microsoft/onnxruntime

Pull Requests Commits

Fix gpu_external_data_loader build: avoid provider_api.h and logging.h

chilo-ms committed 49 days ago

1c855ad7

Fix CUDA shared provider logging: use provider_api.h instead of direct logging include

chilo-ms committed 49 days ago

d012279a

Fix undefined symbol for IExternalDataLoader in CUDA shared provider

chilo-ms committed 49 days ago

c5de900f

fix build issue

chilo-ms committed 56 days ago

6553e2b1

chilo-ms committed 56 days ago

e323f281

chilo-ms committed 56 days ago

2cae7f22

Update GatherBlockQuantized to support 2-bits (#28530)

HectorSVC committed 73 days ago

Verified 97cfc86f

Raise protobuf minimum version in Python requirements (#28558)

anzzraju1997-glitch committed 73 days ago

Verified 7a3f9a21

[Annotations] Address cuda orginal matching (#28623)

yuslepukhin committed 73 days ago

Verified b5fb2b89

Optimize MatMulNBits 2-bit + float zero_point CPU dequantization with multi-threaded kernel (#28589)

Copilot committed 74 days ago

Verified c267f8e2

Fix Erf NaN preservation on x86_64 FMA3 path (#28615)

tairenpiao committed 74 days ago

Verified c5afcc5d

Additional diagnostics for DML failure path (#28495)

adrastogi committed 76 days ago

Verified a1fc9165

Add example and documentation for kOrtEpDevice_EpMetadataKey_OSDriverVersion (#28282)

adrastogi committed 76 days ago

Verified d464b2a0

[MLAS] RVV-Optimized LLM Operators for RISC-V (#28518)

velonica0 committed 76 days ago

Verified b2f6e151

Fix: Accept 'CPU' as a valid provider name in SessionOptionsAppendExecutionProvider (#28625)

jiafatom committed 76 days ago

Verified a4f79e82

Fix Reshape with allowzero=1 producing wrong shape for zero-size tensor in chained Reshape (#28455)

Copilot committed 77 days ago

Verified a4f40c11

[WebGPU] LinearAttention: increase tile_v when subgroups are available (#28519)

daijh committed 77 days ago

Verified 1053327e

fix(qdq): skip DQ forward propagation when DQ input is constant (#28521)

Rishi-Dave committed 77 days ago

Verified 359d9ab4

Optimize MLAS quantized KV-cache GEMM kernels (follow-up to #28578) (#28606)

tianleiwu committed 77 days ago

Verified d2836a8f

Fix CUDA build with contrib ops disabled (#28554)

Copilot committed 77 days ago

Verified 5003d932

QMoE CUDA: input validation, prepack cleanups, and packaging pipeline fix (#28607)

tianleiwu committed 77 days ago

Verified 43989a71

[NVEP]: fix test for multi-gpu situation (#27837)

theHamsta committed 77 days ago

Verified b850fcbb

Parallelize CPU ScatterElements kernel via ThreadPool (#28588)

Copilot committed 77 days ago

Verified e3655b35

Validate seqlens_k against cos_cache bounds in GroupQueryAttention to… (#28277)

apsonawane committed 77 days ago

Verified 1e80c291

Add component governance manifest for WebGPU EP (#28599)

adrastogi committed 77 days ago

Verified e10b9a81

[CoreML EP] Support Gather with scalar 'indices' (#28278)

maxwbuckley committed 78 days ago

Verified cf8e4f55

Validate conv bias shape in WordConvEmbedding to prevent OOB read (#28279)

apsonawane committed 78 days ago

Verified ee444bdc

Fix oob bias access for MatMulIntegerToFloat and DynamicQuantizeMatMul (#28499)

adrianlizarraga committed 78 days ago

Verified 158bdef0

[WebGPU plugin EP] Add Win ARM64 Python package (#28526)

edgchen1 committed 78 days ago

Verified 84d03c58

Address MaxUnpool shortcomings msrc116345 (#28550)

yuslepukhin committed 78 days ago

Verified a894893f

Older