Pull Requests microsoft/onnxruntime

[CORE]Fix windows warning with fp4 header inclusion

#26134 opened 2025-09-23 19:32 by hariharans29

Bump transformers from 4.52.1 to 4.53.0 in /onnxruntime/python/tools/transformers/models/llama dependencies python

#26133 opened 2025-09-23 19:12 by dependabot[bot]

[TRT RTX EP] Fix bug for generating the correct subgraph in GetCapability

#26132 opened 2025-09-23 18:31 by chilo-ms

webgpu qmoe support ep:WebGPU

#26130 opened 2025-09-23 17:10 by guschmue

fix np.testing argument order in ORT backend runner

#26128 opened 2025-09-23 07:34 by xiaofeihan1

[QNN EP] Add support for RandomUniformLike Op in QNN EP

#26123 opened 2025-09-22 22:13 by quic-ashwshan

[QNN-EP] Dynamic EP Perf

#26122 opened 2025-09-22 21:38 by quic-calvnguy

[QNN EP] Add QAIRT Licensing file in QNN EP ep:QNN

#26121 opened 2025-09-22 21:08 by quic-tirupath

Allow vendor IDs in Python IOBinding device binds

#26119 opened 2025-09-22 19:47 by Jonahcb

Fix python test issues for the 1.23.0 release

#26118 opened 2025-09-22 19:33 by snnn

[NV TensorRT RTX EP] Package CUDA Runtime and TRT RTX with NV TensorRT RTX EP python wheel

#26115 opened 2025-09-22 07:46 by keshavv27

Bump actions/download-artifact from 4 to 5 dependencies github_actions

#26112 opened 2025-09-22 01:04 by dependabot[bot]

Bump ruff from 0.12.12 to 0.13.1 dependencies python

#26111 opened 2025-09-22 01:03 by dependabot[bot]

Bump clang-format from 20.1.8 to 21.1.1 dependencies python

#26110 opened 2025-09-22 01:03 by dependabot[bot]

Consistent with the configuration in the packaged cmake

#26104 opened 2025-09-20 18:26 by WangFengtu1996

[DO NOT REVIEW OR MERGE] Test PR

#26103 opened 2025-09-20 04:31 by hariharans29

Add BF16 kernels in several ops for Gemma-3

#26102 opened 2025-09-20 00:53 by kunal-vaishnavi

[C#] Implement missing APIs release:1.23.1

#26101 opened 2025-09-19 22:17 by yuslepukhin

[js/node] enable handling EP options for WebGPU EP

#26099 opened 2025-09-19 09:20 by fs-eire

CuFFTPlanCache: add destructor and Clear() for memory release

#26098 opened 2025-09-19 08:35 by cocotdf

[web] support JSPI build

#26095 opened 2025-09-19 03:41 by fs-eire

[js/webgpu] enable handling EP options for WebGPU EP

#26093 opened 2025-09-19 02:55 by fs-eire

Write parameters based on output shapes not the amount of names

#26092 opened 2025-09-19 02:18 by TedThemistokleous

Update QMoE kernel with optimizations

#26091 opened 2025-09-19 01:25 by apsonawane

MultiheadAttention CUDA BF16 Support

#26083 opened 2025-09-18 17:16 by nenad1002

Declare Shape, Reshape, Transpose, Squeeze, Unsqueeze for opsets 21, 23 on CUDA

#26075 opened 2025-09-18 08:58 by xadupre

[webgpu] Use DoTranspose directly in Conv Op

#26074 opened 2025-09-18 08:29 by wenqinI

[WebGPU] add unidirectional support for MHA ep:WebGPU

#26071 opened 2025-09-17 23:13 by fs-eire

[QNN-EP] Support alternate Layernorm fusion pattern in QNN preprocess ep:QNN

#26060 opened 2025-09-16 19:48 by qti-mattsinc

cherry-pick Fix Attention GQA implementation on CPU

#26057 opened 2025-09-16 17:47 by xadupre