You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@opennlp.apache.org by "dependabot[bot] (via GitHub)" <gi...@apache.org> on 2026/09/21 04:14:44 UTC
[PR] [2.x]: Bump onnxruntime.version from 1.29.0 to 1.30.0 (opennlp)
dependabot[bot] opened a new pull request, #1327:
URL: https://github.com/apache/opennlp/pull/1327
Bumps `onnxruntime.version` from 1.29.0 to 1.30.0.
Updates `com.microsoft.onnxruntime:onnxruntime` from 1.29.0 to 1.30.0
<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 v1.30.0</h2>
<p>ONNX Runtime 1.30.0 expands generative AI inference, improves CPU and GPU performance, adds Go bindings, and strengthens runtime reliability. These notes cover changes since ONNX Runtime 1.29.1.</p>
<h2>Highlights</h2>
<ul>
<li>Expanded CUDA inference support with variable-length causal convolution for continuous batching, speculative decoding in paged XQA, and INT4 paged KV caches with per-channel scales (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32168">#32168</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32340">#32340</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32515">#32515</a>).</li>
<li>Improved WebGPU PagedAttention, added GPT-OSS support and INT8 KV-cache block quantization, and extended convolution optimizations (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31727">#31727</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32277">#32277</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32284">#32284</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32420">#32420</a>).</li>
<li>Added fused CPU LinearAttention kernels for AVX-512, Arm64 NEON, and SVE, plus AVX2 LayerNorm/RMSNorm acceleration (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31674">#31674</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31973">#31973</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32178">#32178</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32356">#32356</a>).</li>
<li>Added Go bindings for the ONNX Runtime C API and DeepSeek Engram contrib operators (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29615">#29615</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32268">#32268</a>).</li>
</ul>
<h2>Announcements & Compatibility</h2>
<ul>
<li>FP4 QMoE kernels are now enabled by default in CUDA builds, with Windows build support added in this release. Source builds can opt out with <code>-Donnxruntime_USE_FP4_QMOE=OFF</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32096">#32096</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32163">#32163</a>).</li>
<li>CUDA fpA-intB builds now default to a compact kernel set for FP16 activations, INT4/INT8 weights, scale-only quantization, and <code>block_size=32</code>. Set <code>-Donnxruntime_USE_FPA_INTB_GEMM_FULL=ON</code> when building from source to retain the full kernel set, including BF16, zero-point, bias, larger-block-size, and native Hopper variants (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32324">#32324</a>).</li>
<li>CPU FP16 <code>Gemm</code> and <code>MatMul</code> execution is gated on hardware acceleration. CPU-assigned FP16 nodes without a matching kernel now fall back to FP32 (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32301">#32301</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32197">#32197</a>).</li>
<li>WebGPU plugin EP packaging now supports Linux AArch64. Plugin versions were advanced to WebGPU 0.4.0 and CUDA 0.2 (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32287">#32287</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31960">#31960</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31970">#31970</a>).</li>
</ul>
<h2>Security & Reliability</h2>
<h3>Model Loading, Memory, and Input Validation</h3>
<ul>
<li>Limited nested model-graph depth and canonicalized external-data locations to harden model loading (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32344">#32344</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32135">#32135</a>).</li>
<li>Added checked rounding for BFC arena allocations and fixed prepacked-weight reference lifetimes (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32010">#32010</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32040">#32040</a>).</li>
<li>Strengthened shape, rank, and parameter validation for <code>Split</code>, <code>Scan</code>, <code>GatherND</code>, <code>ScatterND</code>, <code>SpaceToDepth</code>/<code>DepthToSpace</code>, <code>Crop</code>, <code>Conv</code>, <code>Normalizer</code>, and pooling (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29461">#29461</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31668">#31668</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32034">#32034</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32039">#32039</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32076">#32076</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32157">#32157</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32160">#32160</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32161">#32161</a>, <a href="https://redirect.github.com/microsoft/onnxruntim
e/pull/32345">#32345</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32349">#32349</a>).</li>
<li>Hardened generation and attention input handling, including attention-attribute narrowing, <code>BifurcationDetector</code> inputs, generation subgraph shapes, and QEmbed segment inputs. BeamSearch buffer expansion now uses dynamic shape storage (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31648">#31648</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31701">#31701</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32009">#32009</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32078">#32078</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32144">#32144</a>).</li>
<li>Validated <code>TreeEnsemble</code> node references and bounded subtree comparison, rejected non-finite CPU <code>RoiAlign</code> coordinates, and required <code>ImageScaler</code> bias to match the channel count (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32031">#32031</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32043">#32043</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32011">#32011</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32002">#32002</a>).</li>
<li>Added an allowlist of safe LoRA adapter parameter data types, validated <code>MatMulFpQ4</code> shape inputs, and checked MLAS blockwise quantization/dequantization index ranges (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31682">#31682</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32032">#32032</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32007">#32007</a>).</li>
</ul>
<h3>GPU Bounds and Resource Lifetimes</h3>
<ul>
<li>Hardened CUDA indexing and buffer-size arithmetic in <code>MatMulNBits</code>, <code>RemovePadding</code>, <code>RotaryEmbedding</code>, <code>SparseAttention</code>, Whisper beam search, NMS, QDQ, and <code>GatherElements</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31643">#31643</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31994">#31994</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31995">#31995</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31996">#31996</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31998">#31998</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32014">#32014</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32029">#32029</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32030">#32030</a>).</li>
<li>Fixed overflow in CUDA reduction scans and Softmax offset arithmetic, and handled zero-sized outputs in CUDA random-generator kernels (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32137">#32137</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32330">#32330</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31997">#31997</a>).</li>
<li>Fixed CUDA MultiHeadAttention shared-cache scratch lifetimes and kept <code>CudaAsyncBuffer</code> staging storage alive across CUDA graph replay (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31968">#31968</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32121">#32121</a>).</li>
<li>Fixed WebGPU out-of-bounds subgroup-matrix loads for partial tiles, zero-initialized writable device-allocator buffers, and rejected foreign GPU handles in built-in data transfers (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32364">#32364</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32063">#32063</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32317">#32317</a>).</li>
</ul>
<h3>Dependencies and Tooling</h3>
<ul>
<li>Upgraded Protobuf to 33.6 and refreshed Python documentation dependencies, including an ONNX security-related update (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29906">#29906</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32190">#32190</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32424">#32424</a>).</li>
<li>Updated JavaScript dependencies including <code>js-yaml</code>, <code>joi</code>, <code>fast-uri</code>, and the Next.js end-to-end fixture (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32397">#32397</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32486">#32486</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32488">#32488</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32505">#32505</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32508">#32508</a>).</li>
<li>Pinned GitHub Actions to full-length commit SHAs and strengthened packaging infrastructure with authenticated package feeds and NPM network isolation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32176">#32176</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32005">#32005</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32440">#32440</a>).</li>
</ul>
<h2>New Features</h2>
<h3>Core APIs & Runtime</h3>
<ul>
<li>Added Go bindings for the ONNX Runtime C API (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29615">#29615</a>).</li>
<li>Extended memory importing with host-pointer support and added access to preallocated outputs through <code>KernelContext::GetPreallocatedOutput</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29726">#29726</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32089">#32089</a>).</li>
<li>Added packed-attention workspace recipes and estimates, and made workspace input-shape handling aware of optional inputs (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32283">#32283</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32321">#32321</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32312">#32312</a>).</li>
<li>Added DeepSeek Engram contrib operators, <code>EngramGate</code> and <code>NGramHashMapping</code>, and expanded kernel coverage for Qwen-3.5 operators (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32268">#32268</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32106">#32106</a>).</li>
</ul>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a href="https://github.com/microsoft/onnxruntime/commit/f2c39fe2f838cf35ce7da92824f5a5e3ee6e88a7"><code>f2c39fe</code></a> [CUDA] Add INT4 paged KV cache with per-channel scales (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32515">#32515</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/5894ba8a53526b3aeb702bd86e6b117c9df8fa4f"><code>5894ba8</code></a> Add portable random-access file reads to Env (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32503">#32503</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/a2ee3eb43052de307003e6256fc9258ce19d9c34"><code>a2ee3eb</code></a> Fix CUDA plugin device discovery on WSL (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32517">#32517</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/b652e595b04d202b7f71af1d9105a183fac2b100"><code>b652e59</code></a> [WebGPU] Prepack Conv weights for the im2col-matmul path (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32420">#32420</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/0f0f29f5015f16926912ae47061e6a3eedcfbe04"><code>0f0f29f</code></a> Get rid of spurious warning about not being able to find spectre mitigation (...</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/23dd6510fd395b4556f474c1c0079be2a8b7a8c8"><code>23dd651</code></a> Register ONNX schemas only when static registration is disabled (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32353">#32353</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/33af5d32801b2e631ebece4f382f4ba775f17d8c"><code>33af5d3</code></a> Release external data loaders after graph initialization (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32502">#32502</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/2e3c24d5963d62d0bd7c9d306433b19b7304d5d2"><code>2e3c24d</code></a> Clarify external initializer and EP context path interaction (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32442">#32442</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/e76036b8e49515ee7dd2dd0ac39c9d8b7533d38a"><code>e76036b</code></a> [CUDA] Pin FP8 GEMV residency for grids just past two blocks per SM (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32433">#32433</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/82583c5b6398a9c37815f91e9cd1d7c165a132a7"><code>82583c5</code></a> Add session option for a BNHS GroupQueryAttention Value cache layout (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32139">#32139</a>)</li>
<li>Additional commits viewable in <a href="https://github.com/microsoft/onnxruntime/compare/v1.29.0...v1.30.0">compare view</a></li>
</ul>
</details>
<br />
Updates `com.microsoft.onnxruntime:onnxruntime_gpu` from 1.29.0 to 1.30.0
<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 v1.30.0</h2>
<p>ONNX Runtime 1.30.0 expands generative AI inference, improves CPU and GPU performance, adds Go bindings, and strengthens runtime reliability. These notes cover changes since ONNX Runtime 1.29.1.</p>
<h2>Highlights</h2>
<ul>
<li>Expanded CUDA inference support with variable-length causal convolution for continuous batching, speculative decoding in paged XQA, and INT4 paged KV caches with per-channel scales (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32168">#32168</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32340">#32340</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32515">#32515</a>).</li>
<li>Improved WebGPU PagedAttention, added GPT-OSS support and INT8 KV-cache block quantization, and extended convolution optimizations (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31727">#31727</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32277">#32277</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32284">#32284</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32420">#32420</a>).</li>
<li>Added fused CPU LinearAttention kernels for AVX-512, Arm64 NEON, and SVE, plus AVX2 LayerNorm/RMSNorm acceleration (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31674">#31674</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31973">#31973</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32178">#32178</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32356">#32356</a>).</li>
<li>Added Go bindings for the ONNX Runtime C API and DeepSeek Engram contrib operators (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29615">#29615</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32268">#32268</a>).</li>
</ul>
<h2>Announcements & Compatibility</h2>
<ul>
<li>FP4 QMoE kernels are now enabled by default in CUDA builds, with Windows build support added in this release. Source builds can opt out with <code>-Donnxruntime_USE_FP4_QMOE=OFF</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32096">#32096</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32163">#32163</a>).</li>
<li>CUDA fpA-intB builds now default to a compact kernel set for FP16 activations, INT4/INT8 weights, scale-only quantization, and <code>block_size=32</code>. Set <code>-Donnxruntime_USE_FPA_INTB_GEMM_FULL=ON</code> when building from source to retain the full kernel set, including BF16, zero-point, bias, larger-block-size, and native Hopper variants (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32324">#32324</a>).</li>
<li>CPU FP16 <code>Gemm</code> and <code>MatMul</code> execution is gated on hardware acceleration. CPU-assigned FP16 nodes without a matching kernel now fall back to FP32 (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32301">#32301</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32197">#32197</a>).</li>
<li>WebGPU plugin EP packaging now supports Linux AArch64. Plugin versions were advanced to WebGPU 0.4.0 and CUDA 0.2 (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32287">#32287</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31960">#31960</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31970">#31970</a>).</li>
</ul>
<h2>Security & Reliability</h2>
<h3>Model Loading, Memory, and Input Validation</h3>
<ul>
<li>Limited nested model-graph depth and canonicalized external-data locations to harden model loading (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32344">#32344</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32135">#32135</a>).</li>
<li>Added checked rounding for BFC arena allocations and fixed prepacked-weight reference lifetimes (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32010">#32010</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32040">#32040</a>).</li>
<li>Strengthened shape, rank, and parameter validation for <code>Split</code>, <code>Scan</code>, <code>GatherND</code>, <code>ScatterND</code>, <code>SpaceToDepth</code>/<code>DepthToSpace</code>, <code>Crop</code>, <code>Conv</code>, <code>Normalizer</code>, and pooling (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29461">#29461</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31668">#31668</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32034">#32034</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32039">#32039</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32076">#32076</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32157">#32157</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32160">#32160</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32161">#32161</a>, <a href="https://redirect.github.com/microsoft/onnxruntim
e/pull/32345">#32345</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32349">#32349</a>).</li>
<li>Hardened generation and attention input handling, including attention-attribute narrowing, <code>BifurcationDetector</code> inputs, generation subgraph shapes, and QEmbed segment inputs. BeamSearch buffer expansion now uses dynamic shape storage (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31648">#31648</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31701">#31701</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32009">#32009</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32078">#32078</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32144">#32144</a>).</li>
<li>Validated <code>TreeEnsemble</code> node references and bounded subtree comparison, rejected non-finite CPU <code>RoiAlign</code> coordinates, and required <code>ImageScaler</code> bias to match the channel count (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32031">#32031</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32043">#32043</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32011">#32011</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32002">#32002</a>).</li>
<li>Added an allowlist of safe LoRA adapter parameter data types, validated <code>MatMulFpQ4</code> shape inputs, and checked MLAS blockwise quantization/dequantization index ranges (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31682">#31682</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32032">#32032</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32007">#32007</a>).</li>
</ul>
<h3>GPU Bounds and Resource Lifetimes</h3>
<ul>
<li>Hardened CUDA indexing and buffer-size arithmetic in <code>MatMulNBits</code>, <code>RemovePadding</code>, <code>RotaryEmbedding</code>, <code>SparseAttention</code>, Whisper beam search, NMS, QDQ, and <code>GatherElements</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31643">#31643</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31994">#31994</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31995">#31995</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31996">#31996</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31998">#31998</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32014">#32014</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32029">#32029</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32030">#32030</a>).</li>
<li>Fixed overflow in CUDA reduction scans and Softmax offset arithmetic, and handled zero-sized outputs in CUDA random-generator kernels (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32137">#32137</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32330">#32330</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/31997">#31997</a>).</li>
<li>Fixed CUDA MultiHeadAttention shared-cache scratch lifetimes and kept <code>CudaAsyncBuffer</code> staging storage alive across CUDA graph replay (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/31968">#31968</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32121">#32121</a>).</li>
<li>Fixed WebGPU out-of-bounds subgroup-matrix loads for partial tiles, zero-initialized writable device-allocator buffers, and rejected foreign GPU handles in built-in data transfers (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32364">#32364</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32063">#32063</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32317">#32317</a>).</li>
</ul>
<h3>Dependencies and Tooling</h3>
<ul>
<li>Upgraded Protobuf to 33.6 and refreshed Python documentation dependencies, including an ONNX security-related update (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29906">#29906</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32190">#32190</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32424">#32424</a>).</li>
<li>Updated JavaScript dependencies including <code>js-yaml</code>, <code>joi</code>, <code>fast-uri</code>, and the Next.js end-to-end fixture (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32397">#32397</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32486">#32486</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32488">#32488</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32505">#32505</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32508">#32508</a>).</li>
<li>Pinned GitHub Actions to full-length commit SHAs and strengthened packaging infrastructure with authenticated package feeds and NPM network isolation (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32176">#32176</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32005">#32005</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32440">#32440</a>).</li>
</ul>
<h2>New Features</h2>
<h3>Core APIs & Runtime</h3>
<ul>
<li>Added Go bindings for the ONNX Runtime C API (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29615">#29615</a>).</li>
<li>Extended memory importing with host-pointer support and added access to preallocated outputs through <code>KernelContext::GetPreallocatedOutput</code> (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/29726">#29726</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32089">#32089</a>).</li>
<li>Added packed-attention workspace recipes and estimates, and made workspace input-shape handling aware of optional inputs (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32283">#32283</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32321">#32321</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32312">#32312</a>).</li>
<li>Added DeepSeek Engram contrib operators, <code>EngramGate</code> and <code>NGramHashMapping</code>, and expanded kernel coverage for Qwen-3.5 operators (<a href="https://redirect.github.com/microsoft/onnxruntime/pull/32268">#32268</a>, <a href="https://redirect.github.com/microsoft/onnxruntime/pull/32106">#32106</a>).</li>
</ul>
<!-- raw HTML omitted -->
</blockquote>
<p>... (truncated)</p>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a href="https://github.com/microsoft/onnxruntime/commit/f2c39fe2f838cf35ce7da92824f5a5e3ee6e88a7"><code>f2c39fe</code></a> [CUDA] Add INT4 paged KV cache with per-channel scales (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32515">#32515</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/5894ba8a53526b3aeb702bd86e6b117c9df8fa4f"><code>5894ba8</code></a> Add portable random-access file reads to Env (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32503">#32503</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/a2ee3eb43052de307003e6256fc9258ce19d9c34"><code>a2ee3eb</code></a> Fix CUDA plugin device discovery on WSL (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32517">#32517</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/b652e595b04d202b7f71af1d9105a183fac2b100"><code>b652e59</code></a> [WebGPU] Prepack Conv weights for the im2col-matmul path (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32420">#32420</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/0f0f29f5015f16926912ae47061e6a3eedcfbe04"><code>0f0f29f</code></a> Get rid of spurious warning about not being able to find spectre mitigation (...</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/23dd6510fd395b4556f474c1c0079be2a8b7a8c8"><code>23dd651</code></a> Register ONNX schemas only when static registration is disabled (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32353">#32353</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/33af5d32801b2e631ebece4f382f4ba775f17d8c"><code>33af5d3</code></a> Release external data loaders after graph initialization (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32502">#32502</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/2e3c24d5963d62d0bd7c9d306433b19b7304d5d2"><code>2e3c24d</code></a> Clarify external initializer and EP context path interaction (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32442">#32442</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/e76036b8e49515ee7dd2dd0ac39c9d8b7533d38a"><code>e76036b</code></a> [CUDA] Pin FP8 GEMV residency for grids just past two blocks per SM (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32433">#32433</a>)</li>
<li><a href="https://github.com/microsoft/onnxruntime/commit/82583c5b6398a9c37815f91e9cd1d7c165a132a7"><code>82583c5</code></a> Add session option for a BNHS GroupQueryAttention Value cache layout (<a href="https://redirect.github.com/microsoft/onnxruntime/issues/32139">#32139</a>)</li>
<li>Additional commits viewable in <a href="https://github.com/microsoft/onnxruntime/compare/v1.29.0...v1.30.0">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: commits-unsubscribe@opennlp.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [2.x]: Bump onnxruntime.version from 1.29.0 to 1.30.0 (opennlp)
Posted by "krickert (via GitHub)" <gi...@apache.org>.
krickert commented on code in PR #1327:
URL: https://github.com/apache/opennlp/pull/1327#discussion_r4061602135
##########
pom.xml:
##########
@@ -191,7 +191,7 @@
<junit5-system-exit.version>2.0.3</junit5-system-exit.version>
<uimaj.version>3.6.1</uimaj.version>
<morfologik.version>2.2.0</morfologik.version>
- <onnxruntime.version>1.29.0</onnxruntime.version>
+ <onnxruntime.version>1.30.0</onnxruntime.version>
Review Comment:
Tested locally - works (minus any known issues already filed)
--
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: commits-unsubscribe@opennlp.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
Re: [PR] [2.x]: Bump onnxruntime.version from 1.29.0 to 1.30.0 (opennlp)
Posted by "mawiesne (via GitHub)" <gi...@apache.org>.
mawiesne merged PR #1327:
URL: https://github.com/apache/opennlp/pull/1327
--
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: commits-unsubscribe@opennlp.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org