Pull Requests microsoft/onnxruntime

Fix plugin EP allocator deleter lifetime

#29663 opened 2026-07-11 00:58 by yen-shi

Migrate transformers fusions to onnx-ir + onnxscript rewriter (QuickGelu, BiasAdd, GroupNorm)

#29662 opened 2026-07-10 23:57 by justinchuby

Add duplicate domain dedup check in AddCustomOpDomain release:1.28.0

#29661 opened 2026-07-10 23:31 by chilo-ms

[QNN-EP] Add reshape handler for split axes reshape

#29660 opened 2026-07-10 22:36 by quic-muchhsu

[CUDA] Fix null allocator passed to plugin EP kernel PrePack release:1.28.0

#29658 opened 2026-07-10 18:22 by tianleiwu

[MLAS] Integrate Q4 asymmetric KleidiAI kernels into MatMulNBits

#29654 opened 2026-07-10 13:47 by martin-klacer-arm

[js/webgpu] Gate subgroup feature on Electron+Intel to prevent conv mispredictions

#29653 opened 2026-07-10 12:41 by Copilot

[DO NOT MERGE]: Loosen tolerance for QKV fusion tests on WebGPU

#29650 opened 2026-07-10 02:43 by hariharans29

Validate CPU kernel registrations at compile time

#29649 opened 2026-07-10 00:09 by GopalakrishnanN

Move PropagateCastOps metadata to compile time

#29648 opened 2026-07-09 23:44 by GopalakrishnanN

Make CPU fallback type policy compile-time

#29647 opened 2026-07-09 23:21 by GopalakrishnanN

Validate tensor element enum conversion at compile time

#29646 opened 2026-07-09 22:42 by GopalakrishnanN

implements online softmax for webgpu ep

#29645 opened 2026-07-09 22:22 by prathikr

Make TensorProto element sizes compile-time

#29644 opened 2026-07-09 22:22 by GopalakrishnanN

Make matrix layout invariants compile-time

#29643 opened 2026-07-09 21:47 by GopalakrishnanN

Support cross-compiling onnxruntime for Windows from a Linux host

#29642 opened 2026-07-09 21:09 by Penguinwizzard

Deduplicate ONNX graph navigation helpers

#29641 opened 2026-07-09 21:07 by GopalakrishnanN

feat(ci): ccache and vcpkg cache for ADO pipelines

#29639 opened 2026-07-09 19:55 by sanaa-hamel-microsoft

[CUDA] Fix MaxPool wrong values/indices for dilation>1 with non-zero begin padding ep:CUDA

#29638 opened 2026-07-09 18:00 by Copilot

Use the latest version of AnyChart in the perf tool

#29636 opened 2026-07-09 09:48 by jacquerie

Fix structure of EpLibraryPlugin::Load so that we call Unload on error.

#29634 opened 2026-07-09 08:13 by skottmckay

Fix Qwen3 RotaryEmbedding fusion emitting invalid position_ids shape

#29632 opened 2026-07-09 00:48 by Sammy-Dabbas

Fix CPU AvgPool ceil_mode + count_include_pad wrong average (opset 7-18 / MLAS path) core runtime

#29629 opened 2026-07-08 23:22 by titaiwangms

Harden React Native iOS Detox CI against simulator shell crashes

#29626 opened 2026-07-08 21:08 by Copilot

Enable Spectre-mitigated MSVC libs for BinSkim builds

#29624 opened 2026-07-08 20:54 by tianleiwu

Enable fpA_intB GEMM in CUDA builds and add configurable options release:1.28.0

#29622 opened 2026-07-08 20:24 by tianleiwu

Fix CPU FlashAttention disabled on Linux/aarch64 by detecting L2 cache via cpuinfo

#29621 opened 2026-07-08 18:16 by Sammy-Dabbas

[CUDA] Add cuDNN-free ArgMax/ArgMin/ReduceSum and fix LogSoftmax on plugin EP release:1.28.0

#29620 opened 2026-07-08 18:12 by tianleiwu

[MLAS] Add AVX2 (+VNNI) 2-bit weight CPU kernels

#29619 opened 2026-07-08 17:47 by blazingphoenix7

Fix type/shape inferencing double recursion for subgraphs

#29617 opened 2026-07-08 08:28 by skottmckay