%2F%2Fwww.w3.org%2FTR%2F2023%2FCR-webnn-20230330%2F&doc2=https%3A%2F%2Fwww.w3.org%2FTR%2Fwebnn%2F>
  *   (Informative) Wide Review<https://www.w3.org/2023/Process-20231103/#wide-review> and CR Snapshot<https://www.w3.org/2023/Process-20231103/#publishing-crrs> requirements

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-01-11-wg-agenda.md#%E2%84%B9%EF%B8%8F-enhancements>ℹ️ Enhancements

Subject to available agenda time, we'll discuss select enhancements.

  *   Remove builder.constant(value, dataType) variant

     *    issue: webmachinelearning/webnn#475<https://github.com/webmachinelearning/webnn/issues/475>
  *   Need to restrict the value of alpha to be positive for elu operation

     *    issue: webmachinelearning/webnn#383<https://github.com/webmachinelearning/webnn/issues/383>
  *   Allow 0 size dimensions

     *    issue: webmachinelearning/webnn#391<https://github.com/webmachinelearning/webnn/issues/391>