Re: WebML WG Teleconference – 24 August 2023 - 14:00-15:00 UTC

Hi Joshua,

Thanks for this update. I’ll schedule time for the WG to discuss in general the modern spec conventions with an aim to make editors’ and other spec contributors' work easier.

Now that we’re building on top of the modernized spec any future updates are expected to follow the modern conventions — now is a good time to look at these best practices and remove any friction there may be in the toolchain or the workflow, demystify Bikeshed’s lesser-known features.

Thanks,

-Anssi

On 28. Aug 2023, at 21.52, Joshua Bell <jsbell@google.com<mailto:jsbell@google.com>> wrote:

Hey folks! We didn't end up with time to talk about modern spec conventions in the call last week. I wrote up some of what I wanted to share and merged it into this doc:

https://garykac.github.io/procspec/ - look for the "Processing Model" section

That doc is an introduction for folks new to writing specs with the recent conventions in Web IDL and Infra, as well as getting the most out of Bikeshed. I should also call out the excellent https://dlaliberte.github.io/bikeshed-intro/ which focuses more on the mechanics of getting started with Bikeshed, but has plenty of overlap around best practices and how to think about defining behavior in specs.

(Why two docs? Well, two authors = two docs, of course. Expect these to grow/move/etc over time.)

On Thu, Aug 24, 2023 at 12:10 AM Kostiainen, Anssi <anssi.kostiainen@intel.com<mailto:anssi.kostiainen@intel.com>> wrote:
Hi Joshua,

Your suggestions for conventions modernization have been well received by the WG. You’re welcome to contribute to this discussion on our call. We have an exciting (and a busy) agenda lined up.

Thanks,

-Anssi


On 23. Aug 2023, at 20.35, Joshua Bell <jsbell@google.com<mailto:jsbell@google.com>> wrote:

As part of the "WebIDL and Infra standard conventions" section would it be helpful for me to spend a couple minutes discussing some of the motivations behind modern spec style, and the "processing model" (i.e. how JS types pass through Web IDL to become Infra types), etc?




On Thu, Aug 17, 2023 at 4:08 AM Kostiainen, Anssi <anssi.kostiainen@intel.com<mailto:anssi.kostiainen@intel.com>> wrote:
Latest version: https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md

WebML WG Teleconference – 24 August 2023 - 14:00-15:00 UTC
See the timezone table ...
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#logistics>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/2023/08/24-webmachinelearning-minutes.html


<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#agenda>Agenda
<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#%E2%84%B9%EF%B8%8F-google-chrome-teams-feedback-on-webnn-api>ℹ️ Google Chrome team's feedback on WebNN API

Vivek and Joshua B have gathered feedback from various Google teams interested in WebNN and will share their thoughts with the WG for discussion.

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#%E2%84%B9%EF%B8%8F-webnn-v2-review-proposed-new-ops-and-data-types>ℹ️ WebNN v2: review proposed new ops and data types

Review and discuss the proposed new ops and data types informed by v2 model targets and recent prototyping efforts.

Proposed models:

  *   Text-to-image: Stable Diffusion unet/VAE/text encoder
  *   Image segmentation: Segment Everything decoder
  *   Speech-to-text: Whisper Tiny
  *   Text-to-text? Summarization, translation, code completion demonstrated by Transformers.js?

Proposed new ops:

  *   Logical elementwise comparison/selection operations: equal, greater, lesser, logicalNot, elementwiseIf/ternary, greaterOrEqual/lesserOrEqual
  *   More elementwise unary operations: identity, sqrt, erf (Gauss err func), reciprocal
  *   Reshaping operations: squeeze, unsqueeze, flattenTo2d
  *   Data rearrangement operations: expand, gather
  *   Normalization operations: meanVarianceNormalization
  *   Index seeking operations: argMin/argMax
  *   Misc: cast, fillSequence, triangularMatrix, shape
  *   Others?

Proposed new data types: int64, uint64

Details: webmachinelearning/webnn#375 (comment)<https://github.com/webmachinelearning/webnn/issues/375#issuecomment-1674224992>

Background material:

  *   Transformers.js presentation by Joshua Lochner<https://lists.w3.org/Archives/Public/www-archive/2023Jun/att-0000/Transformers_js.pdf>
  *   Transformer models presentation by Dwayne Robinson<https://lists.w3.org/Archives/Public/www-archive/2023Jun/att-0005/2023-06-29_WebNN_and_Transformers_Progress_W3C.pdf>
  *   Transformer-based models and use cases discussion: webmachinelearning/webnn#375<https://github.com/webmachinelearning/webnn/issues/375>
  *   Proposed v2 ops<https://github.com/webmachinelearning/webnn/issues?q=is%3Aopen+label%3A%22operation+set%22+label%3A%22v2%22> (open new issues for v2 op proposals)

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#%E2%84%B9%EF%B8%8F-webidl-and-infra-standard-conventions>ℹ️ WebIDL and Infra standard conventions

Decide whether we're ready to merge the zk-conventions-integration integration branch to main.

These changes align the entire specification with modern specification conventions and add stylistic improvements on top that make navigating this specification more delightful experience.

The following resources are made available to the group to assist in this review task:

  *   Preview<https://pr-preview.s3.amazonaws.com/zolkis/webnn/pull/446.html> of zk-conventions-integration PR
  *   Diff<https://pr-preview.s3.amazonaws.com/zolkis/webnn/pull/446.html> between main and zk-conventions-integration PR
  *   Issues addressed and PRs merged to<https://github.com/webmachinelearning/webnn/issues/210#issuecomment-1326361748> zk-conventions-integration (see "🟡 review feedback" annotations)
  *   Commits<https://github.com/webmachinelearning/webnn/compare/main...zk-conventions-integration> ahead of main

<https://github.com/webmachinelearning/meetings/blob/main/telcons/2023-08-24-wg-agenda.md#%E2%84%B9%EF%B8%8F-enhancements>ℹ️ Enhancements

  *   split() into sizes not widely supported
     *    issue: webmachinelearning/webnn#392<https://github.com/webmachinelearning/webnn/issues/392>
  *   Clarify the restriction for minValue and maxValue of MLClampOptions
     *    issue: webmachinelearning/webnn#396<https://github.com/webmachinelearning/webnn/issues/396>