Conversation
1. Change `starts` and `sizes` parameters to be sequence of unsigned long. 2. Remove `MLSliceOptions.axes`. Fix webmachinelearning#369
fdwr
left a comment
There was a problem hiding this comment.
👍 Thanks Ningxin for simplifying the backend implementations, reducing test complexity, and increasing explicit clarity.
anssiko
left a comment
There was a problem hiding this comment.
(Publication to https://www.w3.org/TR/webnn/ upon PR merge will resume once speced/bikeshed-boilerplate#38 lands. I expect that to happen soon. I suggest hold on merging until that happens so we can validate the auto-publish job works as planned.)
|
@huningxin @wchao1115 feel free to merge when you’re settled with the remaining discussion. speced/bikeshed-boilerplate#38 was merged unblocking publications to https://www.w3.org/TR/webnn/ This is the first time we publish a Candidate Recommendation Draft (CRD) with this automated workflow so if any issues occur @dontcallmedom will be the expert to assist. |
|
Thanks @anssiko . I think we've resolved all conversations. I am going to merge it. Thanks all for the reviews and approvals. Let's see whether the automated workflow works. |
SHA: 72751c2 Reason: push, by huningxin Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
|
Works as expected: https://www.w3.org/TR/webnn/ 🚀 Thanks @dontcallmedom for reconfiguring this worflow. |
This PR is based on the discussion of #369 as the first step. It includes the changes:
startsandsizesparameters to be of a sequence of unsigned long.MLSliceOptions.axes.The
MLSliceOptions.stepsmay need more discussion and will be left to a separate PR.Fix #369
@wchao1115 @anssiko @pyu10055 @wacky6 @fdwr, PTAL!
Preview | Diff