microsoft/onnxruntime

Pull Requests Commits

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

edgchen1 committed 94 days ago

7a349d49

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

qti-jkilpatrick committed 94 days ago

Verified 669128ac

remove lint for /js/ folder (#26984)

prathikr committed 95 days ago

Verified f86a0ed8

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

tirupath-qti committed 95 days ago

Verified e5e91743

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

justinchuby committed 95 days ago

Verified cfd5667a

Compute cpabality as kCURRENT by default (#26663)

umangb-09 committed 95 days ago

Verified 912f6523

Cuda Graph on by default (#26929)

umangb-09 committed 95 days ago

Verified 0a93edb0

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

qti-mattsinc committed 95 days ago

Verified 65e694ea

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

qti-mattsinc committed 95 days ago

Verified 15efe1c1

Minimize binary footprint of the FL nuget (#26961)

baijumeswani committed 95 days ago

Verified a5831a3d

Integration with ONNX==1.20.1 (#26579)

titaiwangms committed 95 days ago

Verified de398c41

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

qti-mattsinc committed 95 days ago

Verified addc8120

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

fs-eire committed 95 days ago

Verified ce4a7c58

coreml: Fix Gather fp16 support (#26442)

borg323 committed 95 days ago

Verified 412233f9

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

justinchuby committed 96 days ago

Verified 1d149f5a

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

pkubaj committed 96 days ago

Verified e0d5b585

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

tianleiwu committed 97 days ago

Verified 5d24c954

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

adrianlizarraga committed 98 days ago

Verified 1ed8fd90

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

tianleiwu committed 98 days ago

Verified 39d8520b

Allow choosing memory type for CudaPinned allocator (#26892)

AndreyOrb committed 98 days ago

Verified 5b88e4ed

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

quic-tirupath committed 98 days ago

Verified c1da27c2

Update protobuf references from 3.20.3 to 4.25.8 (#26910)

eserscor committed 98 days ago

Verified e6110d03

Add OrtExternalResourceImporter API for D3D12 shared resource import (#26828)

nieubank committed 99 days ago

Verified c54be3c7

[OVEP] Update openvino version to 2025.4.1 (#26943)

preetha-intel committed 99 days ago

Verified a14231a5

[QNN EP] Update tensor naming logic in DebugNodeInputsOutputs (#26660)

yuhuchua-qti committed 99 days ago

Verified 3924b6e8

[ICM CXX] Address Pad kernel shortcomings (#26864)

yuslepukhin committed 99 days ago

Verified 5d596088

[webgpu] Simplify the signature of CanApplyFlashAttention (#26926)

qjia7 committed 100 days ago

Verified 56c984ff

[webgpu] Support `Split-K` in more situations (#26806)

Jiawei-Shao committed 100 days ago

Verified 4a858a82

[CUDA] Speed up flash attention build (#26924)

tianleiwu committed 100 days ago

Verified 541d5daf

GPG changed default install location on 12/31/2025 for Windows (#26917)

eserscor committed 100 days ago

Verified 751af64a

Older