dependabot[bot] opened a new pull request, #1491: URL: https://github.com/apache/opennlp/pull/1491
Bumps `onnxruntime.version` from 1.30.0 to 1.31.0. Updates `com.microsoft.onnxruntime:onnxruntime` from 1.30.0 to 1.31.0 <details> <summary>Release notes</summary> <p><em>Sourced from <a href="https://github.com/microsoft/onnxruntime/releases";>com.microsoft.onnxruntime:onnxruntime's releases</a>.</em></p> <blockquote> <h2>ONNX Runtime v1.31.0</h2> <p>ONNX Runtime 1.31.0 stabilizes model-package and EPContext data APIs, improves CPU model loading and quantized MoE inference, expands device-based execution-provider selection, and strengthens model-loading and runtime reliability. These notes cover changes since ONNX Runtime 1.30.0. CUDA and WebGPU kernel updates are summarized here; detailed provider notes are covered by their separate plugin EP releases.</p> <h2>Highlights</h2> <ul> <li>Promoted the model-package API and EPContext data callbacks to stable C and C++ APIs, with EPContext callback support added across language bindings (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li>Reduced CPU session-initialization overhead by parallelizing eligible weight prepacking, and accelerated block-wise INT4/INT8 QMoE experts with MLAS QNBit kernels and grouped expert dispatch (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31691";>#31691</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32644";>#32644</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32668";>#32668</a>).</li> <li>Added x86 FP16 LayerNorm/RMSNorm acceleration, Arm KleidiAI SVE2.1 FP16 GEMM support, and additional RISC-V vector kernels (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32715";>#32715</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32670";>#32670</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32710";>#32710</a>).</li> <li>Enabled CoreML participation in device-based EP selection, including Apple Neural Engine selection through <code>PREFER_NPU</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31975";>#31975</a>).</li> <li>Added workspace-memory accounting and verification for constrained-memory graph partitioning, alongside broader model, graph, and tensor validation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31962";>#31962</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32189";>#32189</a>).</li> </ul> <h2>Announcements & Compatibility</h2> <ul> <li><a href="https://github.com/microsoft/onnxruntime/releases/tag/plugin-ep-cuda/v0.3.0";>CUDA Plugin EP v0.3.0</a> and WebGPU Plugin EP v0.5.0 are matched with ONNX Runtime 1.31.0. For <strong>contrib-op compatibility</strong>, core and the plugin must use the same contributed-operator schemas; building both from the <strong>same ONNX Runtime revision</strong> is recommended. Schema mismatches can cause incorrect execution or crashes and may not be detected during plugin registration.</li> <li><strong><code>GatherBlockQuantized</code> behavior change:</strong> out-of-range indices now produce zero output slices across CPU, CUDA, and WebGPU for both integer and floating-point quantized formats. This deliberately differs from ONNX <code>Gather</code> error semantics (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32480";>#32480</a>).</li> <li><strong>Experimental API migration:</strong> model-package consumers should use <code>OrtApi::GetModelPackageApi</code> and the stable <code>OrtModelPackageApi</code> table. EPContext callback registration now uses stable APIs; the corresponding experimental lookup entries were removed. The stable native callback contract carries callback/state pointers, with payload limits enforced by application callbacks, bindings, or EPs rather than a native read-options policy (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li><strong>ACL EP is deprecated</strong> and will be removed in a future release. Source builds using <code>--use_acl</code> now receive a deprecation warning (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32564";>#32564</a>).</li> <li>The ONNX dependency remains at <strong>1.22.0</strong>. The initial ONNX 1.23 integration was reverted pending upstream fixes; this release does not ship that integration's new opset or FLOAT6 support (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33073";>#33073</a>).</li> <li>Supported native builds now default to <strong>1DS telemetry</strong>. Windows TraceLogging remains selectable with <code>--use_windows_telemetry</code> or <code>onnxruntime_USE_WINDOWS_TELEMETRY=ON</code>; telemetry can be disabled with <code>--no_telemetry</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32384";>#32384</a>).</li> </ul> <h2>New Features</h2> <h3>Core APIs & Runtime</h3> <ul> <li>Stabilized the model-package C and C++ APIs for loading packages, selecting compatible model variants, and creating sessions (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>).</li> <li>Stabilized named EPContext data read/write callbacks for application-managed storage. Added EP support negotiation so registered callbacks do not silently fall back to filesystem I/O (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li>Extended the Compile API to write external initializers and EP context data to buffers, enabling in-memory compilation workflows for models larger than 2 GB (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31347";>#31347</a>).</li> <li>Added workspace-memory accounting, source reporting, and post-partition reservation verification. Set <code>session.strict_workspace_verification=1</code> to fail initialization on workspace declaration overruns or nonzero orphaned reservations; strict verification is not supported for ORT-format model loads (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31962";>#31962</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32189";>#32189</a>).</li> <li>Added the opt-in source-build option <code>onnxruntime_DISABLE_DEVICE_DISCOVERY=ON</code> to skip platform accelerator probing and report only the CPU device (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32592";>#32592</a>).</li> </ul> <h3>Shared Contrib Operators</h3> <p>These updates affect core operator schemas and cross-provider behavior, not only the separately released GPU plugins.</p> <ul> <li>Extended <code>EngramGate</code> and <code>NGramHashMapping</code> for Qwen4-Exp, and added <code>VarlenNGramHashMapping</code> for packed DeepSeek Engram batching with request-local history. Packed hashing also supports Qwen4-Exp head offsets, EOS resets, and segment boundaries (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32285";>#32285</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32358";>#32358</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32467";>#32467</a>).</li> <li>Added the <code>activation</code> attribute to <code>GatedRMSNorm</code> for SiLU, its Swish alias, and Sigmoid gating across CPU, CUDA, and WebGPU. SiLU remains the default (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32512";>#32512</a>).</li> <li>Added weightless hyper-connection operators <code>BranchwiseRMSNorm</code>, <code>ScaledSiLU</code>, <code>HyperConnectionPreMix</code>, and <code>HyperConnectionPostMix</code>, with CPU, CUDA, and WebGPU implementations (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32687";>#32687</a>).</li> <li>Extended <code>GatherBlockQuantized</code> to FP8/FP4 data, full-axis blocks, and scale broadcasting, and improved CPU UINT8 dequantization. See the out-of-range-index behavior change above (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32480";>#32480</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32621";>#32621</a>).</li> </ul> <h3>Plugin Execution Provider Infrastructure</h3> <ul> <li>Added support for statically linked plugin EPs, with a static WebGPU plugin build configuration (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32395";>#32395</a>).</li> <li>Preserved caller streams in single-tensor plugin data transfers and handled null plugin allocator returns (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32666";>#32666</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32683";>#32683</a>).</li> <li>Deduplicated AutoEP custom-op domains and corrected EP assignment reporting after NHWC layout transformation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32711";>#32711</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/33184";>#33184</a>).</li> <li>Added reference guidance for compiled-model compatibility validation, including multi-device scenarios (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29168";>#29168</a>).</li> </ul> <h2>Execution Provider Updates</h2> <h3>CUDA EP</h3> <!-- raw HTML omitted --> </blockquote> <p>... (truncated)</p> </details> <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/microsoft/onnxruntime/commit/402979ac1bef89dcd55add41eaa7277c407599eb";><code>402979a</code></a> Revert "Bump version to 1.32.0 (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33197";>#33197</a>)" (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33212";>#33212</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/660125b35916c134c00bc746c34503321e17150f";><code>660125b</code></a> [Compile API] Add support for writing external initializers and EP context to...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/06cab16f5727506b6013ff7f77264e7ba5607651";><code>06cab16</code></a> Update parameterized requirement from >=0.8.1 to >=0.9.0 (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33022";>#33022</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/864d87126864c64651f58a9b51d9006ad71b2c78";><code>864d871</code></a> Fix isolated provider tests in Windows GPU CI (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33203";>#33203</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/174470241c9c62da79f34c80ab56826c0844ae7f";><code>1744702</code></a> Bump baseline-browser-mapping from 2.10.31 to 2.11.26 in /js/react_native/e2e...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/be3b8e8864f295f95c9a8d4092bd4b6fa2e9de8e";><code>be3b8e8</code></a> Fix complex STFT window handling (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33136";>#33136</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/40a7167d5f7c4de069ea898189db6c53975ea094";><code>40a7167</code></a> [CPU] Avoid degenerate GQA FlashAttention tiles when L2 cache size is unknown...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/430f563dec9c7c15db5b0e867f542bf41f2cc11f";><code>430f563</code></a> Fix EP assignment reporting after NHWC layout transformation (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33184";>#33184</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/26ab6c7951314bbfe82d80e36cd6ad49725fc2de";><code>26ab6c7</code></a> Improve GQA shared-cache and sliding-window handling (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33175";>#33175</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/d8024cdcd35022b184504ad55d019bb69443ec1d";><code>d8024cd</code></a> Improve CUDA RMSNorm reduction and aligned half2 output (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33178";>#33178</a>)</li> <li>Additional commits viewable in <a href="https://github.com/microsoft/onnxruntime/compare/v1.30.0...v1.31.0";>compare view</a></li> </ul> </details> <br /> Updates `com.microsoft.onnxruntime:onnxruntime_gpu` from 1.30.0 to 1.31.0 <details> <summary>Release notes</summary> <p><em>Sourced from <a href="https://github.com/microsoft/onnxruntime/releases";>com.microsoft.onnxruntime:onnxruntime_gpu's releases</a>.</em></p> <blockquote> <h2>ONNX Runtime v1.31.0</h2> <p>ONNX Runtime 1.31.0 stabilizes model-package and EPContext data APIs, improves CPU model loading and quantized MoE inference, expands device-based execution-provider selection, and strengthens model-loading and runtime reliability. These notes cover changes since ONNX Runtime 1.30.0. CUDA and WebGPU kernel updates are summarized here; detailed provider notes are covered by their separate plugin EP releases.</p> <h2>Highlights</h2> <ul> <li>Promoted the model-package API and EPContext data callbacks to stable C and C++ APIs, with EPContext callback support added across language bindings (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li>Reduced CPU session-initialization overhead by parallelizing eligible weight prepacking, and accelerated block-wise INT4/INT8 QMoE experts with MLAS QNBit kernels and grouped expert dispatch (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31691";>#31691</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32644";>#32644</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32668";>#32668</a>).</li> <li>Added x86 FP16 LayerNorm/RMSNorm acceleration, Arm KleidiAI SVE2.1 FP16 GEMM support, and additional RISC-V vector kernels (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32715";>#32715</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32670";>#32670</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32710";>#32710</a>).</li> <li>Enabled CoreML participation in device-based EP selection, including Apple Neural Engine selection through <code>PREFER_NPU</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31975";>#31975</a>).</li> <li>Added workspace-memory accounting and verification for constrained-memory graph partitioning, alongside broader model, graph, and tensor validation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31962";>#31962</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32189";>#32189</a>).</li> </ul> <h2>Announcements & Compatibility</h2> <ul> <li><a href="https://github.com/microsoft/onnxruntime/releases/tag/plugin-ep-cuda/v0.3.0";>CUDA Plugin EP v0.3.0</a> and WebGPU Plugin EP v0.5.0 are matched with ONNX Runtime 1.31.0. For <strong>contrib-op compatibility</strong>, core and the plugin must use the same contributed-operator schemas; building both from the <strong>same ONNX Runtime revision</strong> is recommended. Schema mismatches can cause incorrect execution or crashes and may not be detected during plugin registration.</li> <li><strong><code>GatherBlockQuantized</code> behavior change:</strong> out-of-range indices now produce zero output slices across CPU, CUDA, and WebGPU for both integer and floating-point quantized formats. This deliberately differs from ONNX <code>Gather</code> error semantics (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32480";>#32480</a>).</li> <li><strong>Experimental API migration:</strong> model-package consumers should use <code>OrtApi::GetModelPackageApi</code> and the stable <code>OrtModelPackageApi</code> table. EPContext callback registration now uses stable APIs; the corresponding experimental lookup entries were removed. The stable native callback contract carries callback/state pointers, with payload limits enforced by application callbacks, bindings, or EPs rather than a native read-options policy (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li><strong>ACL EP is deprecated</strong> and will be removed in a future release. Source builds using <code>--use_acl</code> now receive a deprecation warning (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32564";>#32564</a>).</li> <li>The ONNX dependency remains at <strong>1.22.0</strong>. The initial ONNX 1.23 integration was reverted pending upstream fixes; this release does not ship that integration's new opset or FLOAT6 support (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33073";>#33073</a>).</li> <li>Supported native builds now default to <strong>1DS telemetry</strong>. Windows TraceLogging remains selectable with <code>--use_windows_telemetry</code> or <code>onnxruntime_USE_WINDOWS_TELEMETRY=ON</code>; telemetry can be disabled with <code>--no_telemetry</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32384";>#32384</a>).</li> </ul> <h2>New Features</h2> <h3>Core APIs & Runtime</h3> <ul> <li>Stabilized the model-package C and C++ APIs for loading packages, selecting compatible model variants, and creating sessions (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/33166";>#33166</a>).</li> <li>Stabilized named EPContext data read/write callbacks for application-managed storage. Added EP support negotiation so registered callbacks do not silently fall back to filesystem I/O (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32265";>#32265</a>).</li> <li>Extended the Compile API to write external initializers and EP context data to buffers, enabling in-memory compilation workflows for models larger than 2 GB (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31347";>#31347</a>).</li> <li>Added workspace-memory accounting, source reporting, and post-partition reservation verification. Set <code>session.strict_workspace_verification=1</code> to fail initialization on workspace declaration overruns or nonzero orphaned reservations; strict verification is not supported for ORT-format model loads (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31962";>#31962</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32189";>#32189</a>).</li> <li>Added the opt-in source-build option <code>onnxruntime_DISABLE_DEVICE_DISCOVERY=ON</code> to skip platform accelerator probing and report only the CPU device (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32592";>#32592</a>).</li> </ul> <h3>Shared Contrib Operators</h3> <p>These updates affect core operator schemas and cross-provider behavior, not only the separately released GPU plugins.</p> <ul> <li>Extended <code>EngramGate</code> and <code>NGramHashMapping</code> for Qwen4-Exp, and added <code>VarlenNGramHashMapping</code> for packed DeepSeek Engram batching with request-local history. Packed hashing also supports Qwen4-Exp head offsets, EOS resets, and segment boundaries (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32285";>#32285</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32358";>#32358</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32467";>#32467</a>).</li> <li>Added the <code>activation</code> attribute to <code>GatedRMSNorm</code> for SiLU, its Swish alias, and Sigmoid gating across CPU, CUDA, and WebGPU. SiLU remains the default (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32512";>#32512</a>).</li> <li>Added weightless hyper-connection operators <code>BranchwiseRMSNorm</code>, <code>ScaledSiLU</code>, <code>HyperConnectionPreMix</code>, and <code>HyperConnectionPostMix</code>, with CPU, CUDA, and WebGPU implementations (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32687";>#32687</a>).</li> <li>Extended <code>GatherBlockQuantized</code> to FP8/FP4 data, full-axis blocks, and scale broadcasting, and improved CPU UINT8 dequantization. See the out-of-range-index behavior change above (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32480";>#32480</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32621";>#32621</a>).</li> </ul> <h3>Plugin Execution Provider Infrastructure</h3> <ul> <li>Added support for statically linked plugin EPs, with a static WebGPU plugin build configuration (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32395";>#32395</a>).</li> <li>Preserved caller streams in single-tensor plugin data transfers and handled null plugin allocator returns (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32666";>#32666</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32683";>#32683</a>).</li> <li>Deduplicated AutoEP custom-op domains and corrected EP assignment reporting after NHWC layout transformation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32711";>#32711</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/33184";>#33184</a>).</li> <li>Added reference guidance for compiled-model compatibility validation, including multi-device scenarios (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29168";>#29168</a>).</li> </ul> <h2>Execution Provider Updates</h2> <h3>CUDA EP</h3> <!-- raw HTML omitted --> </blockquote> <p>... (truncated)</p> </details> <details> <summary>Commits</summary> <ul> <li><a href="https://github.com/microsoft/onnxruntime/commit/402979ac1bef89dcd55add41eaa7277c407599eb";><code>402979a</code></a> Revert "Bump version to 1.32.0 (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33197";>#33197</a>)" (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33212";>#33212</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/660125b35916c134c00bc746c34503321e17150f";><code>660125b</code></a> [Compile API] Add support for writing external initializers and EP context to...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/06cab16f5727506b6013ff7f77264e7ba5607651";><code>06cab16</code></a> Update parameterized requirement from >=0.8.1 to >=0.9.0 (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33022";>#33022</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/864d87126864c64651f58a9b51d9006ad71b2c78";><code>864d871</code></a> Fix isolated provider tests in Windows GPU CI (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33203";>#33203</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/174470241c9c62da79f34c80ab56826c0844ae7f";><code>1744702</code></a> Bump baseline-browser-mapping from 2.10.31 to 2.11.26 in /js/react_native/e2e...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/be3b8e8864f295f95c9a8d4092bd4b6fa2e9de8e";><code>be3b8e8</code></a> Fix complex STFT window handling (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33136";>#33136</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/40a7167d5f7c4de069ea898189db6c53975ea094";><code>40a7167</code></a> [CPU] Avoid degenerate GQA FlashAttention tiles when L2 cache size is unknown...</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/430f563dec9c7c15db5b0e867f542bf41f2cc11f";><code>430f563</code></a> Fix EP assignment reporting after NHWC layout transformation (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33184";>#33184</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/26ab6c7951314bbfe82d80e36cd6ad49725fc2de";><code>26ab6c7</code></a> Improve GQA shared-cache and sliding-window handling (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33175";>#33175</a>)</li> <li><a href="https://github.com/microsoft/onnxruntime/commit/d8024cdcd35022b184504ad55d019bb69443ec1d";><code>d8024cd</code></a> Improve CUDA RMSNorm reduction and aligned half2 output (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/33178";>#33178</a>)</li> <li>Additional commits viewable in <a href="https://github.com/microsoft/onnxruntime/compare/v1.30.0...v1.31.0";>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 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) </details> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
