WebML WG Teleconference – 18 December 2025 - meeting minutes

Hi All,

Thanks for joining us!

Special thanks to Tarek for sharing an exciting early Xmas present with the group: a Python and Rust implementation of the WebNN API aimed at Firefox! 🎁

We will resume these WG meetings after the holiday break on 15th January 2026. Happy Holidays Everyone!

Minutes:

https://www.w3.org/2025/12/18-webmachinelearning-minutes.html


Summary:

  1.  Incubations<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#2a6f>
  2.  New implementation experience and developer feedback<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#1476>
     *   Python and Rust implementation of the WebNN API aimed at Firefox<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#daf1>
     *   New developer feedback from the developer ecosystem<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#5539>
  3.  External weights, learnings from WebGPU<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#60f1>
  4.  New device selection hints for MLContextOptions<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#646f>
  5.  Add minimum data type set and rank range for input, constant, output<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#0d99>
  6.  Remove conformance tests with negative scale of DQ/Q operators<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#ac0d>
  7.  2025 Reflections<https://www.w3.org/2025/12/18-webmachinelearning-minutes.html#7c2b>

Thanks,

-Anssi (WebML WG/CG chair)