- From: Kostiainen, Anssi <anssi.kostiainen@intel.com>
- Date: Wed, 8 May 2024 12:36:13 +0000
- To: "public-webmachinelearning-wg@w3.org" <public-webmachinelearning-wg@w3.org>
- Message-ID: <225F9302-8D33-4152-8EF8-7E3FDBF34F86@intel.com>
Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md
WebML WG Teleconference – 16 May 2024 - 14:00-15:00 UTC
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#webml-wg-teleconference--16-may-2024---1400-1500-utc>
See the timezone table ...
Logistics
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#logistics>
* Chair: Anssi
* Scribe: ? (see howto<https://github.com/webmachinelearning/meetings/blob/main/scribe-howto.md>, 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/2024/05/16-webmachinelearning-minutes.html
Agenda
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#agenda>
ℹ️ NPU support
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%E2%84%B9%EF%B8%8F-npu-support>
We agreed to start formalize the NPU support with the simplest design (option 1: deviceType: "npu") informed by implementation experience. Review a draft PR if available, discuss any new information and considerations brought to the group's attention in the issue. Discuss privacy and fingerprinting considerations of the expanded three bits of entropy.
* ⨀ WebNN should support NPU and QDQ operations webmachinelearning/webnn#623<https://github.com/webmachinelearning/webnn/issues/623>
Related:
* ⛭ Chromium implementation (NPU device type) https://chromium-review.googlesource.com/c/chromium/src/+/5330647
ℹ️ Open issues and PRs
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%E2%84%B9%EF%B8%8F-open-issues-and-prs>
Discuss open issues and review PRs based on your feedback:
* ⨀ All open issues<https://github.com/webmachinelearning/webnn/issues>
* ⛙ All open pull requests<https://github.com/webmachinelearning/webnn/pulls>
A curated shortlist of recently discussed issues below. If any are addressed prior to the meeting, we'll debrief the group on the changes to keep everyone up to date with significant developments. Please refer to the triage guidance<https://github.com/webmachinelearning/webnn/blob/main/docs/IssueTriage.md> for more on our triage practices.
🏷️ Debrief on PRs merged recently
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-debrief-on-prs-merged-recently>
The editors and PR authors will debrief the group on substantive PRs that got merged in the last few weeks, answer question from the group.
* ⛙ Recently merged PRs<https://github.com/webmachinelearning/webnn/pulls?q=is%3Apr+is%3Amerged>
🏷️ process<https://github.com/webmachinelearning/webnn/labels/process>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-process>
* 🆕 Introduce "interop" label
* ⨀ webmachinelearning/webnn#673<https://github.com/webmachinelearning/webnn/issues/673>
* ⛙ webmachinelearning/webnn#682<https://github.com/webmachinelearning/webnn/pull/682>
* ☰ Review PR and triage interop issues
* 🆕 TypeScript Types Declarations for WebNN
* ⨀ webmachinelearning/webnn#677<https://github.com/webmachinelearning/webnn/issues/677>
* ☰ Review a proposal for a possible new WebML CG deliverable (note: Community Group)
* Broad device coverage and maintainability
* ⨀ webmachinelearning/webnn#453<https://github.com/webmachinelearning/webnn/issues/453>
* ☰ Revisit this high-level issue to discuss what has been done, what remains to be done, disseminate any new information
🏷️ bug<https://github.com/webmachinelearning/webnn/labels/bug>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-bug>
* ArgMax/Min selectLastIndex is not supported on CoreML
* ⨀ webmachinelearning/webnn#652<https://github.com/webmachinelearning/webnn/issues/652>
* ☰ Proposal to remove selectLastIndex parameter
* Consider changing output type of ArgMax/Argmin to int32, or allow passing output_type
* ⨀ webmachinelearning/webnn#653<https://github.com/webmachinelearning/webnn/issues/653>
* ☰ Discussion to unblock exploration into option 2) allow passing output data type as a parameter
* Synchronously validate input operands/activations
* ⨀ webmachinelearning/webnn#572<https://github.com/webmachinelearning/webnn/issues/572>
* ⛙ webmachinelearning/webnn#674<https://github.com/webmachinelearning/webnn/pull/674>
* ☰ Review PR
🏷️ question<https://github.com/webmachinelearning/webnn/labels/question>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-question>
* 🆕 Make MLOperandDescriptor.dimensions a FrozenArray
* ⨀ webmachinelearning/webnn#670<https://github.com/webmachinelearning/webnn/issues/670>
* ☰ Discussion on frozen array type subtleties
* 🆕 Reconsider MLOperand methods
* ⨀ webmachinelearning/webnn#666<https://github.com/webmachinelearning/webnn/issues/666>
* ☰ Review 4 options proposed for the MLOperand methods
* 🆕 Why gather indices only accept uint32 or int64?
* ⨀ webmachinelearning/webnn#675<https://github.com/webmachinelearning/webnn/issues/675>
* ☰ Discuss CoreML findings
* Do we need an MLConstantOperand?
* ⨀ webmachinelearning/webnn#668<https://github.com/webmachinelearning/webnn/issues/668>
* ☰ Revisit with new information from XNNPACK
🏷️ feature request<https://github.com/webmachinelearning/webnn/labels/feature%20request>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-feature-request>
* 🆕 Type casting all the things
* ⨀ Scalar values webmachinelearning/webnn#678<https://github.com/webmachinelearning/webnn/issues/678>
* ⨀ fp/int <-> uint webmachinelearning/webnn#489<https://github.com/webmachinelearning/webnn/issues/489>
* ⨀ fp64 -> fp32/fp16 webmachinelearning/webnn#325<https://github.com/webmachinelearning/webnn/issues/325>
* ☰ Discussion to clarify casting behavior, an early proposal in #678
* Allow checking whether operators/types are supported for a backend before creating a graph
* ⨀ webmachinelearning/webnn#463<https://github.com/webmachinelearning/webnn/issues/463>
* ☰ Discuss new information from related uint64/int64 data type #654 and MLConstantOperand #668 issues
* MLNumber
* ⨀ webmachinelearning/webnn#442<https://github.com/webmachinelearning/webnn/issues/442>
* ⛙ webmachinelearning/webnn#647<https://github.com/webmachinelearning/webnn/pull/647>
* ☰ Review (draft) PR as appropriate
* Rename MLOperandDescriptor's dimensions to shape
* ⨀ webmachinelearning/webnn#669<https://github.com/webmachinelearning/webnn/issues/669>
* ⛙ webmachinelearning/webnn#676<https://github.com/webmachinelearning/webnn/pull/676>
* ☰ Review (draft) PR as appropriate
* Consider adding node labels for more diagnosable error messages for async errors
* ⨀ webmachinelearning/webnn#585<https://github.com/webmachinelearning/webnn/issues/585>
* ☰ Review alternatives A) add to options dict B) extend MLOperand's label field C) keep current
🏷️ operator specific<https://github.com/webmachinelearning/webnn/labels/operator%20specific>
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2024-05-16-wg-agenda.md#%EF%B8%8F-operator-specific>
* argMax/Min only support scalar axis in TFLite runtime
* ⨀ webmachinelearning/webnn#629<https://github.com/webmachinelearning/webnn/issues/629>
* ☰ Review other backends and frameworks
* Clarify the restriction for minValue and maxValue of MLClampOptions
* ⨀ webmachinelearning/webnn#396<https://github.com/webmachinelearning/webnn/issues/396>
* ☰ Proposal to remove "Not all implementations support minValue equal to maxValue." note from the spec
Received on Wednesday, 8 May 2024 12:36:25 UTC