De-duplicate 1D scale and zero point tensors to scalars in DML kernels #18629
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#18495) Update TryConvertTensorToBroadcastScalar to fail gracefully if not CPU tensor ### Description Added early exit for in TryConvertTensorToBroadcastScalar if there is no CpuData to prevent failure caused by constExpTensor->GetByteData() throwing. ### Motivation and Context DmlPrototype branch was failing in QLinearConv when there was no CpuData --------- Co-authored-by: Christian Larson <[email protected]>
### Description Fix TryConvertTensorToBroadcastScalar build break, caused by incorrect method call. - else if (!IsCpuData()) + else if (!constExpTensor->IsCpuData()) ### Motivation and Context Previous build causing build break. Co-authored-by: Christian Larson <[email protected]>
chrilaMSFT
reviewed
Nov 29, 2023
| @@ -111,6 +111,8 @@ class DmlOperatorBatchNormalization15 : public DmlOperator, BatchNormalizationHe | |||
| std::vector<DML_TENSOR_DESC> inputDescs = GetDmlInputDescs(); | |||
| std::vector<DML_TENSOR_DESC> outputDescs = GetDmlOutputDescs(); | |||
|
|
|||
| // TODO: Port this to a graph description to enable DML graph optimization | |||
chrilaMSFT
reviewed
Nov 29, 2023
| @@ -85,7 +85,10 @@ namespace Windows::AI::MachineLearning::Adapter | |||
| { | |||
| uint32_t nodeCount = 0; | |||
| std::vector<std::unique_ptr<AbstractOperatorDesc>> nodesAsOperatorDesc; | |||
|
|
|||
| // TODO: Remove this | |||
There was a problem hiding this comment.
should this be removed?
jeffbloo
added a commit
that referenced
this pull request
Jan 2, 2024
#18862) ### Description Cleanup and rebase from [this PR](#18629) ### Motivation and Context --------- Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Jeff Bloomfield <[email protected]> Co-authored-by: Anagha Rao <[email protected]>
|
This is merged in this PR instead: |
jeffbloo
added a commit
that referenced
this pull request
Jan 4, 2024
#18862) ### Description Cleanup and rebase from [this PR](#18629) ### Motivation and Context --------- Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Jeff Bloomfield <[email protected]> Co-authored-by: Anagha Rao <[email protected]>
jslap-ubi
pushed a commit
to jslap-ubi/onnxruntime
that referenced
this pull request
Apr 5, 2024
microsoft#18862) ### Description Cleanup and rebase from [this PR](microsoft#18629) ### Motivation and Context --------- Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Christian Larson <[email protected]> Co-authored-by: Jeff Bloomfield <[email protected]> Co-authored-by: Anagha Rao <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
De-duplicate 1D scale and zero point tensors to scalars in DML kernels
Fix TryConvertTensorToBroadcastScalar build break by chrilaMSFT · Pull Request #18497 · microsoft/onnxruntime (github.com)
Update TryConvertTensorToBroadcastScalar to exit early if constExpTen… · microsoft/onnxruntime@e36cc6e (github.com)
Motivation and Context