Sync with Microsoft ONNX Runtime - 09/01/2026#897
Merged
ankitm3k merged 21 commits intoovep-developfrom Jan 9, 2026
Merged
Conversation
### Description <!-- Describe your changes. --> Add Plugin EP API `OrtEp::IsConcurrentRunSupported()` and connect it to the internal `PluginExecutionProvider::ConcurrentRunSupported()` member function. ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. --> Additional plugin EP support.
Update GQA benchmark to support bfloat16 and default to testing the first configuration (fast mode). Note that test_sparse_attention.py was removed in microsoft#23547. It is referenced by the benchmark script, so I add it back and disable the test in pipeline mode. Example output from H200 GPU: ``` prompt-sm90-Llama3-8B-b1-h32_8x128-float16: sequence_length ORT-GQA-Dense ORT-GQA-Dense-PackedQKV 0 16.0 0.781751 0.571226 1 32.0 0.893813 0.684198 2 64.0 1.434056 1.589263 3 128.0 1.142192 1.681969 4 256.0 1.503483 2.225498 5 512.0 1.045732 1.878660 6 1024.0 2.334924 0.916745 7 2048.0 2.229924 3.001290 8 4096.0 4.309678 3.198855 9 8192.0 7.932211 7.910411 token-sm90-Llama3-8B-b1-h32_8_d128-float16: past_sequence_length ORT-GQA-Dense ORT-GQA-Dense-PackedQKV 0 16.0 1.751966 0.780081 1 32.0 1.302806 0.043939 2 64.0 2.301024 2.207282 3 128.0 2.294556 3.010107 4 256.0 2.931330 1.781768 5 512.0 1.210220 2.799579 6 1024.0 2.767142 2.660434 7 2048.0 1.420229 0.091433 8 4096.0 0.860655 0.801022 9 8191.0 0.749525 0.820858 prompt-sm90-Llama3-8B-b1-h32_8x128-bfloat16: sequence_length ORT-GQA-Dense ORT-GQA-Dense-PackedQKV 0 16.0 1.085427 0.666664 1 32.0 1.714795 0.931262 2 64.0 1.729093 1.438733 3 128.0 1.071263 2.486135 4 256.0 1.957349 1.342417 5 512.0 1.159680 1.591321 6 1024.0 0.743702 2.035150 7 2048.0 1.452736 1.788801 8 4096.0 4.029917 4.041565 9 8192.0 7.934485 7.931600 token-sm90-Llama3-8B-b1-h32_8_d128-bfloat16: past_sequence_length ORT-GQA-Dense ORT-GQA-Dense-PackedQKV 0 16.0 0.044354 0.043983 1 32.0 0.040715 0.044061 2 64.0 0.045586 0.044071 3 128.0 0.062204 0.061418 4 256.0 0.074764 4.874854 5 512.0 2.472094 2.102259 6 1024.0 4.911269 1.396149 7 2048.0 4.898032 1.684034 8 4096.0 2.523432 2.192279 9 8191.0 1.651366 3.427370 ```
### Description ### Motivation and Context For the 1.24 release.
### Description Executive Summary The PR successfully updates the VCPKG baseline and Abseil dependency to version 20250814.0. A significant improvement is the introduction of a dynamic dependency resolution script for Abseil. There are also substantial refactoring changes in the WebGPU provider build logic which appear to be bundled with this update. Key Changes Verified 1. VCPKG & Abseil Update VCPKG Baseline: Updated to 120deac3062162151622ca4860575a33844ba10b in cmake/vcpkg-configuration.json. Abseil Port: cmake/vcpkg-ports/abseil/vcpkg.json and portfile.cmake correctly updated to version 20250814.0. Dependencies: cmake/deps.txt reflects the correct URL and hash for the new Abseil version. 2. Dependency Management New Tool: tools/python/resolve_absl_deps_dynamic.py has been added. Review: This is a robust script that automatically resolves and topologically sorts Abseil dependencies by parsing CMake lists. This is a great addition for maintainability, replacing the manual list management. CMake Integration: cmake/external/abseil-cpp.cmake has been updated to use the new list of components, consistent with the script's output. ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. --> --------- Co-authored-by: Changming Sun <chasun@microsoft.com> Co-authored-by: Gemini <gemini@google.com>
…t#26728) ## Summary - Directly load QnnHtp.dll and query platform info using GetPlatformInfo API in the GTest SetUp function. - Add QNN include files in unittest CMake for QNN enums related to platform. - Skip all FP16 tests via macro for HTP architecture v68 on Windows. - Skip Auto EP tests via macro for HTP architecture v68 on Windows. ## Description This patch introduces a mechanism in QNN-specific unit tests to detect HTP architecture version and related platform details. Based on this information, tests that require unsupported features on certain hardware models can be dynamically disabled. This ensures better compatibility and stability across different SoC configurations. ## Motivation and Context Due to variations in device configuration and feature support across SoC models, some unit tests should only run on specific devices. This patch addresses that by querying platform information from the existing SDK library on Windows, enabling conditional test execution based on hardware capabilities.
### Description <!-- Describe your changes. --> ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. -->
This pull request refactors and streamlines the computation of Q, K, V tensors in the WebGPU BERT Attention operator. The main changes include removing a custom QKV preparation kernel in favor of a more modular approach using a MatMul operation followed by a dedicated split kernel, and generalizing the QKV splitting logic for broader reuse. This improves maintainability, code reuse, and performance since we have done many optimization on MatMul op. With this change, PrepareQKV becomes 128.88 ms from 751.67 ms in phi4-vision model. Before Kernel | Time (ms) | Percentage (%) -- | -- | -- Attention\|AttentionPrepare | 751.67 | 49.91 After Kernel | Time (ms) | Percentage (%) -- | -- | -- Attention\|MatMul | 120.87 | 19.77 Attention\|SplitPackedQKV | 1.94 | 0.32
This pull request adds support for batched matrix multiplication in the DP4A quantized matmul WebGPU kernels and their associated C++ code and tests. The changes update the kernel code, tensor shapes, dispatch logic, and test infrastructure to properly handle a `batch_count` greater than 1, enabling efficient batched execution.
…rosoft#26915) Aligns onnxruntime-foundry `caller-framework` build flag with the windows-ml build flag so telemetry can work through the same infrastructure.
Bumps [qs](https://github.com/ljharb/qs) and [body-parser](https://github.com/expressjs/body-parser). These dependencies needed to be updated together. Updates `qs` from 6.13.0 to 6.14.1 <details> <summary>Changelog</summary> <p><em>Sourced from <a href="https://github.com/ljharb/qs/blob/main/CHANGELOG.md">qs's changelog</a>.</em></p> <blockquote> <h2><strong>6.14.1</strong></h2> <ul> <li>[Fix] ensure arrayLength applies to <code>[]</code> notation as well</li> <li>[Fix] <code>parse</code>: when a custom decoder returns <code>null</code> for a key, ignore that key</li> <li>[Refactor] <code>parse</code>: extract key segment splitting helper</li> <li>[meta] add threat model</li> <li>[actions] add workflow permissions</li> <li>[Tests] <code>stringify</code>: increase coverage</li> <li>[Dev Deps] update <code>eslint</code>, <code>@ljharb/eslint-config</code>, <code>npmignore</code>, <code>es-value-fixtures</code>, <code>for-each</code>, <code>object-inspect</code></li> </ul> <h2><strong>6.14.0</strong></h2> <ul> <li>[New] <code>parse</code>: add <code>throwOnParameterLimitExceeded</code> option (<a href="https://redirect.github.com/ljharb/qs/issues/517">#517</a>)</li> <li>[Refactor] <code>parse</code>: use <code>utils.combine</code> more</li> <li>[patch] <code>parse</code>: add explicit <code>throwOnLimitExceeded</code> default</li> <li>[actions] use shared action; re-add finishers</li> <li>[meta] Fix changelog formatting bug</li> <li>[Deps] update <code>side-channel</code></li> <li>[Dev Deps] update <code>es-value-fixtures</code>, <code>has-bigints</code>, <code>has-proto</code>, <code>has-symbols</code></li> <li>[Tests] increase coverage</li> </ul> <h2><strong>6.13.1</strong></h2> <ul> <li>[Fix] <code>stringify</code>: avoid a crash when a <code>filter</code> key is <code>null</code></li> <li>[Fix] <code>utils.merge</code>: functions should not be stringified into keys</li> <li>[Fix] <code>parse</code>: avoid a crash with interpretNumericEntities: true, comma: true, and iso charset</li> <li>[Fix] <code>stringify</code>: ensure a non-string <code>filter</code> does not crash</li> <li>[Refactor] use <code>__proto__</code> syntax instead of <code>Object.create</code> for null objects</li> <li>[Refactor] misc cleanup</li> <li>[Tests] <code>utils.merge</code>: add some coverage</li> <li>[Tests] fix a test case</li> <li>[actions] split out node 10-20, and 20+</li> <li>[Dev Deps] update <code>es-value-fixtures</code>, <code>mock-property</code>, <code>object-inspect</code>, <code>tape</code></li> </ul> </blockquote> </details> <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/ljharb/qs/commit/3fa11a5f643c76896387bd2d86904a2d0141fdf7"><code>3fa11a5</code></a> v6.14.1</li> <li><a href="https://github.com/ljharb/qs/commit/a62670423c1ccab0dd83c621bfb98c7c024e314d"><code>a626704</code></a> [Dev Deps] update <code>npmignore</code></li> <li><a href="https://github.com/ljharb/qs/commit/3086902ecf7f088d0d1803887643ac6c03d415b9"><code>3086902</code></a> [Fix] ensure arrayLength applies to <code>[]</code> notation as well</li> <li><a href="https://github.com/ljharb/qs/commit/fc7930e86c2264c1568c9f5606830e19b0bc2af2"><code>fc7930e</code></a> [Dev Deps] update <code>eslint</code>, <code>@ljharb/eslint-config</code></li> <li><a href="https://github.com/ljharb/qs/commit/0b06aac566abee45ef0327667a7cc89e7aed8b58"><code>0b06aac</code></a> [Dev Deps] update <code>@ljharb/eslint-config</code></li> <li><a href="https://github.com/ljharb/qs/commit/64951f6200a1fb72cc003c6e8226dde3d2ef591f"><code>64951f6</code></a> [Refactor] <code>parse</code>: extract key segment splitting helper</li> <li><a href="https://github.com/ljharb/qs/commit/e1bd2599cdff4c936ea52fb1f16f921cbe7aa88c"><code>e1bd259</code></a> [Dev Deps] update <code>@ljharb/eslint-config</code></li> <li><a href="https://github.com/ljharb/qs/commit/f4b3d39709fef6ddbd85128d1ba4c6b566c4902e"><code>f4b3d39</code></a> [eslint] add eslint 9 optional peer dep</li> <li><a href="https://github.com/ljharb/qs/commit/6e94d9596ca50dffafcef40a5f64eca89962cf34"><code>6e94d95</code></a> [Dev Deps] update <code>eslint</code>, <code>@ljharb/eslint-config</code>, <code>npmignore</code></li> <li><a href="https://github.com/ljharb/qs/commit/973dc3c51c86da9f4e30edeb4b1725158d439102"><code>973dc3c</code></a> [actions] add workflow permissions</li> <li>Additional commits viewable in <a href="https://github.com/ljharb/qs/compare/v6.13.0...v6.14.1">compare view</a></li> </ul> </details> <br /> Updates `body-parser` from 1.20.3 to 1.20.4 <details> <summary>Release notes</summary> <p><em>Sourced from <a href="https://github.com/expressjs/body-parser/releases">body-parser's releases</a>.</em></p> <blockquote> <h2>1.20.4</h2> <h2>What's Changed</h2> <ul> <li>Remove redundant depth check by <a href="https://github.com/blakeembrey"><code>@blakeembrey</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/538">expressjs/body-parser#538</a></li> <li>ci: add support for Node.js v23 by <a href="https://github.com/Phillip9587"><code>@Phillip9587</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/553">expressjs/body-parser#553</a></li> <li>ci: restore CI for 1.x branch by <a href="https://github.com/bjohansebas"><code>@bjohansebas</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/665">expressjs/body-parser#665</a></li> <li>deps: qs@^6.14.0 by <a href="https://github.com/bjohansebas"><code>@bjohansebas</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/664">expressjs/body-parser#664</a></li> <li>deps: use tilde notation and update certain dependencies by <a href="https://github.com/Phillip9587"><code>@Phillip9587</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/668">expressjs/body-parser#668</a></li> <li>chore: remove SECURITY.md by <a href="https://github.com/Phillip9587"><code>@Phillip9587</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/669">expressjs/body-parser#669</a></li> <li>ci: add CodeQL (SAST) by <a href="https://github.com/Phillip9587"><code>@Phillip9587</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/670">expressjs/body-parser#670</a></li> <li>Release: 1.20.4 by <a href="https://github.com/UlisesGascon"><code>@UlisesGascon</code></a> in <a href="https://redirect.github.com/expressjs/body-parser/pull/672">expressjs/body-parser#672</a></li> </ul> <p><strong>Full Changelog</strong>: <a href="https://github.com/expressjs/body-parser/compare/1.20.3...1.20.4">https://github.com/expressjs/body-parser/compare/1.20.3...1.20.4</a></p> </blockquote> </details> <details> <summary>Changelog</summary> <p><em>Sourced from <a href="https://github.com/expressjs/body-parser/blob/master/HISTORY.md">body-parser's changelog</a>.</em></p> <blockquote> <h1>1.20.4 / 2025-12-01</h1> <ul> <li>deps: qs@~6.14.0</li> <li>deps: use tilde notation for dependencies</li> <li>deps: http-errors@~2.0.1</li> <li>deps: raw-body@~2.5.3</li> </ul> </blockquote> </details> <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/expressjs/body-parser/commit/7db202cac84a001e6566c2dc6516b44db98beff3"><code>7db202c</code></a> 1.20.4 (<a href="https://redirect.github.com/expressjs/body-parser/issues/672">#672</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/d8f8adb898676dfdf997b4455e5f9b689b53e989"><code>d8f8adb</code></a> ci: add CodeQL (SAST) (<a href="https://redirect.github.com/expressjs/body-parser/issues/670">#670</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/6d133c19b3e7c0bb8301959ca1dba283d23d23c3"><code>6d133c1</code></a> chore: remove SECURITY.md (<a href="https://redirect.github.com/expressjs/body-parser/issues/669">#669</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/fcd15355041ada6f37288dd13858d50429016b66"><code>fcd1535</code></a> deps: use tilde notation and update certain dependencies (<a href="https://redirect.github.com/expressjs/body-parser/issues/668">#668</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/ec5fa290d25d85e0049757e240249072331eaee6"><code>ec5fa29</code></a> deps: qs@~6.14.0 (<a href="https://redirect.github.com/expressjs/body-parser/issues/664">#664</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/ffb95c12c7785ec6d3852ce46b8711ac74009252"><code>ffb95c1</code></a> ci: restore CI for 1.x branch (<a href="https://redirect.github.com/expressjs/body-parser/issues/665">#665</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/48a5f074a4db07066087ed8b6ff641825c9c03cf"><code>48a5f07</code></a> ci: add support for Node.js v23 (<a href="https://redirect.github.com/expressjs/body-parser/issues/553">#553</a>)</li> <li><a href="https://github.com/expressjs/body-parser/commit/f20f6adc7118cbf973e927d34bc0bbf2ff177459"><code>f20f6ad</code></a> Remove redundant depth check (<a href="https://redirect.github.com/expressjs/body-parser/issues/538">#538</a>)</li> <li>See full diff in <a href="https://github.com/expressjs/body-parser/compare/1.20.3...1.20.4">compare view</a></li> </ul> </details> <br /> Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting `@dependabot rebase`. [//]: # (dependabot-automerge-start) [//]: # (dependabot-automerge-end) --- <details> <summary>Dependabot commands and options</summary> <br /> You can trigger Dependabot actions by commenting on this PR: - `@dependabot rebase` will rebase this PR - `@dependabot recreate` will recreate this PR, overwriting any edits that have been made to it - `@dependabot merge` will merge this PR after your CI passes on it - `@dependabot squash and merge` will squash and merge this PR after your CI passes on it - `@dependabot cancel merge` will cancel a previously requested merge and block automerging - `@dependabot reopen` will reopen this PR if it is closed - `@dependabot close` will close this PR and stop Dependabot recreating it. You can achieve the same result by closing it manually - `@dependabot show <dependency name> ignore conditions` will show all of the ignore conditions of the specified dependency - `@dependabot ignore this major version` will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this minor version` will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself) - `@dependabot ignore this dependency` will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself) You can disable automated security fix PRs for this repo from the [Security Alerts page](https://github.com/microsoft/onnxruntime/network/alerts). </details> Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
- Fix QNN-EP to support static bias tensors with mismatched quantization encodings - Add requantization logic to ensure bias_scale = weights_scale x activation_scale ### Description The QNN-EP currently lacks support for static bias tensors with quantization encodings that don't match the expected mathematical relationship bias_scale = weights_scale × activation_scale. ### Motivation and Context This leads to HTP backend disregard the encodings of the bias tensor and recompute its own causing accuracy drops --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: qti-ashwshan <ashwshan@qti.qualcomm.com>
…t#26803) ### Description Adds new `OrtKernelInfo` APIs to enable plugin EP kernel implementations to retrieve operator metadata and access the parent EP instance. - Added four new KernelInfo APIs: `KernelInfo_GetOperatorDomain()`, `KernelInfo_GetOperatorType()`, `KernelInfo_GetOperatorSinceVersion()`, and `KernelInfo_GetEp()` - Extended example plugin EP to demonstrate accessing EP configuration from kernel implementations. ### Motivation and Context <!-- - Why is this change required? What problem does it solve? - If it fixes an open issue, please link to the issue here. --> --------- Co-authored-by: Edward Chen <18449977+edgchen1@users.noreply.github.com>
microsoft#26872) ### Description Use `file(GLOB ...)` to allow for the specified file to not exist. ### Motivation and Context If the QNN library is missing, using `list(APPEND ...)` will cause errors during the build process.
…soft#26917) ### Description Update search for GPG to check for both Program Files and Program Files (x86) ### Motivation and Context --------- Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## Summary This pull request aims to significantly reduce the build time for Flash Attention by removing support for less common head dimensions (160 and 224). It also includes a build option for quick build `--cmake_extra_defines onnxruntime_QUICK_BUILD=ON`, which will only build flash attention kernel for float16 and head dimension=128. That could speed up development. ## Key Changes ### 1. Flash Attention Build Optimization - **Removed Head Dimensions:** Deleted source files and kernel instantiations for head dimensions **160** and **224** (both FP16 and BF16). These dimensions are less frequently used, and removing them reduces the number of kernels to be compiled, thereby speeding up the build process. - **Updated Dispatch Logic:** Modified `static_switch.h` and `flash_api.h` to remove the dispatch cases for `kHeadDim = 160` and `kHeadDim = 224`. ### 2. Test Enhancements - **GQA Tests:** Updated `onnxruntime/test/python/transformers/test_gqa.py` to detect whether it is quick build package. If it is, only test supported data type (float16) and head dimension (128 only) for flash attention, and use `has_flash_attention(bf16=True)` when checking for Flash Attention availability in BF16 tests. This ensures that tests are skipped appropriately if BF16 kernels are not compiled/available. ## Impact - **Build Time:** Faster compilation of the CUDA provider due to fewer Flash Attention kernels. - **Functionality:** Head dimensions 160 and 224 are no longer supported for Flash Attention. Models using these specific head dimensions will fall back to next supported head dimension like 192 or 256. ## Verification - Validated that the build completes successfully with the reduced kernel set. - `test_gqa.py` should pass or skip correctly based on hardware support. - Build onnxruntime-gpu package with `--cmake_extra_defines onnxruntime_QUICK_BUILD=ON` option, and the build info has "quick-build=1", like the following python script: ```python import onnxruntime print(onnxruntime.get_build_info()) ``` The output is like ``` ORT Build Info: git-branch=main, git-commit-id=ecf164a945, quick-build=1, build type=Release ```
### Description This patch supports more `dim_inner` (up to 4096) for `Split-K` to optimize more models. This patch also enables `Split-K` on `gen-12lp`. ### Motivation and Context With this PR we can achieve about 30% improvement on `jina-clip-v1-text-fp16` and 20% improvement on `jina-embeddings-v2-base-code-fp16` on Lunar Lake iGPUs.
…26926) This pull request simplifies the logic for handling present key/value tensors in the WebGPU Flash Attention implementation. The main change is that the responsibility for creating internal present key/value tensors is moved from the caller to the `ApplyFlashAttention` function itself. This reduces code duplication and makes the API easier to use. Additionally, the `CanApplyFlashAttention` function is simplified to remove unnecessary checks for present key/value tensors.
### Description Address OOB reads and writes issues. ### Motivation and Context microsoft#11828 microsoft#13332 See also this: onnx/onnx#4294
…oft#26660) ### Description Introduced a new environment variable ORT_DEBUG_NODE_IO_PREPEND_EP_TO_FILE_NAME to control whether the Execution Provider (EP) type is prepended to tensor file names. Updated MakeTensorFileName() and related logic to include the EP type in the format: `<EPType>_<TensorName>` when the variable is set to a non-zero value. This change helps avoid file name conflicts and improves clarity when analyzing outputs from multi-EP runs. ### Motivation and Context Accuracy unit tests may run the same model multiple times using different EPs, such as ORT CPU and QNN HTP. Previously, dumped tensor files used identical names when tensor names matched across sessions, causing file overwrites. ### Example When ORT_DEBUG_NODE_IO_PREPEND_EP_TO_FILE_NAME=1, dumped tensors include EP-specific prefixes: ``` CPU_input.tensorproto CPU_input_token_0.tensorproto CPU_output.tensorproto QNN_constant.tensorproto QNN_input.tensorproto QNN_input_token_0.tensorproto QNN_output.tensorproto ```
### Description Update OV version to latest release in OVEP Co-authored-by: jatinwadhwa921 <jatin.wadhwa@intel.com>
ankitm3k
approved these changes
Jan 9, 2026
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
15 participants
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.
Synchronizing intel/onnxruntime ovep-develop branch with latest changes from microsoft/onnxruntime master branch.