microsoft/onnxruntime

Pull Requests Commits

Aditya Rastogi committed 3 hours ago

66f530f9

Fix build pipeline for CUDA 13.0 (#26983)

tianleiwu committed 11 hours ago

Verified de99059e

Add EP Device Compatibility APIs (#26922)

xhan65 committed 13 hours ago

Verified 081633e9

[QNN EP] Adjust tolerance for Clip and Transpose tests due to FP16 default in QNN HTP (#26499)

yuhuchua-qti committed 19 hours ago

Verified c343143a

[webgpu] Fix the incorrect shapes when profiling (#26928)

qjia7 committed 20 hours ago

Verified aeac7579

[QNN] Disable 7 RMSNorm onnx backend tests (#26999)

titaiwangms committed 23 hours ago

Verified cde7ed48

Allow MSVC to build KleidiAI in Windows on Arm environments. (#26995)

Colm-in-Arm committed 1 day ago

Verified 13774261

Address some issues and fix VS 2026 build (#26905)

skottmckay committed 1 day ago

Verified 4b50b2f1

[QNN EP] Debug info dump of inputs, outputs and initializers in QNN EP (#26622)

quic-ashwshan committed 1 day ago

Verified de2bb397

[QNN-EP] Disable DSPQ polling when needed (#26468)

quic-calvnguy committed 1 day ago

Verified d9f379c1

[QNN EP] Upgrade QNN to 2.42.0 (#26958)

qti-jkilpatrick committed 1 day ago

Verified 669128ac

remove lint for /js/ folder (#26984)

prathikr committed 1 day ago

Verified f86a0ed8

[QNN EP] Add RMSNorm Op support in QNN EP (#26853)

tirupath-qti committed 1 day ago

Verified e5e91743

Enhance stale issue management with new label checks (#26982)

justinchuby committed 2 days ago

Verified cfd5667a

Compute cpabality as kCURRENT by default (#26663)

umangb-09 committed 2 days ago

Verified 912f6523

Cuda Graph on by default (#26929)

umangb-09 committed 2 days ago

Verified 0a93edb0

[QNN-EP] Apply Softmax layout transformation for GPU (#26233)

qti-mattsinc committed 2 days ago

Verified 65e694ea

[QNN-EP] Support pad op pre-opset11 (#26248)

qti-mattsinc committed 2 days ago

Verified 15efe1c1

Minimize binary footprint of the FL nuget (#26961)

baijumeswani committed 2 days ago

Verified a5831a3d

Integration with ONNX==1.20.1 (#26579)

titaiwangms committed 2 days ago

Verified de398c41

[QNN EP] Default to HTP instead of using last provided device (#26696)

qti-mattsinc committed 2 days ago

Verified addc8120

[js/node] fix incorrect output name mapping (#26939)

fs-eire committed 2 days ago

Verified ce4a7c58

coreml: Fix Gather fp16 support (#26442)

borg323 committed 2 days ago

Verified 412233f9

Allow present_key to be empty when past_key is provided in Attention (#26303)

justinchuby committed 2 days ago

Verified 1d149f5a

[MLAS] Fix building on FreeBSD / powerpc64le (#26895)

pkubaj committed 3 days ago

Verified e0d5b585

[CUDA] Update Flash Attention Implementation and APIs (#26937)

tianleiwu committed 4 days ago

Verified 5d24c954

[EP ABI] Add APIs to create control flow kernels for plugin EPs (#26927)

adrianlizarraga committed 4 days ago

Verified 1ed8fd90

[CUDA] GQA CUDA Kernel Fusion and Performance Optimization (#26920)

tianleiwu committed 5 days ago

Verified 39d8520b

Allow choosing memory type for CudaPinned allocator (#26892)

AndreyOrb committed 5 days ago

Verified 5b88e4ed

[QNN EP] Add Case-2 LPBQ pattern support for Gemm and Matmul nodes (#25865)

quic-tirupath committed 5 days ago

Verified c1da27c2

Older