- From: Kostiainen, Anssi <anssi.kostiainen@intel.com>
- Date: Thu, 2 Feb 2023 17:29:14 +0000
- To: "public-webmachinelearning-wg@w3.org" <public-webmachinelearning-wg@w3.org>
Hi All, Please find our meeting minutes at: https://www.w3.org/2023/02/02-webmachinelearning-minutes.html We spent the full hour discussing PR #322 to simplify MLContext creation and its relationship with the upcoming CR milestone. Thank you all for sharing your insights! https://github.com/webmachinelearning/webnn/pull/322 PR #322 introduces a normative dependency on the WebGPU API for the GPU path. We discussed how maturization of this important feature would greatly benefit from tight coordination with WebGPU contributors, implementation experience and feedback from key customers such as JS ML framework. We also discussed pros and cons and logistics of landing this feature in a revised CR vs. in the initial CR. The upcoming WebGPU F2F 2023-02-16/17 in the SF bay area would be a great opportunity to socialise and seek feedback on the WebNN-WebGPU interoperability design overall. As for our charter 2023-2025, I met with Dom following our WG call and we merged the charter 2023-2025 PR. Dom is doing his final pass review before the charter starts its AC review. Here’s the latest draft for your final review: https://w3c.github.io/machine-learning-charter/charter.html Thanks, -Anssi (WebML WG chair)
Received on Thursday, 2 February 2023 17:29:40 UTC