See the timezone table ...
| San Francisco (U.S.A. - California) | Thu, 24 June 2021 | 07:00 | UTC-7 hours |
| Boston (U.S.A. - Massachusetts) | Thu, 24 June 2021 | 10:00 | UTC-4 hours |
| London (United Kingdom - England) | Thu, 24 June 2021 | 15:00 | UTC+1 hours |
| Berlin (Germany) | Thu, 24 June 2021 | 16:00 | UTC+2 hours |
| Helsinki (Finland) | Thu, 24 June 2021 | 17:00 | UTC+3 hours |
| Shanghai (China) | Thu, 24 June 2021 | 22:00 | UTC+8 hours |
| Tokyo (Japan) | Thu, 24 June 2021 | 23:00 | UTC+9 hours |
| Corresponding UTC (GMT) | Thu, 24 June 2021 | 14:00 UTC | |
Other locations: https://www.timeanddate.com/worldclock/fixedtime.html?iso=20210624T14
- Chair: Anssi
- Scribe: ?
- IRC: irc://irc.w3.org:6667/#webmachinelearning
- IRC web client: https://irc.w3.org/?channels=#webmachinelearning
- Call-in details: https://lists.w3.org/Archives/Member/internal-webmachinelearning/2020Apr/0000.html
- Minutes: https://www.w3.org/2021/06/24-webmachinelearning-minutes.html
- First Public Working Draft (FPWD) publication
- Discuss next steps and milestones on the W3C Recommendation Track
webnn-native and webnn-polyfill are software projects of the Web Machine Learning Community Group. These projects informs the Web Neural Network API specification work and we will review the progress and findings of these project on these Working Group calls from time to time.
- webnn-native and webnn-polyfill introduction to the WG, how to get involved, areas welcoming contributions, Q&A (Ningxin)
- https://github.com/webmachinelearning/webnn-native
- https://github.com/webmachinelearning/webnn-polyfill
WebNN API has received W3C Technical Architecture Group (TAG) review, we discuss and address the remaining open issues:
- [tag-tracker] Define a common term for logical tensor changes?
- Issue: webmachinelearning/webnn#150
- PR: WIP (Rama?)
- [tag-tracker] Ergonomics of the JS examples
- Issue: webmachinelearning/webnn#139
- PR: WIP - pending concrete TAG feedback to inform the resolution (Sangwhan)
- [tag-tracker] String enum for activations
- Issue: webmachinelearning/webnn#138
- PR: WIP - add an informative note to be incorporated into the spec to explain current design principles around activations (Chai)
These issues on our issue tracker are welcoming comments and proposed solutions to be reviewed on our post-break meetings. Please review this list for triage purposes, time permitting.
- AI accelerator device selection webmachinelearning/webnn#169
- HardSwish op webmachinelearning/webnn#181
- argMax op webmachinelearning/webnn#184
- Dilated pooling webmachinelearning/webnn#180
- int8 quantized webmachinelearning/webnn#128
- Dynamic shape inference webmachinelearning/webnn#124
- QuantizeLinear and DequantizeLinear webmachinelearning/webnn#93
- [v2] Support output tensor in native memory format webmachinelearning/webnn#173
- Fused activation webmachinelearning/webnn#185
- OperandType of gemm / matmul return webmachinelearning/webnn#84
- Handling unsupported OperandType webmachinelearning/webnn#36
- OperandDescriptor's dimension should match ECMAScript's TypedArray dimension webmachinelearning/webnn#29
- Specify the ModelBuilder.createModel and other ModelBuilder members webmachinelearning/webnn#107
- Buffer sharing between GPU and ML accelerator webmachinelearning/webnn#33
- Workflow from authoring tools to webnn webmachinelearning/webnn#74
These meetings will break for July and will resume mid-August.
Please respond to the meeting scheduling poll to signal your preferences: https://doodle.com/poll/gymnpmtcu4dcknck
The GitHub repos will remain open for contributions while our meetings are on pause. Thank you for your continued contributions!