See the timezone table ...
| San Francisco (U.S.A. - California) | Thu, 19 October 2022 | 07:00 | UTC-7 hours |
| Boston (U.S.A. - Massachusetts) | Thu, 19 October 2022 | 10:00 | UTC-4 hours |
| London (United Kingdom - England) | Thu, 19 October 2022 | 15:00 | UTC+1 hours |
| Berlin (Germany) | Thu, 19 October 2022 | 16:00 | UTC+2 hours |
| Helsinki (Finland) | Thu, 19 October 2022 | 17:00 | UTC+3 hours |
| Shanghai (China) | Thu, 19 October 2022 | 22:00 | UTC+8 hours |
| Tokyo (Japan) | Thu, 19 October 2022 | 23:00 | UTC+9 hours |
| Corresponding UTC (GMT) | Thu, 19 October 2022 | 14:00 UTC | |
Other locations: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20231019T14
- Chair: Anssi
- Scribe: ? (see howto, 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/2023/10/19-webmachinelearning-minutes.html
Discuss and review any spec PRs or contributions for the proposed transformer ops informed by the op breakdown exercise.
References:
- Op breakdown (w/ TOSA, StableHLO mapping): https://docs.google.com/spreadsheets/d/1ELfHuv2UqP2LoXWLgqsC0L8T_qqfBx48KxzFighl8d8/edit#gid=86906292
- Feedback via webmachinelearning/webnn#375
Discuss the recently published WebGPU security technical report, hardening opportunities for the WebNN API based on these insights.
- WebGPU security technical report: https://chromium.googlesource.com/chromium/src/+/main/docs/security/research/graphics/webgpu_technical_report.md
- Softmax axis absent
- issue: webmachinelearning/webnn#466
- split() into sizes not widely supported
- issue: webmachinelearning/webnn#392
- Clarify the restriction for minValue and maxValue of MLClampOptions
- issue: webmachinelearning/webnn#396