WebML WG Teleconference – 25 September 2025 - 15:00-16:00 UTC

Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md


WebML WG Teleconference – 25 September 2025 - 15:00-16:00 UTC
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#webml-wg-teleconference--25-september-2025---1500-1600-utc>
See the timezone table ...
🤝 Logistics
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-logistics>

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

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

  *   Zoom joining instructions: internal-webmachinelearning<https://lists.w3.org/Archives/Member/internal-webmachinelearning/2023Jun/0000.html>
  *   Minutes: https://www.w3.org/2025/09/25-webmachinelearning-minutes.html


🧪 Incubations
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-incubations>

A debrief from WebML CG Teleconference – 18 September 2025<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-18-cg-agenda.md> (minutes<https://www.w3.org/2025/09/18-webmachinelearning-minutes.html>).

🧠 Web Neural Network API
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-web-neural-network-api>

Tip

Specification<https://www.w3.org/TR/webnn/> | Explainer<https://github.com/webmachinelearning/webnn/blob/main/explainer.md> | Implementation Report<https://wpt.fyi/results/webnn> | Implementation Status<https://webmachinelearning.github.io/webnn-status/>

🏷️ New features and operator specific issues
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#%EF%B8%8F-new-features-and-operator-specific-issues>

  *   Support dynamic tensor resizing for slice and resample2d

     *   ☰ Discuss generalizability of a solution to eliminate dynamic nodes in certain decoder models.
     *   ⨀ webmachinelearning/webnn#885<https://github.com/webmachinelearning/webnn/issues/885>
  *   Flexible input sizes

     *   ☰ Check back on the ORT feedback.
     *   ⨀ webmachinelearning/webnn#883<https://github.com/webmachinelearning/webnn/issues/883>

🔤 Core operator set
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-core-operator-set>

Revisit our core operator set effort that aims to identify current primitive gaps by mapping compositional fundamentals to WebNN operators.

Feedback from participants familiar with TOSA/linalg welcome on e.g. conv* variants.

  *   ⨀ webmachinelearning/webnn#573<https://github.com/webmachinelearning/webnn/issues/573>
  *   ☰ Machine Learning Operator Mapping - All Raw Operators<https://onedrive.live.com/edit?id=EE82F5C6F06C7371!345450&resid=EE82F5C6F06C7371!345450&ithint=file%2Cxlsx&authkey=!AK8f-RDTleqlLXE&wdo=2&cid=ee82f5c6f06c7371>

↔️ WebNN-WebGPU interop
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#%EF%B8%8F-webnn-webgpu-interop>

Refresh webgpu interop<https://github.com/webmachinelearning/webnn/labels/webgpu%20interop> spec issues in the light of new implementation experience from ML-GPU interop work in Chromium.

🔒 Privacy considerations
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-privacy-considerations>

Review and discuss any contributions to revised privacy considerations.

  *   ⨀ webmachinelearning/webnn#886<https://github.com/webmachinelearning/webnn/issues/886>

💾 MLGraph Cache
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-mlgraph-cache>

Discuss feedback from Apple, any further implementation experience.

  *   ☰ Explainer https://github.com/webmachinelearning/webnn/blob/main/cache-explainer.md

  *   ⨀ Feeback via webmachinelearning/webnn#807<https://github.com/webmachinelearning/webnn/issues/807>
  *   (context:) ⛙ webmachinelearning/webnn#862<https://github.com/webmachinelearning/webnn/pull/862>

🎲 Query supported devices
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-query-supported-devices>

Before graph compilation

  *   ☰ Review simplified boolean-returning context.accelerated API.
  *   ⨀ webmachinelearning/webnn#815 (comment)<https://github.com/webmachinelearning/webnn/issues/815#issuecomment-3236159329>
  *   ⛙ webmachinelearning/webnn#884<https://github.com/webmachinelearning/webnn/pull/884>

After graph compilation

  *   ☰ Discuss async graph.devices() proposal to account for dynamic post-compile device changes, a Core ML imposed requirement.
  *   ⨀ webmachinelearning/webnn#836<https://github.com/webmachinelearning/webnn/issues/836>
  *   ⛙ webmachinelearning/webnn#854<https://github.com/webmachinelearning/webnn/pull/854>

⛙ Open PRs
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2025-09-25-wg-agenda.md#-open-prs>

Ready to merge? For discussion as needed.

  *   opSupportLimits: Level of detail for output tensor(s)
     *   ⛙ webmachinelearning/webnn#857<https://github.com/webmachinelearning/webnn/pull/857>
  *   Rename pool2d roundingType -> outputShapeRounding
     *   ⛙ webmachinelearning/webnn#770<https://github.com/webmachinelearning/webnn/pull/770>

Review any other open PRs that benefit from high-bandwidth discussion:

  *   ⛙ https://github.com/webmachinelearning/webnn/pulls